Portal de Periódicos da CAPES

Histone octamer function in vivo : mutations in the dimer-tetramer interfaces disrupt both gene activation and repression

1997; Springer Nature; Volume: 16; Issue: 9 Linguagem: Inglês

10.1093/emboj/16.9.2493

1460-2075

Maria Santisteban, Gina Arents, Evangelos N. Moudrianakis, M. Mitchell Smith,

RNA and protein synthesis mechanisms

Article1 May 1997free access Histone octamer function in vivo: mutations in the dimer–tetramer interfaces disrupt both gene activation and repression Maria Soledad Santisteban Maria Soledad Santisteban Department of Microbiology and University of Virginia Cancer Center, Box 441 Jordan Building, School of Medicine, University of Virginia, Charlottesville, VA, 22908 USA Search for more papers by this author Gina Arents Gina Arents Department of Biology, The Johns Hopkins University, Baltimore, MD, 21218 USA Search for more papers by this author Evangelos N. Moudrianakis Evangelos N. Moudrianakis Department of Biology, The Johns Hopkins University, Baltimore, MD, 21218 USA Search for more papers by this author M.Mitchell Smith Corresponding Author M.Mitchell Smith Department of Microbiology and University of Virginia Cancer Center, Box 441 Jordan Building, School of Medicine, University of Virginia, Charlottesville, VA, 22908 USA Search for more papers by this author Maria Soledad Santisteban Maria Soledad Santisteban Department of Microbiology and University of Virginia Cancer Center, Box 441 Jordan Building, School of Medicine, University of Virginia, Charlottesville, VA, 22908 USA Search for more papers by this author Gina Arents Gina Arents Department of Biology, The Johns Hopkins University, Baltimore, MD, 21218 USA Search for more papers by this author Evangelos N. Moudrianakis Evangelos N. Moudrianakis Department of Biology, The Johns Hopkins University, Baltimore, MD, 21218 USA Search for more papers by this author M.Mitchell Smith Corresponding Author M.Mitchell Smith Department of Microbiology and University of Virginia Cancer Center, Box 441 Jordan Building, School of Medicine, University of Virginia, Charlottesville, VA, 22908 USA Search for more papers by this author Author Information Maria Soledad Santisteban1, Gina Arents2, Evangelos N. Moudrianakis2 and M.Mitchell Smith 1 1Department of Microbiology and University of Virginia Cancer Center, Box 441 Jordan Building, School of Medicine, University of Virginia, Charlottesville, VA, 22908 USA 2Department of Biology, The Johns Hopkins University, Baltimore, MD, 21218 USA *E-mail: [email protected] The EMBO Journal (1997)16:2493-2506https://doi.org/10.1093/emboj/16.9.2493 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Within the core histone octamer each histone H4 interacts with each H2A–H2B dimer subunit through two binding surfaces. Tyrosines play a central role in these interactions with H4 tyrosines 72 and 88 contacting one H2A–H2B dimer subunit, and tyrosine 98 contacting the other. To investigate the roles of these interactions in vivo, we made site-directed amino acid substitutions at each of these tyrosine residues. Elimination of either set of interactions is lethal, suggesting that binding of the tetramer to both dimers is essential. Temperature-sensitive mutants were obtained through single amino acid substitutions at each of the tyrosines. The mutants show both strong positive and negative effects on transcription. Positive effects include Spt- and Sin-phenotypes resulting from mutations at each of the three tyrosines. One allele has a strong negative effect on the expression of genes essential for the G1 cell cycle transition. At restrictive temperature, mutant cells fail to express the CLN1, CLN2, SWI4 and SWI6 genes, and have reduced levels of CLN3 mRNA. These results demonstrate the critical role of histone dimer–tetramer interactions in vivo, and define their essential role in the expression of genes regulating G1 cell cycle progression. Introduction The fundamental unit of chromatin organization, the histone core complex, is maintained as an octamer by two types of protein–protein interactions (Eickbush and Moudrianakis, 1978). One class is the strong interaction between H3 and H4 within the tetramer, and between H2A and H2B within the dimer subunits. The second class is the weaker interaction between the [H3–H4]2 tetramer and the two [H2A–H2B] dimer subunits. These dimer–tetramer interactions are of considerable interest because they are proposed to play important roles in many aspects of chromosome function. DNA replication is one such candidate function. Chromatin assembly in vivo is thought to occur through the stepwise deposition of histone subunits onto newly replicated DNA. Biochemical experiments have shown that [H3–H4]2 tetramers are deposited first during assembly, and then nucleosome formation is completed with the later addition of [H2A–H2B] dimers (Worcel et al., 1978; Smith et al., 1984). Several in vitro studies with replication-coupled chromatin assembly systems are consistent with this two-step model (Dilworth et al., 1987; Almouzni et al., 1990). It has also been suggested that pre-existing histone octamers may dissociate into tetramers and dimers during replication (Jackson, 1990). Thus, these interactions could be significant both for the assembly of new nucleosomes, and for the rearrangement of old nucleosomes during replication. Several lines of evidence suggest that RNA transcription is also likely to depend on functional histone dimer–tetramer interactions. In vivo, histone [H2A–H2B] dimers are in dynamic exchange in the chromatin, part of which appears to depend on RNA polymerase activity (Jackson, 1990). Biochemical and physical analyses of nucleosomes from chromatin partially enriched for actively transcribing genes have indicated an altered structure consistent with the depletion of one of the [H2A–H2B] dimers (Baer and Rhodes, 1983; Locklear et al., 1990). Strong support for a role for dimer–tetramer interactions in transcription is also provided by genetic experiments in Saccharomyces cerevisiae where normal transcription depends on the balanced gene dosage of the histone gene sets. Changes in the normal ratio of the histone H2A–H2B gene sets relative to the H3–H4 gene sets, either too high or too low, can alter the selection of specific gene promoters (Clark-Adams et al., 1988) and bypass the need for positive transcriptional activation complexes, presumably by disrupting normal chromatin architecture (Hirschhorn et al., 1992). Recently, bypass mutants in transcriptional activation were identified in histones H3 and H4, and one class of such mutants were in the histone fold facing the [H2A–H2B] interface of the octamer (Kruger et al., 1995). A variety of studies in vitro are consistent with these results. For example, nucleosomes lacking one [H2A–H2B] dimer interact with RNA polymerase more strongly than the complete nucleosome core (Gonzalez et al., 1987) and are transcribed more efficiently (Gonzalez and Palacian, 1989). In addition, the histone-binding protein nucleoplasmin facilitates the binding of the transcription factors GAL4-AH, USF and SP1 to nucleosomal DNA through the sequential displacement of [H2A–H2B] dimers followed by H3–H4 tetramer displacement onto competing DNA (Chen et al., 1994). The role of dimer–tetramer interactions in transcription may not be confined to single nucleosomes. It has been proposed that depletion of [H2A–H2B] dimers may disrupt the ability of nucleosome arrays to fold into higher-order structures, and thereby relieve transcriptional repression (Hansen and Wolffe, 1994). Consistent with this hypothesis, changes in the dosage of the H2A–H2B gene sets in S.cerevisiae have been reported to alter the chromatin over certain chromosomal regions (Norris et al., 1988). Finally the altered stoichiometry of the core histones produced by imbalanced expression of the octamer subunits also impairs the fidelity of mitotic chromosome transmission, suggesting possible roles for these dimer–tetramer interactions in chromosome behavior during mitosis (Meeks-Wagner and Hartwell, 1986; Smith and Stirling, 1988). All of these results suggest that interactions between dimer and tetramer subunits of the protein core of the nucleosome may be critical for normal nuclear functions. Here we report the results of the first direct genetic analysis of these interactions and their roles in gene expression. Based on the X-ray crystal model of the histone octamer, we perturbed the dimer–tetramer interface with site-directed mutations and examined the effects of these mutations on transcription, cell cycle progression and general viability in the budding yeast S.cerevisiae. The phenotypes of these mutants show that disruption of the normal dimer–tetramer interactions can have both strong positive and negative effects on gene transcription, and define a new role for chromatin structure in the regulation of G1 progression in the cell division cycle. Results Selection of tyrosines for mutagenesis In choosing sites for directed mutations that would specifically alter the dimer–tetramer interfaces of the histone octamer, we focused on tyrosine residues in H4 based on the results of both solution and crystallographic studies. Early findings from solution physicochemical studies (Eickbush and Moudrianakis, 1978) demonstrated that the core histone octamer in solution behaves as a tripartite thermodynamic entity in reversible equilibrium with its subunits, one [H3–H4]2 tetramer and two [H2A–H2B] dimers. Analysis of the properties of this equilibrium led directly to the proposal that the centrally located tetramer interacts with the two flanking dimers via a limited number of contacts that include essential tyrosines. The reversible modulation of these interfacial contacts was proposed to be responsible for the functional cycles of chromatin. The importance of tyrosine contacts for histone octamer structure has since been supported by a wide variety of biophysical studies involving both spectroscopic analysis (Butler and Olins, 1982; Michalski-Scrive et al., 1982) and chemical modifications (Chan and Piette, 1982; Kleinschmidt and Martinson, 1984; Zweidler, 1992). The tripartite organization of the histone octamer and the involvement of tyrosines in the integrity of the structure were confirmed by crystallographic studies (Arents et al., 1991; Arents and Moudrianakis, 1993) that also identified a common architectural motif among the four core histones, termed the histone fold (Arents and Moudrianakis, 1993). The core octamer of the nucleosome contains four structural subunits, in three thermodynamic domains, creating a left-hand superhelical protein ramp (Arents et al., 1991). Proceeding into the octamer structure along the superhelical axis, the order of dimer subunits is [H2A1–H2B1], [H31–H41], [H32–H42] and [H2A2–H2B2]. Figure 1A shows a view of the histone octamer, looking down the superhelical axis, with the molecular 2-fold axis horizontal. In this view, the [H2A2–H2B2] dimer subunit (blue) and the [H32–H42] dimer subunit (H3 green and H4 white) comprise most of the visible structure, with a portion of H31 (green) also visible in the lower left quarter of the structure. Figure 1.Tyrosine residues in the dimer–tetramer interfaces of the core histone octamer. (A) View of the histone octamer looking down the superhelical axis with the molecular 2-fold axis horizontal from left to right. (B) Same view as (A), but with the second dimer subunit, [H2A2–H2B2], removed to expose the buried interface. (C) A view of the hexamer substructure model shown in (B), created by rotating 90° around the vertical axis, looking down the molecular 2-fold axis from the 'back'. (D) Same view as (C) with the other dimer subunit, [H2A1–H2B1], also removed leaving just the [H3–H4]2 tetramer. For (A–D), histone H4 is white, histone H3 is green and the H2A–H2B dimer subunits are blue. Histone H4 Tyr72 is yellow, Tyr88 is red and Tyr98 is black. (E) A ribbon representation illustrating the relationship of the two [H2A–H2B] dimer subunits with the [H32–H42] dimer subunit. In this view, the molecular 2-fold axis is vertical, running from bottom to top. The [H31–H41] dimer subunit has been removed from this picture for clarity. Interactions symmetrical with those shown for the [H32–H42] dimer subunit in the panel are formed by the [H31–H41] dimer subunit as well. The [H2A1–H2B1] dimer is dark blue, the [H32–H42] dimer is white and the [H2A2–H2B2] dimer is light blue. The H4 tyrosines Y72, Y88 and Y98 (white) and the H2B tyrosine Y83 (blue) are modeled with the tyrosyl ring shown. Download figure Download PowerPoint The interfaces between the centrally located tetramer and the flanking dimers are formed from both fold and non-fold elements and contain a number of tyrosine residues in two distinct groupings. In the first group, tyrosine 72 (Y72) and tyrosine 88 (Y88) of H4 interact with one of the flanking [H2A–H2B] dimers, while tyrosine 98 (Y98) interacts with the other. This is illustrated in Figure 1B where the [H2A2–H2B2] dimer has been removed, exposing Y98 (black) of H41 which interacts with that dimer. On the other hand, Y72 (yellow) and Y88 (red) of H41 are only partially visible because they are in contact with the remaining [H2A1–H2B1] dimer in the figure. The back view shown in Figure 1C reveals the interactions of the other H42 subunit. In this view, it can be seen that removal of the [H2A2– H2B2] dimer has exposed Y72 (yellow) and Y88 (red) of H42 where they would normally make contact with that dimer. The Y98 residue of H42 is not visible in this view because it is in contact with the remaining [H2A1–H2B1] dimer and covered by that dimer. The symmetry of these interactions is illustrated in Figure 1D in which just the [H3–H4]2 tetramer is shown. Thus, one of the major features of the dimer–tetramer interactions in the histone octamer is that each of the H4 molecules in the [H3–H4]2 tetramer touches both [H2A–H2B] dimer subunits across the dimer–tetramer interfaces. Further details of these interactions are depicted in the ribbon representation shown in Figure 1E. In this view of the octamer, the [H31–H41] dimer subunit has been removed from the model for clarity, leaving the dimers [H2A1–H2B1], [H32–H42] and [H2A2–H2B2]. The two groupings of interactions described above are apparent. In the first group, where [H2A2–H2B2] contacts the tetramer, the histone fold part of H4 interacts with the histone fold part of H2B to form a four-helix bundle. Although a number of residues from both H2B and H4 contribute to this interface, one of the most striking features of the interface is the large hydrophobic domain generated by the contacts between three tyrosines: Y72 and Y88 from H4, and Y83 from H2B. These three tryrosines form a cluster in which the planes of the side chains are approximately perpendicular to each other. In the second group, the contact between H42 and [H2A1–H2B1] arises from a non-fold, extended chain arm of H4 running roughly parallel to a non-fold, extended chain arm of H2A. Most of these contacts arise from main chain interactions. Within this interface, however, Y98 of H42 is exceptional due to the insertion of the large tyrosyl ring into a cleft in the dimer surface. Symmetrical interactions occur with the other [H31–H41] dimer subunit not shown in this representation. Thus, there are two distinct classes of interactions between histone H4 and the [H2A–H2B] dimers, and both biophysical and structural studies have identified tyrosine residues as key participants in each type of interaction. Construction and properties of mutants Based on these structural considerations, we constructed mutant histone H4 alleles with amino acid substitutions at positions 72, 88 and 98. In the initial series of constructs, the tyrosine codons at these positions were changed to glycine codons, creating mutant alleles hhf1-36 (Y72G), hhf1-37 (Y88G) and hhf1-38 (Y98G). The wild-type gene and each of these new mutant alleles were integrated at the LEU2 locus on chromosome III, providing the only source of histone H4 in their respective cells (see Materials and methods). These ectopic integrations produced a set of strains that were isogenic except for their histone H4 alleles. As can be seen in Figure 2, all three mutants have detectable growth phenotypes. Both hhf1-36 (Y72G) and hhf1-37 (Y88G) cells grow somewhat more slowly than the control HHF1 strain at 28°C, and are temperature sensitive (Ts−) for growth at 37°C. Of the two, hhf1-36 is the more severe allele. The third allele, hhf1-38 (Y98G), is lethal. Cells with hhf1-38 integrated at the LEU2 locus could be propagated only in the presence of plasmid pMS329, a single copy URA3 plasmid that expresses the wild-type HHF1 histone H4 gene. When this strain was transferred to medium contain 5-fluoroorotic acid (5-FOA) to select for cells that had lost the URA3 pMS329 plasmid, it failed to give colonies, indicating that hhf1-38 alone does not support cell growth. This result is consistent with previous deletion studies in which an H4 mutant with a carboxy-terminal truncation of residues 100–102 was found to be viable, whereas a truncation of residues 97–102 was dead (Kayne et al., 1988). Together, these results suggest an essential role for Y98 in H4 function. Figure 2.Conditional growth of histone H4 tyrosine mutants. Growth is shown on YPD plates for wild-type and four histone H4 mutant strains after incubation for 5 days at 28 or 37°C. All the strains were isogenic except for the histone H4 alleles: WT (HHT1), Y98W (hhf1-40), Y98H (hhf1-39), Y88G (hhf1-37) and Y72G (hhf1-36). Download figure Download PowerPoint We next constructed additional alleles with point mutants at position 98 by making more conservative changes, replacing tyrosine with either histidine (hhf1-39, Y98H) or tryptophan (hhf1-40, Y98W). Figure 2 shows that hhf1-40 (Y98W) cells grow normally at both 28 and 37°C. In fact, hhf1-40 mutants are wild-type for all phenotypes tested to date. In contrast, hhf1-39 grows poorly at 28°C, and is Ts− lethal at 37°C (Figure 2). Thus, the function of histone H4 is highly sensitive to substitutions at position 98: tyrosine or tryptophan support wild-type function, histidine is only partially functional and glycine is lethal. The phenotypic pattern of the single substitution mutants paralleled the structural interactions of H4 with the [H2A–H2B] dimer subunits. Tyr98 is the sole site of tyrosine interaction with one of the [H2A–H2B] dimers in the core octamer, whereas Y72 and Y88 both interact simultaneously with the other. Based on this model, we predicted that the Y72G,Y88G double mutant might be lethal by completely disrupting the tyrosine interactions with the second [H2A–H2B] dimer. In order to test this hypothesis, we constructed the double substitution allele, hhf1-41. Strains with hhf1-41 integrated at the LEU2 locus could be grown only in the presence of pMS329 carrying the wild-type HHF1 gene, and loss of pMS329 resulted in cell death. Thus, this double substitution allele is lethal as predicted. Chromatin structure of H4 tyrosine mutants The overall chromatin structure of the viable H4 tyrosine mutants was examined by micrococcal nuclease digestions. The result for hhf1-36, which has the most severe phenotype, is shown in Figure 3. At both permissive and restrictive temperatures no differences were detected in the oligonucleosome ladders produced by partial micrococcal nuclease digestion. Similar results were obtained for hhf1-37 and hhf1-38 (data not shown). These results indicate that there is no gross defect in the general organization of nucleosomes in the H4 tyrosine mutants at either the permissive or restrictive temperature, and are similar to those obtained for the overall chromatin structure of histone H2A and H2B gene dosage mutants (Norris et al., 1988). Figure 3.Micrococcal nuclease digestion of chromatin. The ethidium bromide staining of the nuclease-digested chromatin is shown for wild-type (HHT1) and Y72G (hhf1-36) mutant strains, both for cells grown at 28°C and after 3 h at 37°C. Similar nucleosome ladders can be observed for the mutant, the wild-type strain and the mutant at 37°C. Download figure Download PowerPoint HTA1–HTB1 synthetic dosage phenotypes The pattern of viability of the single and double glycine substitution mutants was consistent with the disruption of interactions between histone H4 and the [H2A–H2B] dimers. We reasoned that if point mutations in the H4 tyrosines perturb these interactions, then altered dosage of the H2A and H2B genes might significantly affect the growth phenotype of the mutants. To test the effect of reduced H2A and H2B gene dosage, we deleted the HTA1–HTB1 gene set encoding these proteins (Hereford et al., 1979) by one-step gene disruption (see Materials and methods). Decreased dosage of HTA1–HTB1 did not alter the temperature-dependent growth characteristics of either the wild-type or mutant H4 strains (data not shown). To test the effect of overexpression of histones H2A and H2B, strains expressing hhf1-36 (Y72G), hhf1-37 (Y88G) or hhf1-39 (Y98H) were transformed with a high copy YEp24-derived plasmid expressing the HTA1–HTB1 gene set and examined for their growth at restrictive and semi-permissive temperatures. Overexpression of HTA1–HTB1 had no detectable effect on hhf1-37 (Y88G) at either 34 or 37°C (data not shown). However, as can be seen in Figure 4, at a semi-permissive temperature of 34°C, increased HTA1–HTB1 gene dosage strongly inhibited the growth of hhf1-36 (Y72G) cells, compared with expression of YEp24 alone. A slight inhibition was also detected with hhf1-39 (Y98H) cells (Figure 4). These synthetic dosage phenotypes of high-copy HTA1–HTB1 suggest a functional genetic interaction between [H2A–H2B] dimers and the Y72G and Y98H mutant histone H4 proteins (Kroll et al., 1996). Figure 4.Overexpression of H2A–H2B decreases the growth rate of hhf1-36 (Y72G) and hhf1-39 (Y98H) strains. Strain Y72G (hhf1-36) was transformed with either YEp24 (line 1) or HTA1–HTB1 in YEp24 (line 2). Transformants were grown in SDC-URA liquid medium until log phase. Serial dilutions of cell suspensions, ranging from 1×105 down to 10 cells, were spotted onto SDC-ura plates and incubated at 34°C for 3 days. Lines 3 and 4 show the growth of Y98H (hhf1-39) cells transformed with either YEp24 or HTA1–HTB1 in YEp24. Lines 5 and 6 show the growth of the wild-type HHF1 cells transformed with either YEp24 or HTA1–HTB1 on YEp24. Download figure Download PowerPoint H4 tyrosine mutants are Spt− Alterations in the relative ratio of the H3–H4 and H2A–H2B histone gene pairs can have strong effects on a variety of cell functions, including mitotic chromosome transmission (Meeks-Wagner and Hartwell, 1986; Smith and Stirling, 1988) and mRNA transcription (Clark-Adams et al., 1988; Hirschhorn et al., 1992). These effects of histone gene stoichiometry are presumably mediated by changes in the composition of the chromatin through alterations in dimer–tetramer interactions. If the molecular defect in the H4 tyrosine mutants is in dimer–tetramer interactions, then these mutants should exhibit the same phenotypes as strains with altered gene ratios. To test this prediction, we examined the ability of the mutations to suppress Ty1 solo δ insertions. Insertions of Ty1 solo δ elements in the 5′ regions of the HIS4 (his4-912δ) or LYS2 (lys2-128δ) genes cause alterations in their transcription initiation start sites leading to non-functional transcripts. Mutations in SPT genes suppress these defects by restoring transcription of the affected genes (Simchen et al., 1984; Winston et al., 1984; Fassler and Winston, 1988). Either overproduction or underproduction of histone dimer gene sets produces an Spt− phenotype and restores gene expression (Clark-Adams et al., 1988). As shown in Figure 5, hhf1-36 (Y72G), hhf1-37 (Y88G) and hhf1-39 (Y98H), but not hhf1-40 (Y98W) or wild-type HHF1, were able to suppress his4-912δ and lys2-128δ mutations and support colony formation on SDC-HIS and SDC-LYS plates. All three alleles were strong suppressors of his4-912δ and weaker suppressors of lys2-128δ. The hhf1-39 (Y98H) allele was the weakest suppressor of the three, and growth in medium lacking lysine was not observed until at least 4 days of incubation. Thus, substitution mutations in these H4 tyrosines each result in cells that display Spt− phenotypes. Figure 5.Suppression of his4-912δ and lys2-128δ by histone H4 mutations. The wild-type and histone H4 mutant alleles were introduced by ectopic integration into a his4-912δ lys2-128δ strain and the growth was assayed on synthetic media lacking histidine and lacking lysine. After 5 days of incubation at 28°C, only the strains expressing hhf1-36 (Y72G), hhf1-37 (Y88G) and hhf1-39 (Y98H) showed significant growth. Download figure Download PowerPoint Histone H4 mutants are Sin− A number of histone mutations, including altered stoichiometry of the dimer–tetramer gene sets, suppress the loss of transcriptional activation caused by mutations in genes encoding the SWI–SNF complex proteins, rendering regulated target genes SWI–SNF independent (Sin−) (Happel et al., 1991; Malone et al., 1991; Hirschhorn et al., 1992; Kruger et al., 1995). Since the H4 tyrosine mutations targeted the dimer–tetramer interface, we anticipated that they might have Sin− phenotypes. The SUC2 gene encodes the enzyme invertase necessary for utilization of sucrose and raffinose as carbon sources. The transcriptional activation of SUC2 depends on the function of the SWI–SNF complex to remodel the repressive chromatin structure at the SUC2 promoter (Hirschhorn et al., 1992). Mutants in SWI–SNF complex genes, such as snf2 or snf5, fail to remodel promoter chromatin structure and do not activate SUC2 transcription. Deletion of the histone HTA1–HTB1 locus generates an active chromatin structure over the SUC2 promoter, thus bypassing the requirement for SWI–SNF complex function and activating transcription in swi and snf mutants (Hirschhorn et al., 1992). We tested the tyrosine histone H4 mutants for Sin− phenotypes by knocking out SNF2 in the mutants using a snf2::URA3 disruption allele. As shown in Figure 6A, none of the H4 alleles suppressed the growth defect of the snf2 disruption on raffinose plates. Consistent with these results, there was also no increase in SUC2 mRNA levels in the double mutants (Figure 6B). Figure 6.Histone H4 mutations do not suppress the defect in SUC2 transcription due to a snf2::URA3 disruption. (A) Wild-type and histone H4 Tyr mutant strains, in SNF2 and snf2 backgrounds, were grown on complete media and then replica-plated onto media containing glucose or raffinose as carbon sources. Although all the strains were able to grow in the medium supplemented with glucose, only the SNF2 strains grew in the raffinose medium. (B) mRNA levels of SUC2 in the wild-type and histone H4 Tyr mutant cells grown under repressive (2% glucose) or inducing (0.02% glucose) conditions analyzed by Northern blot. Northern blots were probed with SUC2 and ACT1 probes. Download figure Download PowerPoint We next examined the effect of the histone mutations on INO1 gene expression since it also requires the function of the SWI–SNF complex for activation (Peterson et al., 1991). In contrast to the results at SUC2, hhf1-36 (Y72G) and hhf1-37 (Y88G) relieve the growth defect of snf2 mutants on Ino− plates (Figure 7A). The hhf1-39 (Y98H) allele also relieves repression, but more weakly (Figure 7A). Consistent with its Spt+ phenotype, no suppression was observed for hhf1-40 (Y98W). These results were confirmed by Northern blot analysis of INO1 mRNA levels (Figure 7B). In the SNF2 wild-type strain, the induction of the INO1 mRNA by low inositol was ∼14-fold (lanes 1 and 2). The isogenic snf2 mutant shows about a 35-fold reduction in INO1 transcription under derepressing conditions (lane 6). Under the same conditions, the snf2 double mutants with either hhf1-36 or hhf1-37 increased INO1 mRNA levels ∼26-fold over the snf2 single mutant (lanes 9 and 11). A more modest increase of 16-fold was observed for the hhf1-39 mutant (lane 13). Increased INO1 transcription was also observed under repressing conditions in the three histone mutant strains, resulting in only about a 1.5-fold difference between repressed and induced mRNA levels. Therefore, the histone H4 tyrosine mutations not only suppress mutations in snf2, but also relieve the repression caused by high levels of inositol in the medium (Hirsch and Henry, 1986). Figure 7.Histone H4 mutations suppress the inositol auxotrophy due to a snf2::URA3 disruption. (A) Growth of the wild-type H4 and mutant strains is shown for media lacking inositol or supplemented with 100 μM of myo-inositol. All the strains carried the snf2::URA3 disruption. Wild-type HHT1 and hhf1-40 (Y98W) cells required inositol to grow, whereas hhf1-36 (Y72G), hhf1-37 (Y88G) and hhf1–39 (Y98H) cells could grow without inositol in the medium. (B) Northern analysis of the INO1 for the wild-type and isogenic H4 mutant strains in the snf2 and SNF2 backgrounds. mRNA levels are shown for cells grown under either repressive (100 μM myo-inositol) or inducing conditions (10 μM myo-inositol). Download figure Download PowerPoint We then examined the effect of the histone mutations on the positive regulators of INO1. The products of the INO2 and INO4 genes form a transcription factor complex that binds to the INO1 promoter (Ambroziak and Henry, 1994; Ashburner and Lopes, 1995). As shown in Figure 8, the hhf1-37 allele did not suppress an ino2::TRP1 disruption for growth on media lacking inositol. Identical results were obtained with the other mutant alleles of H4 (data not shown). Consistent with this growth pattern, the levels of INO1 mRNA in ino2::TRP1 mutants were low both in wild-