The Biochemical and Phenotypic Characterization of Hho1p, the Putative Linker Histone H1 of Saccharomyces cerevisiae *

on the isola-tion and definitive identification of histone H1 in Saccharomyces cerevisiae It however, shown that the yeast HHO1 gene codes for a predicted protein homologous to H1 of higher eukaryotes there is no biochemical evidence that shows that Hho1p is, indeed, yeast histone H1. We showed that purified recombinant Hho1p (rHho1p) has electrophoretic and chromatographic properties similar to linker histones. The protein forms a stable ternary complex with a reconstituted core di-nucleosome in vitro at molar rHho1p:core ratios up to 1. Reconstitution of rHho1p with H1-stripped chromatin confers a kinetic pause at ; 168 base pairs in the micrococcal nuclease digestion pattern of the chromatin. These results strongly suggest that Hho1p is a bona fide linker histone. We deleted the HHO1 gene and

The basic structural unit of eukaryotic chromatin is the nucleosome core, composed of an octameric complex of two copies of each of the core histones H2A, H2B, H3, and H4 onto which 146 bp 1 of DNA is spooled as 1.75 turns of a left-handed superhelix (1)(2)(3)(4), and is continued to adjacent nucleosome cores through a variable length of linker DNA. In a nucleosome the histone H1 is associated with the outside of the core and protects 168 bp or two full superhelical turns of DNA from MNase cleavage (5,6). Histone H1 was first identified in 1951 as a lysine rich "subsidiary" histone in calf thymus (7) and was subsequently shown to be an extended family of histone isotypes or variants present in a wide variety of eukaryotes (8).
Although the structural role of the nucleosome core and the spatial placement of the core histones within the octamer are well established (1,3,4,9), the function and the precise location of the linker histone has been more elusive (reviewed in Ref. 10).
Electron microscopic and hydrodynamic studies have shown that histone H1 is required for the full salt-dependent condensation of chromatin in vitro (11)(12)(13). These observations have led to the proposal that the function of the linker histone is the partial neutralization of the negatively charged DNA backbone, allowing the close approach of the internucleosomal linker DNA that would normally coloumbically repel in the fully compacted 30 nm fiber (for a review, see Ref. 14 and references cited therein). This role of H1 is supported by the observed reduction in the size of mitotic chromatids and nuclear volume concurrent with the appearance of histone H1 during the mid-blastula transition of the developing Drosophila embryo (15). It is also consistent with the approximately 2-fold expansion of the Tetrahymena micro-or macronucleus in the absence of the four micronuclear-specific micLH polypeptides or macronuclear-specific H1 protein, respectively (16). This implied role of H1 in chromatin condensation is dynamic and is modulated by the cell cycle-dependent reversible phosphorylation of H1 by p34 cdc2 (reviewed in Refs. 17 and 18). This modification presumably allows the interaction of additional factors such as the structural maintenance of chromosomes (SMC) class of proteins with chromatin, affecting full condensation into the mitotic chromosome (see Ref. 19 for a review).
There has been no definitive identification of a linker histone in Saccharomyces cerevisiae, although some studies have reported indirect evidence for the existence of an H1-like protein in yeast (20 -23). However, the absence of a direct biochemical identification of the histone led some investigators to suggest that it is absent in yeast (16). This proposed lack is unusual, since eukaryotes that phylogenetically diverge both before (such as Tetrahymena) and after (such as Psammechinus) yeast have been shown to contain linker histones.
The sequencing of chromosome XVI in yeast (24) revealed the presence of a predicted open reading frame (YPL127C) with regions of significant sequence homology to the H1 histones of higher eukaryotes (25,26). Theoretical model building 2 sug-gested that these regions of homology may assume a structural conformation similar to that of the known single-winged helix structure of the chicken H5 (27) and H1.11L (28) globular domains. Ushinsky et al. (26) have shown that the predicted open reading frame is functional and that a fusion of the gene product with fluorescent green protein is located in the nucleus. These results contributed to the assignment of YPL127C as the HHO1 gene which was proposed to encode the yeast linker histone H1 (Hho1p) (25,26). There are no other genes in the yeast genome that code for proteins with significant sequence similarities to H1 (25).
In this study, we biochemically address the assignment of Hho1p as the yeast H1 linker histone. In particular, we investigated the ability of this protein to stably associate with a reconstituted core di-nucleosome and to confer an MNase kinetic pause at approximately 168 bp in H1-stripped chromatin in vitro. The phenotype of a yeast strain in which the HHO1 gene has been deleted was also studied. We specifically investigated the effect of the absence of Hho1p on cell viability, growth rate, mating efficiency, basal transcription levels, telomeric repression, sporulation efficiency, chromatin structure, and Sin and Spt phenotypes.

EXPERIMENTAL PROCEDURES
Construction of Plasmids and Yeast Strains-Genomic DNA was isolated from yeast strain FY24 as described by Zhu et al. (29). A 1500-bp region between positions 308,523 and 310,024 on yeast chromosome XVI (24), incorporating the entire HHO1 open reading frame and 303 bp upstream and 420 bp downstream, was amplified with Taq DNA polymerase and template mismatched primers to introduce EcoRI and XbaI sites at the fragment ends, as described (30). The recovered fragment was digested with EcoRI and XbaI and ligated into the EcoRI and XbaI sites of the plasmid pRS413 (31) lacking the 2688-bp PmlI-EcoRV fragment. The resulting plasmid was denoted pHHO1, and the sequence of the insert confirmed by nucleotide sequencing. This plasmid was further modified by introducing a 1185-bp fragment containing the HIS3 gene into the BamHI and SphI sites of the HHO1 sequence, and the proper insertion of the HIS3 gene in the plasmid, denoted pHHO1-HIS3, was confirmed by restriction enzyme analysis. This manipulation removes the sequence coding for amino acids 63-187 of Hho1p, including the entire primary globular domain C-terminal to helix I, the lysine-rich interglobular region, and the assigned secondary globular domain N-terminal of helix II (25).
An Escherichia coli expression vector, pET20-HHO1, was constructed by ligating an NdeI/XhoI restricted 789-bp fragment containing the HHO1 coding sequence into the NdeI/XhoI sites of pET20b(ϩ) (Novagen). This construct codes for the entire Hho1p protein and introduces six histidine residues at the Hho1p C terminus.
The HHO1 deletion strain was constructed by electroporating strain YPH500⌬L (32) with the 1.5-kilobase pair EcoRI/XbaI fragment from pHHO1-HIS3. The recombination of the partially replaced HHO1 open reading frame to the proper locus in the recovered histidine prototrophs was confirmed by polymerase chain reaction and Southern hybridization. This strain was denoted YHGP101.
Genetic crosses and tetrad dissections were performed as described in Ref. 33. A complete list of the yeast strains used in this study and their relevant genotypes is shown in Table I.
Purification of Recombinant Hho1p-A 1-liter culture of E. coli strain BL21(DE3) containing pET20-HHO1 was induced with 0.4 mM isopropyl-1-thio-␤-D-galactopyranoside for 2 h, and the cells were recovered in 10 ml of 20 mM imidazole, 200 mM NaCl, and 20 mM Tris⅐Cl (pH 7.9) (binding buffer). The cells were lysed by sonication (20% duty cycle at setting 6 on a Branson model 450 sonicator), and the cellular debris were pelleted by centrifugation (18,000 rpm for 20 min at 4°C in a Sorval SS34 rotor). The recovered supernatant was loaded onto a nickel-agarose column (1-ml bed volume), equilibrated in binding buffer. The column was washed with 25 ml of binding buffer, followed by 15 ml of 40 mM imidazole, 200 mM NaCl, and 20 mM Tris⅐Cl (pH 7.9). The recombinant protein was eluted from the column with 300 mM imidazole, 200 mM NaCl, and 20 mM Tris⅐Cl (pH 7.9), and 500-l fractions were collected. Fractions enriched in recombinant Hho1p were identified by SDS-PAGE electrophoresis, pooled, and loaded onto a CM-Sephadex column (1-ml bed volume) equilibrated in phosphate buffer (10 mM sodium phosphate (pH 7.0), 0.25 mM EDTA, and 0.25 mM phenylmethylsulfonyl fluoride) containing 200 mM NaCl. The column was washed with 10 ml of the same buffer and developed with 10 ml of a 200 -1000 mM NaCl linear gradient in phosphate buffer. Fractions (500 l) containing pure recombinant Hho1p were identified by SDS-PAGE, pooled, and concentrated, and the buffer was replaced with 10 mM sodium phosphate (pH 7.0), 0.25 mM EDTA, and 0.25 mM phenylmethylsulfonyl fluoride with a P10 centricon device (Amicon). The samples were stored frozen at Ϫ70°C. The rHho1p concentration was determined by the absorbance at 230 nm using an E 1% of 18.5 (34).
Phenotypic Characterizations-The 5-fluoroorotic acid (5-FOA) sensitivity of a strain was determined by diluting 1 ml of an overnight culture to an A 600 of 1.0, and 10 l of this culture, serially diluted 10-fold, was applied to a complete synthetic medium (CSM, 6.7 g/liter yeast nitrogen base without amino acids, 20 g/liter glucose, 0.7 g/liter amino acid supplement (Bio 101), and 20 g/liter bacto agar) plate lacking the appropriate amino acid in the absence or presence of 0.1% (w/v) 5-fluoroorotic acid (Jersey Laboratories). The plates were incubated at 30°C for 2-3 days.
The sporulation efficiency of a strain was determined by inoculating a single diploid colony into 10 ml of pre-sporulation medium (0.25% (w/v) glucose, 1% (w/v) potassium acetate, 0.6% (w/v) yeast nitrogen base without amino acids, 0.5% (w/v) yeast extract, and 0.5% (w/v) bacto peptone) and incubating the culture with vigorous shaking at 30°C to an A 600 of 0.5-0.6. The cells were pelleted, washed in sporulation medium (1% (w/v) potassium acetate, all auxotrophic supplements at  33), resuspended in 5-ml volumes of sporulation medium to an A 600 of 0.2, and incubated at 22°C for 5 days. The sporulation efficiency of a culture was quantified by counting the number of tetrads relative to the total number of cells and tetrads in a hemocytometer.
The ␤-galactosidase assays were performed as described in Ref. 35.
Chromatin Reconstitution and Nuclease Digestions-The 390-bp 32 Plabeled fragment (10 ng) containing a tandem dimer of the sea urchin 5 S rRNA gene was incubated with 40 g of H1-stripped HeLa chromatin (a gift from M. Vignali and T. Owen-Hughes) in a 100-l volume of reconstitution buffer (10 mM Tris⅐Cl (pH 8.0), 0.25 mM EDTA and 0.25 mM phenylmethylsulfonyl fluoride) containing 1000 mM NaCl at 37°C for 30 min. The ionic strength was reduced to 800 and 600 mM NaCl by the sequential addition of reconstitution buffer at 30-min intervals at 37°C. Where rHho1p was added, the sample was split into aliquots following the dilution to 600 mM NaCl, and rHho1p was added in the amounts indicated in the text. The dilution of fractions in the presence or absence of rHho1p was continued to 400 mM NaCl after 30 min, followed by a stepwise 2-fold dilution with reconstitution buffer at 37°C at 30-min intervals to a final concentration of 12.5 mM NaCl. Aliquots (20 l) were run on a 0.7% (w/v) agarose gel in 0.5ϫ TBE at 100 V, and the gel was dried and autoradiographed. The reconstitution of H1stripped HeLa chromatin with Hho1p was performed exactly as described, except that the 5 S rRNA gene dimer fragment was omitted, and rHho1p was added at a molar ratio of 1:1 (Hho1p:core). This and a similarly treated sample lacking Hho1p were digested with concentrations of MNase indicated in the text. The digestion products were purified and electrophoresed on an 8% (w/v) polyacrylamide gel in 1ϫ TBE at 150 V. The preparation of nuclei, digestion of chromatin and free DNA, and the primer extension of the recovered DNA template were performed as described previously (30).

RESULTS
Is Hho1p the Yeast Linker Histone H1?-We investigated the assigned identity of Hho1p as a linker histone in yeast by studying the properties of a recombinant Hho1p (rHho1p). Histidine-tagged rHho1p was overexpressed in E. coli cells and isolated from the sonicated cell lysate by passage over a nickelagarose column (Fig. 1A). The identity of the band indicated as rHho1p ( Fig. 1A) was confirmed by its absence in identically performed mock isolations from E. coli containing only the vector without the HHO1 coding sequence. Although the amino acid sequence of Hho1p predicts a molecular mass of approximately 28 kDa, the protein migrates at a size equivalent to approximately 33 kDa during SDS-PAGE ( Fig. 1, A and B). This anomalously slow electrophoretic behavior closely resembles that of linker histones from other species (36) and is likely due to the basicity of the protein, which has a predicted isoelectric point of 10.2. The recovered rHho1p was next purified from contaminating nickel-agarose binding E. coli proteins by passage over a cation exchange resin, eluted by a linear salt gradient. Recombinant Hho1p eluted off the cation exchange column at approximately 700 mM NaCl (Fig. 1B). This chromatographic trait of rHho1p is also very similar to that of linker histones from higher eukaryotes (37). These properties, however, reflect only the amino acid composition of the protein and not its proposed function as a linker histone. We therefore tested the ability of the purified recombinant protein to conform to two common characteristics of a linker histone: (i) to form a stable ternary complex with a reconstituted nucleosome core in vitro, and (ii) to produce a kinetic pause at approximately 168 bp in the MNase digestion pattern of H1-stripped native chromatin.
rHho1p Forms a Stable Ternary Complex with a Reconstituted Core Di-nucleosome in Vitro-Several investigators have shown that both the full-length and globular domain of recombinant and native linker histones can be reconstituted with nucleosome cores in vitro (13,38). Since the proposed role of histone H1 is the neutralization of the charge of the internucleosomal linker DNA, we chose to investigate the association of rHho1p with a reconstituted core di-nucleosome, which con-tains a short length of linker DNA and is expected to resemble a natural H1 substrate more closely. A 390-bp radiolabeled fragment containing a tandem repeat of sea urchin 5 S DNA (39) was reconstituted into a di-nucleosome in the presence of a range of rHho1p concentrations, and the reconstitution products were electrophoretically resolved on an agarose gel ( Fig.  2). Addition of rHho1p to the core di-nucleosome resulted in the formation of two slower migrating species. The faster migrating of these two species (N1) appears at lower molar Hho1p input ratios and decreases toward higher input ratios, concurrent with an increase in the slower migrating complex (N2). At a molar input ratio of one molecule of rHho1p per core, essentially all of the core di-nucleosome exists as the N2 species. At higher molar input ratios or at ionic strengths in excess of approximately 100 mM NaCl, the reconstitute aggregates and does not enter the gel matrix. This result clearly demonstrates that rHho1p forms a stable ternary complex with a reconstituted core di-nucleosome. The apparent conversion of the N1 to the N2 species as a function of rHho1p input suggests that at lower rHho1p ratios a single molecule of rHho1p binds to the core di-nucleosome forming the N1 complex. At higher rHho1p: core ratios a second molecule of rHho1p binds to the N1 complex, resulting in the appearance of the N2 complex.
MNase Digestion of H1-stripped Chromatin Reconstituted with rHho1p Shows a 168-bp Kinetic Pause-The presence of histone H1 in chromatin protects an additional 10 bp of DNA on either side of the nucleosome core from exonucleolytic MNase digestion, resulting in the appearance of a digestion intermediate of approximately 168 bp prior to trimming to 146 bp and subsequent sub-nucleosomal length fragments (5, 6). The appearance of the kinetic pause at ϳ168 bp was shown to require the basic amino acid residues Lys-40 and Arg-42 be- tween helix I and helix II and Lys-52 on helix II (40) in the proposed secondary DNA-binding site of the globular domain of H5 (27). The alignment of the proposed primary globular domain of Hho1p (25) with that of several H1 isotypes from different species (Fig. 3) shows that all three of the predicted DNA-binding residues in the secondary DNA-binding site (27) that are highly conserved among the different H1 proteins are also conserved in yeast Hho1p. Also, at least two of the three most conserved basic residues in the proposed primary DNAbinding site (27) are conserved in Hho1p, as is also the case for the pea H1.
The ability of the proposed primary globular domain of yeast Hho1p (25) to assume a secondary structure similar to that found in the crystal structure of chicken H5 (27) or NMR structure of H1.11L (28) was investigated using the nnPredict structure prediction algorithm of Kneller et al. (41). It is clear from Fig. 3B that the algorithm correctly predicts the presence of extended ␣-helical segments in regions where ␣-helixes were identified in the crystal and NMR structures of the H5 and H1.11L globular domains, respectively (27,28). Similar helical sections are also predicted in the corresponding regions of the Drosophila H1 sequence (see Fig. 3B), previously identified as a linker histone (42). In the case of yeast Hho1p, ␣-helical stretches are predicted in the regions of the aligned sequence similar to that of H5 and H1.11L (see Fig. 3B). This result strongly suggests that the proposed globular domain of Hho1p can assume a secondary structure similar to that of a linker histone, in agreement with the study of Baxevanis and Landsman. 3 To test whether the conservation of the proposed DNA-binding basic amino acid residues and the predicted secondary structural conservation of the assigned globular domain of Hho1p will confer a nucleosome core binding specificity to Hho1p similar to that of histone H1, H1-stripped HeLa chromatin was reconstituted with rHho1p and digested with various concentrations of MNase. A control sample was treated and digested identically, except that it was not reconstituted with rHho1p. The purified digestion products were electrophoresed on a polyacrylamide gel, shown in Fig. 4. MNase digestion of H1-stripped chromatin in the absence of rHho1p results in a limit digestion product that resolves at approximately 146 bp, corresponding to the 1.75 turns of nucleosomal DNA in the core particle. In the case of chromatin reconstituted with rHho1p, a fragment with a length of approximately 168 bp is clearly visible (indicated by the arrowhead in Fig. 4) apart from the fragment resolving at approximately 146 bp. This result clearly shows that rHho1p extends the protection of nucleosomal DNA to two full superhelical turns, conferring a structural stability to a nucleosome core analogous to that caused by a linker histone.
Yeast Lacking Hho1p Are Viable and Have Normal Growth and Mating Properties-We constructed a haploid yeast strain (YHGP101) in which the single HHO1 coding sequence was partially replaced with the HIS3 selectable marker. The absence of the poly(A) ϩ HHO1 mRNA transcript, and therefore the Hho1p protein, was confirmed in the YHGP101 strain by Northern analysis (data not shown). There was no detectable difference in the growth rate of the hho1⌬ strain compared with the WT strain. Similarly, yeast cells lacking Hho1p mated as efficiently as WT cells, suggesting that both silent mating type loci and the appropriate cell type-specific genes are properly repressed in the absence of Hho1p.
Repression of Basal Transcription-The repression of basal transcription by nucleosomes is well established (reviewed in Ref. 43). Several studies have also shown that the basal transcription of chromatin templates in vitro is reduced in an H1dependent manner (44 -46). We investigated the possible involvement of Hho1p in basal transcriptional repression in situ using a reporter gene that consists of a minimal PHO5 promoter fused to the URA3 coding sequence (47). Cells that harbor this plasmid do not express URA3 and are resistant to the drug 5-FOA which kills cells that express the URA3 gene product. If basal transcription is elevated, however, the cells become sensitive to 5-FOA. The transcriptional activity of this episomal reporter gene was tested in the WT and hho1⌬ strains by measuring the growth of transformants in the presence of 5-FOA (Fig. 5A). At equivalent dilution of cells, comparable numbers of WT and hho1⌬ cells survive in the presence of 5-FOA. This result demonstrates that in contrast to the core histones (43), Hho1p does not appear to be involved in the general repression of basal polymerase II transcription in situ. This finding does not, however, exclude the possible involvement of Hho1p in specialized regulatory mechanisms at specific genes.
Hho1p Is Not Required for Telomeric Repression-Several studies have demonstrated the transcriptionally repressive nature of heterochromatin-like structures at yeast telomeres and at the silent mating-type loci (Ref. 48 and references cited therein). The involvement of chromatin in the establishment of these repressive structures was clearly shown by the requirement for the N terminus of histone H4 for full repression at both loci (49) and histone H3 for repression at the telomeric ends and at a partially crippled silent mating-type locus (50). Since Bedoyan et al. (51) have also shown that telomeric chromatin fragments isolated from rat liver nuclei contain H1, we investigated whether Hho1p was required for telomeric silencing in yeast. We constructed an isogenic pair of WT and hho1⌬ strains that contained a URA3 gene integrated at a telomeric locus. In these strains, telomeric repression of the URA3 gene allows the cells to grow on 5-FOA medium, whereas defective repression results in an increased sensitivity to 5-FOA. As shown in Fig. 5B, the URA3 gene was strongly repressed in both the WT and hho1⌬ cells. In contrast, a strain bearing a sir3 deletion, previously implicated in telomeric silencing (49), was highly sensitive to 5-FOA, indicating a loss of repression of the telomeric URA3 gene. These results indicate that Hho1p does not play a detectable role in telomeric silencing.
SIN Phenotype-SWI/SNF, which is believed to function as a chromatin remodeling complex, is required for the activated  transcription of a set of yeast genes (35). Previous studies have shown that the reduced transcription of the HO gene in swi/snf mutants is partially relieved by SIN mutations (52). This class of mutation appears to function at the level of chromatin, since the SIN2 gene was subsequently shown to be identical to HHT1, one of two copies of the gene coding for histone H3. Mutations in either of the two histone H3 genes or substitution of two highly conserved amino acid residues in the histone fold domain of histone H4 was shown to result in a Sin phenotype (53). We therefore asked whether the absence of Hho1p would similarly result in a Sin phenotype.
The SWI1-dependent activity of the HO promoter was inves-tigated in strains containing an HO-lacZ gene fusion integrated at the ho locus (54). In the presence of the ␤-galactosidase substrate, 5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside, WT SWI1 strains produce blue colonies, swi1 colonies remain white, and mutations allowing activation of the HO-lacZ gene in swi1 strains result in blue colonies, defined to confer a Sin phenotype. Analysis of three hho1⌬ swi1 HO-lacZ segregants showed no restoration of lacZ expression (data not shown). Also, the hho1⌬ mutant did not alleviate the slow growth defect of a swi1 mutant (data not shown). Together these results show that deletion of the HHO1 gene does not result in a Sin phenotype.
Analysis of a hho1⌬ sin1 Double Mutant-The SIN1 gene was identified as a mutation that alleviated transcriptional defects due to inactivation of the SWI⅐SNF complex. The predicted Sin1p protein has an amino acid composition similar to the chromatin-associated mammalian non-histone HMG1 protein, and it has been suggested to be involved in the creation of a proper chromatin context for transcription (55). Strains containing a deletion of the SIN1 gene are viable and show no detectable growth or transcriptional defects. We asked whether the absence of Hho1p causes a synthetic phenotype in conjunction with a deletion of SIN1. To this end, basal expression from an episomal CYC1-lacZ fusion gene that contains only a minimal promoter (56) was measured in a WT, hho1⌬, sin1, and a hho1⌬ sin1 strain. The hho1⌬ sin1 strain was viable and showed no detectable growth defects (data not shown). Consistent with the basal transcription results presented above, neither the hho1⌬, sin1, nor the hho1⌬ sin1 strain exhibited any elevation in the basal transcription levels of the reporter gene as assayed by ␤-galactosidase activity (Table II). The repressed state of a URA3 gene integrated at a telomeric locus was also unaffected in the hho1⌬ sin1 strain (data not shown). SPT Phenotype-The SPT genes (suppressor of Ty) were identified by selection for extragenic suppressors of the transcriptional defects caused by ␦ and Ty insertions in the 5Ј regions of the HIS4 and LYS2 genes (57,58). Distinct classes of SPT genes have been identified, one of which includes SPT11/ HTA1 and SPT12/HTB1 encoding histones H2A and H2B, respectively (59). To determine whether loss of Hho1p causes Spt phenotype-like mutations of these core histones, we investigated the effect of the hho1⌬ mutation on transcription of the his4-912␦ and lys2-128␦ alleles. In an Spt ϩ his4-912␦ strain, the predominant HIS4 transcript is initiated in the solo ␦ insertion, producing an abnormally long HIS4 transcript in which the normal HIS4 translation start is not used, resulting in a His Ϫ phenotype. In an Spt Ϫ his4-912␦ strain, a wild-type HIS4 transcript is present in addition to the solo ␦-mediated transcript, resulting in histidine prototrophy (His ϩ ). Similarly, for Spt ϩ lys2-128␦, LYS2 transcription initiates at the ␦ sequence within the 5Ј-exon of the LYS2 coding region, resulting in a nonproductive short transcript and lysine auxotrophy. The presence of an spt mutation results in transcription initiation at the wild-type start site, resulting in lysine prototrophy. The ability of an SPT mutation to modulate the transcriptional activation of the solo ␦ insertion and adjacent gene is not currently understood at a mechanistic level.
To determine if deletion of HHO1 causes an Spt Ϫ phenotype, a hho1⌬ strain was crossed to a strain carrying the his4-912␦ and lys2-128␦ alleles, sporulated, and 88 spores from 22 fourspored tetrads were scored for both His and Lys phenotypes. Since both of the parental strains are Lys Ϫ , the ability of the hho1⌬ mutation to act as an Spt allele should be exhibited as a deviation from a 0:4 Lys ϩ :Lys Ϫ segregation pattern. Of the 88 spores analyzed, none were Lys ϩ . Analysis of the Spt phenotype for the his4-912␦ allele was consistent with the results observed for lys2-128␦ (data not shown). Together these results indicate that a deletion of HHO1 does not result in an Spt phenotype.
Hho1p Is Not Required for Efficient Sporulation-Since we have shown above that the rate of growth of a mitotically dividing yeast cell is unaffected by the absence of Hho1p, we asked whether Hho1p may be required for meiosis. This was investigated by determining the sporulation efficiency of diploid strains. The sporulation efficiencies, expressed as the percentage of tetrads relative to the total number of cells, is shown in Table III. Referring to Table III, it is seen that approximately 83% of wild-type diploid cells sporulated. In the case of the homozygous hho1⌬/hho1⌬ strain, approximately 60% of the diploids sporulated over the same period. To investigate whether this decrease was due to the absence of Hho1p or due to a genetic difference between the two congenic diploid strains, a wild-type copy of the HHO1 gene was reintroduced at the URA3 locus in the hho1⌬/hho1⌬ diploid strain. This strain sporulated at approximately 72% efficiency (see Table III). Although these results suggest a minor involvement of Hho1p in sporulation, a WT/hho1⌬ strain sporulated at only approximately 37% efficiency. Thus, we cannot exclude the contribution of genotypic differences between the compared congenic strains. A similar result was obtained with a sin1/sin1 hho1⌬/ hho1⌬ double mutant strain (see Table III). In all cases examined, the spores germinated and grew normally (data not shown). These data suggest that although Hho1p may make a minor contribution to the efficiency of sporulation, it is not absolutely required for meiosis or spore germination.
There Is No Detectable Change in the Chromatin Structure of a hho1⌬ Strain-The association of H1 with chromatin, in addition to changing the degree of compaction, has also been shown to change the spacing between adjacent nucleosomes (60). We therefore compared the nucleosome repeat length of bulk chromatin in the isogenic WT and hho1⌬ strains. No differences were detected (data not shown). To ensure that minor structural differences are not overlooked, we investigated the chromatin structure of selected regions at single nucleotide resolution. In Fig. 6 we show the primer extension mapping of the micrococcal nuclease scissions in the STE6 gene and the centromeric region of chromosome III in vivo. A comparison of the MNase cleavage sites in chromatin and free DNA at the STE6 locus (Fig. 6A) slows a clear repetitive, nucleosomal pattern in both the WT and the hho1⌬ strain. There is no readily detectable change in the MNase accessibility at the pseudo-dyad axis or within the short internucleosomal linker in the absence of Hho1p. Nor is there evidence of a change in the extent of the nucleosome footprint or the nucleosome repeat length throughout the STE6 gene. The primer extension footprinting of the centromeric region of chromosome III also shows a nucleosomal organization abutting the centromeric region which remains relatively nuclease-resistant in the hho1⌬ strain. To address the possibility that Hho1p is degraded dur-  ing nuclei isolation, we repeated the mapping of these regions using the more rapid procedure of preparing permeabilized spheroplasts (61). Identical results were obtained (data not shown).

DISCUSSION
Is Hho1p the Yeast Linker Histone H1?-We have shown above that recombinant Hho1p has electrophoretic and chromatographic properties similar to the linker histones of higher eukaryotes. Although there are several other lysine-rich proteins in the yeast genome with expected biochemical properties similar to Hho1p, we have also shown that rHho1p forms a stable ternary complex with a core di-nucleosome in vitro. This association appears to be specific, since the titration of the core di-nucleosome with rHho1p leads to the stepwise appearance of two well defined supershifted complexes, most likely a core di-nucleosome containing one and two rHho1p molecules, respectively. These complexes are formed at rHho1p:core ratios similar to those found for the linker histones of higher eukaryotes (38). Although we have not shown direct chromatin association in vivo, Ushinsky et al. (26) have shown that a fusion of Hho1p with the fluorescent green protein is located in the nucleus, strongly suggesting that the nucleosome core binding Hho1p is chromatin-associated.
The predicted secondary structure of the assigned globular domain of Hho1p suggests a single-winged helix protein fold similar to that of the chicken H5 (27) and H1.11L (28) globular domains. Several other DNA-binding proteins such as HNF-3␥ (62) and CAP (63) exhibit a similar winged helix fold. Although these proteins do bind to DNA, they differ in an important aspect from linker histones. Virtually all linker histone isotypes from a wide variety of organisms and tissue types have three conserved basic amino acid residues at positions corresponding to Lys-40, Arg-42, and Lys-52 in chicken histone H5. These three amino acid residues, which form a predicted secondary DNA-binding site (27), was shown to be essential for the proper binding of H1 to a nucleosome core, protecting ϳ168 bp of nucleosomal DNA from MNase digestion (40). We have shown above that yeast Hho1p has perfectly conserved amino acid residues at each of these three positions. Furthermore, the reconstitution of H1-stripped HeLa chromatin with rHho1p caused the protection of two full superhelical turns of nucleosomal DNA from exonucleolytic cleavage by MNase. Taken together, these data show that Hho1p acts like a true linker histone.
What Is the Role of Histone H1?-A substantial body of evidence exists that implicates histone H1 in chromatin condensation (reviewed in Ref. 14). We have systematically investigated an extensive list of possible phenotypes that may reflect the aberrant organization or improper condensation of chromatin in the hho1⌬ strain and have shown that a yeast cell lacking Hho1p appears to function as efficiently as a WT cell. This result is not unexpected, since mice, homozygous for an H1 0 gene disruption, were found to grow and reproduce normally with no anatomical or histological abnormality, although other H1 variants may have compensated for the absence of H1 0 (64). Similarly, Gorovsky and colleagues (16,65) have shown in Tetrahymena that vegetative growth, general polymerase I, II, and III transcription, protein synthesis, and general nucleosome repeat length are all unaffected by the absence of the four micronucleus-specific micLH peptides or macronucleus-specific H1. Linker histone-dependent changes were, however, observed in the nuclear volume, the transcriptional regulation of specific genes, and the efficiency of meiotic division (65). Although we did observe a difference in the sporulation efficiency of a WT versus a hho1⌬ yeast strain, we could not exclude the possible contribution of minor genetic background differences to this observation. Thus, although Hho1p may have an effect, it is not required for sporulation and spore formation.
It was previously shown that histone H1 represses basal transcription in vitro (44 -46). We could not detect any dere-  Yeast Histone H1 pression in basal polymerase II transcription of a reporter gene driven by either a PHO5 or a CYC1 promoter in the hho1⌬ strain. It is possible that in vivo the regulation of only selected genes are affected, as was found for the ngoA and CyP genes in Tetrahymena lacking a linker histone (65). The regulatory effect on these two starvation-specific genes is intriguing, since Roth et al. (66) have shown that starvation of Tetrahymena is accompanied by dephosphorylation of histone H1 and presumably changes in the condensation state of chromatin. It is not clear whether the differential transcriptional effect is mechanistically direct or indirect. In the former case it may be related to a requirement for a compact structure placing transcription regulatory components within a required spatial proximity. Alternatively, structural features of the compacted fiber or a region of chromatin-associated histone H1 itself may be directly recognized by components involved in transcription of select genes. In the latter case, improperly condensed chromatin may disrupt the general nuclear architecture, influencing compartmentalization of specialized structures. Interestingly, a differential effect on transcription was also observed by Linder and Thoma (67) who reported that the overexpression of the sea urchin H1␣ protein in yeast repressed polymerase I-transcribed rRNA genes and the polymerase II-transcribed ACT1 and URA3 genes but not the polymerase II-transcribed Ty gene. Two previous studies have shown that expression of an exogenous H1 in yeast at a stoichiometry well below that of the core histones resulted in a marked decrease in cell viability (23,67). The evident interpretation of this lethality is that the overexpression of H1 at moderate levels results in an excess of linker histones in yeast that already contains Hho1p. However, it was shown that the overexpression of the mouse H1c and H1e variants in 3T3 cells had little effect on the growth properties and viability of the cells in culture (68). It is not entirely clear why overexpression of sea urchin H1 in yeast and mouse H1 in 3T3 cells should differ so drastically in their effects. One possibility is that the different states of differentiation of cultured cells and vegetatively growing yeast result in differences in the general chromatin organization. Alternatively, the contrasting results in yeast and 3T3 cells may be due to the specific histone variants. The expression of H1 variants is tissue-specific and developmentally regulated (69) and has been shown to differ in both efficacy of chromatin condensation (70) and modulation of gene expression (71). It is also possible that the sea urchin H1 overexpressed in yeast does not localize properly to appropriate regions of chromatin or interferes with the localization or post-translational modification of the endogenous Hho1p.
Absence of Hho1p did not result in any bulk translational reorganization of nucleosomes or a change in the chromatin organization of specific regions. The only major structural change H1 was previously shown to confer on chromatin in vitro, apart from condensation, was an increase in the nucleosomal repeat length (60). Since the bulk nucleosome repeat length of yeast in the presence of Hho1p is approximately 160 bp, adjacent nucleosomes are closely stacked and joined by a linker of negligible length. The absence of a linker histone is therefore not expected to result in a further reduction of the nucleosome repeat length. It is also possible that the Hho1p protein is present at very low levels in mitotically cycling cells or is only associated with specific regions of chromosomes, in which case the absence of Hho1p is not expected to cause a detectable change in the structure of bulk chromatin.
Given that the absence of Hho1p in yeast or micLH/H1 in Tetrahymena (16) does not appear to affect cell viability and growth rate, one may ask why the linker histones are evolu-tionary conserved? We note that all measurements were performed under optimal growth conditions in the laboratory. It is possible that undetected effects may become much more pronounced under sub-optimal conditions of a natural environment. Such effects, even if minor, may play a significant role over evolutionary periods, thus maintaining selective pressure on H1 and Hho1p.