|
Advertisement | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
J. Biol. Chem., Vol. 280, Issue 10, 8629-8632, March 11, 2005
Minireview Toward Biochemical Understanding of a Transcriptionally Silenced Chromosomal Domain in Saccharomyces cerevisiae*![]() From the Department of Biomolecular Chemistry, University of Wisconsin Medical School, Madison, Wisconsin 53706
A few discrete portions of the Saccharomyces cerevisiae genome, such as the chromosome ends within 510 kilobase pairs (kb) of telomeres and the silent mating-type regions named HMR and HML, are defined as transcriptionally silent because they cause repression of an RNA polymerase II gene placed within them (reviewed comprehensively in Ref. 1). S. cerevisiae silencing is a form of transcriptional repression that is long range, occurring over distances larger than a single gene. Silencing occurs in only certain chromosomal regions and depends more on a gene's position in the genome than its promoter sequence. Another defining characteristic is a stable inheritance pattern during cell division, indicating that the repressive state is duplicated along with the underlying DNA sequences during DNA replication. These features make S. cerevisiae transcriptional silencing functionally akin to position-effect-variegation in Drosophila melanogaster and X-chromosome inactivation in female mammals (for recent reviews see Refs. 2 and 3). It is now established that these functionally similar chromosomal phenomena share a requirement for a distinct chromatin structure. The precise molecules involved in forming repressive chromatin vary between organisms and even different chromosomal regions within an organism (for examples see review in Ref. 4). However all silent chromosomal domains contain particular post-translational modifications of nucleosomes, relatively ordered nucleosome positioning and specialized chromatin-binding proteins that distinguish them from actively transcribed chromosomal domains (reviewed in Ref. 5).
The transcriptionally silent HMR locus in S. cerevisiae serves as an important model for the structure, assembly, and function of a repressive chromatin domain. HMR and HML (HM loci; silent mating-type loci) are two regions of
Here we focus on how HMR (Fig. 1B), which shares fundamental molecular requirements for silencing with HML, assembles into a silent chromatin domain. Over the past 10 years the primarily genetic focus of this field has shifted toward more molecular analyses, and several remaining questions will benefit from biochemical approaches. It is now common to present transcriptional silencing in terms of a molecular model that we have adapted for this discussion of HMR (810) (Figs. 2, 3, 4).
The HMR-E silencer is a small 150-bp DNA element required for silent chromatin to form at HMR (Fig. 2). HMR-E contains binding sites for three different sequence-specific DNA-binding proteins, the origin recognition complex (ORC),1 Rap1p, and Abf1p (only ORC and Rap1p are shown in Fig. 2). These silencer-binding proteins bound to the silencer DNA provide a unique protein-DNA interaction surface (the silencer protein complex, SPC, in Fig. 3) that is recognized by four specialized chromatin-binding proteins, the Sir (silent information regulator) proteins.
All three silencer-binding proteins have additional non-silencing functions when bound elsewhere in the genome. The ORC is a 6-subunit heteromeric protein complex that binds to each of the Part of the answer comes from the distinct arrangement of silencer-binding proteins at HMR-E. The close juxtaposition of ORC and Rap1p binding sites, for example, is unique to silencers. Considering only two of the Sir proteins, Sir1p and Sir4p, illustrates how this juxtaposition contributes to the stability and specificity of HMR silencing. The Sir4p physically contacts Rap1p (15), whereas the Sir1p contacts ORC (16, 17). In addition, Sir1p and Sir4p physically interact (16, 18). The unique arrangement of ORC and Rap1p at the silencer allows these individually weak interactions to occur simultaneously, contributing to a stable complex. Several other Sir-Sir and Sir protein-SPC interactions work similarly to contribute to HMR-E silencer function (reviewed in Ref. 19) (Fig. 2). Indeed, the final complex is so stable that multiple interaction defects must be created in vivo simultaneously, through combinations of specific mutations, to cause a detectable HMR silencing defect. Distinct interactions between the silencer-binding proteins and the HMR-E silencer-DNA may also contribute to silencer function. For example, the interaction of ORC with its binding site at the HMR-E silencer is substantially stronger than its interaction with many of its binding sites present in non-silencer DNA replication origins (20). This strong interaction, which is greater than needed for occupancy of ORC at replication origins, contributes to HMR-E silencer function, raising the possibility that a silencer-specific ORC-DNA interaction enhances the binding of Sir1p to ORC. A mechanistic understanding of how individual silencer-binding protein-DNA interactions at the HMR-E silencer control the binding of Sir proteins requires further investigation. For example, the ORC-silencer DNA interaction could enhance Sir1p binding either through promoting a unique conformation of ORC or by presenting silencer DNA for direct contacts by Sir1p. The other silencer-binding proteins may be influenced by their interactions with their silencer binding site. Indeed, Rap1p function may be influenced by the nature of its binding site (21, 22). It is unclear how Abf1p functions at the silencer because no distinct Abf1p-Sir interactions have been published although there are specific silencing-defective alleles of ABF1 (19). Finally, the SPC functions with a distinct orientation, nucleating silent chromatin more efficiently in one direction (Fig. 3), but the mechanisms controlling this directionality are unknown. Structural studies of the SPC should reveal how it promotes the stability and directionality of silent chromatin at HMR.
A Sir-protein complex bound to the SPC is positioned to enhance the binding of additional Sir24 complexes to neighboring nucleosomes (see below and Fig. 3). The Sir1p clearly binds to the SPC through interactions with ORC (Fig. 2), but its association with regions distal to HMR-E is reduced (9). Sir1p's association contrasts with that of the other three Sir proteins that bind similarly to regions throughout HMR (Fig. 3) (1). Therefore the Sir24 proteins, in contrast to Sir1p, are viewed as inherent structural components of the silent chromatin domain. Sir24 associations with the silent domains are interdependent, suggesting that each Sir helps stabilize binding of the other Sirs (8, 9, 23). A current view is that Sir24 proteins function together in a complex that may be stabilized in part through associations with chromatin (8, 10, 24). The Sir24 complex is proposed to play two different but interdependent roles in forming a silent chromatin domain. First, the complex modifies nucleosomes, playing an enzymatic role in post-translationally modifying chromatin. Second, the complex binds directly to these modified nucleosomes and plays a direct structural role in silent chromatin. A common feature of all silent chromatin domains is that lysine residues present on the unstructured tails of histones within nucleosomes are hypoacetylated (for reviews see Refs. 25 and 26). Arguably the most significant recent biochemical discovery relevant to S. cerevisiae silencing is that Sir2p is an NAD-dependent protein deacetylase that removes acetyl groups from lysine residues (for a short review see Ref. 27), and Sir2 is proposed to deacetylate nucleosomes at HMR (28). In contrast to the other Sir proteins that show little conservation, Sir2p is the founding member of a family of NAD-dependent protein deacetylases (sirtuins); multiple members of this family, which share a conserved catalytic domain but diverge beyond this core, are found in prokaryotes and eukaryotes, and protein substrates probably vary considerably between individual sirtuins (reviewed in Ref. 29). Silencing represents a small subsection of the broader fields of sirtuin enzymology and chromatin modification for which there are excellent reviews (27, 30). This discussion will confine itself to considering Sir2p deacetylase in terms of the model for silent chromatin at HMR (Fig. 3). In addition to the Sir proteins, relatively hypoacetylated histones can also be detected at all regions within the defined HMR domain. Critically, this depends on the enzymatic activity of Sir2p (9) that removes, minimally, an acetyl group from lysine 16 of histone H4 (3133). In cells expressing a catalytically inactive Sir2p, HMR fails to repress transcription, and Sir proteins are not detected at regions of HMR distal to the silencer (9). In addition, nucleosomes within HMR in such mutants are acetylated similarly to actively transcribed regions such as MAT. However, even in yeast cells harboring catalytically inactive Sir2p, a Sir24 complex can be detected at the HMR-E silencer itself, indicating that the deacetylase activity of Sir2 is not required for Sir24 complex binding to the SPC. Based on these data, the idea is that Sir2p within the Sir24 complex bound to the SPC deacetylates a nucleosome adjacent to the silencer. This modification in turn promotes binding of another Sir24 complex to the nucleosome. Consistent with this proposal is that both Sir3 and Sir4 can bind the N-terminal tails of histone H3 and H4 (34) and Sir3p has a higher affinity for hypoacetylated histone tail peptides than acetylated peptides (35). The new Sir24 complex is now positioned to modify the next adjacent nucleosome and so on, until the entire domain is hypoacetylated and stably associated with Sir24 protein complexes (Fig. 2). The model is consistent with the available data, but many questions remain. For example, the model is often presented as involving one Sir24 complex per nucleosome and depicts these complexes spreading two-dimensionally from the silencer, along an unstructured array of nucleosomes (Fig. 2). However, it is conceivable that a submolar ratio of Sir complexes to nucleosomes at HMR would be sufficient to modify and affect the domain's chromatin structure in a way that would sufficiently repress transcription and create Sir protein contacts with the entire HMR region (36). In addition, it is highly probable that a more complicated "supramolecular" chromatin complex is formed due to both the effects of deacetylation and Sir protein binding on the nucleosome array (reviewed in Ref. 37). A more descriptive model of silent chromatin, in terms of number and types of individual proteins and protein modifications will require refining techniques for the purification and analysis of a defined HMR chromatin domain (38). Also more accurate structural models will require both high resolution structural studies of Sir proteins and relevant complexes, which have begun (39), as well as lower resolution studies of the effect Sir proteins have on defined chromatin fibers (36). Nevertheless, in the absence of such data, the two-dimensional picture is a simple way to visualize aspects of silencing and helps one consider mechanisms that establish boundaries between silent and active chromatin domains (Fig. 4).
The action of distinct chromosomal regions called boundary elements prevent Sir-dependent chromatin from advancing into active chromatin domains and inappropriately repressing transcription (reviewed in Ref. 40). A boundary element creates a permissive chromatin domain with reduced affinity for Sir proteins that is in opposition to the silent chromatin emanating from HMR-E. The boundary element on the right side of HMR is the better understood and probably more relevant because silent chromatin spreads from the HMR-E silencer primarily toward the right telomere (41). This element consists of a promoter for an actively transcribed tRNA gene. However, the act of transcription per se is not critical for its boundary function. Rather, the ability of the promoter to recruit proteins, such as histone acetyltransferases that make chromatin in this region permissive for transcription, is important (42, 43). As a boundary element, the job of the tRNA gene promoter is to act as an "anti-silencer" promoting an active chromatin domain in direct competition with the silent chromatin domain promoted by the HMR-E silencer. Transcription is a normal consequence of formation of this active chromatin domain but not the inherent cause of silent chromatin inhibition. Another mechanism influencing the right boundary element at HMR relies on alternative nucleosomes that contain the histone 2A variant H2A.Z (encoded by HTZ1) in place of the more abundant H2A (encoded by two redundant genes, HTA1 and HTA2) (44). In the absence of H2A.Z, which unlike H2A is not essential for yeast viability, transcription of active genes located near silent domains is reduced. Further, Sir proteins and the associated hypoacetylated nucleosomes move beyond their normal limits at HMR, binding to neighboring regions containing active genes (Fig. 4). Thus the H2A.Z-containing nucleosomes positioned over certain gene promoters function to promote transcription by preventing the encroachment of Sir24-dependent complexes onto active gene promoters. Why do some promoters act as boundaries, preventing Sir-chromatin formation, whereas others are repressed by Sir chromatin? The answer is that not all gene promoters are equivalent. Sir-dependent silent chromatin is more effective at repressing genes normally transcribed at low (basal) levels but significantly less effective at repressing genes expressed at high (induced) levels (45). Silent chromatin in other organisms behaves similarly (46). Therefore, an effective boundary requires a nucleation center (in yeast essentially a gene promoter) that forms transcriptionally permissive chromatin at a level that can out-compete the ability of the HMR-E silencer to form silent chromatin. Whether all promoters that can act as Sir-chromatin boundary elements also use HZA.Z-containing nucleosomes in this context is an open question. Many important biochemical issues are raised by the ability of boundary elements to prevent the spread of Sir-dependent silent chromatin. In terms of H2A.Z it will be interesting to learn whether nucleosomes containing H2A.Z bind Sir proteins weakly compared with nucleosomes containing H2A. If this simple explanation is not correct, perhaps the presence of H2A.Z promotes additional changes to chromatin or recruits additional proteins that in turn inhibit binding of a Sir24 complex. Regardless, if yeast boundary function can be studied as something distinct from promoter function, as current analysis of HZA.Z suggests, the challenge will be to continue to identify clearly those factors and mechanisms important for transcription of genes near Sir-dependent chromatin but relatively dispensable for transcription of genes within the active chromatin domains that make up most of the yeast genome.
Although it is well established that the Sir proteins participate directly in the formation of silent chromatin at HMR, it is unclear how this structure actually represses transcription. A long held view was that Sir-dependent chromatin simply prevented access of transcription factors and the RNA polymerase II holoenzyme to gene promoters. This idea was attractive in part because the simplest subunit of chromatin, the nucleosome, can inhibit binding of these factors (reviewed in Ref. 47). In addition, Sir-dependent chromatin is relatively insensitive to the activities of several DNA-modifying enzymes, indicating that the DNA within silent chromatin is indeed generally inaccessible compared with active chromatin (reviewed in Ref. 48). However, relatively recent studies provide compelling evidence against the idea that Sir-dependent chromatin simply inhibits access of initiation factors to gene promoters (49, 50). In particular, normal levels of TBP, the general transcription factor that binds the TATA box in RNA polymerase II promoters, and the core subunit of RNA polymerase II, polII, are associated with a gene promoter repressed by Sir-dependent silent chromatin as measured by chromatin immunoprecipitation assays (ChIPs) (50). Indeed, transcription of a gene can be repressed by Sir24-dependent chromatin without a substantial reduction in the binding of key components of the transcriptional machinery to the gene's promoter. These data eliminate the simplest mechanism by which Sir24 silent chromatin inhibits transcription because clearly factors required for transcription can co-exist with Sir proteins. However, the ChIP assay is limited in its ability to distinguish between several remaining possibilities. For example, it is possible that key transcription initiation factors are associated with the silenced gene's promoter without being bound in a productive form. Alternatively, RNA polII holoenzyme may begin initiation but be incapable of a promoter-escape step that allows it to form an elongation complex (50). More refined reagents and assays for the study of promoter escape and elongation may help address aspects of this latter alternative by ChIP analysis. However it is likely that additional in vitro transcription assays on preassembled chromatin templates, as well as biochemical purification and analysis of a discrete Sir-dependent silenced HMR domain, will help distinguish among several possible explanations for how transcription is repressed within silent chromatin.
Silencing of the HMR domain in S. cerevisiae is achieved through the action of a DNA element called a silencer that binds sequence-specific DNA-binding proteins to form a SPC. The SPC recruits specialized chromatin proteins, the Sir proteins that deacetylate and bind nucleosomes within HMR. Sir-dependent chromatin is absolutely necessary to repress transcription of a gene within HMR and causes repression at a step after recruitment of the RNA polII holoenzyme to the gene's promoter. Specific mechanisms at boundary elements that include specialized nucleosomes help confine Sir-dependent chromatin to discrete chromosomal domains. Genetics will continue to play a role in defining important features of HMR silencing. However future mechanistic advances will depend on a combination of structural and functional biochemical approaches.
* This minireview will be reprinted in the 2005 Minireview Compendium, which will be available in January, 2006.
1 The abbreviations used are: ORC, origin recognition complex; SPC, silencer protein complex; ChIP, chromatin immunoprecipitation assay.
This article has been cited by other articles:
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Advertisement | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||