JBC Anatrace, Inc.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M700623200 on July 13, 2007

J. Biol. Chem., Vol. 282, Issue 38, 27610-27621, September 21, 2007
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
282/38/27610    most recent
M700623200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Barbaric, S.
Right arrow Articles by Korber, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Barbaric, S.
Right arrow Articles by Korber, P.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Redundancy of Chromatin Remodeling Pathways for the Induction of the Yeast PHO5 Promoter in Vivo*

Slobodan Barbaric{ddagger}1, Tim Luckenbach§1, Andrea Schmid§, Dorothea Blaschke§, Wolfram Hörz§{dagger}, and Philipp Korber§2

From the §Adolf-Butenandt-Institut, Universität München, 44 Schillerstrasse, 80336 München, Germany and the {ddagger}Laboratory of Biochemistry, Faculty of Food Technology and Biotechnology, University of Zagreb, 10000 Zagreb, Croatia

Received for publication, January 22, 2007 , and in revised form, June 21, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Induction of the yeast PHO5 and PHO8 genes leads to a prominent chromatin transition at their promoter regions as a prerequisite for transcription activation. Although induction of PHO8 is strictly dependent on Snf2 and Gcn5, there is no chromatin remodeler identified so far that would be essential for the opening of PHO5 promoter chromatin. Nonetheless, the nonessential but significant involvement of cofactors can be identified if the chromatin opening kinetics are delayed in the respective mutants. Using this approach, we have tested individually all 15 viable Snf2 type ATPase deletion mutants for their effect on PHO5 promoter induction and opening. Only the absence of Snf2 and Ino80 showed a strong delay in chromatin remodeling kinetics. The snf2 ino80 double mutation had a synthetic kinetic effect but eventually still allowed strong PHO5 induction. The same was true for the snf2 gcn5 and ino80 gcn5 double mutants. This strongly suggests a complex network of redundant and mutually independent parallel pathways that lead to the remodeling of the PHO5 promoter. Further, chromatin remodeling kinetics at a transcriptionally inactive TATA box-mutated PHO5 promoter were affected neither under wild type conditions nor in the absence of Snf2 or Gcn5. This demonstrates the complete independence of promoter chromatin opening from downstream PHO5 transcription processes. Finally, the histone variant Htz1 has no prominent role for the kinetics of PHO5 promoter chromatin remodeling.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Eukaryotic DNA is folded in the nucleus into a compact structure called chromatin. On the first level of compaction, the DNA is assembled with basic histone proteins to form the nucleosome, which constitutes the fundamental building block of chromatin (1). The nucleosome interferes with binding of many factors to the underlying DNA, so the cell needs mechanisms that allow access to the thus packaged DNA. These chromatin dynamics provide an important level of gene regulation. For example, the chromatin structure at gene promoters may be remodeled between a repressive and an induced state (2, 3), and nucleosome occupancy influences the usage of potential factor binding sites in the genome (4). The cofactors that aid in rendering chromatin dynamic are therefore likely to play a role in gene induction, most notably chromatin-remodeling complexes, histone-modifying enzymes, histone chaperones, and histone variants (5, 6).

The chromatin remodeling complexes utilize the energy of ATP hydrolysis to disrupt histone-DNA interactions (7, 8). They all contain an ATPase subunit, which is related to the Snf2 subfamily of helicases (9–11). Based on the sequence homology of their ATPase subunit, different groups of remodeling complexes can be distinguished (e.g. the SWI/SNF, ISWI, Mi-2, or Ino80 complexes) (8, 12, 13).

At present, it is hard to predict which remodeler is involved in a given chromatin transition. However, although nucleosomal DNA displays significant spontaneous movements in vitro (14), it seems reasonable to assume that most larger scale changes of chromatin structure in the cell are catalyzed by chromatin remodeling complexes or other energy-dependent machines like polymerases. It is less clear if chromatin transitions always involve histone modifications, but several cases were reported where chromatin-remodeling complexes and histone-modifying enzymes as well as other factors, such as histone chaperones, can work together in order to modulate chromatin structure (6, 15–20).

The yeast PHO5 promoter is a well studied example of a eukaryotic promoter where the chromatin structure undergoes an extensive transition in the process of induction (2, 21). We and others showed that chromatin remodeling at the PHO5 promoter results in the loss of histones from this region (18, 22) by an eviction mechanism in trans (23, 24). This way an activator binding site that is protected by a nucleosome in the repressed state (25) becomes accessible. The intriguing mechanism of histone eviction from active promoter regions appears to be common in yeast, as shown by two genome-wide studies (26, 27), and in Drosophila active gene promoter regions were found to be depleted of histones as well (28). It is therefore of general interest to identify the chromatin remodeler and putative other chromatin cofactors that are necessary and/or sufficient for such a mechanism.

So far, no cofactor has been identified that would be essential for the chromatin transition at the PHO5 promoter. The search for the putative dedicated chromatin remodeler or other cofactor for opening of the PHO5 promoter has been ongoing for a long time. It was fueled in part by the finding that the PHO8 promoter, which is activated via the same signaling pathway and the same specific activator, is strictly dependent on both the chromatin remodeling complex SWI/SNF and the histone acetyltransferase Gcn5 (29). However, the PHO5 promoter does not require the histone acetyltransferase Gcn5 (30) or the SWI/SNF remodeling complex ATPase subunit Snf2 (31) for chromatin remodeling and activation under conditions of maximal induction. Although these cofactors are not essential for full induction of PHO5, our laboratory and others showed that Gcn5 (32), Snf2 (18, 33, 34), and also the histone chaperone Asf1 (19, 35) contribute importantly to the rate of chromatin remodeling, since the respective deletion mutants showed a strong delay in PHO5 induction kinetics.

Here, we report a comprehensive search for possible effects on PHO5 induction kinetics or extent in all viable chromatin remodeler mutants as well as in some double mutants. Only the absence of Snf2 and Ino80 showed a clear delay in chromatin remodeling kinetics. In addition to the previously reported kinetic delay of an snf2 deletion mutant, we show now that a point mutation in the ATPase domain of Snf2 is already sufficient to cause this delay but does not display a dominant negative phenotype.

The snf2 ino80 as well as snf2 gcn5 and ino80 gcn5 double mutants can still induce PHO5, although with a synthetic kinetic delay. This indicates that all of the respective cofactors can work independently of each other in parallel pathways. Apparently, histone eviction from the PHO5 promoter is mediated by very redundant pathways. This calls into question the initial concept of a dedicated remodeling pathway and/or remodeler for the PHO5 promoter.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Strains and Media—For a complete list of the Saccharomyces cerevisiae strains used in this study, see Table 1. Strains CY337 ino80 and CY407 ino80 (snf2 ino80) are derivatives of CY337 and CY407, respectively, in which the INO80 gene was disrupted by linear transformation with a DNA fragment of the INO80 locus with a URA3 marker gene cassette inserted into the INO80 ORF. Strain CY337 swr1 was generated accordingly. Yeast strains were grown in YPDA plus 1 g/liter KH2PO4 or in YNB selection medium supplemented with the required amino acids for plasmid-bearing strains, as repressive conditions (high phosphate) and in phosphate-free synthetic medium for induction (2). This phosphate-free medium corresponds to a respective YNB medium but without any phosphate. For galactose induction of strains carrying the plasmids pPpho5v33-lacZ or pPgal1-lacZ, cells were pregrown in YNB medium with 2% (w/v) raffinose and induced by the addition of galactose to a final concentration of 2% (w/v).


View this table:
[in this window]
[in a new window]

  		TABLE 1
S. cerevisiae strains used in this study

 
Plasmids—The plasmid pPpho5v33-lacZ is a derivative of the pPHO5lacZ reporter plasmid (36) in which the two Pho4 binding sites, UASp1 and UASp2, were replaced by Gal4 binding sites (32, 37). The pPgal1-lacZ plasmid is the same as p416-GAL1-lacZ in Ref. 32. The Pho4 overexpression plasmid pP4-70L corresponds to YEpP4 (38) but carries the LEU2 marker. The plasmid pTAP5C{Delta}TATA containing the TATA-deleted PHO5 promoter variant was described before (24).

Functional Assays and Chromatin Analysis—Acid phosphatase (39) and beta-galactosidase assays (36) were done as described. The preparation of yeast nuclei (40) and chromatin analysis of nuclei by restriction nucleases and DNase I digestion with indirect end labeling were as described (41, 42).

Chromatin Immunoprecipitation Analysis—Yeast cultures with a density of 1–2 x 107 cells/ml were treated with 1% formaldehyde for 20 min at room temperature. Cross-linking was quenched by adding glycine to a final concentration of 125 mM. The cells were washed two times with ice-cold 0.9% NaCl, resuspended in HEG150 buffer (150 mM NaCl, 50 mM HEPES, pH 7.6, 10% glycerol, 1% Triton X-100, 1 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride), and lysed with a French press (three times at 1, 100 pounds/inch2) or by sonication (Bioruptor, Diagenode; three times for 30 s with a 60-s pause, position high, ice water bath). In this last step, chromatin was sheared to an average size of 500 bp. Immunoprecipitation was performed as described before (43). The anti-Pho4 antibody was commercially prepared by Eurogentec upon our request, using purified Pho4. The anti-histone H3 C-terminal antibody was obtained from Abcam (ab1791-100). Immunoprecipitated DNA was quantitatively measured in triplicates with the ABI PRISM 7000 sequence detection system using the following amplicons: PHO5 UASp2-A, GAATAGGCAATCTCTAAATGAATCGA; PHO5 UASp2-B, GAAAACAGGGACCAGAATCATAAATT; PHO5 UASp2-probe, 5'-FAM3-ACCTTGGCACTCACACGTGGGACTAGC-3'-TAM; PHO5 HS1-A, CCTTTACCGTAATTTTCAATTGCTAA; PHO5 HS1-B, TCGCTTCTTCAACAGTGGTAAAAATA; PHO5 HS1-probe, 5'-FAM-CAATGTTCCTTGGTTATCCCATCGCCA-3'-TAM; TEL-A, TCCGAACGCTATTCCAGAAAGT; TEL-B, CCATAATGCCTCCTATATTTAGCCTTT; TEL-probe, 5'-FAMTCCAGCCGCTTGTTAACTCTCCGACA-3'-TAM; ACT1-A, TGGATTCCGGTGATGGTGTT; ACT1-B, TCAAAATGGCGTGAGGTAGAGA; ACT1-probe, 5'FAM-CTCACGTCGTTCCAATTTACGCTGGTTT-3'-TAM; PHO5 {Delta}TATA-A, CGCACATGCCAAATTATCAAA; PHO5 {Delta}TATA-B, ACATCAGCGCTTATATACGTTTCATT; PHO5 {Delta}TATA-C, CGACTTAGCAAAACATGGAATTCC; PHO5 {Delta}TATA-probe, 5'-FAM-TGGTCACCTTACTTGGCAAGGCATATACCC-3'-TAM; PGK1-for, GAATGGCGGGAAAGGGTTTA; PGK1-rev, TGTTTGAAAGAGAGAGAGTAACAGTACGA; PGK1-probe, 5'-FAMTACCACATGCTATGATGCCCACTGTGATC-3'-TAM.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
snf2 and ino80 Strains Are the Only Viable Remodeler Mutants with a Major Effect on PHO5 Induction—There are 17 Snf2 type helicases in the yeast genome (13) that could potentially be involved in the remodeling of PHO5 promoter chromatin upon gene induction. Two of these, Sth1 and Mot1, are encoded by essential genes (44;87) and are therefore difficult to study in vivo (see "Discussion"). For the case of Mot1, there is a temperature sensitive mutant available, mot1-1, that shows clear phenotypes already at the permissive temperature of 30°C (85;87) and was used in this study. The deletion of the remaining 15 remodeler ATPase genes yields viable mutants. We monitored Pho5 acid phosphatase induction kinetics in no phosphate medium with all these 15 deletion mutants (snf2, isw1, isw2, chd1, lsh1, swr1, ino80, fun30, rad54, rdh54, rad5, rad16, ris1, YLR247C, rad26) and the mot1-1 mutant. Only two of the mutants showed a rather strong effect on the PHO5 induction kinetics: snf2 and ino80 (Fig. 1, A and B). The absence of either Snf2 or Ino80 caused a strong delay in PHO5 induction kinetics, but the observed delay in the snf2 mutant was somewhat more pronounced (see also Fig. 4A for a direct comparison). All other remodeler mutants showed no significant effect, as exemplified by the PHO5 induction kinetics in the swr1 and chd1 mutants (Fig. 1, C and D).


Figure 1
View larger version (19K):
[in this window]
[in a new window]

  		FIGURE 1.
Only Snf2 and Ino80, of 16 nonessential remodeler ATPases, strongly affect the kinetics of PHO5 induction. A, PHO5 induction kinetics after a shift to phosphate-free medium were followed by acid phosphatase activity in a WT strain (CY337, closed circles) and the corresponding snf2 mutant (CY407, open circles). B, asin A but comparing WT (CY337, closed circles) and the corresponding ino80 mutant (CY337 ino80, open circles). C, asin A but comparing WT (CY337, closed circles) and the corresponding swr1 mutant (CY337 ino80, open circles). D, asin A, but comparing WT (Y00000 (= BY4741), closed circles) and the corresponding chd1 mutant (Y06160, open circles). Error bars, S.D. of 2–5 independent experiments. o/n, overnight induction times.

 
The ATPase Activity Is Critically Required for Snf2 Function in Chromatin Remodeling at the PHO5 Promoter—Our laboratory showed before that chromatin remodeling at the PHO5 promoter was delayed in an snf2 mutant (18). Recently, a similar kinetic delay was observed by measuring acid phosphatase activity during induction (33, 34), and this is confirmed in our systematic screen (Fig. 1A). In order to test if the ATP-dependent remodeling activity is critically important for Snf2 function at the PHO5 promoter, we now used a strain with a point mutation in the SNF2 gene, snf2K798A, which specifically abolishes the catalytic activity of the ATPase domain (45). This mutation led to a similar if not somewhat stronger delay in PHO5 activation kinetics as the snf2 deletion (Fig. 2A).

We wanted to confirm that the kinetic delay as observed by measuring Pho5 phosphatase activity was due to a delay in the level of chromatin remodeling. Measuring the accessibility of the ClaI site, which is located within nucleosome -2 of the repressed PHO5 promoter (Fig. 2B), is a reliable and quantitative assay for chromatin opening (2, 42). The ClaI site was mostly inaccessible under repressive conditions (+Pi) in wild type (Fig. 2C) and mutant cells (data not shown) but almost completely open in the wild type strain after 4 h of phosphate starvation (Fig. 2C). In contrast, the snf2 mutant was less than 50% open, and the snf2K798A mutant was open even less (Fig. 2C). Thus, both mutants showed a strong kinetic delay in chromatin remodeling.

Following the histone density over the PHO5 promoter region by chromatin immunoprecipitation (ChIP) is another way of measuring chromatin opening in the course of induction (18). Also in this assay we observed the kinetic delay of an snf2 mutant (Fig. 2D), comparable with that detected by ClaI accessibility. Nucleosome -2 precludes the binding of Pho4 to UASp2 in the repressed state of the PHO5 promoter (25). Thus, the binding of Pho4 should also be affected by delayed remodeling in the snf2 or snf2K798A strains. We assayed Pho4 binding to the PHO5 promoter by ChIP. At early time points of induction, we found much less Pho4 bound in both mutants than in the WT strain (Fig. 2E).4 A region that does not bind Pho4 and is located about 700 base pairs upstream of the Pho4 binding sites (HS1) was tested as a control for the specificity of the ChIP analysis and confirmed to be negative for the Pho4 binding signal.

Early time points of kinetics in phosphate-free medium should correspond in effect to submaximal induction conditions, as for example generated by growing cells in low phosphate rather than in phosphate-free medium (48, 49). It is shown that such submaximal conditions allow the detection of effects of cofactors that are not essential for induction of PHO5 under maximally inducing conditions but contribute to the process nonetheless (32, 47). Alternative conditions of submaximal induction can be achieved by overexpression of Pho4 in high phosphate medium. Under such conditions, PHO5 promoter chromatin is largely remodeled in WT cells, although transcription is only poorly activated as compared with full induction in phosphate-free medium (40). We now examined the effect of Snf2 on chromatin opening also under such submaximal steady-state induction conditions in continuously growing cells with the snf2K798A mutant and found much lower ClaI accessibility than in WT cells (Fig. 2F). This underscores the critical requirement for the ATPase activity of Snf2 for chromatin remodeling at the PHO5 promoter and shows that the Snf2-independent remodeling pathway, which brings about complete promoter remodeling after prolonged incubation under fully inducing conditions, is not efficient enough at certain submaximal induction conditions.


Figure 2
View larger version (27K):
[in this window]
[in a new window]

  		FIGURE 2.
The absence of Snf2 or its ATPase activity delays chromatin remodeling at the PHO5 promoter. A, PHO5 induction kinetics after shift to phosphate-free medium were followed by acid phosphatase activity in a WT strain (CY396, closed circles) and the corresponding Snf2 ATPase domain mutant snf2K798A (CY397, open circles). B, schematic of the nucleosomal organization of the PHO5 promoter in the repressed (+Pi) and induced (-Pi) state. Large circles, positioned nucleosomes with the numbering relative to the translation start site of the open reading frame (broken box). Open circles, nucleosomes that become remodeled upon induction; broken circles in the induced state represent less extensively remodeled nucleosomes. Small circles, Pho4 (gray) binding sites: the linker site UASp1 (open) and the intranucleosomal UASp2 (closed). The position of the ClaI and BstEII sites used for chromatin remodeling assays and the TATA box (T) is shown. C, nuclei isolated from the indicated strains at repressive conditions (+Pi) or after 4 h of induction in phosphate-free media (4 h, -Pi) were digested with 50 or 200 units of ClaI (left and right lane of each condition, respectively) and probed for the PHO5 promoter. The ratio of the upper band (uncut) to the lower band (cut) reflects the accessibility of the ClaI site. D, histone loss kinetics over the PHO5 promoter (amplicon PHO5-UASp2) during induction of WT (CY337, closed circles) and snf2 (CY407, open circles) strains as monitored by ChIP with the histone H3 C-terminal antibody (Upstate%20Biotechnology">Upstate Biotechnology, Inc.). ChIP data were normalized to input DNA and to an amplicon at the telomere. Error bars show the S.D. of at least three independent experiments. E, kinetics of Pho4 binding to UASp2 of the PHO5 promoter (amplicon PHO5-UASp2) were measured by ChIP with the Pho4 antibody during a time course of induction in WT (CY396, closed circles), snf2K798A (CY397, closed triangles), and snf2 (CY407, open circles) strains. ChIP data were normalized to input DNA and to an amplicon at the PGK1 promoter. Amplicon PHO5-HS1 (open triangles) corresponds to a control region upstream of the PHO5 promoter, encompassing hypersensitive site 1 (HS1) (40) and is shown for the WT strain. F, nuclei of WT (CY337) and snf2K798A (CY397) strains both carrying the Pho4 overexpression plasmid pP4–70L and growing in repressive high phosphate media, were isolated and subjected to ClaI digestion as in C. o/n, overnight induction times.

 
Altogether, an active site mutation in the ATPase domain of Snf2 is sufficient to delay PHO5 induction on the levels of promoter chromatin remodeling, Pho4 binding, and Pho5 phosphatase activity. This strongly suggests that the ATPase activity of Snf2, and not possible additional functions of the SWI/SNF complex, is critically required for Snf2 function in the process of PHO5 induction.

Snf2 and Gcn5 Can Function Independently of Each Other during PHO5 Induction—The kinetic delay in the snf2 mutant was reminiscent of the delay previously observed in the gcn5 mutant (32), demonstrating a role for both cofactors in the wild type pathway of PHO5 promoter chromatin remodeling. We therefore asked if there was functional redundancy between the two activities so that the absence of both Gcn5 and Snf2 would result in a much stronger delay or even prevent PHO5 induction all together or if otherwise both activities belonged to the same remodeling pathway.

After overnight induction, the snf2 gcn5 strain achieved a considerable final level of phosphatase activity5 (Fig. 3A). The DNase I pattern was indistinguishable from the WT under repressive conditions (Fig. 3B), as was shown before for the two single mutants (30, 31). Again, similar to the single mutants, the snf2 gcn5 double mutant showed a largely open PHO5 promoter under fully inducing conditions (Fig. 3C). This indicates that both Snf2 and Gcn5 together are dispensable for the chromatin transition under these conditions.

Although the snf2 gcn5 double mutant could eventually open up the PHO5 promoter, we observed a much more pronounced kinetic delay in phosphatase induction for the double mutant as compared with both single deletion strains (Fig. 3A). This synthetic phenotype shows that the two factors do not work exclusively in the same chromatin remodeling pathway but can function independently of each other in parallel pathways.


Figure 3
View larger version (60K):
[in this window]
[in a new window]

  		FIGURE 3.
The snf2 gcn5 double mutant shows a strong synthetic effect on the kinetics of PHO5 induction but still achieves an open PHO5 promoter chromatin pattern after overnight induction. A, same experiment as in Fig. 1 but with WT (FY1353, closed circles), snf2 (FY1360, open circles), gcn5 (FY1354, closed triangles), and snf2 gcn5 (FY1352, open triangles) strains. Error bars, S.D. of 2–5 independent experiments. o/n, overnight induction times. B, DNase I mapping of the PHO5 promoter chromatin structure in WT (CY337) and snf2 gcn5 (FY1352) strains in high phosphate medium. Ramps on top of the lanes designate increasing DNase I concentrations. The marker fragments (lane M) were generated by double digests with ApaI/DraI, ApaI/ClaI, and ApaI/BamHI (from top to bottom in the lane). The schematic on the left is analogous to the repressed state in Fig. 2B. The arrow marked with HS points to the short hypersensitive site between nucleosomes -2 and -3. C, asin B, but with gcn5 (FY1354), snf2 gcn5 (FY1352), and snf2 (FY1360) strains after overnight induction in phosphate-free medium. The marker fragments (lane M) are the same as in B, but the topmost band corresponds to an ApaI/BstEII fragment. The schematic on the left is analogous to the induced state in Fig. 2B. HS, the extended hypersensitive region characteristic for the induced promoter state.

 
PHO5 as Well as PHO8 Promoter Chromatin Remodeling Kinetics Are Delayed in the Absence of Ino80—As mentioned above, the ino80 deletion mutant was also delayed in PHO5 induction (Fig. 1B) but not as compromised as the snf2 mutant (Fig. 4A). A similar delay due to the ino80 deletion was also seen in another strain background (BY background; Table 1) (data not shown). We confirmed also for the ino80 deletion that the promoter induction was indeed slowed down and compromised on the level of chromatin opening. After 2 h of induction in no phosphate medium, the ClaI site in the PHO5 promoter was significantly more accessible in the WT than in the ino80 mutant (Fig. 4B). There was no chromatin remodeling under conditions of submaximal induction (Fig. 4C), similar to the snf2K798A mutant (Fig. 2F). However, under prolonged fully inducing conditions, the promoter chromatin was completely remodeled in the absence of Ino80 (data not shown for the endogenous PHO5 promoter; Fig. 5C for the Ppho5v33-lacZ construct; see below).

Since the PHO5 promoter could eventually be remodeled in the absence of either Snf2 or Ino80, we asked whether one remodeler was compensated by the other in the respective deletion mutants. Therefore, we constructed an snf2 ino80 double mutant and tested if induction of PHO5 was abolished completely in these cells. The double mutant showed a significantly more pronounced delay in the PHO5 induction kinetics than either single mutant, but nevertheless, strong induction was achieved after prolonged incubation (Fig. 4A). We conclude that Snf2 and Ino80 both participate independently of each other in the opening of the PHO5 promoter and that additional remodeling activities could be involved.

To find out if Ino80 functions through the same or parallel chromatin remodeling pathway as Gcn5, we also tested the effect of an ino80 gcn5 double mutant on PHO5 induction kinetics and obtained analogous results as with the snf2 gcn5 double mutant (i.e. a synthetic kinetic delay) (data not shown), suggesting that both factors can work independently of each other.

We also monitored chromatin opening at the PHO8 promoter during induction in an ino80 mutant using the HpaI site that becomes accessible upon induction (49). Also here the chromatin was significantly less open than in WT cells at the 2-h time point (Fig. 4B). This finding is rather interesting, since chromatin remodeling at the PHO8 promoter is strictly Snf2-dependent (29), and it was not necessarily to be expected that another remodeler like Ino80 would contribute as well.

The Kinetic Effect on PHO5 Induction in the Absence of Snf2 or Ino80 Does Not Depend on the Pho4 Activator or the PHO Signaling Pathway—To rule out the possibility that steps in the PHO signal transduction pathway or any function of the specific activator Pho4 are affected by an snf2 or ino80 deletion, we made use of the PHO5 promoter variant 33 (Ppho5v33), which is under the control of the Gal4 activator and can be remodeled and activated in response to galactose addition very much in the same way as the wild type PHO5 promoter upon phosphate starvation (32, 37). This PHO5 promoter variant drives the expression of the lacZ reporter gene (Ppho5v33-lacZ); thus, its induction upon galactose addition was followed by measuring beta-galactosidase activity. Importantly, the galactose induction pathway is largely independent of Snf2 (15, 31, 50). The effect of the snf2 deletion on the induction kinetics of the Gal4-activated PHO5 promoter variant (Fig. 5A) was similar to that observed with the wild type PHO5 promoter (Fig. 1A), arguing against indirect effects of Snf2 on the PHO pathway. At the same time, this shows that the Snf2 requirement for efficient chromatin remodeling does not depend strictly on Pho4 as the trigger and recruitment factor but appears to be rather an inherent property of the promoter chromatin structure.


Figure 4
View larger version (19K):
[in this window]
[in a new window]

  		FIGURE 4.
Chromatin remodeling kinetics at the PHO5 and PHO8 promoters are delayed in an ino80 strain, and the snf2 ino80 mutant shows a synthetic kinetic effect in PHO5 induction. A, same experiment as in Fig. 1 but with WT (CY337, closed circles), ino80 (CY337 ino80, open circles), snf2 (CY407, closed triangles), and snf2 ino80 (CY407 ino80, open triangles) strains. The data for the WT and ino80 strains are the same as shown in Fig. 1B. Error bars, S.D. of at least three independent experiments. o/n, overnight induction times. B, the accessibility of chromatin at the PHO5 or PHO8 promoter was assayed by nuclei digestion with ClaI or HpaI, respectively, in the same kind of experiment as in Fig. 2C but with WT (BY4741, black bars) and ino80 (BY4741 ino80, gray bars) strains induced for 2 h in phosphate-free medium. Error bars give the variation of two independent experiments. C, same experiment as in Fig. 2F but with overexpression of PHO4 (PHO4 o/x) in strain BY4741 ino80.

 
Similarly, we tested the Gal4-driven PHO5 promoter variant in the ino80 mutant and found again a kinetic delay compared with the WT cells (Fig. 5B). This kinetic delay was, however, very strong, and the final level of beta-galactosidase activity was not close to the wild type level. Therefore, we tested if Ppho5v33 chromatin was opened efficiently at all in the ino80 mutant. This was the case, since the ClaI and BstEII sites in the PHO5 promoter (Fig. 2B) showed typical accessibilities for the repressed state prior to induction and were consecutively as accessible after overnight induction as the WT (Fig. 5C). It was possible that the very strong kinetic delay and the decreased final level of beta-galactosidase activity was due to an effect of the ino80 deletion on the GAL pathway, lacZ transcription, or other steps in the expression pathway not linked to PHO5 promoter chromatin remodeling. To control for such effects, we used a plasmid containing the lacZ gene driven by the GAL1 promoter. This construct uses the same GAL pathway for induction and the same downstream processes for the expression of beta-galactosidase. It also showed a kinetic delay in the ino80 mutant but less so compared with the Ppho5v33-lacZ construct (Fig. 5D). Due to the very different final levels of induction in WT and ino80 cells for the Ppho5v33-lacZ construct and in order to get a clearer picture of the kinetic delay in the ino80 mutant, we normalized the induction kinetics data to the respective final (i.e. overnight) induction levels (Fig. 5, E and F). After this normalization, only the delay with the Ppho5v33-lacZ construct was statistically significant at early time points. Therefore, although a GAL1 promoter-driven lacZ construct is compromised in the ino80 background, there is a more significant effect specific for the PHO5 promoter variant. We conclude that the kinetic delay of PHO5 induction in an ino80 mutant is not due to indirect effects on PHO signaling or Pho5 processing but rather due to a delay in PHO5 promoter chromatin remodeling (also see "Discussion").

Kinetics of Chromatin Remodeling at the PHO5 Promoter under Wild Type Conditions and Even in the Absence of Snf2 or Gcn5 Are Independent of PHO5 Transcription—Full chromatin remodeling can be achieved at a TATA box-deleted PHO5 promoter variant, making the important point that PHO5 transcription is not essentially required for the remodeling process (51). However, we now wanted to examine if the process of PHO5 transcription could affect the rate of remodeling by a kind of positive feedback mechanism, for example through remodeling machines recruited together with the RNA polymerase. To this end, we introduced a plasmid harboring the transcriptionally inactive TATA box-deleted PHO5 promoter variant (24, 51) into a WT strain and monitored in parallel the kinetics of histone eviction at the unaltered chromosomal and the TATA-less plasmid locus by ChIP assay. No major difference in the rate of histone eviction at the two loci was observed (Fig. 6A).

We further tested if the alternative pathways, which lead to chromatin remodeling in the absence of Snf2 or Gcn5, were supported by or even depended on PHO5 transcription. The same kind of experiment as described above was performed in snf2 and gcn5 mutants. Again, no difference in the kinetics of histone eviction at the wild type and TATA-deleted promoters was detected (Fig. 6, B and C). Therefore, neither the wild type nor the alternative remodeling pathways are dependent on PHO5 transcription.

The Histone Variant Htz1 Plays Only a Minor Role in PHO5 Induction—Recently, several groups mapped the occurrence of Htz1 across the yeast genome and found generally an enrichment of Htz1 at promoters, where it is discussed to play a role either in priming the promoter for opening or in resetting the promoter after induction (52–59). With regard to PHO5, there are three studies that report the presence of the histone variant Htz1 in the promoter region (58–60). However, its function stays somewhat elusive. Santisteban et al. (60) showed an effect of Htz1 on final expression levels but not on the final extent of chromatin opening, and Millar et al. (59) showed that lysine 14 of Htz1 is important for the deposition of Htz1 at the repressed promoter. In our remodeler mutant screen, we observed that the strain deleted for SWR1, the ATPase subunit of the remodeling complex involved in Htz1 deposition (61–63), did not show an appreciable effect on PHO5 induction (Fig. 1C).

These data and the previously published observations by others left the question about a role of Htz1 in PHO5 promoter chromatin opening kinetics rather undecided. We tested directly if Htz1 may be important for the remodeling kinetics. The effect of the htz1 deletion on the level of Pho5 phosphatase activity was rather slight (Fig. 7A). We checked also another strain background (MSY background; Table 1) and found a slightly stronger effect (data not shown). However, when we examined this kinetic delay in the MSY background on the level of chromatin opening, the effect was still not very pronounced (Fig. 7B). Therefore, Htz1 seems to have no prominent function in the kinetics of PHO5 promoter chromatin remodeling.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Ever since the PHO5 promoter became a classical model system for the role of chromatin remodeling in gene regulation (2, 21, 64), it has remained an unresolved question what cofactors are necessary for the chromatin transition upon PHO5 induction. This question became even more tantalizing, since the essential remodeler for the coregulated PHO8 gene, SWI/SNF, was identified (29), but SWI/SNF was found not to be essential for PHO5 promoter remodeling (31). In order to dissect the remodeler requirements for PHO5 promoter opening, we examined the role of all nonessential chromatin remodelers and some other chromatin-related factors.

The main outcome of this in vivo study is that none of the tested chromatin cofactor mutations, not even several double mutations, prevented PHO5 promoter opening under fully inducing conditions (for a complete list of tested mutants, see Table 1). It is clear now that Ino80, although it was already suggested to play a role in PHO5 promoter chromatin remodeling (see below), is not the dedicated and only remodeling ATPase alternative to Snf2. Despite the lack of a chromatin cofactor mutation that would abolish induction of the PHO5 promoter, there are numerous mutations leading to a pronounced kinetic delay in chromatin remodeling. These are summarized, together with previously published observations, in Table 2.


View this table:
[in this window]
[in a new window]

  		TABLE 2
Summary of effects in chromatin-related cofactor mutants on PHO5 and PHO8 induction or chromatin remodeling as reported in this study or previously published

Note that the effects of mutations that directly affect Pho4 binding, like pho2 (48, 79, 80) or esa1 (81), are not listed here. ND, not determined.

 
The absence of the ATPase subunit of the chromatin remodeling complex SWI/SNF, Snf2, led to a delay in chromatin remodeling during induction kinetics (see also Refs. 18, 33, and 34). Now we pinpoint here that the ATPase activity is critically required for Snf2 function at the PHO5 promoter, since a point mutation in Snf2, snf2K798A, which specifically abolishes the ATPase activity, resulted in the same phenotype as the complete deletion of the SNF2 gene. Since the point mutation leaves the SWI/SNF complex intact, we can exclude the possibility that any putative additional function of the SWI/SNF complex was responsible for the snf2 phenotype. In addition, the phenotype of the point mutation was only somewhat more severe than that of the deletion mutant, indicating that there is no strong dominant negative effect of an inactive SWI/SNF complex that could be stably and exclusively recruited by the transactivator Pho4. This indicates the ability of Pho4 to interact with a variety of remodelers and other cofactors, and such promiscuity could be the basis for the observed redundancy of pathways that lead to the chromatin transition at the PHO5 promoter (see below).


Figure 5
View larger version (26K):
[in this window]
[in a new window]

  		FIGURE 5.
The kinetic effect on PHO5 induction in the absence of Snf2 or Ino80 does not depend on the Pho4 activator or the PHO signaling pathway. A, beta-galactosidase activity of WT (CY337, closed circles) and snf2 (CY407, open circles) strains both harboring the plasmid pPpho5v33-lacZ, containing a PHO5 promoter variant that is activated by Gal4, was measured during a time course of galactose induction. Error bars, the S.D. of at least three independent experiments. B, same as in A but with the WT (BY4741, closed circles) and ino80 (BY4741 ino80, open circles) strains. C, nuclei of WT (BY4741) or ino80 (BY4741 ino80) strains harboring plasmid pPpho5v33-lacZ were prepared after growing in raffinose medium (raff) or after overnight induction with galactose (o/n + gal), digested with 50 and 200 units (left and right lane, respectively) of ClaI and BstEII and probed for the PHO5 promoter. The ratio of the upper band to the lower band reflects the accessibility of the indicated restriction site. D, same as in B but with the plasmid pPgal1-lacZ. E, same data as in B but after normalization to the corresponding overnight values. F, same data as in D but after normalization to the corresponding overnight values. o/n, overnight induction times.

 
We obtained similar results as for Snf2 for the chromatin remodeler ATPase subunit Ino80, which was also necessary for the wild type rate but not for the final extent of chromatin remodeling at the PHO5 promoter. This suggests that these two remodeling complexes probably work together in the wild type situation and that remodeling is less efficient in the absence of one of them. Importantly, however, even the combination of both the snf2 and the ino80 mutation did not abolish eventual remodeling, but only the rate of remodeling was even further decreased. Such synthetic effects were also the result of combining the snf2 or the ino80 mutation with a deletion of the gene for Gcn5. This latter mutation was already shown by us to lead to a kinetic delay of chromatin remodeling without affecting the final level (32).

Several chromatin remodeler mutants were already reported not to be essential for final PHO5 induction (23, 65), but in these studies, there was no information about possible effects on induction kinetics (i.e. about a nonessential but important role of the respective remodelers). Recently, the O'Shea group published a systematic high throughput screen of all viable deletion mutants for effects on PHO5 induction kinetics (34). Regarding chromatin-related cofactors, this screen identified only subunits of the SAGA and the SWI/SNF complex, which fits with our earlier findings on the role of SAGA (32, 66) and the known role of Snf2 (see above). This screen, however, did not identify Ino80, maybe due to statistical cut-off criteria, although an earlier study from the same group showed a significant requirement for Ino80 as well as for the SWI/SNF complex in chromatin remodeling at the PHO5 promoter (65). The more severe effects found in the study by Steger et al. (65) compared with the screen by Huang and O'Shea (34) may be explained by the use of different induction conditions. The chemically induced inactivation of the negative regulator Pho85 for 1 h (65) probably corresponds to submaximal induction, such as early time points of induction kinetics or conditions of Pho4 overexpression in high phosphate medium (48). Dhasarathy and Kladde (47) showed recently that submaximal induction conditions, in their case controlled by varying levels of phosphate in the medium, led to more stringent cofactor requirements for remodeling of PHO5 promoter chromatin. This study directly shows how the extent of induction conditions leads to more or less pronounced cofactor requirements, and we described the equivalent observation concerning the requirement of the histone chaperone Asf1 (19). In the same line of argument, we show here that submaximal induction conditions, achieved by the overexpression of Pho4 in high phosphate medium, lead to strong deficiencies in chromatin opening in the snf2K798A or ino80 mutants compared with the WT strain.


Figure 6
View larger version (13K):
[in this window]
[in a new window]

  		FIGURE 6.
PHO5 transcription has no effect on PHO5 promoter chromatin remodeling kinetics, either in WT, snf2, or gcn5 strains. Shown is the same experiment as in Fig. 2D, but with WT (CY337) (A), snf2 (CY407) (B), and gcn5 (CY53379) (C) strains all harboring plasmid pTAP5C{Delta}TATA. The amplicon for the WT PHO5 promoter PHO5-UASp2 (Ppho5, closed circles) and the additional amplicon PHO5--TATA (Ppho5{Delta}TATA, open circles), which specifically recognizes the TATA-less PHO5 promoter on the plasmid locus, were used. ChIP data were normalized to input DNA and an amplicon in the ACT1 coding region. Error bars show the variation of two independent immunoprecipitation experiments. o/n, overnight induction times.

 


Figure 7
View larger version (13K):
[in this window]
[in a new window]

  		FIGURE 7.
Htz1 has only a minor role in PHO5 promoter chromatin remodeling kinetics. A, PHO5 induction kinetics after a shift to phosphate-free medium as in Fig. 1 were followed by acid phosphatase activity in a WT strain (Y00000 (= BY4741), closed circles) and the corresponding htz1 mutant (Y01703, open circles). Error bars give the S.D. of two or three independent experiments. B, the accessibility of chromatin at the PHO5 promoter was assayed by nuclei digestion with ClaI in the same kind of experiment as in Fig. 2C but with WT (MSY599) and htz1 (MSY1147) strains induced for 2 h in phosphate-free medium. Error bars give the variation of two independent experiments.

 
One could think that simply the slower growth of some mutants could lead to a delay in PHO induction, since the intracellular (poly) phosphate pools determine and buffer the signal strength for the PHO regulon (67, 68) and are mainly depleted by replication. In addition, several factors were found to be involved in the signal transduction pathway leading to PHO5 induction (34) and could in principle be affected by any of the mutations that showed the kinetic effect in PHO5 induction. Addressing such concerns, there are strong arguments, both from our own study as well as from published data, for a direct role of the cofactors Snf2 and Ino80 in PHO5 promoter chromatin remodeling and against indirect effects of the respective mutations on the PHO signal transduction pathway.

First, our experiments under submaximal induction conditions (i.e. overexpression of Pho4 in high phosphate medium) are performed at steady state conditions in continuously growing cells and bypass all PHO signaling steps upstream of Pho4. Nonetheless, they still show severe defects in PHO5 promoter chromatin remodeling in snf2K798A and ino80 mutants.

Second, we control for possible indirect effects by using the Gal4 activated variant of the PHO5 promoter (variant 33), which also shows a kinetic delay in PHO5 induction in the snf2 and ino80 mutants, although it is independent of the PHO regulon signal transduction pathway and of Pho4. For the case of the ino80 mutant, this control was more difficult to interpret, since the GAL1 induction is also affected (69, 70). Nonetheless, the effect on the Gal4-activated PHO5 promoter variant was still more significant than the effect on GAL1 induction. The use of this PHO5 promoter variant also demonstrates that the requirement for a chromatin-related cofactor is not necessarily determined by the type of transactivator and its putative recruitment specificity but rather by the particular chromatin structure at the promoter. In accord with this conclusion are the different inherent nucleosome stabilities at the Pho4-coregulated PHO5 and PHO8 promoters (71), and a similar conclusion was drawn in a study comparing the Gcn5 dependence of chromatin remodeling at various promoters using various activation domains (72). This latter study also showed that remodeling of the same chromatin structure could become differently dependent on Gcn5 when different activation domains were used. Also here the redundancy of cofactors for chromatin remodeling was discussed (see below).

Third, not all strains with growth defects will automatically show a delay in PHO5 induction kinetics. For example, we tested the two mutants rsc1 and rsc2 (see below) that have a similar slow growth phenotype, and the rsc1 strain showed no delay and the rsc2 mutant only a very slight delay in PHO5 induction (data not shown).

Fourth, Steger et al. (65) and Jonsson et al. (69) demonstrated by ChIP analysis the recruitment of the Ino80 complex, and Dhasarathy and Kladde (47) demonstrated the recruitment of the SWI/SNF complex to the PHO5 promoter, both under inducing conditions. This corroborates the direct role of Snf2 and Ino80 in PHO5 promoter opening.

We showed previously that active transcription of the PHO5 gene does not play a role in the final extent of promoter chromatin remodeling (51), but at that time an effect on remodeling kinetics was not assessed. Therefore, it was interesting to find out if the kinetics of the remodeling process at the PHO5 promoter are influenced by PHO5 transcription. Even more intriguing was the possibility that alternative pathways in cofactor mutants were actually supported or even fully dependent on active PHO5 transcription. We compared the kinetics of histone loss at a wild type as well as a transcriptionally inactive TATA box-deleted PHO5 promoter in the same WT cell but found no difference. While this manuscript was in the review process, Uhler et al. (73) also published that the rate of PHO5 promoter remodeling is unaffected in WT cells by a TATA box mutation. Furthermore, we show here that this was also true in an snf2 or gcn5 mutant, clearly demonstrating that even the alternative, Snf2- or Gcn5-independent remodeling pathway does not make use of cofactors that could be recruited via the TATA box-binding protein (TBP) and does not depend on active PHO5 transcription.

The inability to find a cofactor mutant where PHO5 promoter opening is not possible under fully inducing conditions and the existence of several mutants with delayed remodeling as well as synthetic kinetic effects in double mutants speak for a redundancy of pathways for chromatin remodeling at the PHO5 promoter. Kinetic effect mutants comprise different types of chromatin-related cofactors like remodelers (Snf2 and Ino80), a histone acetyltransferase (Gcn5), and a histone chaperone (Asf1). In the absence of any of these factors, an alternative but less efficient pathway can still support chromatin remodeling at the PHO5 promoter, and this pathway may still use any of the other factors. For example, the Snf2-independent pathway still uses the cofactors Gcn5, Ino80, and Asf1, since double mutants combining the snf2 deletion with a deletion of the genes of any of the other cofactors showed synthetic kinetic effects (Table 2). We speculate that the search for "the essential cofactor" for PHO5 promoter chromatin remodeling may be in vain, at least by in vivo techniques. PHO5 promoter opening appears to be sufficiently supported by so many cofactors that multiple deletions would be lethal before preventing PHO5 activation.

An obvious candidate for the ultimate back up system for PHO5 promoter opening is the RSC chromatin remodeling complex (44). It was shown recently to completely disassemble nucleosomes in cooperation with histone chaperones (74), just as it would be necessary for remodeling of the PHO5 promoter chromatin structure. It is essential for mitotic growth, since its ATPase subunit Sth1 and some of the other subunits are encoded by essential genes (44). The RSC complex exists in two types, containing either the subunit Rsc1 or Rsc2 (75). As mentioned above, we checked the rsc1 and rsc2 deletions (see Table 1) for their effect on PHO5 induction but found not much of an effect (data not shown), speaking against a specific role of one of the two RSC complexes, or maybe one type can compensate for the other.

Since RSC activity is essential for cell growth, it is very difficult to induce PHO5 induction in no phosphate medium after ablating RSC activity. Due to the intracellular phosphate pools, some rounds of replication are necessary in order to obtain full induction of the PHO system in no phosphate medium (see above and Ref. 76). Moreira et al. (77) used a conditional mutant of an essential subunit of the RSC complex, swh3ts, and under restrictive conditions it showed substantial induction of PHO5 mRNA in low phosphate medium. This experiment speaks against a critical role of RSC for PHO5 induction. However, one may argue that apparently there was significant RSC activity left in order to allow some rounds or replication, otherwise PHO5 could not have been induced (76).

In order to get around the problem of sustaining replication while ablating an essential protein, one could alternatively induce PHO5 by deletion or induced inactivation of the negative regulator complex Pho80-Pho85 under otherwise repressive conditions (32, 76, 78). However, some cofactors may seem essentially required under pho80 conditions, although they are not under full induction conditions in no phosphate medium, as we have seen previously for Gcn5 (30, 32).

Taken together, the apparent redundancy of remodeling pathways for the PHO5 promoter makes it possible in our view that the RSC complex takes part in PHO5 promoter chromatin remodeling and is the remodeler that opens the PHO5 promoter, for example in the snf2 ino80 double mutant. However, if RSC is "the essential cofactor" for the PHO5 promoter, chromatin remodeling is for the discussed reasons not straightforward to test in vivo and remains to be assessed.

As an alternative to back up remodeler activities, it is conceivable that the process of replication may play an important role in remodeler mutant backgrounds, although replication is not necessary for PHO5 induction in the WT situation (76).


   FOOTNOTES
 
This paper is dedicated to Gerda and Reiner Luckenbach.

* This work was supported by the Deutsche Forschungsgemeinschaft (Transregio 5), the 6th Framework Programme of the European Union (Epigenome Network of Excellence), and Ministry of Education, Science, and Technology of the Republic of Croatia Grant 58025 (to S. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

1 These authors contributed equally to this work. Back

{dagger} Died Nov. 13, 2005. Back

2 To whom correspondence should be addressed. Tel.: 49-89-218075435; Fax: 49-89-218075425; E-mail: pkorber{at}lmu.de.

3 The abbreviations used are: FAM, carboxyfluorescein; TAM, carboxytetramethylrhodamine; ChIP, chromatin immunoprecipitation; WT, wild type. Back

4 Although the ClaI accessibility differs between the snf2 and the snf2K798A mutant after 4 h of induction (Fig. 2C), both show a very similar low degree of Pho4 binding in the ChIP assay at this time point (Fig. 2E). The binding of Pho4 to the PHO5 promoter is determined by the cooperativity between both binding sites UASp1 and UASp2 (81). The full ChIP signal is critically dependent on UASp2, since no large increase in signal during induction is observed in a mutant strain deleted for UASp2 (H. Reinke and W. Hörz, unpublished observations), but some Pho4 binding may stem from the contribution of UASp1, and the resolution of the ChIP analysis cannot distinguish between both sites. UASp1 is located in the extended linker between nucleosomes -2 and -3 (Fig. 2B), and therefore its accessibility for Pho4 is not primarily influenced by chromatin remodeling. So there is no exact quantitative correlation to be expected between an anti-Pho4 ChIP and the ClaI assay. Back

5 For presently unknown reasons, the snf2 mutant in the FY strain background gave very high final phosphatase activity values (Fig. 3A). Since this high activity does not correspond to changes in the chromatin structure (Fig. 3C) and was not observed in other strain backgrounds (Figs. 1A and 2A), it seems to be the result of an effect on some step downstream of promoter chromatin remodeling in this strain background and is of no further relevance to our study here. Back


   ACKNOWLEDGMENTS
 
We thank Brad Cairns, Martine Collart, Craig Peterson, Xuetong Shen, Mitch Smith, Toshio Tsukiyama, and Fred Winston for the kind gifts of strains.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Kornberg, R. D., and Lorch, Y. (1999) Cell 98, 285–294[CrossRef][Medline] [Order article via Infotrieve]
  2. Almer, A., Rudolph, H., Hinnen, A., and Hörz, W. (1986) EMBO J. 5, 2689–2696[Medline] [Order article via Infotrieve]
  3. Moreira, J. A., and Holmberg, S. (1998) EMBO J. 17, 6028–6038[CrossRef][Medline] [Order article via Infotrieve]
  4. Liu, X., Lee, C. K., Granek, J. A., Clarke, N. D., and Lieb, J. D. (2006) Genome Res. 16, 1517–1528[Abstract/Free Full Text]
  5. Narlikar, G. J., Fan, H. Y., and Kingston, R. E. (2002) Cell 108, 475–487[CrossRef][Medline] [Order article via Infotrieve]
  6. Loyola, A., and Almouzni, G. (2004) Biochim. Biophys. Acta 1677, 3–11[Medline] [Order article via Infotrieve]
  7. Becker, P. B., and Hörz, W. (2002) Annu. Rev. Biochem. 71, 247–273[CrossRef][Medline] [Order article via Infotrieve]
  8. Eberharter, A., and Becker, P. B. (2004) J. Cell Sci. 117, 3707–3711[Free Full Text]
  9. Eisen, J. A., Sweder, K. S., and Hanawalt, P. C. (1995) Nucleic Acids Res. 23, 2715–2723[Abstract/Free Full Text]
  10. Peterson, C. L. (2000) FEBS Lett. 476, 68–72[CrossRef][Medline] [Order article via Infotrieve]
  11. Vignali, M., Hassan, A. H., Neely, K. E., and Workman, J. L. (2000) Mol. Cell. Biol. 20, 1899–1910[Free Full Text]
  12. Tsukiyama, T. (2002) Nat. Rev. Mol. Cell. Biol. 3, 422–429[CrossRef][Medline] [Order article via Infotrieve]
  13. Flaus, A., Martin, D. M., Barton, G. J., and Owen-Hughes, T. (2006) Nucleic Acids Res. 34, 2887–2905[Abstract/Free Full Text]
  14. Li, G., Levitus, M., Bustamante, C., and Widom, J. (2005) Nat. Struct. Mol. Biol. 12, 46–53[CrossRef][Medline] [Order article via Infotrieve]
  15. Krebs, J. E., Fry, C. J., Samuels, M. L., and Peterson, C. L. (2000) Cell 102, 587–598[CrossRef][Medline] [Order article via Infotrieve]
  16. Cosma, M. P., Tanaka, T., and Nasmyth, K. (1999) Cell 97, 299–311[CrossRef][Medline] [Order article via Infotrieve]
  17. Reinke, H., Gregory, P. D., and Hörz, W. (2001) Mol. Cell 7, 529–538[CrossRef][Medline] [Order article via Infotrieve]
  18. Reinke, H., and Hörz, W. (2003) Mol. Cell 11, 1599–1607[CrossRef][Medline] [Order article via Infotrieve]
  19. Korber, P., Barbaric, S., Luckenbach, T., Schmid, A., Schermer, U. J., Blaschke, D., and Horz, W. (2006) J. Biol. Chem. 281, 5539–5545[Abstract/Free Full Text]
  20. Moshkin, Y. M., Armstrong, J. A., Maeda, R. K., Tamkun, J. W., Verrijzer, P., Kennison, J. A., and Karch, F. (2002) Genes Dev. 16, 2621–2626[Abstract/Free Full Text]
  21. Reinke, H., and Hörz, W. (2004) Biochim. Biophys. Acta 1677, 24–29[Medline] [Order article via Infotrieve]
  22. Boeger, H., Griesenbeck, J., Strattan, J. S., and Kornberg, R. D. (2003) Mol. Cell 11, 1587–1598[CrossRef][Medline] [Order article via Infotrieve]
  23. Boeger, H., Griesenbeck, J., Strattan, J. S., and Kornberg, R. D. (2004) Mol. Cell 14, 667–673[CrossRef][Medline] [Order article via Infotrieve]
  24. Korber, P., Luckenbach, T., Blaschke, D., and Hörz, W. (2004) Mol. Cell. Biol. 24, 10965–10974[Abstract/Free Full Text]
  25. Venter, U., Svaren, J., Schmitz, J., Schmid, A., and Hörz, W. (1994) EMBO J. 13, 4848–4855[Medline] [Order article via Infotrieve]
  26. Bernstein, B. E., Liu, C. L., Humphrey, E. L., Perlstein, E. O., and Schreiber, S. L. (2004) Genome Biol. 5, R62[CrossRef][Medline] [Order article via Infotrieve]
  27. Lee, C. K., Shibata, Y., Rao, B., Strahl, B. D., and Lieb, J. D. (2004) Nat. Genet. 36, 900–905[CrossRef][Medline] [Order article via Infotrieve]
  28. Mito, Y., Henikoff, J. G., and Henikoff, S. (2005) Nat. Genet. 37, 1090–1097[CrossRef][Medline] [Order article via Infotrieve]
  29. Gregory, P. D., Schmid, A., Zavari, M., Münsterkötter, M., and Hörz, W. (1999) EMBO J. 18, 6407–6414[CrossRef][Medline] [Order article via Infotrieve]
  30. Gregory, P. D., Schmid, A., Zavari, M., Lui, L., Berger, S. L., and Hörz, W. (1998) Mol. Cell 1, 495–505[CrossRef][Medline] [Order article via Infotrieve]
  31. Gaudreau, L., Schmid, A., Blaschke, D., Ptashne, M., and Hörz, W. (1997) Cell 89, 55–62[CrossRef][Medline] [Order article via Infotrieve]
  32. Barbaric, S., Walker, J., Schmid, A., Svejstrup, J. Q., and Hörz, W. (2001) EMBO J. 20, 4944–4951[CrossRef][Medline] [Order article via Infotrieve]
  33. Neef, D. W., and Kladde, M. P. (2003) Mol. Cell. Biol. 23, 3788–3797[Abstract/Free Full Text]
  34. Huang, S., and O'Shea, E. K. (2005) Genetics 169, 1859–1871[Abstract/Free Full Text]
  35. Adkins, M. W., Howar, S. R., and Tyler, J. K. (2004) Mol. Cell 14, 657–666[CrossRef][Medline] [Order article via Infotrieve]
  36. Straka, C., and Hörz, W. (1991) EMBO J. 10, 361–368[Medline] [Order article via Infotrieve]
  37. Ertinger, G. (1998) Rolle der Transkriptionsfaktoren bei deröffnung der Chromatinstruktur am PHO5-Promoter in Saccharomyces cerevisiae, Ph.D. thesis, Universität München
  38. Svaren, J., Schmitz, J., and Hörz, W. (1994) EMBO J. 13, 4856–4862[Medline] [Order article via Infotrieve]
  39. Haguenauer-Tsapis, R., and Hinnen, A. (1984) Mol. Cell. Biol. 4, 2668–2675[Abstract/Free Full Text]