Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Global nucleosome occupancy in yeast

Global nucleosome occupancy in yeast Background: Although eukaryotic genomes are generally thought to be entirely chromatin- associated, the activated PHO5 promoter in yeast is largely devoid of nucleosomes. We systematically evaluated nucleosome occupancy in yeast promoters by immunoprecipitating nucleosomal DNA and quantifying enrichment by microarrays. Results: Nucleosome depletion is observed in promoters that regulate active genes and/or contain multiple evolutionarily conserved motifs that recruit transcription factors. The Rap1 consensus was the only binding motif identified in a completely unbiased search of nucleosome-depleted promoters. Nucleosome depletion in the vicinity of Rap1 consensus sites in ribosomal protein gene promoters was also observed by real-time PCR and micrococcal nuclease digestion. Nucleosome occupancy in these regions was increased by the small molecule rapamycin or, in the case of the RPS11B promoter, by removing the Rap1 consensus sites. Conclusions: The presence of transcription factor-binding motifs is an important determinant of nucleosome depletion. Most motifs are associated with marked depletion only when they appear in combination, consistent with a model in which transcription factors act collaboratively to exclude nucleosomes and gain access to target sites in the DNA. In contrast, Rap1-binding sites cause marked depletion under steady-state conditions. We speculate that nucleosome depletion enables Rap1 to define chromatin domains and alter them in response to environmental cues. the distribution of nucleosomes, the fundamental units of Background Global gene-expression patterns are established and main- chromatin, is poorly understood on a gene-specific basis, tained by the concerted actions of transcription factors and much less a global basis [2]. the proteins that constitute chromatin. The global network of interactions between transcription factors and promoters in The nucleosome consists of approximately 146 base-pairs yeast is increasingly being characterized [1]. The role of chro- (bp) of DNA wrapped around an octamer of histone proteins matin in gene regulation is less clear, however. For example, - two each of histones H2A, H2B, H3 and H4. Eukaryotic Genome Biology 2004, 5:R62 R62.2 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 genomes are packaged into repeating units of nucleosomes are remarkably similar as shown by a genome-wide correla- separated by around 10-80 bp of linker DNA. High occupancy tion of 0.83 (Figure 1a-c). This correlation is comparable to by nucleosomes is thought to be generally repressive [3], and that observed when comparing replicate H3 datasets (or H2B extensive remodeling (and loss) of nucleosomes occurs in the datasets), and suggests that both assays measure similar phe- promoters of genes undergoing activation [4]. In the case of nomena. In the H3 and H2B datasets, respectively, there are the PHO5 promoter in yeast, this remodeling proceeds until 347 and 214 regions depleted at least 1.5-fold relative to the essentially no nucleosomes are detected across a region of average over all intergenics. In contrast, there are just 84 and several hundred base-pairs [5,6]. 6 regions in the respective datasets enriched at least 1.5-fold relative to this average. The relatively narrow range of ChIP Transcription factors and chromatin proteins each form com- enrichment and the negative skew of the data (Figure 1b) are plex regulatory networks that interact in a variety of ways consistent with the conventional view that the majority of the [1,7]. Transcription factors modify chromatin structure by genome is packaged into nucleosomes with intervening recruiting enzymes that remodel nucleosomes or posttransla- stretches of free DNA such as the activated PHO5 promoter tionally modify histones (by acetylation or methylation, for [5,6]. example) [8-10]. The modifications can be maintained through cell division and propagated to proximal nucleo- Despite these consistencies, a possible caveat to using ChIP to somes by positive-feedback mechanisms [7,11,12]. Hence, a evaluate nucleosome occupancy is that immunoprecipitation signal such as the activation of a transcription factor can be efficiency can depend on epitope accessibility. Rather than temporally and spatially transmitted through chromatin. having low occupancies, genomic regions depleted in the H3 Conversely, chromatin can influence transcription factor ChIP might be inaccessible as a result of association with function by modulating the accessibility of target binding large protein complexes in chromatin. To investigate this pos- sites in the DNA [13,14]. sibility, we examined a published chromatin fractionation dataset in which cross-linked chromatin fragments were sub- We used chromatin immunoprecipitation (ChIP) and DNA jected to phenol-chloroform extraction and DNA that parti- microarrays to evaluate nucleosome occupancy levels for tioned into the aqueous phase was quantified by microarrays essentially all promoters in yeast. Promoters that regulate [15]. Given the polar nature of DNA and the hydrophobic active genes, contain multiple conserved motifs or recruit nature of denatured protein, aqueous extraction should gen- Rap1 tend to be relatively nucleosome-depleted. We also used erally enrich for free DNA. We found that regions depleted in real-time PCR and micrococcal nuclease digestion to show the H3 ChIP assay overlap extensively with regions enriched that nucleosomes are depleted in the vicinity of Rap1 consen- by aqueous extraction, but not with regions depleted by aque- sus sites. This depletion can be partially reversed by the ous extraction (Figure 1d). Overall, there is a negative corre- actions of the small molecule rapamycin or by removing lation of -0.54 between the H3 ChIP and aqueous-extraction Rap1-binding sites. We suggest that other transcription fac- datasets. Although the fractionation data may partially reflect tors have less robust nucleosome-depleting activities than differential cross-linking of lysines in the histone tails [15], Rap1 and must therefore act collaboratively to gain access to this analysis suggests that regions depleted in the H3 ChIP their cognate sites in the DNA. experiment are relatively protein-free, as would be expected of non-nucleosomal DNA. Nucleosome occupancy correlates inversely with Results ChIP-based assay for nucleosome occupancy promoter strength Histones are essential components of the nucleosome and As previous studies show that PHO5 activation is accompa- efficiently cross-link to nucleosomal DNA. Antibodies against nied by marked nucleosome loss in the promoter region [5,6], invariant portions of histones have been used previously in we sought to determine whether nucleosome depletion is a ChIP assays to follow nucleosome loss at the yeast PHO5 pro- general attribute of active promoters. A total of 4,365 inter- moter [5,6]. We extended this approach to evaluate relative genic regions that reside immediately upstream of one or nucleosome occupancy at essentially all promoters and other more validated yeast genes were assigned as promoters. Rel- intergenic regions in yeast. DNA associated in vivo with his- ative transcription rates were determined for each yeast gene tone H3 was isolated by ChIP using antibody against the car- from transcript levels measured by array and previously col- boxy terminus of histone H3 (no posttranslational lected mRNA half-life data [16]. We found an inverse correla- modifications are thought to occur in this region). ChIP DNA tion of -0.39 between the enrichment of promoters in the H3 and unenriched control DNA were amplified by in vitro tran- and H2B ChIP assays and the transcription rates of down- scription and evaluated using microarrays. DNA associated stream genes (Figure 2a). Under the conditions examined, with histone H2B was evaluated in a similar fashion using PHO5 is not induced and its promoter has an average nucleo- anti-FLAG antibody and a FLAG-H2B strain. H3 and H2B some occupancy according to these datasets. To evaluate fur- datasets were compiled by averaging four and three inde- ther the relationship between nucleosome depletion and pendent biological experiments, respectively. These datasets transcription, we collated a set of 308 nucleosome-depleted Genome Biology 2004, 5:R62 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.3 (a) (c) R = 0.83 n = 48 H2B-depleted H3-depleted (> 1.5-fold) (> 1.5-fold) (d) −1 n = 2 Aqueous- −2 198 149 161 depleted (bottom 5%) −2 −10 1 Enrichment ratio H3 ChIP (log ) Aqueous- H3-depleted enriched (> 1.5-fold) (top 5%) (b) H3 ChIP H2B ChIP Enrichment ratio (log ) C Fio grrelatio ure 1 n between H3 and FLAG-H2B ChIP datasets Correlation between H3 and FLAG-H2B ChIP datasets. DNA associated with histones in vivo was enriched in ChIP assays using antibodies against histone H3 or FLAG-H2B, and quantified by microarrays. (a) Relative enrichment of promoters and other non-coding regions in the H3 and H2B ChIP assays is shown. (b) Histogram showing distributions of enrichment for promoter regions in the H3 and H2B ChIP assays. (c) Overlap between regions depleted in the H3 and FLAG-H2B assays is shown. Overall, there is an 0.83 correlation between these ChIP datasets. (d) Overlap between regions depleted in the H3 ChIP assay and regions enriched by aqueous extraction is shown [62]. promoters on the basis of their relative depletion across the between promoter strength and nucleosome depletion. How- replicate H3 and H2B experiments. Of these nucleosome- ever, as this correspondence is not complete there are likely to depleted promoters, 42% regulate highly active genes (Figure be other determinants of nucleosome occupancy. 2b). These data suggest that there is a systematic relationship Genome Biology 2004, 5:R62 −2 −1.8 −1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0.2 0.4 0.6 0.8 1.2 Enrichment ratio H2B ChIP (log ) Frequency 2 R62.4 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 The second sequence element enriched in nucleosome- (a) depleted promoters corresponds to the consensus motif for 0.25 the Rap1 transcription factor. This motif commonly occurs in the promoters of ribosomal proteins genes and is required for Rap1 binding in vitro and in vivo [19,20]. Some variant of this motif appears in 22% of nucleosome-depleted promoters, H2B ChIP compared with just 8% of promoters overall (hypergeometric H3 ChIP -5 −0.25 p < 10 ). Furthermore, multiple Rap1 sites are found in 19% of nucleosome-depleted promoters with Rap1 sites, com- pared to 8% of promoters with Rap1 sites overall (hypergeo- −0.5 -3 metric p < 10 ). These data suggest that Rap1 recruitment may lead to nucleosome loss. −0.75 Because only the Rap1 consensus site was identified in an 0 102030405060708090 100 Percentile rank of transcription rate unbiased search, we sought to identify additional sequence motifs by incorporating species conservation data. Specifi- (b) cally, we evaluated a set of 71 conserved motifs identified by −5 p < 10 Kellis and colleagues, a majority of which function in tran- scription factor recruitment [21]. Nearly half of these 71 motifs are over-represented in nucleosome-depleted promot- ers relative to promoters overall, as defined by a hypergeo- 179 129 308 metric p < 0.001. However, many of the implicated motifs appear in the same promoters. For example, nine of the over- Nucleosome- Promoters 5′ of represented motifs are associated with filamentation gene depleted promoters active genes (top 10%) promoters [21]. We therefore considered the possibility that the total number of conserved motifs might be a more rele- Istren Fig nveu rs r ge e as th 2 sociation between nucleosome occupancy and promoter vant predictor of nucleosome depletion. Indeed, we found Inverse association between nucleosome occupancy and promoter that 31% of nucleosome-depleted promoters contain at least strength. (a) Relative enrichment of promoter regions in the H3 and eight motifs, compared with 11% of promoters overall (hyper- FLAG-H2B ChIP assays plotted against transcription rate of downstream -5 geometric p <10 ; Figure 3b). Furthermore, nucleosome- genes (moving average, window 50). (b) Overlap between promoters upstream of active genes and the set of nucleosome-depleted promoters depleted promoters contain an average of 6.1 motifs, whereas defined on the basis of depletion across the replicate H3 and FLAG-H2B the average promoter contains 3.1 (permutation p < 0.001; experiments. Figure 3c). Next, we sought motifs associated with nucleo- some depletion in the absence of multiple motifs, by confin- ing our analysis to promoters containing a maximum of four Transcription factor binding motifs are over- motifs. This analysis identified just two over-represented represented in nucleosome-depleted promoters motifs, which correspond to the Rap1 and Swi4 binding sites. To identify additional determinants of occupancy, we sought Hence, although a large number of conserved motifs are sequence elements associated with nucleosome depletion. enriched in nucleosome-depleted promoters, most appear to Specifically, we carried out an unbiased search for elements be relevant mainly when occurring in combination. up to 10 bp in length that occur with higher frequency in nucleosome-depleted promoters. Two distinct categories of Functionally cooperative transcription factors sequences emerged (Figure 3a). The first includes associate with nucleosome-depleted promoters poly(dA.dT) elements. Stretches of 10 or more dA.dT nucle- As a majority of the conserved motifs recruit transcription otides appear in 38% of depleted promoters, compared with factors [21], we examined the relationship between transcrip- -5 26% of promoters overall (hypergeometric p < 10 ). dA.dT tion factor binding and nucleosome occupancy more directly. stretches destabilize nucleosome formation in vitro and in Lee and colleagues combined ChIP and microarrays to iden- vivo [17,18]. The enrichment of poly(dA.dT) elements in tify target promoters for essentially all yeast transcription nucleosome-depleted promoters probably reflects, at least in factors under the same conditions used here to evaluate part, this destabilizing influence. As a high proportion of the nucleosome occupancy [1]. For each factor, we determined poly(dA.dT) elements identified in nucleosome-depleted pro- the significance of overlap between its target promoters and moters are more than 10 bp long (30% are at least 14 bp), the set of nucleosome-depleted promoters. Of the 113 tran- these data do not address the minimum length required for scription factors in their database, 31 tend to associate with destabilization. However, in vitro studies show that a 16-bp nucleosome-depleted promoters as defined by a hypergeo- insertion leads to a 1.7-fold increase in accessibility of nucle- metric p < 0.001. Rap1 has the most significant association osomal target sites [18]. (Figure 4a), consistent with the enrichment of its binding Genome Biology 2004, 5:R62 Enrichment ratio (log ) 2 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.5 (a) Sequence Hits in nucleosome- Hits expected P-value depleted promoters at random CACCCGTACA 38 4 5.1E-20 ACACCCGTAC 33 3 2.6E-15 CATCCGTACA 40 6 6.0E-14 TTCTTTTTTT 218 133 4.0E-12 TTTTTTTTCC 177 97 5.1E-12 CACCCATACA 43 8 2.3E-11 GTTTTTTTTC 159 84 2.7E-11 TTTTTTTCTC 170 93 3.0E-11 TTTTTTTCTG 150 77 5.1E-11 TTTTTTTTTT 211 131 3.8E-10 (b) p < 10 5 213 95 381 Nucleosome- Promoters with depleted promoters ≥ 8 conserved motifs (c) Nucleosome- depleted Conserved motifs per promoter (average) Seque Figure nce motifs o 3 ver-represented in nucleosome-depleted promoters Sequence motifs over-represented in nucleosome-depleted promoters. (a) An unbiased search for sequences up to 10 bp in length over-represented in nucleosome-depleted promoters (relative to promoters overall) identified the poly(dA.dT) sequence element and variants of the Rap1 consensus motif ACACCCATACAT [21]. (b) Overlap between nucleosome-depleted promoters and promoters that contain multiple conserved motifs is shown [21]. (c) Histogram showing average numbers of motifs in 1,000 randomly generated promoter sets. Nucleosome-depleted promoters contain an average of 6.1 conserved motifs, significantly higher than in these randomly generated sets. motif (see above). Other top-ranked factors include Fhl1, However, these factors utilize a variety of binding domains, which associates with many Rap1-bound promoters, and regulate different pathways, and only a minority have signifi- Swi4, whose binding motif is also enriched (Table 1). cant associations with promoters of highly active genes. Nonetheless, a commonality does emerge when transcription We sought an underlying binding mechanism or function factor cooperativity is considered. A recent informatics study common to the transcription factors we had identified. by Banerjee and Zhang identified 31 functionally cooperative Genome Biology 2004, 5:R62 2.5 3.5 4.5 5.5 6.5 Number of random sets R62.6 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 (a) −5 (d) p < 10 Rap1 motifs RPS11B 207 101 190 RPS11B∆ RAP1 Nucleosome- Promoters depleted promoters bound by Rap1 (b) H3 ChIP (mutant) Ratio (relative to wild-type) RPS22A RPS15 RPS11B RPL23A RPS11B∆ RAP1 +2.1-fold −2 (e) −4 H3 ChIP Ratio (relative to untreated) −6 (1 hr rapamycin) −8 RPS22A +3.0-fold RPS15 +4.7-fold −10 RPS11B +2.1-fold Histone H3 ChIP RPL23A +1.9-fold Histone H2B ChIP (c) TUB2 RPS11B RPS15 Nu Figcleo ureso 4me depletion in the vicinity of Rap1-binding sites Nucleosome depletion in the vicinity of Rap1-binding sites. (a) Overlap between the 308 most nucleosome-depleted promoters and promoters found to recruit Rap1 in a global ChIP study [1]. (b) Nucleosome depletion in the vicinity of Rap1-binding sites in ribosomal gene promoters evaluated by ChIP. Fold-enrichment was determined by real-time PCR using primers that span Rap1-binding motifs in the RPS22A, RPS15, RPS11B and RPL23A promoters. (c) Southern blots showing DNA from yeast spheroplasts digested with increasing concentrations of micrococcal nuclease probed with labeled PCR products spanning the TUB2 promoter and the Rap1 sites in the RPS11B and RPS15 promoters. (d) Nucleosome occupancy for a mutant RPS11B promoter lacking Rap1 consensus sites was determined by H3 ChIP and real-time PCR. The mutant promoter is enriched 2.1-fold relative to wild type. (e) Nucleosome occupancy at Rap1-binding sites in ribosomal protein gene promoters after treatment with rapamycin evaluated by H3 ChIP and real-time PCR. transcription factor pairs (representing a total of 33 factors) Table 1). Of the 31 factors we found to associate with nucleo- on the basis of comprehensive binding and expression data some-depleted promoters, 17 were found to be functionally -5 [22]. Only a fraction of these are known to interact physically, cooperative by Banerjee and Zhang (p < 10 ). Furthermore, suggesting that other mechanisms also confer cooperative an evaluation of nucleosome occupancy at promoters bound function. There is a remarkable correspondence between by both members of a cooperative pair revealed a significant these functionally cooperative factors and those that prefer- association with nucleosome-depletion for 18 of the 31 pairs entially associate with nucleosome-depleted promoters (see (hypergeometric p < 0.01). Together, these findings suggest Genome Biology 2004, 5:R62 Fold-enrichment (relative to TUB2) comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.7 Table 1 Transcription factors that tend to associate with nucleosome-depleted promoters Transcription factor Pathway Number of targets Nucleosome-depleted Functionally cooperative Rap1 Biosynthesis 291 35% Fhl1 Biosynthesis 137 48% Swi4 Cell cycle 165 36% Hsf1 Environmental response 114 35% Gat3 Metabolism 119 31% Cin5 Environmental response 200 23% Phd1 Metabolism 138 25% Dal81 Metabolism 70 34% Ndd1 Cell cycle 122 26% Yap6 Environmental response 123 26% Fkh2 Cell cycle 145 24% Pdr1 Environmental response 103 27% Ino4 Metabolism 118 25% Smp1 Environmental response 99 27% Yap5 Environmental response 113 26% Ash1 Development 41 41% Transcription factors are ranked according to the significance of their association with nucleosome-depleted promoters, as determined by a hypergeometric model. Shown are the 16 top-ranked factors along with relevant physiologic pathway, number of promoters bound [1], and percent of target promoters that are nucleosome-depleted. Factors found previously to be functionally cooperative are indicated [22]. that binding motifs and transcription factors act in combina- marked nucleosome-depletion attributed to this region by tion to deplete nucleosomes and suggest a role for nucleo- global ChIP and real-time PCR analysis. The region sur- somes in transcription factor cooperativity [23-25]. rounding the RAP1 sites in RPS11B exhibits weak nuclease protection, consistent with the modest nucleosome-depletion attributed to this region by global ChIP and real-time PCR. Conditional nucleosome depletion at Rap1 consensus motifs Although these focused analyses specifically addressed Rap1 Although a number of transcription factors appear to act in sites in ribosomal protein genes, our global analyses indicate defining promoter nucleosome occupancy, only the Rap1 con- that approximately 30% of nucleosome-depleted promoters sensus motif was identified in an unbiased search of nucleo- containing Rap1 motifs do not regulate ribosomal protein some-depleted promoters. Furthermore, there is a highly genes. Together these data confirm that nucleosomes are significant association between nucleosome-depleted pro- markedly depleted in the vicinity of Rap1 consensus sites in moters and promoters bound by this factor in vivo [1] (Figure vivo, and thus extend previous studies showing that Rap1 4a). To investigate the relationship between Rap1 recruit- induces local alterations in chromatin structure that, for ment and nucleosome depletion further, we used ChIP and example, result in increased nuclease sensitivity [27-29]. real-time PCR to evaluate nucleosome occupancy at several Rap1 binding sites in ribosomal protein promoters. We found To gain further insight into the relationship between Rap1 that these regions are depleted 3- to 10-fold in H3 and FLAG- and nucleosome depletion, we examined a mutant RPS11B H2B ChIP assays, relative to a control promoter (TUB2) with promoter lacking its Rap1 consensus sites. We found that average occupancy by global analysis (Figure 4b). We also removal of these sites, which completely abrogates Rap1 used an orthogonal approach in which micrococcal nuclease binding [30], causes nucleosomes to return to the region, as digestion [26] was used to probe for nucleosomes at the reflected by a greater than twofold change in H3 ChIP enrich- TUB2, RPS11B and RPS15 promoters (Figure 4c). A pattern of ment (Figure 4d). We also examined the effect of rapamycin nuclease protection indicative of a regular nucleosome array treatment on nucleosome occupancy in the vicinity of these is evident at the TUB2 promoter, consistent with the average consensus sites. Although ribosomal protein gene expression nucleosome occupancy attributed to this promoter by global is dramatically reduced by rapamycin [31,32], Rap1 remains ChIP analysis. In contrast, nuclease protection is not evident bound to its target promoters ([30,33], and B.B., E.P. and at the RAP1 sites in the RPS15 promoter, consistent with the S.S., unpublished results). We found that rapamycin treat- Genome Biology 2004, 5:R62 R62.8 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 ment causes nucleosomes to return to the vicinity of Rap1 depletion in the vicinity of Rap1 sites in the promoters of sites, as reflected by twofold and greater increases in H3 ChIP ribosomal protein genes. Moreover, nucleosomes appeared to enrichment (Figure 4e). Together these data show that Rap1 return when the Rap1 consensus sites in one of these promot- consensus sites are required for conditional nucleosome ers were removed. These findings are consistent with previ- depletion at ribosomal protein gene promoters. ously described roles for Rap1 in opening chromatin and altering nucleosome positioning [27,28]. However, Rap1 recruitment is not equally associated with nucleosome deple- tion under all conditions. We find that nucleosomes partially Discussion To gain further insight into the role of nucleosomes in gene return to the vicinity of Rap1 sites during a rapamycin- regulation, we systematically evaluated promoter nucleo- induced starvation response [34], even though Rap1 remains some occupancy in yeast by immunoprecipitating nucleo- bound ([30,33], and B.B, E.P. and S.S., unpublished results). somal DNA and quantifying enrichment with microarrays. Hence, the nucleosome loss associated with Rap1 recruitment Promoters that are inefficiently immunoprecipitated by gen- is most likely to require additional proteins, such as Esa1, a eral anti-histone antibodies, and are therefore presumed to histone acetyltransferase recruited by Rap1 under exponen- be relatively nucleosome-depleted, tend to regulate active tial growth conditions but released in stress [30]. genes (Figure 2). This is consistent with the previous observa- tion that the activated PHO5 promoter is largely devoid of These findings may also offer insight into the barrier activity nucleosomes [5,6]. However, as not all nucleosome-depleted previously documented for Rap1 [35]. Heterochromatin promoters regulate active genes, there are most likely to be propagation involves the sequential modification of histones additional determinants of depletion. An unbiased search for in adjacent nucleosomes through positive-feedback mecha- sequence elements enriched in nucleosome-depleted promot- nisms [7,11]. Certain factors such as Rap1 are able to block ers revealed poly(dA.dT) elements, previously shown to this propagation by largely unknown mechanisms [36]. One destabilize nucleosome formation [17,18], and the Rap1 con- model speculates that these barriers create nucleosome-free sensus motif. By incorporating sequence conservation data 'holes' lacking the histone substrate required for heterochro- [21], more than 30 other enriched motifs could be identified. matin propagation [29,35]. By identifying such a 'hole' in the However, most of these appear to be relevant mainly when vicinity of Rap1-binding sites in vivo our data support this occurring in combination. When we limited this analysis to model. Remarkably, the nucleosomal hole and the barrier promoters containing four or fewer motifs, all but two of function ascribed to Rap1 may be conditional, as nucleosomes these additional motifs drop out (only the Rap1 and Swi4 con- return following treatment with the small molecule rapamy- sensus sites remain). As the majority of conserved motifs cin, which activates a starvation response. Heterochromatic incorporated in this analysis recruit transcription factors silencing has been shown previously to moderate under these [21], these data suggest that multiple transcription factors act conditions [37]. Hence, we speculate that dynamic influences in combination to deplete nucleosomes. This possibility is on nucleosome occupancy may enable Rap1 to define further supported by our finding that functionally coopera- chromatin domains and vary them in response to environ- tive transcription factors tend to bind nucleosome-depleted mental cues. promoters. These associations may reflect a mechanistic model in which transcription factors compete collaboratively More broadly, the widespread nucleosome loss observed in to displace nucleosomes in order to gain access to target sites the promoters of active genes provides a general caveat for in the DNA [23]. This model was formulated to explain why ChIP studies examining posttranslational histone modifica- certain pairs of transcription factors bind cooperatively to tions, as a decrease in signal for a histone modification at a proximal target sites in vivo and on a chromatin template, but promoter undergoing activation may actually reflect nucleo- not to naked DNA [23-25]. This view invokes a broad role for some loss. Similarly, regions that appear relatively hypo- nucleosomes as ubiquitous negative regulators of transcrip- modified by ChIP may actually be nucleosome-depleted. tion factor binding and function. We speculate that by pro- However, this is not the case for low levels of acetylation [38] moting synergy among multiple transcription factors and and H3 lysine 4 methylation [39] observed at yeast telomeres, impeding the activities of individual ones, nucleosomes facil- as these regions have high occupancy. The data also provide itate threshold behavior and filter noise (for example, genetic insight into the maintenance of epigenetic information by his- variation in motif sequence) in the transcriptional regulatory tone modifications. Whereas epigenetic memory of a network. repressed state can be maintained on histones in promoters, memory of an activated state must be maintained on histones Although many factors appear to act in defining promoter outside the promoters, for example in transcribed regions, nucleosome occupancy, our data indicate that Rap1 has a which may not undergo significant nucleosome loss during uniquely important role. Rap1 and its consensus motif are activation [5,6]. Methylation of histone H3 at lysines 4 and both markedly enriched in nucleosome-depleted promoters. 36, targeted to transcribed regions in yeast via interactions Follow-up studies using real-time PCR and micrococcal between RNA polymerase and the methylases [39-47], may nuclease digestion also demonstrate marked nucleosome represent such 'activating' marks. Genome Biology 2004, 5:R62 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.9 composite Cy5:Cy3 ratios determined according to protocols Materials and methods at the Stanford Microarray Database [56]. Correlations Chromatin immunoprecipitation (ChIP) DNA associated with histone H3 in vivo was immunoprecipi- between replicate datasets were ~0.8 for all experiments. tated with antibodies against the invariant H3 carboxy termi- Composite datasets were log transformed and zero centered nus using a ChIP protocol described previously [39,48,49]. before further analysis. The histone H3 ChIP dataset was Briefly, 45 ml log-phase w303a yeast (OD ~ 1.0) growing in determined from four independent immunoprecipitations yeast extract/peptone/dextrose (YPD) were cross-linked in and hybridizations (two each using antibodies from Cell Sig- 1% formaldehyde for 15 min, washed twice in PBS, resus- naling or Abcam). The FLAG-H2B ChIP dataset was deter- pended in 400 µl lysis buffer (50 mM Hepes-KOH pH 7.5, 140 mined from three independent immunoprecipitations and mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deox- hybridizations. The mRNA dataset was determined from ycholate) and lysed with glass beads. The resulting extract three independent extractions and hybridizations of mRNA was sonicated to fragment chromatin (4 × 20 sec burst/30 sec against genomic DNA. Relative transcription rates were rest with a Branson Sonifier 250 at 70% duty, power 3) and determined by dividing transcript levels by half-life data col- centrifuged for 15 min. Solubilized chromatin was then lected by Wang and colleagues [16]. A set of activated promot- immunoprecipitated with polyclonal antibodies against the ers was defined as those in the top 10% by mRNA expression carboxy terminus of histone H3 (Abcam or Cell Signaling). A level of associated gene, with divergent promoters assigned to unenriched whole-cell extract sample (WCE) was also the more highly expressed gene. Complete datasets are avail- retained as a control. After enrichment, cross-links were able online [57]. reversed by incubating samples in 10 mM Tris-HCl pH 8.0, 1 mM EDTA, 1.0% SDS, 150 mM NaCl at 65°C overnight. DNA Analysis of nucleosome-depleted promoters was purified from ChIP and WCE samples by proteinase K Z-scores were assigned to each intergenic that reflect deple- treatment, phenol/chloroform extraction, ethanol precipita- tion across the four H3 and three H2B ChIP experiments, tion, and incubation with RNAse. DNA associated with his- using the formula Z = (x - µ)/σ where x is the average of the tone H2B in vivo was isolated in a similar manner from yeast replicate measurements, µ is the average of all intergenics containing epitope-tagged H2B [50] using anti-FLAG M2 and σ is the standard error of the replicate measurements. We monoclonal antibodies (Sigma). defined as nucleosome-depleted the 410 features with the highest Z-scores. This set, which includes 308 promoters, DNA amplification and hybridization contains nucleosome-depleted outliers and is not inclusive of To obtain sufficient quantities for hybridization, immunopre- all promoters that immunoprecipitate with average or lower cipitated DNA (from approximately 10 cells) and whole-cell efficiency. The average aqueous enrichment ratio [15] for extract DNA (unenriched control) were amplified in a linear these 308 depleted promoters is 1.7-fold, significantly higher fashion as described [51]. Briefly, terminal transferase was than expected by chance (permutation p < 0.001), consistent used to add poly(T) tails to DNA fragment and a T7-poly(A) with the premise that these promoters are relatively free of adaptor primer was used to incorporate T7 promoters. The nucleosomes. reaction products were used as template for an in vitro tran- scription reaction carried out with the T7 Megascript Kit Sequence elements common to nucleosome-depleted pro- (Ambion) and RNA samples were purified using an RNeasy moters were identified by searching between 10 and 500 bp Mini Kit (Qiagen). Amplified RNA was reverse-transcribed, upstream of gene start sites for over-represented sequences incorporating amino-allyl dUTP, and the resulting DNA was up to 10 bp in length using the GeneSpring program suite (Sil- fluorescently labeled by incubation with monofunctional icon Genetics). Enrichment was confirmed by evaluating the reactive Cy5 (enriched sample) or Cy3 (unenriched control) significance of overlap between the set of nucleosome- dye as described [52]. Microarrays containing 6,438 PCR- depleted promoters and the set of promoters containing Rap1 amplified intergenic regions were prepared as described pre- consensus motifs (ACACCCATACAT with up to two mis- viously [39,53,54]. Mixed Cy5-/Cy3-labeled probe was matches) or poly dA.dT stretches at least 10 bp in length hybridized to intergenic microarrays for 12-14 h at 60°C, (identified using PatMatch, Saccharomyces Genome Data- washed and then scanned using a GenePix 4000A scanner base [58]). Statistical significances of overlaps between sets with GenePix Pro software (Axon Instruments) as described are expressed as P-values calculated by a hypergeometric [55]. In addition, transcript levels were determined by probability model. The P-values reflect the extent to which hybridizing Cy5-labeled mRNA extracted from log phase observed overlaps exceed that expected under the null w303a yeast against Cy3-labeled genomic DNA on microar- hypothesis that there is no relationship between the sets [59]. rays containing 6,218 open reading frames (ORFs), as Where specified, permutation analyses were carried out by described previously [16]. generating 1,000 random but representative promoter sets with an Excel macro and used to confirm statistical signifi- cance. Lists of promoters containing the 71 conserved motifs Microarray data processing Cy5 and Cy3 fluorescence were integrated for each feature [21] were collated from gene sets available online [60]. Lists using GenePix Pro Software (Axon). Data were processed and Genome Biology 2004, 5:R62 R62.10 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 of promoters bound by transcription factors at a significance extracted with phenol twice and chloroform once and precip- of p < 0.001 [1] were collated from data available at [61]. itated in ethanol. Samples were washed, resuspended in 10 mM Tris pH 7.5, subjected to RNAse treatment, cleaned up Real-time PCR with the MinElute kit (Qiagen) and run out in a 1% agarose Regions approximately 200 bp in size that span one or more gel. Following depurination, denaturation and neutralization Rap1 consensus sites in ribosomal protein gene promoters of the gel, DNA was transferred onto nylon membranes by were amplified from ChIP and unenriched control samples capillary action and covalently linked to the membranes by using SYBR green PCR mix (Qiagen) in an MJ Research real- UV irradiation. Southern blotting was carried out using a DIG time PCR machine according to the manufacturers' instruc- Luminescent Detection Kit (Roche) and DIG-labeled probe tions. Fold-ratios that reflect relative enrichment or depletion generated by PCR using the TUB2, RPS11B and RPS15 prim- of a given region in the H3 or FLAG-H2B ChIP assays were ers described above. -∆∆C determined using the 2 method described in the Applied Biosystems User Bulletin. For each region examined, the TUB2 promoter was used as the normalizer (this promoter is Acknowledgements We thank Mary Ann Osley and Kevin Struhl for generously providing yeast used as a control because its occupancy approximates that of strains, and Jay Bradner, Jeff McMahon, Aly Shamji, Jianping Cui, Manolis the average promoter by global analysis), and the unenriched Kellis, Vamsi Mootha, Mike Kamal and Oliver Rando for insightful discus- control sample was used as the calibrator. Each reported ratio sions. This study was supported by a grant from NIGMS (GM38627, awarded to S.L.S.). B.E.B. is supported by a K08 Development Award from represents the average of three independent ChIP experi- the National Cancer Institute. C.L.L. is supported by a Graduate Research ments analyzed in duplicate by real-time PCR. The following Fellowship from the National Science Foundation. S.L.S. is an Investigator primer pairs were used: at the Howard Hughes Medical Institute. RPS22A promoter: 5'-GCCTAAAACGCCCATAAGTT-3' and References 5'-ACTGCAAACCCATATTCAAGA-3' 1. Lee TI, Rinaldi NJ, Robert F, Odom DT, Bar-Joseph Z, Gerber GK, Hannett NM, Harbison CT, Thompson CM, Simon I, et al.: Tran- RPS15 promoter: 5'-TACACCGCGCGTATAAATCA-3' and 5'- scriptional regulatory networks in Saccharomyces cerevisiae. Science 2002, 298:799-804. CCCAGCAAGGAGTTTCTCAG-3' 2. Kornberg RD, Lorch Y: Twenty-five years of the nucleosome, fundamental particle of the eukaryote chromosome. Cell RPS11B promoter: 5'-GAAGAAATATTTCCTTGCTGCACC-3' 1999, 98:285-294. 3. Han M, Grunstein M: Nucleosome loss activates yeast down- and 5'-AAGGGAAACGTAAAGCTATTGGAC-3' stream promoters in vivo. Cell 1988, 55:1137-1145. 4. Narlikar GJ, Fan HY, Kingston RE: Cooperation between RPL23A promoter: 5'-ATTAACATCTGTACACCCCCAACT-3' complexes that regulate chromatin structure and transcription. Cell 2002, 108:475-487. and 5'-TACAGTTCGTTTCCTGCC ATATTA-3' 5. Boeger H, Griesenbeck J, Strattan JS, Kornberg RD: Nucleosomes unfold completely at a transcriptionally active promoter. Mol Cell 2003, 11:1587-1598. TUB2 promoter: 5'-GGCCTAACAGTAAAGATATCCTCC-3' 6. Reinke H, Horz W: Histones are first hyperacetylated and then and 5'-GTTGTAGTAGCTGCTATGT CACTC-3' lose contact with the activated PHO5 promoter. Mol Cell 2003, 11:1599-1607. 7. Schreiber SL, Bernstein BE: Signaling network model of Centromeric vectors containing either a mutant RPS11B pro- chromatin. Cell 2002, 111:771-778. moter lacking the two Rap1 consensus motifs [30] or an 8. Hassan AH, Neely KE, Vignali M, Reese JC, Workman JL: Promoter targeting of chromatin-modifying complexes. Front Biosci 2001, essentially wild-type allele were transformed into wild-type 6:D1054-D1064. yeast and used in an H3 ChIP assay to evaluate the conse- 9. Kouzarides T: Histone methylation in transcriptional control. quence of removing Rap1 binding sites on nucleosome occu- Curr Opin Genet Dev 2002, 12:198-209. 10. Kurdistani SK, Grunstein M: Histone acetylation and deacetyla- pancy. Enrichment was evaluated by real-time PCR using the tion in yeast. Nat Rev Mol Cell Biol 2003, 4:276-284. following primer pair that selectively amplifies the plasmid 11. Grewal SI, Elgin SC: Heterochromatin: new possibilities for the alleles but not the endogenous RPS11B promoter: 5'- inheritance of structure. Curr Opin Genet Dev 2002, 12:178-187. 12. Turner BM: Cellular memory and the histone code. Cell 2002, CTGGAAGAAATATTTCCTT GCTCTAG-3' and 5'-AAG- 111:285-291. GGAAACGTAAAGCTATTGGAC-3'. 13. Gregory PD, Horz W: Chromatin and transcription - how tran- scription factors battle with a repressive chromatin environment. Eur J Biochem 1998, 251:9-18. Micrococcal nuclease assay 14. Morse RH: Getting into chromatin: how do transcription fac- Log-phase cultures of W303a yeast grown in 450 ml YPD to tors get past the histones? Biochem Cell Biol 2003, 81:101-112. 15. Nagy PL, Cleary ML, Brown PO, Lieb JD: Genomewide demarca- OD of 1.0 were spheroplasted with zymolase (10 mg in 40 tion of RNA polymerase II transcription units revealed by ml volume of 1 M sorbitol, 50 mM Tris pH 7.4, 10 mM β-mer- physical fractionation of chromatin. Proc Natl Acad Sci USA 2003, 100:6364-6369. captoethanol (β-ME), at 30°C for 38 min shaking at 300 16. Wang Y, Liu CL, Storey JD, Tibshirani RJ, Herschlag D, Brown PO: rpm), divided into five aliquots, and digested with increasing Precision and functional specificity in mRNA decay. Proc Natl concentrations (20 U to 320 U) of micrococcal nuclease Acad Sci USA 2002, 99:5860-5865. 17. Iyer V, Struhl K: Poly(dA:dT), a ubiquitous promoter element (Worthington Biochem) in 600 µl 0.5 mM spermidine, 1 mM that stimulates transcription via its intrinsic DNA structure. β-ME, 0.075% NP-40. DNA from digested samples was EMBO J 1995, 14:2570-2579. Genome Biology 2004, 5:R62 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.11 18. Anderson JD, Widom J: Poly(dA-dT) promoter elements 42. Krogan NJ, Kim M, Tong A, Golshani A, Cagney G, Canadien V, Rich- increase the equilibrium accessibility of nucleosomal DNA ards DP, Beattie BK, Emili A, Boone C, et al.: Methylation of his- target sites. Mol Cell Biol 2001, 21:3830-3839. tone H3 by Set2 in Saccharomyces cerevisiae is linked to 19. Graham IR, Chambers A: Use of a selection technique to iden- transcriptional elongation by RNA polymerase II. Mol Cell Biol tify the diversity of binding sites for the yeast RAP1 tran- 2003, 23:4207-4218. scription factor. Nucleic Acids Res 1994, 22:124-130. 43. Schaft D, Roguev A, Kotovic KM, Shevchenko A, Sarov M, Neuge- 20. Buchman AR, Kimmerly WJ, Rine J, Kornberg RD: Two DNA-bind- bauer KM, Stewart AF: The histone 3 lysine 36 methyltrans- ing factors recognize specific sequences at silencers, ferase, SET2, is involved in transcriptional elongation. Nucleic upstream activating sequences, autonomously replicating Acids Res 2003, 31:2475-2482. sequences, and telomeres in Saccharomyces cerevisiae. Mol Cell 44. Xiao T, Hall H, Kizer KO, Shibata Y, Hall MC, Borchers CH, Strahl Biol 1988, 8:210-225. BD: Phosphorylation of RNA polymerase II CTD regulates 21. Kellis M, Patterson N, Endrizzi M, Birren B, Lander ES: Sequencing H3 methylation in yeast. Genes Dev 2003, 17:654-663. and comparison of yeast species to identify genes and regu- 45. Li B, Howe L, Anderson S, Yates JR 3rd, Workman JL: The Set2 his- latory elements. Nature 2003, 423:241-254. tone methyltransferase functions through the phosphor- 22. Banerjee N, Zhang MQ: Identifying cooperativity among tran- ylated carboxyl-terminal domain of RNA polymerase II. J Biol scription factors controlling the cell cycle in yeast. Nucleic Chem 2003, 278:8897-8903. Acids Res 2003, 31:7024-7031. 46. Gerber M, Shilatifard A: Transcriptional elongation by RNA 23. Miller JA, Widom J: Collaborative competition mechanism for polymerase II and histone methylation. J Biol Chem 2003, gene activation in vivo. Mol Cell Biol 2003, 23:1623-1632. 278:26303-26306. 24. Vashee S, Melcher K, Ding WV, Johnston SA, Kodadek T: Evidence 47. Hampsey M, Reinberg D: Tails of intrigue: phosphorylation of for two modes of cooperative DNA binding in vivo that do RNA polymerase II mediates histone methylation. Cell 2003, not involve direct protein-protein interactions. Curr Biol 1998, 113:429-432. 8:452-458. 48. Kuo MH, Allis CD: In vivo cross-linking and 25. Adams CC, Workman JL: Binding of disparate transcriptional immunoprecipitation for studying dynamic Protein:DNA activators to nucleosomal DNA is inherently cooperative. associations in a chromatin environment. Methods 1999, Mol Cell Biol 1995, 15:1405-1421. 19:425-433. 26. Gregory PD, Horz W: Mapping chromatin structure in yeast. 49. Bernstein BE, Humphrey EL, Liu CL, Schreiber SL: The use of chro- Methods Enzymol 1999, 304:365-76. matin immunoprecipitation assays in genome-wide analyses 27. Yu L, Morse RH: Chromatin opening and transactivator poten- of histone modifications. Methods Enzymol 2004, 376:349-360. tiation by RAP1 in Saccharomyces cerevisiae. Mol Cell Biol 1999, 50. Robzyk K, Recht J, Osley MA: Rad6-dependent ubiquitination of 19:5279-88. histone H2B in yeast. Science 2000, 287:501-504. 28. De Sanctis V, La Terra S, Bianchi A, Shore D, Burderi L, Di Mauro E, 51. Liu CL, Schreiber SL, Bernstein BE: Development and validation Negri R: In vivo topography of Rap1p-DNA complex at Sac- of a T7 based linear amplification for genomic DNA. BMC charomyces cerevisiae TEF2 UAS(RPG) during transcrip- Genomics 2003, 4:19. tional regulation. J Mol Biol 2002, 318:333-349. 52. Carroll AS, Bishop AC, DeRisi JL, Shokat KM, O'Shea EK: Chemical 29. Yu Q, Qiu R, Foland TB, Griesen D, Galloway CS, Chiu YH, Sand- inhibition of the Pho85 cyclin-dependent kinase reveals a meier J, Broach JR, Bi X: Rap1p and other transcriptional regu- role in the environmental stress response. Proc Natl Acad Sci lators can function in defining distinct domains of gene USA 2001, 98:12578-12583. expression. Nucleic Acids Res 2003, 31:1224-1233. 53. Iyer VR, Horak CE, Scafe CS, Botstein D, Snyder M, Brown PO: 30. Reid JL, Iyer VR, Brown PO, Struhl K: Coordinate regulation of Genomic binding sites of the yeast cell-cycle transcription yeast ribosomal protein genes is associated with targeted factors SBF and MBF. Nature 2001, 409:533-538. recruitment of Esa1 histone acetylase. Mol Cell 2000, 54. Ren B, Robert F, Wyrick JJ, Aparicio O, Jennings EG, Simon I, Zeitlin- 6:1297-1307. ger J, Schreiber J, Hannett N, Kanin E, et al.: Genome-wide location 31. Powers T, Walter P: Regulation of ribosome biogenesis by the and function of DNA binding proteins. Science 2000, rapamycin-sensitive TOR-signaling pathway in Saccharomy- 290:2306-2309. ces cerevisiae. Mol Biol Cell 1999, 10:987-1000. 55. Microarrays.org [http://www.microarrays.org] 32. Hardwick JS, Kuruvilla FG, Tong JK, Shamji AF, Schreiber SL: 56. Gollub J, Ball CA, Binkley G, Demeter J, Finkelstein DB, Hebert JM, Rapamycin-modulated transcription defines the subset of Hernandez-Boussard T, Jin H, Kaloper M, Matese JC, et al.: The nutrient-sensitive signaling pathways directly controlled by Stanford Microarray Database: data access and quality the Tor proteins. Proc Natl Acad Sci USA 1999, 96:14866-14870. assessment tools. Nucleic Acids Res 2003, 31:94-96. 33. Rohde JR, Cardenas ME: The tor pathway regulates gene 57. Stuart Schreiber group website [http://www.broad.broad.mit.edu/chem expression by linking nutrient sensing to histone acetylation. bio/lab_schreiber/pubs/protocols/global%20nucleosome/globalnuc.html] Mol Cell Biol 2003, 23:629-635. 58. Saccharomyces Genome Database [http://www.yeastge 34. Shamji AF, Kuruvilla FG, Schreiber SL: Partitioning the transcrip- nome.org] tional program induced by rapamycin among the effectors of 59. Tavazoie S, Hughes JD, Campbell MJ, Cho RJ, Church GM: System- the Tor proteins. Curr Biol 2000, 10:1574-1581. atic determination of genetic network architecture. Nat Genet 35. Bi X, Broach JR: UASrpg can function as a heterochromatin 1999, 22:281-285. boundary element in yeast. Genes Dev 1999, 13:1089-1101. 60. Yeast comparative genomics [http://www.broad.mit.edu/anno 36. Donze D, Kamakaka RT: Braking the silence: how heterochro- tation/fungi/comp_yeasts] matic gene repression is stopped in its tracks. BioEssays 2002, 61. Young lab [http://web.wi.mit.edu/young] 24:344-349. 62. Lieb JD, Liu X, Botstein D, Brown PO: Promoter-specific binding 37. Ai W, Bertram PG, Tsang CK, Chan TF, Zheng XF: Regulation of of Rap1 revealed by genome-wide maps of protein-DNA subtelomeric silencing during stress response. Mol Cell 2002, association. Nat Genet 2001, 28:327-334. 10:1295-1305. 38. Braunstein M, Rose AB, Holmes SG, Allis CD, Broach JR: Transcrip- tional silencing in yeast is associated with reduced nucleo- some acetylation. Genes Dev 1993, 7:592-604. 39. Bernstein BE, Humphrey EL, Erlich RL, Schneider R, Bouman P, Liu JS, Kouzarides T, Schreiber SL: Methylation of histone H3 Lys 4 in coding regions of active genes. Proc Natl Acad Sci USA 2002, 99:8695-8700. 40. Ng HH, Robert F, Young RA, Struhl K: Targeted recruitment of Set1 histone methylase by elongating Pol II provides a local- ized mark and memory of recent transcriptional activity. Mol Cell 2003, 11:709-719. 41. Li J, Moazed D, Gygi SP: Association of the histone methyltrans- ferase Set2 with RNA polymerase II plays a role in transcrip- tion elongation. J Biol Chem 2002, 277:49383-49388. Genome Biology 2004, 5:R62 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Genome Biology Springer Journals

Global nucleosome occupancy in yeast

Download PDF
11 pages

Loading next page...
 
/lp/springer-journals/global-nucleosome-occupancy-in-yeast-YnFTuKJr71

References (70)

Publisher
Springer Journals
Copyright
2004 Bernstein et al.; licensee BioMed Central Ltd.
eISSN
1474-760X
DOI
10.1186/gb-2004-5-9-r62
Publisher site
See Article on Publisher Site

Abstract

Background: Although eukaryotic genomes are generally thought to be entirely chromatin- associated, the activated PHO5 promoter in yeast is largely devoid of nucleosomes. We systematically evaluated nucleosome occupancy in yeast promoters by immunoprecipitating nucleosomal DNA and quantifying enrichment by microarrays. Results: Nucleosome depletion is observed in promoters that regulate active genes and/or contain multiple evolutionarily conserved motifs that recruit transcription factors. The Rap1 consensus was the only binding motif identified in a completely unbiased search of nucleosome-depleted promoters. Nucleosome depletion in the vicinity of Rap1 consensus sites in ribosomal protein gene promoters was also observed by real-time PCR and micrococcal nuclease digestion. Nucleosome occupancy in these regions was increased by the small molecule rapamycin or, in the case of the RPS11B promoter, by removing the Rap1 consensus sites. Conclusions: The presence of transcription factor-binding motifs is an important determinant of nucleosome depletion. Most motifs are associated with marked depletion only when they appear in combination, consistent with a model in which transcription factors act collaboratively to exclude nucleosomes and gain access to target sites in the DNA. In contrast, Rap1-binding sites cause marked depletion under steady-state conditions. We speculate that nucleosome depletion enables Rap1 to define chromatin domains and alter them in response to environmental cues. the distribution of nucleosomes, the fundamental units of Background Global gene-expression patterns are established and main- chromatin, is poorly understood on a gene-specific basis, tained by the concerted actions of transcription factors and much less a global basis [2]. the proteins that constitute chromatin. The global network of interactions between transcription factors and promoters in The nucleosome consists of approximately 146 base-pairs yeast is increasingly being characterized [1]. The role of chro- (bp) of DNA wrapped around an octamer of histone proteins matin in gene regulation is less clear, however. For example, - two each of histones H2A, H2B, H3 and H4. Eukaryotic Genome Biology 2004, 5:R62 R62.2 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 genomes are packaged into repeating units of nucleosomes are remarkably similar as shown by a genome-wide correla- separated by around 10-80 bp of linker DNA. High occupancy tion of 0.83 (Figure 1a-c). This correlation is comparable to by nucleosomes is thought to be generally repressive [3], and that observed when comparing replicate H3 datasets (or H2B extensive remodeling (and loss) of nucleosomes occurs in the datasets), and suggests that both assays measure similar phe- promoters of genes undergoing activation [4]. In the case of nomena. In the H3 and H2B datasets, respectively, there are the PHO5 promoter in yeast, this remodeling proceeds until 347 and 214 regions depleted at least 1.5-fold relative to the essentially no nucleosomes are detected across a region of average over all intergenics. In contrast, there are just 84 and several hundred base-pairs [5,6]. 6 regions in the respective datasets enriched at least 1.5-fold relative to this average. The relatively narrow range of ChIP Transcription factors and chromatin proteins each form com- enrichment and the negative skew of the data (Figure 1b) are plex regulatory networks that interact in a variety of ways consistent with the conventional view that the majority of the [1,7]. Transcription factors modify chromatin structure by genome is packaged into nucleosomes with intervening recruiting enzymes that remodel nucleosomes or posttransla- stretches of free DNA such as the activated PHO5 promoter tionally modify histones (by acetylation or methylation, for [5,6]. example) [8-10]. The modifications can be maintained through cell division and propagated to proximal nucleo- Despite these consistencies, a possible caveat to using ChIP to somes by positive-feedback mechanisms [7,11,12]. Hence, a evaluate nucleosome occupancy is that immunoprecipitation signal such as the activation of a transcription factor can be efficiency can depend on epitope accessibility. Rather than temporally and spatially transmitted through chromatin. having low occupancies, genomic regions depleted in the H3 Conversely, chromatin can influence transcription factor ChIP might be inaccessible as a result of association with function by modulating the accessibility of target binding large protein complexes in chromatin. To investigate this pos- sites in the DNA [13,14]. sibility, we examined a published chromatin fractionation dataset in which cross-linked chromatin fragments were sub- We used chromatin immunoprecipitation (ChIP) and DNA jected to phenol-chloroform extraction and DNA that parti- microarrays to evaluate nucleosome occupancy levels for tioned into the aqueous phase was quantified by microarrays essentially all promoters in yeast. Promoters that regulate [15]. Given the polar nature of DNA and the hydrophobic active genes, contain multiple conserved motifs or recruit nature of denatured protein, aqueous extraction should gen- Rap1 tend to be relatively nucleosome-depleted. We also used erally enrich for free DNA. We found that regions depleted in real-time PCR and micrococcal nuclease digestion to show the H3 ChIP assay overlap extensively with regions enriched that nucleosomes are depleted in the vicinity of Rap1 consen- by aqueous extraction, but not with regions depleted by aque- sus sites. This depletion can be partially reversed by the ous extraction (Figure 1d). Overall, there is a negative corre- actions of the small molecule rapamycin or by removing lation of -0.54 between the H3 ChIP and aqueous-extraction Rap1-binding sites. We suggest that other transcription fac- datasets. Although the fractionation data may partially reflect tors have less robust nucleosome-depleting activities than differential cross-linking of lysines in the histone tails [15], Rap1 and must therefore act collaboratively to gain access to this analysis suggests that regions depleted in the H3 ChIP their cognate sites in the DNA. experiment are relatively protein-free, as would be expected of non-nucleosomal DNA. Nucleosome occupancy correlates inversely with Results ChIP-based assay for nucleosome occupancy promoter strength Histones are essential components of the nucleosome and As previous studies show that PHO5 activation is accompa- efficiently cross-link to nucleosomal DNA. Antibodies against nied by marked nucleosome loss in the promoter region [5,6], invariant portions of histones have been used previously in we sought to determine whether nucleosome depletion is a ChIP assays to follow nucleosome loss at the yeast PHO5 pro- general attribute of active promoters. A total of 4,365 inter- moter [5,6]. We extended this approach to evaluate relative genic regions that reside immediately upstream of one or nucleosome occupancy at essentially all promoters and other more validated yeast genes were assigned as promoters. Rel- intergenic regions in yeast. DNA associated in vivo with his- ative transcription rates were determined for each yeast gene tone H3 was isolated by ChIP using antibody against the car- from transcript levels measured by array and previously col- boxy terminus of histone H3 (no posttranslational lected mRNA half-life data [16]. We found an inverse correla- modifications are thought to occur in this region). ChIP DNA tion of -0.39 between the enrichment of promoters in the H3 and unenriched control DNA were amplified by in vitro tran- and H2B ChIP assays and the transcription rates of down- scription and evaluated using microarrays. DNA associated stream genes (Figure 2a). Under the conditions examined, with histone H2B was evaluated in a similar fashion using PHO5 is not induced and its promoter has an average nucleo- anti-FLAG antibody and a FLAG-H2B strain. H3 and H2B some occupancy according to these datasets. To evaluate fur- datasets were compiled by averaging four and three inde- ther the relationship between nucleosome depletion and pendent biological experiments, respectively. These datasets transcription, we collated a set of 308 nucleosome-depleted Genome Biology 2004, 5:R62 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.3 (a) (c) R = 0.83 n = 48 H2B-depleted H3-depleted (> 1.5-fold) (> 1.5-fold) (d) −1 n = 2 Aqueous- −2 198 149 161 depleted (bottom 5%) −2 −10 1 Enrichment ratio H3 ChIP (log ) Aqueous- H3-depleted enriched (> 1.5-fold) (top 5%) (b) H3 ChIP H2B ChIP Enrichment ratio (log ) C Fio grrelatio ure 1 n between H3 and FLAG-H2B ChIP datasets Correlation between H3 and FLAG-H2B ChIP datasets. DNA associated with histones in vivo was enriched in ChIP assays using antibodies against histone H3 or FLAG-H2B, and quantified by microarrays. (a) Relative enrichment of promoters and other non-coding regions in the H3 and H2B ChIP assays is shown. (b) Histogram showing distributions of enrichment for promoter regions in the H3 and H2B ChIP assays. (c) Overlap between regions depleted in the H3 and FLAG-H2B assays is shown. Overall, there is an 0.83 correlation between these ChIP datasets. (d) Overlap between regions depleted in the H3 ChIP assay and regions enriched by aqueous extraction is shown [62]. promoters on the basis of their relative depletion across the between promoter strength and nucleosome depletion. How- replicate H3 and H2B experiments. Of these nucleosome- ever, as this correspondence is not complete there are likely to depleted promoters, 42% regulate highly active genes (Figure be other determinants of nucleosome occupancy. 2b). These data suggest that there is a systematic relationship Genome Biology 2004, 5:R62 −2 −1.8 −1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4 −0.2 0.2 0.4 0.6 0.8 1.2 Enrichment ratio H2B ChIP (log ) Frequency 2 R62.4 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 The second sequence element enriched in nucleosome- (a) depleted promoters corresponds to the consensus motif for 0.25 the Rap1 transcription factor. This motif commonly occurs in the promoters of ribosomal proteins genes and is required for Rap1 binding in vitro and in vivo [19,20]. Some variant of this motif appears in 22% of nucleosome-depleted promoters, H2B ChIP compared with just 8% of promoters overall (hypergeometric H3 ChIP -5 −0.25 p < 10 ). Furthermore, multiple Rap1 sites are found in 19% of nucleosome-depleted promoters with Rap1 sites, com- pared to 8% of promoters with Rap1 sites overall (hypergeo- −0.5 -3 metric p < 10 ). These data suggest that Rap1 recruitment may lead to nucleosome loss. −0.75 Because only the Rap1 consensus site was identified in an 0 102030405060708090 100 Percentile rank of transcription rate unbiased search, we sought to identify additional sequence motifs by incorporating species conservation data. Specifi- (b) cally, we evaluated a set of 71 conserved motifs identified by −5 p < 10 Kellis and colleagues, a majority of which function in tran- scription factor recruitment [21]. Nearly half of these 71 motifs are over-represented in nucleosome-depleted promot- ers relative to promoters overall, as defined by a hypergeo- 179 129 308 metric p < 0.001. However, many of the implicated motifs appear in the same promoters. For example, nine of the over- Nucleosome- Promoters 5′ of represented motifs are associated with filamentation gene depleted promoters active genes (top 10%) promoters [21]. We therefore considered the possibility that the total number of conserved motifs might be a more rele- Istren Fig nveu rs r ge e as th 2 sociation between nucleosome occupancy and promoter vant predictor of nucleosome depletion. Indeed, we found Inverse association between nucleosome occupancy and promoter that 31% of nucleosome-depleted promoters contain at least strength. (a) Relative enrichment of promoter regions in the H3 and eight motifs, compared with 11% of promoters overall (hyper- FLAG-H2B ChIP assays plotted against transcription rate of downstream -5 geometric p <10 ; Figure 3b). Furthermore, nucleosome- genes (moving average, window 50). (b) Overlap between promoters upstream of active genes and the set of nucleosome-depleted promoters depleted promoters contain an average of 6.1 motifs, whereas defined on the basis of depletion across the replicate H3 and FLAG-H2B the average promoter contains 3.1 (permutation p < 0.001; experiments. Figure 3c). Next, we sought motifs associated with nucleo- some depletion in the absence of multiple motifs, by confin- ing our analysis to promoters containing a maximum of four Transcription factor binding motifs are over- motifs. This analysis identified just two over-represented represented in nucleosome-depleted promoters motifs, which correspond to the Rap1 and Swi4 binding sites. To identify additional determinants of occupancy, we sought Hence, although a large number of conserved motifs are sequence elements associated with nucleosome depletion. enriched in nucleosome-depleted promoters, most appear to Specifically, we carried out an unbiased search for elements be relevant mainly when occurring in combination. up to 10 bp in length that occur with higher frequency in nucleosome-depleted promoters. Two distinct categories of Functionally cooperative transcription factors sequences emerged (Figure 3a). The first includes associate with nucleosome-depleted promoters poly(dA.dT) elements. Stretches of 10 or more dA.dT nucle- As a majority of the conserved motifs recruit transcription otides appear in 38% of depleted promoters, compared with factors [21], we examined the relationship between transcrip- -5 26% of promoters overall (hypergeometric p < 10 ). dA.dT tion factor binding and nucleosome occupancy more directly. stretches destabilize nucleosome formation in vitro and in Lee and colleagues combined ChIP and microarrays to iden- vivo [17,18]. The enrichment of poly(dA.dT) elements in tify target promoters for essentially all yeast transcription nucleosome-depleted promoters probably reflects, at least in factors under the same conditions used here to evaluate part, this destabilizing influence. As a high proportion of the nucleosome occupancy [1]. For each factor, we determined poly(dA.dT) elements identified in nucleosome-depleted pro- the significance of overlap between its target promoters and moters are more than 10 bp long (30% are at least 14 bp), the set of nucleosome-depleted promoters. Of the 113 tran- these data do not address the minimum length required for scription factors in their database, 31 tend to associate with destabilization. However, in vitro studies show that a 16-bp nucleosome-depleted promoters as defined by a hypergeo- insertion leads to a 1.7-fold increase in accessibility of nucle- metric p < 0.001. Rap1 has the most significant association osomal target sites [18]. (Figure 4a), consistent with the enrichment of its binding Genome Biology 2004, 5:R62 Enrichment ratio (log ) 2 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.5 (a) Sequence Hits in nucleosome- Hits expected P-value depleted promoters at random CACCCGTACA 38 4 5.1E-20 ACACCCGTAC 33 3 2.6E-15 CATCCGTACA 40 6 6.0E-14 TTCTTTTTTT 218 133 4.0E-12 TTTTTTTTCC 177 97 5.1E-12 CACCCATACA 43 8 2.3E-11 GTTTTTTTTC 159 84 2.7E-11 TTTTTTTCTC 170 93 3.0E-11 TTTTTTTCTG 150 77 5.1E-11 TTTTTTTTTT 211 131 3.8E-10 (b) p < 10 5 213 95 381 Nucleosome- Promoters with depleted promoters ≥ 8 conserved motifs (c) Nucleosome- depleted Conserved motifs per promoter (average) Seque Figure nce motifs o 3 ver-represented in nucleosome-depleted promoters Sequence motifs over-represented in nucleosome-depleted promoters. (a) An unbiased search for sequences up to 10 bp in length over-represented in nucleosome-depleted promoters (relative to promoters overall) identified the poly(dA.dT) sequence element and variants of the Rap1 consensus motif ACACCCATACAT [21]. (b) Overlap between nucleosome-depleted promoters and promoters that contain multiple conserved motifs is shown [21]. (c) Histogram showing average numbers of motifs in 1,000 randomly generated promoter sets. Nucleosome-depleted promoters contain an average of 6.1 conserved motifs, significantly higher than in these randomly generated sets. motif (see above). Other top-ranked factors include Fhl1, However, these factors utilize a variety of binding domains, which associates with many Rap1-bound promoters, and regulate different pathways, and only a minority have signifi- Swi4, whose binding motif is also enriched (Table 1). cant associations with promoters of highly active genes. Nonetheless, a commonality does emerge when transcription We sought an underlying binding mechanism or function factor cooperativity is considered. A recent informatics study common to the transcription factors we had identified. by Banerjee and Zhang identified 31 functionally cooperative Genome Biology 2004, 5:R62 2.5 3.5 4.5 5.5 6.5 Number of random sets R62.6 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 (a) −5 (d) p < 10 Rap1 motifs RPS11B 207 101 190 RPS11B∆ RAP1 Nucleosome- Promoters depleted promoters bound by Rap1 (b) H3 ChIP (mutant) Ratio (relative to wild-type) RPS22A RPS15 RPS11B RPL23A RPS11B∆ RAP1 +2.1-fold −2 (e) −4 H3 ChIP Ratio (relative to untreated) −6 (1 hr rapamycin) −8 RPS22A +3.0-fold RPS15 +4.7-fold −10 RPS11B +2.1-fold Histone H3 ChIP RPL23A +1.9-fold Histone H2B ChIP (c) TUB2 RPS11B RPS15 Nu Figcleo ureso 4me depletion in the vicinity of Rap1-binding sites Nucleosome depletion in the vicinity of Rap1-binding sites. (a) Overlap between the 308 most nucleosome-depleted promoters and promoters found to recruit Rap1 in a global ChIP study [1]. (b) Nucleosome depletion in the vicinity of Rap1-binding sites in ribosomal gene promoters evaluated by ChIP. Fold-enrichment was determined by real-time PCR using primers that span Rap1-binding motifs in the RPS22A, RPS15, RPS11B and RPL23A promoters. (c) Southern blots showing DNA from yeast spheroplasts digested with increasing concentrations of micrococcal nuclease probed with labeled PCR products spanning the TUB2 promoter and the Rap1 sites in the RPS11B and RPS15 promoters. (d) Nucleosome occupancy for a mutant RPS11B promoter lacking Rap1 consensus sites was determined by H3 ChIP and real-time PCR. The mutant promoter is enriched 2.1-fold relative to wild type. (e) Nucleosome occupancy at Rap1-binding sites in ribosomal protein gene promoters after treatment with rapamycin evaluated by H3 ChIP and real-time PCR. transcription factor pairs (representing a total of 33 factors) Table 1). Of the 31 factors we found to associate with nucleo- on the basis of comprehensive binding and expression data some-depleted promoters, 17 were found to be functionally -5 [22]. Only a fraction of these are known to interact physically, cooperative by Banerjee and Zhang (p < 10 ). Furthermore, suggesting that other mechanisms also confer cooperative an evaluation of nucleosome occupancy at promoters bound function. There is a remarkable correspondence between by both members of a cooperative pair revealed a significant these functionally cooperative factors and those that prefer- association with nucleosome-depletion for 18 of the 31 pairs entially associate with nucleosome-depleted promoters (see (hypergeometric p < 0.01). Together, these findings suggest Genome Biology 2004, 5:R62 Fold-enrichment (relative to TUB2) comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.7 Table 1 Transcription factors that tend to associate with nucleosome-depleted promoters Transcription factor Pathway Number of targets Nucleosome-depleted Functionally cooperative Rap1 Biosynthesis 291 35% Fhl1 Biosynthesis 137 48% Swi4 Cell cycle 165 36% Hsf1 Environmental response 114 35% Gat3 Metabolism 119 31% Cin5 Environmental response 200 23% Phd1 Metabolism 138 25% Dal81 Metabolism 70 34% Ndd1 Cell cycle 122 26% Yap6 Environmental response 123 26% Fkh2 Cell cycle 145 24% Pdr1 Environmental response 103 27% Ino4 Metabolism 118 25% Smp1 Environmental response 99 27% Yap5 Environmental response 113 26% Ash1 Development 41 41% Transcription factors are ranked according to the significance of their association with nucleosome-depleted promoters, as determined by a hypergeometric model. Shown are the 16 top-ranked factors along with relevant physiologic pathway, number of promoters bound [1], and percent of target promoters that are nucleosome-depleted. Factors found previously to be functionally cooperative are indicated [22]. that binding motifs and transcription factors act in combina- marked nucleosome-depletion attributed to this region by tion to deplete nucleosomes and suggest a role for nucleo- global ChIP and real-time PCR analysis. The region sur- somes in transcription factor cooperativity [23-25]. rounding the RAP1 sites in RPS11B exhibits weak nuclease protection, consistent with the modest nucleosome-depletion attributed to this region by global ChIP and real-time PCR. Conditional nucleosome depletion at Rap1 consensus motifs Although these focused analyses specifically addressed Rap1 Although a number of transcription factors appear to act in sites in ribosomal protein genes, our global analyses indicate defining promoter nucleosome occupancy, only the Rap1 con- that approximately 30% of nucleosome-depleted promoters sensus motif was identified in an unbiased search of nucleo- containing Rap1 motifs do not regulate ribosomal protein some-depleted promoters. Furthermore, there is a highly genes. Together these data confirm that nucleosomes are significant association between nucleosome-depleted pro- markedly depleted in the vicinity of Rap1 consensus sites in moters and promoters bound by this factor in vivo [1] (Figure vivo, and thus extend previous studies showing that Rap1 4a). To investigate the relationship between Rap1 recruit- induces local alterations in chromatin structure that, for ment and nucleosome depletion further, we used ChIP and example, result in increased nuclease sensitivity [27-29]. real-time PCR to evaluate nucleosome occupancy at several Rap1 binding sites in ribosomal protein promoters. We found To gain further insight into the relationship between Rap1 that these regions are depleted 3- to 10-fold in H3 and FLAG- and nucleosome depletion, we examined a mutant RPS11B H2B ChIP assays, relative to a control promoter (TUB2) with promoter lacking its Rap1 consensus sites. We found that average occupancy by global analysis (Figure 4b). We also removal of these sites, which completely abrogates Rap1 used an orthogonal approach in which micrococcal nuclease binding [30], causes nucleosomes to return to the region, as digestion [26] was used to probe for nucleosomes at the reflected by a greater than twofold change in H3 ChIP enrich- TUB2, RPS11B and RPS15 promoters (Figure 4c). A pattern of ment (Figure 4d). We also examined the effect of rapamycin nuclease protection indicative of a regular nucleosome array treatment on nucleosome occupancy in the vicinity of these is evident at the TUB2 promoter, consistent with the average consensus sites. Although ribosomal protein gene expression nucleosome occupancy attributed to this promoter by global is dramatically reduced by rapamycin [31,32], Rap1 remains ChIP analysis. In contrast, nuclease protection is not evident bound to its target promoters ([30,33], and B.B., E.P. and at the RAP1 sites in the RPS15 promoter, consistent with the S.S., unpublished results). We found that rapamycin treat- Genome Biology 2004, 5:R62 R62.8 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 ment causes nucleosomes to return to the vicinity of Rap1 depletion in the vicinity of Rap1 sites in the promoters of sites, as reflected by twofold and greater increases in H3 ChIP ribosomal protein genes. Moreover, nucleosomes appeared to enrichment (Figure 4e). Together these data show that Rap1 return when the Rap1 consensus sites in one of these promot- consensus sites are required for conditional nucleosome ers were removed. These findings are consistent with previ- depletion at ribosomal protein gene promoters. ously described roles for Rap1 in opening chromatin and altering nucleosome positioning [27,28]. However, Rap1 recruitment is not equally associated with nucleosome deple- tion under all conditions. We find that nucleosomes partially Discussion To gain further insight into the role of nucleosomes in gene return to the vicinity of Rap1 sites during a rapamycin- regulation, we systematically evaluated promoter nucleo- induced starvation response [34], even though Rap1 remains some occupancy in yeast by immunoprecipitating nucleo- bound ([30,33], and B.B, E.P. and S.S., unpublished results). somal DNA and quantifying enrichment with microarrays. Hence, the nucleosome loss associated with Rap1 recruitment Promoters that are inefficiently immunoprecipitated by gen- is most likely to require additional proteins, such as Esa1, a eral anti-histone antibodies, and are therefore presumed to histone acetyltransferase recruited by Rap1 under exponen- be relatively nucleosome-depleted, tend to regulate active tial growth conditions but released in stress [30]. genes (Figure 2). This is consistent with the previous observa- tion that the activated PHO5 promoter is largely devoid of These findings may also offer insight into the barrier activity nucleosomes [5,6]. However, as not all nucleosome-depleted previously documented for Rap1 [35]. Heterochromatin promoters regulate active genes, there are most likely to be propagation involves the sequential modification of histones additional determinants of depletion. An unbiased search for in adjacent nucleosomes through positive-feedback mecha- sequence elements enriched in nucleosome-depleted promot- nisms [7,11]. Certain factors such as Rap1 are able to block ers revealed poly(dA.dT) elements, previously shown to this propagation by largely unknown mechanisms [36]. One destabilize nucleosome formation [17,18], and the Rap1 con- model speculates that these barriers create nucleosome-free sensus motif. By incorporating sequence conservation data 'holes' lacking the histone substrate required for heterochro- [21], more than 30 other enriched motifs could be identified. matin propagation [29,35]. By identifying such a 'hole' in the However, most of these appear to be relevant mainly when vicinity of Rap1-binding sites in vivo our data support this occurring in combination. When we limited this analysis to model. Remarkably, the nucleosomal hole and the barrier promoters containing four or fewer motifs, all but two of function ascribed to Rap1 may be conditional, as nucleosomes these additional motifs drop out (only the Rap1 and Swi4 con- return following treatment with the small molecule rapamy- sensus sites remain). As the majority of conserved motifs cin, which activates a starvation response. Heterochromatic incorporated in this analysis recruit transcription factors silencing has been shown previously to moderate under these [21], these data suggest that multiple transcription factors act conditions [37]. Hence, we speculate that dynamic influences in combination to deplete nucleosomes. This possibility is on nucleosome occupancy may enable Rap1 to define further supported by our finding that functionally coopera- chromatin domains and vary them in response to environ- tive transcription factors tend to bind nucleosome-depleted mental cues. promoters. These associations may reflect a mechanistic model in which transcription factors compete collaboratively More broadly, the widespread nucleosome loss observed in to displace nucleosomes in order to gain access to target sites the promoters of active genes provides a general caveat for in the DNA [23]. This model was formulated to explain why ChIP studies examining posttranslational histone modifica- certain pairs of transcription factors bind cooperatively to tions, as a decrease in signal for a histone modification at a proximal target sites in vivo and on a chromatin template, but promoter undergoing activation may actually reflect nucleo- not to naked DNA [23-25]. This view invokes a broad role for some loss. Similarly, regions that appear relatively hypo- nucleosomes as ubiquitous negative regulators of transcrip- modified by ChIP may actually be nucleosome-depleted. tion factor binding and function. We speculate that by pro- However, this is not the case for low levels of acetylation [38] moting synergy among multiple transcription factors and and H3 lysine 4 methylation [39] observed at yeast telomeres, impeding the activities of individual ones, nucleosomes facil- as these regions have high occupancy. The data also provide itate threshold behavior and filter noise (for example, genetic insight into the maintenance of epigenetic information by his- variation in motif sequence) in the transcriptional regulatory tone modifications. Whereas epigenetic memory of a network. repressed state can be maintained on histones in promoters, memory of an activated state must be maintained on histones Although many factors appear to act in defining promoter outside the promoters, for example in transcribed regions, nucleosome occupancy, our data indicate that Rap1 has a which may not undergo significant nucleosome loss during uniquely important role. Rap1 and its consensus motif are activation [5,6]. Methylation of histone H3 at lysines 4 and both markedly enriched in nucleosome-depleted promoters. 36, targeted to transcribed regions in yeast via interactions Follow-up studies using real-time PCR and micrococcal between RNA polymerase and the methylases [39-47], may nuclease digestion also demonstrate marked nucleosome represent such 'activating' marks. Genome Biology 2004, 5:R62 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.9 composite Cy5:Cy3 ratios determined according to protocols Materials and methods at the Stanford Microarray Database [56]. Correlations Chromatin immunoprecipitation (ChIP) DNA associated with histone H3 in vivo was immunoprecipi- between replicate datasets were ~0.8 for all experiments. tated with antibodies against the invariant H3 carboxy termi- Composite datasets were log transformed and zero centered nus using a ChIP protocol described previously [39,48,49]. before further analysis. The histone H3 ChIP dataset was Briefly, 45 ml log-phase w303a yeast (OD ~ 1.0) growing in determined from four independent immunoprecipitations yeast extract/peptone/dextrose (YPD) were cross-linked in and hybridizations (two each using antibodies from Cell Sig- 1% formaldehyde for 15 min, washed twice in PBS, resus- naling or Abcam). The FLAG-H2B ChIP dataset was deter- pended in 400 µl lysis buffer (50 mM Hepes-KOH pH 7.5, 140 mined from three independent immunoprecipitations and mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deox- hybridizations. The mRNA dataset was determined from ycholate) and lysed with glass beads. The resulting extract three independent extractions and hybridizations of mRNA was sonicated to fragment chromatin (4 × 20 sec burst/30 sec against genomic DNA. Relative transcription rates were rest with a Branson Sonifier 250 at 70% duty, power 3) and determined by dividing transcript levels by half-life data col- centrifuged for 15 min. Solubilized chromatin was then lected by Wang and colleagues [16]. A set of activated promot- immunoprecipitated with polyclonal antibodies against the ers was defined as those in the top 10% by mRNA expression carboxy terminus of histone H3 (Abcam or Cell Signaling). A level of associated gene, with divergent promoters assigned to unenriched whole-cell extract sample (WCE) was also the more highly expressed gene. Complete datasets are avail- retained as a control. After enrichment, cross-links were able online [57]. reversed by incubating samples in 10 mM Tris-HCl pH 8.0, 1 mM EDTA, 1.0% SDS, 150 mM NaCl at 65°C overnight. DNA Analysis of nucleosome-depleted promoters was purified from ChIP and WCE samples by proteinase K Z-scores were assigned to each intergenic that reflect deple- treatment, phenol/chloroform extraction, ethanol precipita- tion across the four H3 and three H2B ChIP experiments, tion, and incubation with RNAse. DNA associated with his- using the formula Z = (x - µ)/σ where x is the average of the tone H2B in vivo was isolated in a similar manner from yeast replicate measurements, µ is the average of all intergenics containing epitope-tagged H2B [50] using anti-FLAG M2 and σ is the standard error of the replicate measurements. We monoclonal antibodies (Sigma). defined as nucleosome-depleted the 410 features with the highest Z-scores. This set, which includes 308 promoters, DNA amplification and hybridization contains nucleosome-depleted outliers and is not inclusive of To obtain sufficient quantities for hybridization, immunopre- all promoters that immunoprecipitate with average or lower cipitated DNA (from approximately 10 cells) and whole-cell efficiency. The average aqueous enrichment ratio [15] for extract DNA (unenriched control) were amplified in a linear these 308 depleted promoters is 1.7-fold, significantly higher fashion as described [51]. Briefly, terminal transferase was than expected by chance (permutation p < 0.001), consistent used to add poly(T) tails to DNA fragment and a T7-poly(A) with the premise that these promoters are relatively free of adaptor primer was used to incorporate T7 promoters. The nucleosomes. reaction products were used as template for an in vitro tran- scription reaction carried out with the T7 Megascript Kit Sequence elements common to nucleosome-depleted pro- (Ambion) and RNA samples were purified using an RNeasy moters were identified by searching between 10 and 500 bp Mini Kit (Qiagen). Amplified RNA was reverse-transcribed, upstream of gene start sites for over-represented sequences incorporating amino-allyl dUTP, and the resulting DNA was up to 10 bp in length using the GeneSpring program suite (Sil- fluorescently labeled by incubation with monofunctional icon Genetics). Enrichment was confirmed by evaluating the reactive Cy5 (enriched sample) or Cy3 (unenriched control) significance of overlap between the set of nucleosome- dye as described [52]. Microarrays containing 6,438 PCR- depleted promoters and the set of promoters containing Rap1 amplified intergenic regions were prepared as described pre- consensus motifs (ACACCCATACAT with up to two mis- viously [39,53,54]. Mixed Cy5-/Cy3-labeled probe was matches) or poly dA.dT stretches at least 10 bp in length hybridized to intergenic microarrays for 12-14 h at 60°C, (identified using PatMatch, Saccharomyces Genome Data- washed and then scanned using a GenePix 4000A scanner base [58]). Statistical significances of overlaps between sets with GenePix Pro software (Axon Instruments) as described are expressed as P-values calculated by a hypergeometric [55]. In addition, transcript levels were determined by probability model. The P-values reflect the extent to which hybridizing Cy5-labeled mRNA extracted from log phase observed overlaps exceed that expected under the null w303a yeast against Cy3-labeled genomic DNA on microar- hypothesis that there is no relationship between the sets [59]. rays containing 6,218 open reading frames (ORFs), as Where specified, permutation analyses were carried out by described previously [16]. generating 1,000 random but representative promoter sets with an Excel macro and used to confirm statistical signifi- cance. Lists of promoters containing the 71 conserved motifs Microarray data processing Cy5 and Cy3 fluorescence were integrated for each feature [21] were collated from gene sets available online [60]. Lists using GenePix Pro Software (Axon). Data were processed and Genome Biology 2004, 5:R62 R62.10 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. http://genomebiology.com/2004/5/9/R62 of promoters bound by transcription factors at a significance extracted with phenol twice and chloroform once and precip- of p < 0.001 [1] were collated from data available at [61]. itated in ethanol. Samples were washed, resuspended in 10 mM Tris pH 7.5, subjected to RNAse treatment, cleaned up Real-time PCR with the MinElute kit (Qiagen) and run out in a 1% agarose Regions approximately 200 bp in size that span one or more gel. Following depurination, denaturation and neutralization Rap1 consensus sites in ribosomal protein gene promoters of the gel, DNA was transferred onto nylon membranes by were amplified from ChIP and unenriched control samples capillary action and covalently linked to the membranes by using SYBR green PCR mix (Qiagen) in an MJ Research real- UV irradiation. Southern blotting was carried out using a DIG time PCR machine according to the manufacturers' instruc- Luminescent Detection Kit (Roche) and DIG-labeled probe tions. Fold-ratios that reflect relative enrichment or depletion generated by PCR using the TUB2, RPS11B and RPS15 prim- of a given region in the H3 or FLAG-H2B ChIP assays were ers described above. -∆∆C determined using the 2 method described in the Applied Biosystems User Bulletin. For each region examined, the TUB2 promoter was used as the normalizer (this promoter is Acknowledgements We thank Mary Ann Osley and Kevin Struhl for generously providing yeast used as a control because its occupancy approximates that of strains, and Jay Bradner, Jeff McMahon, Aly Shamji, Jianping Cui, Manolis the average promoter by global analysis), and the unenriched Kellis, Vamsi Mootha, Mike Kamal and Oliver Rando for insightful discus- control sample was used as the calibrator. Each reported ratio sions. This study was supported by a grant from NIGMS (GM38627, awarded to S.L.S.). B.E.B. is supported by a K08 Development Award from represents the average of three independent ChIP experi- the National Cancer Institute. C.L.L. is supported by a Graduate Research ments analyzed in duplicate by real-time PCR. The following Fellowship from the National Science Foundation. S.L.S. is an Investigator primer pairs were used: at the Howard Hughes Medical Institute. RPS22A promoter: 5'-GCCTAAAACGCCCATAAGTT-3' and References 5'-ACTGCAAACCCATATTCAAGA-3' 1. Lee TI, Rinaldi NJ, Robert F, Odom DT, Bar-Joseph Z, Gerber GK, Hannett NM, Harbison CT, Thompson CM, Simon I, et al.: Tran- RPS15 promoter: 5'-TACACCGCGCGTATAAATCA-3' and 5'- scriptional regulatory networks in Saccharomyces cerevisiae. Science 2002, 298:799-804. CCCAGCAAGGAGTTTCTCAG-3' 2. Kornberg RD, Lorch Y: Twenty-five years of the nucleosome, fundamental particle of the eukaryote chromosome. Cell RPS11B promoter: 5'-GAAGAAATATTTCCTTGCTGCACC-3' 1999, 98:285-294. 3. Han M, Grunstein M: Nucleosome loss activates yeast down- and 5'-AAGGGAAACGTAAAGCTATTGGAC-3' stream promoters in vivo. Cell 1988, 55:1137-1145. 4. Narlikar GJ, Fan HY, Kingston RE: Cooperation between RPL23A promoter: 5'-ATTAACATCTGTACACCCCCAACT-3' complexes that regulate chromatin structure and transcription. Cell 2002, 108:475-487. and 5'-TACAGTTCGTTTCCTGCC ATATTA-3' 5. Boeger H, Griesenbeck J, Strattan JS, Kornberg RD: Nucleosomes unfold completely at a transcriptionally active promoter. Mol Cell 2003, 11:1587-1598. TUB2 promoter: 5'-GGCCTAACAGTAAAGATATCCTCC-3' 6. Reinke H, Horz W: Histones are first hyperacetylated and then and 5'-GTTGTAGTAGCTGCTATGT CACTC-3' lose contact with the activated PHO5 promoter. Mol Cell 2003, 11:1599-1607. 7. Schreiber SL, Bernstein BE: Signaling network model of Centromeric vectors containing either a mutant RPS11B pro- chromatin. Cell 2002, 111:771-778. moter lacking the two Rap1 consensus motifs [30] or an 8. Hassan AH, Neely KE, Vignali M, Reese JC, Workman JL: Promoter targeting of chromatin-modifying complexes. Front Biosci 2001, essentially wild-type allele were transformed into wild-type 6:D1054-D1064. yeast and used in an H3 ChIP assay to evaluate the conse- 9. Kouzarides T: Histone methylation in transcriptional control. quence of removing Rap1 binding sites on nucleosome occu- Curr Opin Genet Dev 2002, 12:198-209. 10. Kurdistani SK, Grunstein M: Histone acetylation and deacetyla- pancy. Enrichment was evaluated by real-time PCR using the tion in yeast. Nat Rev Mol Cell Biol 2003, 4:276-284. following primer pair that selectively amplifies the plasmid 11. Grewal SI, Elgin SC: Heterochromatin: new possibilities for the alleles but not the endogenous RPS11B promoter: 5'- inheritance of structure. Curr Opin Genet Dev 2002, 12:178-187. 12. Turner BM: Cellular memory and the histone code. Cell 2002, CTGGAAGAAATATTTCCTT GCTCTAG-3' and 5'-AAG- 111:285-291. GGAAACGTAAAGCTATTGGAC-3'. 13. Gregory PD, Horz W: Chromatin and transcription - how tran- scription factors battle with a repressive chromatin environment. Eur J Biochem 1998, 251:9-18. Micrococcal nuclease assay 14. Morse RH: Getting into chromatin: how do transcription fac- Log-phase cultures of W303a yeast grown in 450 ml YPD to tors get past the histones? Biochem Cell Biol 2003, 81:101-112. 15. Nagy PL, Cleary ML, Brown PO, Lieb JD: Genomewide demarca- OD of 1.0 were spheroplasted with zymolase (10 mg in 40 tion of RNA polymerase II transcription units revealed by ml volume of 1 M sorbitol, 50 mM Tris pH 7.4, 10 mM β-mer- physical fractionation of chromatin. Proc Natl Acad Sci USA 2003, 100:6364-6369. captoethanol (β-ME), at 30°C for 38 min shaking at 300 16. Wang Y, Liu CL, Storey JD, Tibshirani RJ, Herschlag D, Brown PO: rpm), divided into five aliquots, and digested with increasing Precision and functional specificity in mRNA decay. Proc Natl concentrations (20 U to 320 U) of micrococcal nuclease Acad Sci USA 2002, 99:5860-5865. 17. Iyer V, Struhl K: Poly(dA:dT), a ubiquitous promoter element (Worthington Biochem) in 600 µl 0.5 mM spermidine, 1 mM that stimulates transcription via its intrinsic DNA structure. β-ME, 0.075% NP-40. DNA from digested samples was EMBO J 1995, 14:2570-2579. Genome Biology 2004, 5:R62 comment reviews reports deposited research refereed research interactions information http://genomebiology.com/2004/5/9/R62 Genome Biology 2004, Volume 5, Issue 9, Article R62 Bernstein et al. R62.11 18. Anderson JD, Widom J: Poly(dA-dT) promoter elements 42. Krogan NJ, Kim M, Tong A, Golshani A, Cagney G, Canadien V, Rich- increase the equilibrium accessibility of nucleosomal DNA ards DP, Beattie BK, Emili A, Boone C, et al.: Methylation of his- target sites. Mol Cell Biol 2001, 21:3830-3839. tone H3 by Set2 in Saccharomyces cerevisiae is linked to 19. Graham IR, Chambers A: Use of a selection technique to iden- transcriptional elongation by RNA polymerase II. Mol Cell Biol tify the diversity of binding sites for the yeast RAP1 tran- 2003, 23:4207-4218. scription factor. Nucleic Acids Res 1994, 22:124-130. 43. Schaft D, Roguev A, Kotovic KM, Shevchenko A, Sarov M, Neuge- 20. Buchman AR, Kimmerly WJ, Rine J, Kornberg RD: Two DNA-bind- bauer KM, Stewart AF: The histone 3 lysine 36 methyltrans- ing factors recognize specific sequences at silencers, ferase, SET2, is involved in transcriptional elongation. Nucleic upstream activating sequences, autonomously replicating Acids Res 2003, 31:2475-2482. sequences, and telomeres in Saccharomyces cerevisiae. Mol Cell 44. Xiao T, Hall H, Kizer KO, Shibata Y, Hall MC, Borchers CH, Strahl Biol 1988, 8:210-225. BD: Phosphorylation of RNA polymerase II CTD regulates 21. Kellis M, Patterson N, Endrizzi M, Birren B, Lander ES: Sequencing H3 methylation in yeast. Genes Dev 2003, 17:654-663. and comparison of yeast species to identify genes and regu- 45. Li B, Howe L, Anderson S, Yates JR 3rd, Workman JL: The Set2 his- latory elements. Nature 2003, 423:241-254. tone methyltransferase functions through the phosphor- 22. Banerjee N, Zhang MQ: Identifying cooperativity among tran- ylated carboxyl-terminal domain of RNA polymerase II. J Biol scription factors controlling the cell cycle in yeast. Nucleic Chem 2003, 278:8897-8903. Acids Res 2003, 31:7024-7031. 46. Gerber M, Shilatifard A: Transcriptional elongation by RNA 23. Miller JA, Widom J: Collaborative competition mechanism for polymerase II and histone methylation. J Biol Chem 2003, gene activation in vivo. Mol Cell Biol 2003, 23:1623-1632. 278:26303-26306. 24. Vashee S, Melcher K, Ding WV, Johnston SA, Kodadek T: Evidence 47. Hampsey M, Reinberg D: Tails of intrigue: phosphorylation of for two modes of cooperative DNA binding in vivo that do RNA polymerase II mediates histone methylation. Cell 2003, not involve direct protein-protein interactions. Curr Biol 1998, 113:429-432. 8:452-458. 48. Kuo MH, Allis CD: In vivo cross-linking and 25. Adams CC, Workman JL: Binding of disparate transcriptional immunoprecipitation for studying dynamic Protein:DNA activators to nucleosomal DNA is inherently cooperative. associations in a chromatin environment. Methods 1999, Mol Cell Biol 1995, 15:1405-1421. 19:425-433. 26. Gregory PD, Horz W: Mapping chromatin structure in yeast. 49. Bernstein BE, Humphrey EL, Liu CL, Schreiber SL: The use of chro- Methods Enzymol 1999, 304:365-76. matin immunoprecipitation assays in genome-wide analyses 27. Yu L, Morse RH: Chromatin opening and transactivator poten- of histone modifications. Methods Enzymol 2004, 376:349-360. tiation by RAP1 in Saccharomyces cerevisiae. Mol Cell Biol 1999, 50. Robzyk K, Recht J, Osley MA: Rad6-dependent ubiquitination of 19:5279-88. histone H2B in yeast. Science 2000, 287:501-504. 28. De Sanctis V, La Terra S, Bianchi A, Shore D, Burderi L, Di Mauro E, 51. Liu CL, Schreiber SL, Bernstein BE: Development and validation Negri R: In vivo topography of Rap1p-DNA complex at Sac- of a T7 based linear amplification for genomic DNA. BMC charomyces cerevisiae TEF2 UAS(RPG) during transcrip- Genomics 2003, 4:19. tional regulation. J Mol Biol 2002, 318:333-349. 52. Carroll AS, Bishop AC, DeRisi JL, Shokat KM, O'Shea EK: Chemical 29. Yu Q, Qiu R, Foland TB, Griesen D, Galloway CS, Chiu YH, Sand- inhibition of the Pho85 cyclin-dependent kinase reveals a meier J, Broach JR, Bi X: Rap1p and other transcriptional regu- role in the environmental stress response. Proc Natl Acad Sci lators can function in defining distinct domains of gene USA 2001, 98:12578-12583. expression. Nucleic Acids Res 2003, 31:1224-1233. 53. Iyer VR, Horak CE, Scafe CS, Botstein D, Snyder M, Brown PO: 30. Reid JL, Iyer VR, Brown PO, Struhl K: Coordinate regulation of Genomic binding sites of the yeast cell-cycle transcription yeast ribosomal protein genes is associated with targeted factors SBF and MBF. Nature 2001, 409:533-538. recruitment of Esa1 histone acetylase. Mol Cell 2000, 54. Ren B, Robert F, Wyrick JJ, Aparicio O, Jennings EG, Simon I, Zeitlin- 6:1297-1307. ger J, Schreiber J, Hannett N, Kanin E, et al.: Genome-wide location 31. Powers T, Walter P: Regulation of ribosome biogenesis by the and function of DNA binding proteins. Science 2000, rapamycin-sensitive TOR-signaling pathway in Saccharomy- 290:2306-2309. ces cerevisiae. Mol Biol Cell 1999, 10:987-1000. 55. Microarrays.org [http://www.microarrays.org] 32. Hardwick JS, Kuruvilla FG, Tong JK, Shamji AF, Schreiber SL: 56. Gollub J, Ball CA, Binkley G, Demeter J, Finkelstein DB, Hebert JM, Rapamycin-modulated transcription defines the subset of Hernandez-Boussard T, Jin H, Kaloper M, Matese JC, et al.: The nutrient-sensitive signaling pathways directly controlled by Stanford Microarray Database: data access and quality the Tor proteins. Proc Natl Acad Sci USA 1999, 96:14866-14870. assessment tools. Nucleic Acids Res 2003, 31:94-96. 33. Rohde JR, Cardenas ME: The tor pathway regulates gene 57. Stuart Schreiber group website [http://www.broad.broad.mit.edu/chem expression by linking nutrient sensing to histone acetylation. bio/lab_schreiber/pubs/protocols/global%20nucleosome/globalnuc.html] Mol Cell Biol 2003, 23:629-635. 58. Saccharomyces Genome Database [http://www.yeastge 34. Shamji AF, Kuruvilla FG, Schreiber SL: Partitioning the transcrip- nome.org] tional program induced by rapamycin among the effectors of 59. Tavazoie S, Hughes JD, Campbell MJ, Cho RJ, Church GM: System- the Tor proteins. Curr Biol 2000, 10:1574-1581. atic determination of genetic network architecture. Nat Genet 35. Bi X, Broach JR: UASrpg can function as a heterochromatin 1999, 22:281-285. boundary element in yeast. Genes Dev 1999, 13:1089-1101. 60. Yeast comparative genomics [http://www.broad.mit.edu/anno 36. Donze D, Kamakaka RT: Braking the silence: how heterochro- tation/fungi/comp_yeasts] matic gene repression is stopped in its tracks. BioEssays 2002, 61. Young lab [http://web.wi.mit.edu/young] 24:344-349. 62. Lieb JD, Liu X, Botstein D, Brown PO: Promoter-specific binding 37. Ai W, Bertram PG, Tsang CK, Chan TF, Zheng XF: Regulation of of Rap1 revealed by genome-wide maps of protein-DNA subtelomeric silencing during stress response. Mol Cell 2002, association. Nat Genet 2001, 28:327-334. 10:1295-1305. 38. Braunstein M, Rose AB, Holmes SG, Allis CD, Broach JR: Transcrip- tional silencing in yeast is associated with reduced nucleo- some acetylation. Genes Dev 1993, 7:592-604. 39. Bernstein BE, Humphrey EL, Erlich RL, Schneider R, Bouman P, Liu JS, Kouzarides T, Schreiber SL: Methylation of histone H3 Lys 4 in coding regions of active genes. Proc Natl Acad Sci USA 2002, 99:8695-8700. 40. Ng HH, Robert F, Young RA, Struhl K: Targeted recruitment of Set1 histone methylase by elongating Pol II provides a local- ized mark and memory of recent transcriptional activity. Mol Cell 2003, 11:709-719. 41. Li J, Moazed D, Gygi SP: Association of the histone methyltrans- ferase Set2 with RNA polymerase II plays a role in transcrip- tion elongation. J Biol Chem 2002, 277:49383-49388. Genome Biology 2004, 5:R62

Journal

Genome Biology – Springer Journals

Published: Aug 1, 2004

Keywords: Animal Genetics and Genomics; Human Genetics; Plant Genetics and Genomics; Microbial Genetics and Genomics; Bioinformatics; Evolutionary Biology

There are no references for this article.