Rad6-Dependent Ubiquitination of Histone H2B in Yeast

See allHide authors and affiliations

Science  21 Jan 2000:
Vol. 287, Issue 5452, pp. 501-504
DOI: 10.1126/science.287.5452.501


Although ubiquitinated histones are present in substantial levels in vertebrate cells, the roles they play in specific biological processes and the cellular factors that regulate this modification are not well characterized. Ubiquitinated H2B (uH2B) has been identified in the yeast Saccharomyces cerevisiae, and mutation of the conserved ubiquitination site is shown to confer defects in mitotic cell growth and meiosis. uH2B was not detected in rad6mutants, which are defective for the ubiquitin-conjugating enzyme Ubc2, thus identifying Rad6 as the major cellular activity that ubiquitinates H2B in yeast.

Modulation of chromatin structure by the posttranslational modification of histones has emerged as an important mechanism for regulating chromosome function in eukaryotes. Although acetylation of the histone NH2-termini has been shown to be intimately connected to transcriptional regulation, the biological roles of other histone modifications remain obscure (1). A noteworthy modification is the conjugation of ubiquitin to the COOH-termini of the core histones H2A, H2B, and H3 (2). Ubiquitinated forms of these histones are stable in vivo, and their incorporation into nucleosomes has been proposed to alter chromatin structure locally (2, 3). Although the precise cellular roles of histone ubiquitination are unclear, this modification has been correlated with increased transcriptional activity, replication, and meiosis in higher eukaryotes (3,4).

Ubiquitin is transferred to target proteins in a reaction catalyzed by members of a large group of ubiquitin-conjugating enzymes (Ubc's), which donate ubiquitin to the ɛ-amino group of specific lysine residues, often in a substrate-specific manner (5). Two evolutionarily conserved Ubc's, Rad6/Ubc2 and Cdc34/Ubc3, are able to ubiquitinate histones in vitro without the mediation of an E3 ubiquitin ligase (6, 7). However, neither protein has been demonstrated to ubiquitinate histones in vivo, and the Ubc that targets histones in cells remains to be identified. Here, we present evidence that histone H2B is ubiquitinated in yeast. We also show that attachment of ubiquitin to this core histone depends primarily on the activity of Rad6/Ubc2 and is required for both optimal mitotic cell growth and meiosis.

It has been reported that Saccharomyces cerevisiae contains little, if any, uH2A or uH2B (8). We reinvestigated this issue using combined genetic and immunological approaches. Lysine-to-arginine (K → R) substitutions were introduced at the conserved ubiquitination sites of both H2A and H2B (9). A single K → R substitution at Lys123 in H2B (htb1-K123R) was combined with four K → R substitutions at Lys119, Lys120, Lys123, and Lys126 of H2A [hta1-K119R, K120R,K123R, K126R (hereafter hta1-4K/R)] to eliminate the possibility that, in the absence of the preferred H2A ubiquitination site (Lys119), adjacent lysine residues (Lys120, Lys123, or Lys126) could serve as acceptors for ubiquitin conjugation. Strains that contained the fully mutant forms of H2A plus H2B (hta1-4K/R +htb1-K123R) were viable (9, 10) but showed pronounced mitotic and meiotic defects. The mitotic phenotype was characterized by a small colony size on plates (11) and an ∼30% increase in doubling time in liquid medium (Table 1). A slight increase (∼10%) in the number of large budded cells additionally suggested a delay in the S or G2/M phases of the cell cycle (11). The meiotic defect resulted in a failure of homozygous mutant diploids to form spores (Table 1). Mutant diploids had a single mass of DNA, indicating that neither meiotic division occurred and that the meiotic program was blocked at an early stage (Fig. 1). These results indicate that the lysine residues corresponding to the evolutionarily conserved ubiquitination sites in H2A and/or H2B are required for specific cellular processes in yeast.

Figure 1

Meiotic phenotype of H2B ubiquitination site mutant. Homozygous diploid strains YKR55 (hta1-htb1Δhta2-htb2Δ +HTA1-HTB1) and YKR57 (hta1-htb1Δhta2-htb2Δ +HTA1-htb1-K123R) were grown to stationary phase in YPD medium and induced to sporulate by transfer to sporulation medium (SPM) (1% potassium acetate). At 48 hours after transfer to SPM, aliquots of each culture were fixed with 3.7% formaldehyde for 1 hour, washed with distilled water, and resuspended in 70% ethanol for 30 min. After washing with distilled water, cells were resuspended in 1 μg/ml of the fluorescent dye 4′,6-diamidino-2-phenylindole for 30 min, then washed and resuspended in distilled water. Cells were spotted onto polylysine-coated glass slides and photographed with a Zeiss Axiophot fluorescence microscope. The top panels were imaged with Nomarski optics, and the bottom panels were imaged with DNA fluorescence. The wild-type diploid sporulated with ∼50% efficiency, and the mutant diploid sporulated with 0% efficiency.

Table 1

Phenotypes of H2A and H2B ubiquitination site mutants. Growth and sporulation were monitored inhta1-htb1Δhta2-htb2Δ homozygous diploids that carried the indicated HTA1 or HTB1 alleles on a plasmid. Growth was followed by measuring the optical density at 600 nm of cultures grown at 30°C. Sporulation was induced by transfer of stationary-phase YPD cells to SPM (1% potassium acetate). Percent sporulation was determined by counting the number of asci present among 200 to 500 cells using light microscopy. UV sensitivity was monitored by spotting 10-fold serial dilutions of cultures growing exponentially in YPD medium onto YPD plates, followed by irradiation at 50 J/m2 and incubation at 30°C for 2 days in the dark.

View this table:

To identify which K → R substitutions conferred these phenotypes, we examined the H2A and H2B mutants individually. The H2B mutant on its own showed the mitotic and meiotic defects of the quintuple mutant, whereas the quadruple H2A mutant was phenotypically wild type (Fig. 1and Table 1) (11). Together, the data strongly support a role for the conserved ubiquitination site of H2B, but not that of H2A, in optimal cell growth and meiosis in yeast.

To determine if ubiquitin is conjugated to H2B in vivo, we developed a sensitive immunoassay to detect modified forms of this histone. We employed a yeast strain that contained Flag-H2B as the only source of this core histone (12). Flag-H2B can be quantitatively immunoprecipitated from yeast cell extracts (11, 12), allowing detection of even low levels of modified H2B. To further sensitize detection of ubiquitin-H2B conjugates, we next introduced a plasmid encoding a galactose-inducible, hemagglutinin (HA)-tagged ubiquitin gene (HA-Ub) into the Flag-H2B strain (13). When this strain was grown in galactose, high levels of HA-Ub were induced. Growth in dextrose prevented expression of tagged ubiquitin and served as a control.

Using a method that preserved only covalent interactions during cell lysis (14), we immunoprecipitated Flag-H2B from cells in which HA-Ub was either present (galactose) or absent (dextrose). After SDS–polyacrylamide gel electrophoresis (SDS-PAGE), we analyzed the immunoprecipitates by Western blot analysis with α-Flag or α-HA antibodies (Fig. 2A). Three bands were specifically detected by the α-Flag antibody when HA-Ub was expressed (Fig. 2A, lanes 1 and 3). The most abundant and rapidly migrating band represented unmodified Flag-H2B. The two less abundant and more slowly migrating bands represented ubiquitinated conjugates of Flag-H2B. The upper band was identified as HA-Ub conjugates of Flag-H2B (Flag-HA/uH2B) because it was also detected by the α-HA antibody (Fig. 2A, lanes 8 and 10) and was absent when HA-Ub was not expressed (Fig. 2A, lanes 2 and 4 and 9 and 11). The lower band represents endogenous ubiquitin conjugates of Flag-H2B (Flag-uH2B). This band was detected only by the α-Flag antibody, and its presence was independent of HA-Ub expression (Fig. 2A, lanes 1 through 4). These designations were confirmed by probing Western blots of α-Flag immunoprecipitates with polyclonal antibodies against native ubiquitin (Fig. 2B): Each of the two putative ubiquitinated Flag-H2B species, but not Flag-H2B itself, selectively reacted with this antibody.

Figure 2

Detection of uH2B in mitotically dividing yeast cells. (A) Haploid htb1-1 htb2-1 strains contained a CEN-URA3 (YKR30) or 2μm-URA3(YKR28) Flag-HTB1 plasmid, a 2μm-HIS3 Flag-htb1-K123R plasmid (YKR42), or a CEN-HIS3plasmid with an untagged HTB1 gene (YKR33). A 2μm-TRP1 plasmid that contained aGAL1-regulated HA-UBI4 gene (pKR41) was also present in each strain. Expression of HA-Ub was induced or repressed by growth in the presence of 2% galactose (G) or 2% dextrose (D), respectively. Lysates were prepared from equivalent numbers of cells and incubated with α-Flag monoclonal antibody coupled to Sepharose beads (14). Immunoprecipitates were collected, eluted with Flag peptide, and 20-μl (α-Flag) or 10-μl (α-HA) aliquots were subjected to 15% SDS-PAGE. After transfer to filters, the blots were probed with α-Flag (lanes 1 through 7) or α-HA (lanes 8 through 14) monoclonal antibodies. Flag-H2B migrates as a doublet, and preliminary data indicate that the top band of the doublet represents a phosphorylated species of H2B. HC, immunoglobulin heavy chain; NS, nonspecific band. (B) Haploid htb1-1 htb2-1 strain YKR28 containing a Flag-HTB1 gene on a 2μm-URA3 plasmid and GAL1-regulated HA-Ub was grown in the presence of galactose (G) or dextrose (D). After the preparation of lysates and immunoprecipitation with α-Flag monoclonal antibody, the immunoprecipitates were eluted by boiling, and after SDS-PAGE, the blots were probed with α-Flag monoclonal antibody (lanes 1 and 2) or α-ubiquitin polyclonal antibody (lanes 3 and 4). LC, immunoglobulin light chain.

To determine if ubiquitin was attached to H2B at Lys123, we introduced the K123R mutation into a Flag-HTB1 gene (12) and immunoprecipitated Flag-H2B (K123R) in the presence or absence of HA-Ub. Although unmodified Flag-H2B (K123R) was present in these immunoprecipitates, no ubiquitin conjugates could be detected (Fig. 2A, lanes 5 and 6 and 12 and13). Together, the results support the conclusion that ubiquitin is attached to yeast H2B at the same lysine residue to which ubiquitin is conjugated in vertebrate H2B (2).

Although proteins can be either mono- or polyubiquitinated (5), the mobility of the two ubiquitinated H2B species is consistent with monoubiquitination; moreover, a longer exposure of the immunoblots failed to reveal species that migrated more slowly than either Flag-uH2B or Flag-HA/uH2B (11). However, it is also possible that polyubiquitinated forms of H2B are unstable or present at levels too low to be detected by the immunoassay. Titration experiments suggest that the levels of monoubiquitinated H2B might represent as much as 10% of the total cellular H2B (15). This is significantly higher than the levels in fly or vertebrate cells, where uH2B accounts for 1 to 2% of the total cellular H2B (16).

Rad6/Ubc2 shows marked specificity for histones H2A and H2B in vitro, catalyzing predominantly monoubiqitinated forms of these histones (7). Because the H2B ubiquitination site mutant shares the sporulation defect of rad6 mutants (17), it seemed likely that Rad6 might be the cellular Ubc that ubiquitinates H2B. To test this, we transferred the following three rad6 mutations into the strain that was tagged with double epitopes (10):rad6Δ, which lacks Rad6 protein;rad6-C88A, which is defective in ubiquitin-conjugating activity because the site of ubiquitin linkage has been destroyed (C88A, Cys88 → Ala88); andrad6-149, which is catalytically active but unable to ubiquitinate histones efficiently in vitro because of the absence of acidic COOH-terminal residues (18, 19). Only unmodified Flag-H2B was present in these strains, despite the use of HA-Ub to sensitize the detection of uH2B (Fig. 3A). Together with the analysis of the H2B ubiquitin site mutant, the results support the conclusions that Rad6 ubiquitinates yeast histone H2B on Lys123 in mitotically growing cells and that, in its absence, no other Ubc's are able to substitute efficiently. This identifies H2B as a physiologically relevant substrate of Rad6/Ubc2 in yeast.

Figure 3

Rad6/Ubc2 is required for ubiquitination of H2B in mitotically dividing cells and during meiosis. (A) CEN-URA3 plasmids carryingRAD6 (pKR57), rad6-C88A (pKR58), orrad6-149 (pKR97) alleles were introduced into strain YKR36, a rad6Δhtb1-1 htb2-1 haploid strain that contained both Flag-HTB1 on a CEN-TRP1 plasmid and GAL1-regulated HA-Ub on a 2μm-HIS3plasmid. After growth in the presence of galactose (G) or dextrose (D), lysates were prepared from equivalent cell numbers and incubated with α-Flag monoclonal antibody. The immunoprecipitates were eluted with Flag peptide, and after SDS-PAGE, the blots were probed with α-Flag (lanes 1 through 8) or α-HA (lanes 9 through 12) monoclonal antibodies. (B) RAD6(YKR49) or rad6Δ (YKR64) htb1-1 htb2-1 homozygous diploids that contained Flag-HTB1 on a 2μm-URA3 or 2μm-LEU2 plasmid were grown vegetatively in YPD medium (veg), allowed to reach stationary phase in the same medium, and then were induced to sporulate by transfer into SPM (1% potassium acetate). During vegetative growth and at the indicated times after sporulation was induced, lysates were prepared from equivalent numbers of cells and incubated with α-Flag monoclonal antibody. The immunoprecipitates were eluted with Flag peptide, and after SDS-PAGE, the Western blots were probed with α-Flag monoclonal antibody. At 48 hours after transfer into SPM, 52% of the RAD6/RAD6 cells and 0% of the rad6Δ/rad6Δ cells had formed spores.

We next asked if Rad6 is also required to ubiquitinate H2B during meiosis. The presence of uH2B in meiotic cells was monitored at various times after homozygous Flag-H2B diploids had been induced to sporulate (Fig. 3B). In a RAD6 diploid, Flag-uH2B was present throughout sporulation and decreased in amount only late in the meiotic program (Fig. 3B, lanes 2 through 6), when mature spores appeared (11). In a rad6Δ diploid, however, Flag-uH2B could not be detected in either premeiotic vegetative cells or in sporulating cultures (Fig. 3B, lanes 7 through 12), and spores were not formed (11). A meiosis-defective rad6-149 diploid also contained no Flag-uH2B when induced to sporulate (11). Thus, the Rad6 dependence of budding yeast meiosis might result in part from the ubiquitination of H2B. Rad6 homologs are also required for meiosis in fission yeast and vertebrates (4), but it is not known whether uH2B is present in meiotic cells from these organisms.

Although these results indicate that H2B is a major target of Rad6 in both mitotic and meiotic cells, Rad6 must have additional targets as well. For example, the failure to ubiquitinate H2B (and H2A) cannot account for the ultraviolet (UV) sensitivity of a rad6Δ mutant because the quintuple H2A plus H2B ubiquitin site mutant is not UV sensitive (Table 1).

In higher eukaryotes, where ubiquitinated histones were first discovered, uH2A is the most abundant species (3,16). This is apparently reversed in yeast, where uH2B is abundant and uH2A is either absent or present at levels that are too low to be detected by the Flag/HA–based detection system (8,11). Thus, in yeast, uH2B might assume some of the roles played by uH2A in higher eukaryotic cells. Covalent attachment of ubiquitin to the H2B COOH-terminus could modulate the interaction of the histone tail with linker DNA, adjacent nucleosomes, or nonhistone regulatory proteins (20). This could lead to a more open chromatin structure or mark chromatin for recognition by regulatory proteins and, in turn, permit the transcription of key genes during mitosis and meiosis.

The pleiotropic roles of yeast Rad6 in the DNA-mediated events of repair, mutagenesis, meiosis, retrotransposition, and gene silencing has long prompted the view that these events might result from Rad6-dependent histone ubiquitination (7, 19, 21, 22). Ubiquitin conjugation is required for all Rad6-dependent functions (23), and ubiquitination of H2B by Rad6 might indeed constitute an essential meiotic function because rad6mutants (17, 24) and the H2B K123R mutant fail to sporulate. The slow growth phenotype of the H2B K123R mutant also suggests that Rad6-dependent H2B ubiquitination helps to promote optimal mitotic cell growth. Although the H2B ubiquitination-defectiverad6-149 mutant has no growth defect (24), residual activity of rad6-149 protein might ubiquitinate H2B at low levels that are sufficient for cell growth (11, 19). At least two other Rad6-dependent processes, however, are unlikely to result from H2B ubiquitination: DNA repair, which is unaffected by the H2B K123R mutation, and telomeric silencing, which is maintained in tailless rad6 mutants (22). Thus, the multifunctional Rad6 protein probably ubiquitinates other chromosomal proteins besides histones.

  • * Present address: Department of Microbiology, Harvard Medical School, Boston, MA 02115, USA.

  • To whom correspondence should be addressed. E-mail: m-osley{at}


Stay Connected to Science

Navigate This Article