Histone variant H2A.Z and linker histone H1 influence chromosome condensation in Saccharomyces cerevisiae

Abstract

Chromosome condensation is essential for the fidelity of chromosome segregation during mitosis and meiosis. Condensation is associated both with local changes in nucleosome structure and larger-scale alterations in chromosome topology mediated by the condensin complex. We examined the influence of linker histone H1 and variant histone H2A.Z on chromosome condensation in budding yeast cells. Linker histone H1 has been implicated in local and global compaction of chromatin in multiple eukaryotes, but we observe normal condensation of the rDNA locus in yeast strains lacking H1. However, deletion of the yeast HTZ1 gene, coding for variant histone H2A.Z, causes a significant defect in rDNA condensation. Loss of H2A.Z does not change condensin association with the rDNA locus or significantly affect condensin mRNA levels. Prior studies reported that several phenotypes caused by loss of H2A.Z are suppressed by eliminating Swr1, a key component of the SWR complex that deposits H2A.Z in chromatin. We observe that an htz1Δ swr1Δ strain has near-normal rDNA condensation. Unexpectedly, we find that elimination of the linker histone H1 can also suppress the rDNA condensation defect of htz1Δ strains. Our experiments demonstrate that histone H2A.Z promotes chromosome condensation, in part by counteracting activities of histone H1 and the SWR complex.

Introduction

Chromosome structure in eukaryotes is determined at the local level by nucleosome placement, histone modifications, and nucleosome–nucleosome interactions and on larger scales by the formation of chromatin loops and long-range chromosome–chromosome associations. Large-scale chromosome organization is mediated principally by the cohesin and condensin complexes; cohesins maintain the association between sister chromatids following DNA replication, while condensins orchestrate the complex reorganization of chromosome structure that occurs as eukaryotic cells prepare for mitosis. Chromosome condensation requires the five subunit condensin complex, comprised in budding yeast of subunits Smc2, Smc4, Brn1, Ycs4, and Ycg1 (see Kalitsis et al. 2017; Skibbens 2019 for review). Condensins bind throughout the genome in yeast but are enriched at the rDNA, tRNA genes, and centromeres (Freeman et al. 2000; Strunnikov et al. 2001; D'Ambrosio et al. 2008; Swygert et al. 2019). The condensin complex creates chromosome loops via asymmetric loop extrusion, anchoring to DNA and reeling it in from one side of the anchor point (Terakawa et al. 2017; Ganji et al. 2018; Ryu et al. 2020).

Histone H1 has been linked to both local and large-scale changes in chromosome structure. In vitro, H1 promotes the salt-dependent condensation of nucleosome arrays into higher-order configurations (Thoma et al. 1979; Butler and Thomas 1980; Allan et al. 1981; Zlatanova and van Holde 1996), while in Drosophila and Tetrahymena, the presence of histone H1 correlates with decreased nuclear volume and chromosome size (Ner and Travers 1994; Shen et al. 1995). A reduction in histone H1 levels in Mus musculus and Xenopus leads to decreased chromatin compaction (Fan et al. 2005; Maresca et al. 2005). In yeast cells, increased association of H1 with chromatin in stationary phase cells is associated with reduced accessibility to micrococcal nuclease, suggesting that H1 induces a more compact chromatin state (Schäfer et al. 2008), while cells lacking H1 during meiosis have slightly expanded nuclear volume, consistent with chromosome decondensation (Bryant et al. 2012).

Histone H2A.Z, a highly conserved variant of canonical histone H2A, is preferentially found in nucleosomes near the transcriptional start site in eukaryotes (Santisteban et al. 2000; Raisner et al. 2005; Albert et al. 2007); several lines of evidence suggest that H2A.Z-containing nucleosomes are more easily displaced, possibly permitting a more rapid transcriptional response (Santisteban et al. 2000; Zhang et al. 2005; Venters et al. 2011; Tramantano et al. 2016). In addition to altered transcription patterns, budding yeast cells lacking H2A.Z exhibit chromosome instability (Krogan et al. 2004; Keogh et al. 2006) and defects in chromosome cohesion (Sharma et al. 2013). In fission yeast, cells lacking H2A.Z exhibit chromosome entanglements in anaphase and premature dissociation of condensin complexes (Kim et al. 2009).

Overall, these prior observations suggest that H1 and H2A.Z independently contribute to a chromatin structure that promotes successful transmission of chromosomes in mitosis. To directly assess the roles of H2A.Z and H1 in chromosome condensation in budding yeast, we examined condensation of the rDNA locus during mitosis in strains lacking H1, H2A.Z, or both. We find that cells lacking H2A.Z fail to efficiently condense the rDNA locus; this defect appears to be independent of condensin recruitment to chromatin and is unlikely to be due to changes in condensin gene transcription. Strains lacking only histone H1 have no alterations in rDNA condensation, but surprisingly, the loss of H1 can rescue the condensation defects caused by H2A.Z loss. Our results suggest that chromosome condensation at the rDNA locus in budding yeast relies on H2A.Z to counteract an anticondensation activity of linker histone H1.

Materials and methods

Media and cell culture

Cells were grown in YPD liquid media (1% Bacto yeast extract, 2% Bacto peptone extract, 2% dextrose) at 30°C unless otherwise noted.

Yeast strains

Yeast strains used in this study are described in Table 1 and Supplementary Table 1. Strain construction was generally performed using PCR-mediated gene replacement, using the MX-series plasmids as templates for selectable markers (Wach et al. 1994; Goldstein and McCusker 1999). The wild-type NET1-GFP strain, YSH1185, is from the Invitrogen Yeast GFP Clone Collection (Huh et al. 2003). Addition of GFP-HIS3MX6 to NET1 in additional strains was accomplished by PCR amplification of a DNA fragment from YSH1185 DNA containing NET1 and HIS3MX6 sequences and standard transformation and selection. YSH1433, YSH1434, and YSH1435 were derived from strains YAG77, YAG83, and YAG85, respectively; YAG77, YAG83, and YAG85 were generously provided by Jennifer Benanti (UMass Medical School). To delete the HHO1 open reading frame, strains were transformed with XmnI-digested plasmid pSLPS1, which contains HHO1 5′ and 3′ sequences flanking the hphMX4 selectable marker. To make strains with mutated alleles of HTZ1, strain YSH1186 (Δhtz1::kanMX6) was transformed with kanMX6-targeting CRISPR-Cas9 plasmid pRS425-Cas9-kanMX (created by Gang Zhao in Bruce Futcher's laboratory at Stony Brook University) and healing fragments with wild-type HTZ1, or alleles of HTZ1 derived from plasmids pJR2659, pJR2973, and pJR2974 (Babiarz et al. 2006), creating strains YSH1325, YSH1326, and YSH1332, respectively. Strain YSH1319 (HTZ1) was made from YSH1186 (Δhtz1::kanMX6) using pRS425-Cas9-kanMX and a healing fragment containing wild-type HTZ1. Strains YSH1341 and YSH1342, bearing HA-tagged alleles of SMC2 and BRN1, were constructed using plasmid pYM1, which contains 3HA linked to kanMX6 (Knop et al. 1999). TIR1-containing strain ySB036 was generously provided by David Tollervey (Bresson et al. 2017). To construct AID strain YSH1578, the AID* tagging cassette was amplified from pKan-AID*-9myc (Morawska and Ulrich 2013) and integrated at the 3′ end of the HTZ1 open reading frame. Strains D16, D1817, D1821, and D2563, containing temperature-sensitive condensin alleles, were generously provided by Damien D'Amours (University of Ottawa) (Ouspenski et al. 2000; Robellet et al. 2015; Palou et al. 2018) and renamed YSH1409, YSH1410, YSH1411, and YSH1412, respectively. Transplaced condensin alleles were amplified via PCR from strain D16/YSH1409 (brn1-60), strain CH2524 (brn1-9) (Lavoie et al. 2000) or YAG80 (from J. Benanti; aka YSH1418), and 1bAS330 (smc2-8) (Freeman et al. 2000). All yeast strains described in this study are available from the authors upon request.

Table 1.

Strains used in this study.

Strain	Genotype	Source or reference
YSH1185	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 NET1-GFP-HIS3MX6	(Huh et al. 2003)
YSH1186	YSH1185 htz1Δ::kanMX6
YSH1302	YSH1185 hhol1Δ::hphMX4
YSH1303	YSH1186 hhol1Δ::hphMX4
YSH1319	YSH1186 kanMX6::HTZ1
YSH1323	YSH1185 BRN1::brn1-9-natMX4
YSH1324	YSH1185 SMC2::smc2-8-natMX4
YSH1325	YSH1186 HTZ1-FLAG
YSH1326	YSH1186 htz1-K14R-FLAG
YSH1330	YSH1186 swr1Δ::hphMX4
YSH1331	YSH1185 swr1Δ::hphMX4
YSH1332	YSH1186 htz1-K3,8,10,14R-FLAG
YSH1341	YSH1185 SMC2-HA-kanMX6
YSH1342	YSH1185 BRN1-HA-kanMX6
YSH1344	YSH1341 htz1Δ::hphMX4
YSH1345	YSH1342 htz1Δ::hphMX4
YSH1397	YSH1323 hho1Δ::hphMX4
YSH298	ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 GAL+	D. Shore (BY2096)
YSH1488	YSH298 hho1Δ::hphMX4
YSH1409 (D16)	ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 brn1-60	(Ouspenski et al. 2000)
YSH1463	YSH1409 hho1Δ::hphMX4
ySB036	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 TIR1:his3 MET15	(Bresson et al. 2017)
YSH1578	ySB036 HTZ1-kanMX-AID*-9myc NET1-GFP-HIS3MX6
YSH1552	YSH298 NET1-GFP-HIS3MX6
YSH1553	YSH1488 NET1-GFP-HIS3MX6
YSH1554	YSH1409 NET1-GFP-HIS3MX6
YSH1555	YSH1463 NET1-GFP-HIS3MX6

Condensation assay and fluorescence microscopy

Samples were prepared as previously described (Sharma et al. 2013) with minor modifications. Cultures were grown to early log phase (optical density at 600 nm was 0.1–0.2). Mitotic arrests were performed using a two-step block: cells were first synchronized in G1 with α-factor for 2.5 h, washed with prewarmed media, and released into prewarmed media containing 15 µg/mL of nocodazole (EMD Millipore) for 3 h. The efficiency of cell cycle arrests was initially assessed by budding index and flow cytometry (see Supplementary Fig. 4). Condensation assays were performed if at least 80% of the culture was blocked. Cells arrested in mitosis were fixed with fresh paraformaldehyde at a final concentration of 4% for 1 h at room temperature. Cells were then briefly spheroplasted with zymolyase 20 T for 10 min at 37°C.

Immunostaining was performed as previously described (Rockmill 2009). Tubulin was detected with rat monoclonal antibody (YOL1/34, Santa Cruz, SC-53030) at a 1:100 dilution. Alexa Fluor 647 antirat antibody (Jackson ImmunoResearch, 712-605-153) was used as a secondary antibody at a 1:200 dilution. Z-stacked images were acquired on a Deltavision RT imaging system (Applied Precision) that was adapted to an Olympus IX70 microscope running Softworx software. To ensure only metaphase-blocked cells were assessed, we restricted our analysis to large-budded cells, in which DAPI staining was restricted to 1 bud and which had fully blocked, unelaborated tubulin. Less than 5% of nocodazole-blocked cells were excluded from the analysis based on these criteria. For experiments using the HTZ1-degron strain, cells were grown to early-mid log phase (optical density at 600 nm was ∼0.3), synchronized at G1 with α-factor, and then released into media with nocodazole at 23°C. Cultures were then supplemented with 1 mM auxin (indole-3-acetic acid; Sigma-Aldrich) or a solvent control and incubated for 2 h. In all individual experiments, a minimum of 50 metaphase-arrested cells were scored; data were accumulated over at least 3 independent trials, except for the brn1-9 and smc2-8 strains reported in Fig. 1 and the strains used to assess the role of HTZ1 acetylation, in which data from 2 independent trials are reported. All scoring was done without experimenter knowledge of the strain being scored.

Fig. 1.

Strains with temperature-sensitive condensin alleles exhibit rDNA condensation defects at elevated temperatures. a) Cells were synchronously released from an α-factor block and then arrested in mitosis with nocodazole. b) Metaphase-arrested cells were stained with DAPI (blue), an antibody to tubulin (pink), and imaged for Net1-GFP (green). Cells were assigned to the puff, cluster, or line/loop categories based on the GFP pattern. c) Strains expressing temperature-sensitive alleles of the BRN1 or SMC2 condensin genes were blocked with α-factor at 23°C and then released from the block into media containing nocodazole and allowed to progress to metaphase at either 23°C or 37°C. Only large-budded cells with single DAPI masses and unelaborated tubulin spindles were assessed. The percentage of cells exhibiting each rDNA morphology (puff, cluster, or line/loop) is shown in the bar graphs for each genotype and temperature. Strains examined were wild-type (YSH1185, at 23°C, N = 56, at 37°C N = 57), brn1-9 (YSH1323, N = 108 at 23°C, N = 94 at 37°C), and smc2-8 (YSH1324, N = 112 at 23°C, N = 85 at 37°C). Using Fisher's exact test, the P-value comparing loop and nonloop rDNA for wild-type at 23°C vs 37°C is 0.24, brn1-9 at 23°C vs 37°C is <0.00001, and smc2-8 at 23°C vs 37°C is <0.00001. P-values are reported on the figure as ns (not significant; P > 0.05) or *** (P < 0.0001). Error bars represent the standard error of the mean.

Western blots

H2A.Z-AID-9xMyc was detected using a mouse anti-c-Myc (9E10) monoclonal antibody (sc-40; Santa Cruz Biotechnology) at a 1:5,000 dilution. Tubulin was detected using a rat antitubulin monoclonal antibody (sc-53030; Santa Cruz Biotechnology) at a 1:2,250 dilution. Antibodies were applied to the membranes in 5 mL blocking solution and incubated overnight at 4°C. Indicated values of H2A.Z are normalized to the tubulin signal and expressed relative to the no auxin control.

Chromatin immunoprecipitation

Chromatin immunoprecipitation (ChIP) was performed as previously described (Martins-Taylor et al. 2011) using 2.5 μL anti-HA (12CA5, Roche Applied Science) to detect HA-tagged Smc2, Brn1, or Htz1 and anti-Hho1p from abcam (ab71833) to detect histone H1. Quantification was performed using real-time PCR (primer sequences are listed in Supplementary Table 2). Condensin binding at the RFB region of the rDNA was monitored (Freeman et al. 2000; Strunnikov et al. 2001); binding at TUB2 was assessed as a negative control (Wang et al. 2004).

Serial dilution assays

To assay temperature-sensitive strain viability in response to the loss of HHO1, cultures were grown to log phase in YPD at 23°C or 30°C, and 10x serial dilutions were spotted to YPD media. The plates were grown at 23, 30, 33, and 37°C and imaged after 2–3 days of growth.

Results

H2A.Z promotes chromosome condensation

To determine if H2A.Z or H1 influence chromosome condensation, we examined rDNA morphology in metaphase-blocked cells expressing a Net1-GFP fusion protein. As cells progress through DNA replication into mitosis the condensation state of the rDNA proceeds from a highly decondensed puff, to an intermediate small cluster, and then into a line/loop conformation, corresponding to fully condensed rDNA (see Fig. 1; Guacci et al. 1994; Lavoie et al. 2004; Machín et al. 2004). Net1 binds the highly repetitive rDNA, allowing determination of rDNA morphology by assessing the GFP signal (D'Ambrosio et al. 2008). To demonstrate that this assay effectively assesses condensation, we examined the Net1-GFP pattern in strains bearing temperature-sensitive alleles of condensin genes. Strains expressing either the brn1-9 or smc2-8 allele were grown at the permissive temperature (PT, 23°C) and blocked in G1 phase with α-factor. Following washes, cultures were split, one culture maintained at PT and the other shifted to the nonpermissive temperature (NPT, 37°C) and grown in media containing nocodazole to arrest cell cycle progression in metaphase. Cells were fixed; DNA was imaged using DAPI, and the GFP signal was assessed with experimenter blind to the identity of the strain. We assigned metaphase-blocked cells into 1 of the 3 categories of condensation state. Approximately 77% of wild-type cells exhibited the loop morphology at 23°C, consistent with prior reports (Lavoie et al. 2004; D'Ambrosio et al. 2008). This value decreased to ∼58% in cells shifted to 37°C (Fig. 1c). In contrast, we observed a failure of strains bearing temperature-sensitive condensin alleles to condense the rDNA at the NPT (Fig. 1; see also Fig. 8b), consistent with prior studies assessing chromosome condensation at the rDNA via DNA FISH (Lavoie et al. 2000, 2002).

Fig. 8.

The temperature-sensitive growth defect in brn1-60 strains is partially rescued by loss of H1. a) Serial dilution of strains grown at the indicated temperature. Strains examined in the top panel were wild-type (YSH298), Δhho1 (YSH1488), brn1-60 (YSH1409), and brn1-60 hho1Δ (YSH1463). Strains examined in the bottom panel were wild-type (YSH1185), hho1Δ (YSH1302), brn1-9 (YSH1323), and brn1-9 hho1Δ (YSH1397). b) Condensation assays were performed as described in the Fig. 1 legend. Strains examined were wild-type (YSH1552; at 23°C N = 112, at 37°C N = 120), hho1Δ (YSH1553; at 23°C N = 104, at 37°C N = 111), brn1-60 (YSH1554; at 23°C N = 57, at 37°C N = 116), and brn1-60 hho1Δ (YSH1555; at 23°C N = 112, at 37°C N = 121). Using Fisher's exact test, the P-value comparing loop and nonloop rDNA for wild-type at 23°C vs 37°C is 0.07, Δhho1 at 23°C vs 37°C is 0.08, brn1-60 at 23°C vs 37°C is <0.00001, brn1-60 Δhho1 at 23°C vs 37°C is <0.00001, brn1-60 at 23°C vs brn1-60 Δhho1 at 23°C is 0.20, and brn1-60 at 37°C vs brn1-60 Δhho1 at 37°C is 0.24. P-values are reported on the figure as ns (not significant; P > 0.05). Error bars represent the standard error of the mean.

To determine if histone H1 influences rDNA condensation in budding yeast, we performed Net1-GFP assays on strains lacking HHO1. Again, we see that ∼70% of wild-type cells exhibit rDNA in the line or loop morphology when blocked in metaphase, with few cells exhibiting the highly decondensed puff morphology (Fig. 2). Strains lacking linker histone H1 have essentially wild-type Net1-GFP patterns, indicating that histone H1 is not required for efficient condensation of the rDNA (Fig. 2). In contrast, metaphase-blocked cells lacking histone variant H2A.Z exhibit a significant decrease in cells with the loop structure and more cells with either the cluster or puff morphologies, compared with the wild-type strain (Fig. 2). H2A.Z loss is associated with DNA repair defects and chromosome instability (Krogan et al. 2004; Keogh et al. 2006; Kalocsay et al. 2009; Morillo-Huesca et al. 2010; Papamichos-Chronakis et al. 2011). To confirm that the change in condensation state was directly due to the loss of H2A.Z protein, we restored the wild-type HTZ1 gene to its endogenous location in this htz1Δ strain and observed reversion to the wild-type rDNA condensation phenotype (Fig. 2). We conclude that cells lacking H2A.Z are defective in chromosome condensation at the rDNA.

Fig. 2.

H2A.Z is required for rDNA condensation. Strains with the indicated genotypes were grown at 30°C and assessed for rDNA morphology as described in Fig. 1. The HTZ1 strain was created by replacing the htz1Δ with wild-type HTZ1 at its endogenous location. Strains used were wild-type (YSH1185, N = 179), hho1Δ (YSH1302, N = 173), htz1Δ (YSH1186, N = 146), and HTZ1 (YSH1319, N = 168). Using Fisher's exact test, the P-value comparing loop and nonloop rDNA for wild-type vs hho1Δ is 1, wild-type vs htz1Δ is <0.0001, and htz1Δ vs HTZ1 is <0.001. P-values are reported on the figure as ns (not significant; P > 0.05) or *** (P < 0.0001). Error bars represent the standard error of the mean.

H2A.Z is subject to acetylation on its N-terminal tail (Babiarz et al. 2006; Keogh et al. 2006; Millar et al. 2006); this acetylation influences H2A.Z's function in multiple contexts (Babiarz et al. 2006; Keogh et al. 2006; Millar et al. 2006; Sharma et al. 2013). To determine if acetylation affects H2A.Z's contribution to chromosome condensation, we replaced the wild-type HTZ1 gene with alleles coding for lysine to arginine substitutions at defined acetylation sites of H2A.Z (Babiarz et al. 2006). As seen in Fig. 3, strains that cannot be acetylated at these 4 sites perform like wild-type strains in the Net1-GFP condensation assay, suggesting that H2A.Z's acetylation state does not affect chromosome condensation.

Fig. 3.

Histone H2A.Z's acetylation state does not influence condensation. Strains with the indicated genotypes were assessed for rDNA morphology as described in Fig. 1. Strains examined were HTZ1-FLAG (YSH1325, N = 105), htz1-K14R-FLAG (YSH1326, N = 52), and htz1-K3,8,10,14R-FLAG (YSH1332, N = 110). Using Fisher's exact test, the P-value comparing loop and nonloop rDNA for HTZ1-FLAG vs htz1-K14R-FLAG is 1 and HTZ1-FLAG vs htz1-K3,8,10,14R-FLAG is 0.07. The P-value comparing puff rDNA vs nonpuff rDNA in HTZ1-FLAG vs htz1-K3,8,10,14R-FLAG is 0.12. P-values are reported on the figure as ns (not significant; P > 0.05). Error bars represent the standard error of the mean.

H2A.Z may be necessary for the condensin complex to initiate chromosome condensation or to help maintain the structure of fully condensed chromosomes or both. To assess the requirement for H2A.Z at discrete intervals in the cell cycle, we used an auxin-inducible degron (Nishimura et al. 2009; Morawska and Ulrich 2013) to deplete H2A.Z. Using this system, we reduced steady-state H2A.Z protein to 10–20% of wild-type levels. Whether H2A.Z was depleted in cells traversing the G1 to G2/M interval (Fig. 4) or in metaphase-blocked cells (Supplementary Fig. 1), we observed that condensation in these cells was equivalent to untreated cells, indicating that yeast cells can establish and maintain condensation in conditions of reduced H2A.Z.

Fig. 4.

Depletion of H2A.Z in metaphase does not affect rDNA condensation. a) Strain YSH1578 (HTZ1-kanMX-AID*-9myc) was grown as described in the Fig. 1 legend. Following α-factor block cultures were split; one was treated for 2 h with 1 mM auxin and then allowed to resume progress in the cell cycle in the presence of auxin. The second culture was subjected to the same blocks but not exposed to auxin. b) Quantitation of H2A.Z levels in auxin-treated cultures relative to untreated controls (N = 3). c) Condensation assay assessing rDNA morphology in cells with (N = 142) and without (N = 125) auxin. Using Fisher's exact test, the P-value comparing loop and nonloop rDNA for the condition with vs without auxin is 0.9. P-values are reported on the figure as ns (not significant; P > 0.05). Error bars represent the standard error of the mean.

Loss of H2A.Z does not alter condensin association with the rDNA

To determine if H2A.Z influences the association of the condensin complex with DNA we used ChIP to measure the binding of epitope-tagged versions of the Brn1 and Smc2 condensin proteins to the RFB region of the rDNA (Fig. 5). As a negative control, we examined condensin association with TUB2, where prior studies had shown minimal condensin association (Supplementary Fig. 2) (Freeman et al. 2000; Wang et al. 2004, 2005). Consistent with prior reports (Wang et al. 2004, 2005), we readily detect association of Smc2 and Brn1 with the RFB (Fig. 5). We find that this association is not increased or decreased in strains lacking H2A.Z, suggesting that the defect in condensation observed in htz1Δ strains is not due to reduced association of the condensin complex with chromatin.

Fig. 5.

Loss of HTZ1 does not alter condensin association with the rDNA. ChIP analysis of Smc2-HA and Brn1-HA using a probe for the RFB region of the rDNA (the location of the probe is shown by the black bar). Black bars show the RFB signal from anti-HA immunoprecipitations (HA), and lightly shaded bars show signal from the no antibody (NAB) controls. The deletion of H2A.Z did not significantly change condensin association at the RFB (P-value using a t-test for SMC2-HA vs SMC2-HA htz1Δ is 0.39, and for BRN1-HA vs BRN1-HA htz1Δ, the P-value is 0.97). Strains used in ChIP assays are untagged (YSH1185), SMC2-HA (YSH1341), BRN1-HA (YSH1342), SMC2-HA htz1Δ (YSH1344), and BRN1-HA htz1Δ (YSH1345). For both the RFB and TUB2 experiments, N = 3 for the untagged strain and N = 6 for all other strains; error bars are standard error.

Loss of HHO1 or SWR1 can rescue the defect caused by deletion of H2A.Z

Evidence from metazoans is consistent with a role for histone H1 in promoting chromosome condensation (see Introduction). To see if we could detect an influence of H1 in strains partially compromised for condensation activity, we also examined the effects of eliminating HHO1 in strains lacking H2A.Z. Unexpectedly, we find a recovery to the wild-type phenotype in htz1Δ hho1Δ cells (Fig. 6). These results indicate that histone H1 antagonizes chromosome condensation in cells lacking H2A.Z.

Fig. 6.

Deletion of SWR1 or HHO1 suppresses condensation defects in htz1Δ strains. Condensation assays were performed as described in Fig. 1. Strains tested were swr1Δ (YSH1331, N = 155), htz1Δ swr1Δ (YSH1330, N = 218), and htz1Δ hho1Δ (YSH1303, N = 154). The wild-type and htz1Δ data from Fig. 2 is included for comparison. Using Fisher's exact test, the P-value comparing loop and nonloop rDNA for wild-type vs swr1Δ is 0.8, wild-type vs htz1Δ swr1Δ is 0.4, and wild-type vs htz1Δ hho1Δ is 0.4. P-values are reported on the figure as ns (not significant; P > 0.05). Error bars represent the standard error of the mean.

The SWR-C complex exchanges H2A for H2A.Z in chromatin (Kobor et al. 2004; Krogan et al. 2004; Mizuguchi et al. 2004). The deletion of SWR1, which codes for the catalytic subunit of SWR-C, can rescue some phenotypes caused by the absence of H2A.Z, possibly by alleviating SWR-C alteration of chromatin structure in the absence of H2A.Z (Halley et al. 2010; Morillo-Huesca et al. 2010; Hang and Smith 2011). We find that yeast strains lacking SWR1 have normal rDNA condensation (Fig. 6) but observe that loss of Swr1 can suppress the condensation defect in htz1Δ strains.

Histone H1 inhibits chromosome condensation in the absence of H2A.Z, but its loss in otherwise wild-type cells has no effect. Could the loss of H2A.Z alter H1's association with the rDNA in a manner that influences condensation? To examine this possibility, we used ChIP to track H1's association with the rDNA in the presence and absence of H2A.Z, in log phase and nocodazole-blocked cells. As shown in Fig. 7a, we observe that the amount of H1 bound to the rDNA is broadly independent of cell cycle phase or the presence of H2A.Z. We also conducted the reciprocal experiment, examining H2A.Z association with the rDNA in the presence or absence of histone H1 (Fig. 7a). Our data suggests a reduction in H2A.Z at rDNA sequences in mitotic cells; while the loss of H1 reduces the amount of H2A.Z at the rDNA promoter in log phase cells, H2A.Z association with rDNA chromatin in nocodazole-blocked cells is unchanged in the absence of H1.

Fig. 7.

H1 and H2A.Z are components of rDNA chromatin. a) ChIP analysis of H1 using probes for the rDNA promoter, and RFB was performed in HTZ1 and htz1Δ strains, grown to log phase or blocked with nocodazole. b) ChIP analysis of Htz1-HA performed in HHO1 and hho1Δ strains, grown to log phase or blocked with nocodazole. For each experiment, the signal from the IP (dark blue) and the NAB control (light blue) are reported as a percent of the input chromatin sample.

Altered expression of condensin subunits is unlikely to be the cause of rDNA condensation defects in htz1Δ strains

Strains lacking H2A.Z have altered patterns of transcription (Santisteban et al. 2000; Meneghini et al. 2003; Krogan et al. 2004). To examine if the loss of H2A.Z affects condensation by altering the transcription of condensin genes, we measured condensin mRNA levels in strains lacking HTZ1 using RNA-seq. Table 2 lists expression levels of condensin genes in the indicated strains, relative to their expression level in wild-type cells. In general, condensin genes have slightly reduced mRNA levels in cells lacking H2A.Z; however, these levels are equally reduced in strains that do not exhibit condensation defects. In particular, condensin mRNA levels are roughly equal in the htz1Δ strain, which exhibits a condensation defect, and the htz1Δ swr1Δ and htz1Δ hho1Δ strains, which do not. While H2A.Z and H1 could influence the transcription of other genes that affect condensation, we conclude that the condensation defect we observe in strains lacking H2A.Z is unlikely to be due to reduced expression of condensin proteins.

Table 2.

Relative mRNA levels of condensin genes.

Gene	htz1Δ	hho1Δ	swr1Δ	htz1Δ hho1Δ	htz1Δ swr1Δ
BRN1	0.79	0.73	0.70	0.62	0.87
SMC2	0.84	0.74	0.79	0.76	0.84
SMC4	0.81	0.74	0.76	0.69	0.81
YCG1	0.85	0.91	0.85	0.83	0.92
YCS4	0.76	0.89	0.79	0.83	0.87

A genetic interaction between HHO1 and BRN1

Loss of histone H1 can suppress the rDNA condensation defect caused by the absence of variant histone H2A.Z. To examine the specificity of this effect, we assessed the consequences of eliminating HHO1 in cells bearing temperature-sensitive alleles of each of the 5 condensin genes. These strains fail to grow at elevated temperatures due to compromised function of condensin proteins. As shown in Fig. 8a, the absence of H1 does not affect the growth of wild-type strains grown at elevated temperatures. Similarly, the growth of strains bearing temperature-sensitive alleles of the SMC2, SMC4, YCG1, or YCS4 condensin genes is not influenced by the absence of H1 at the temperatures tested (Supplementary Fig. 3). However, we observe that loss of H1 can partially suppress the temperature-sensitive growth defect of strains bearing the brn1-60 allele. This suppression is allele-specific, as H1 loss does not affect growth of the brn1-9 strain (Fig. 8a). Loss of H1 failed to suppress the defect in rDNA condensation caused by the brn1-60 mutation (Fig. 8b), suggesting that the condensation defect at the rDNA locus may not be the direct cause of viability loss in this strain. A prior study determined that the brn1-60 allele codes for a Brn1 protein with 2 amino acid substitutions, K489E and P490S (Ouspenski et al. 2000), which we confirmed by sequencing. Interestingly, these amino acid changes are located in the “safety belt” region, which anchors the condensin complex to the chromatin by encircling and trapping the DNA (Kschonsak et al. 2017). Sequencing brn1-9 alleles revealed 3 common coding changes, E607G, F692S, and T734I, which lie in the C-terminus of the Brn1 subunit, a region not predicted to interact with DNA (Kschonsak et al. 2017).

Discussion

Local chromatin structure allows specificity in the regulation of transcription, recombination, and other discrete DNA transactions. The specific influence of chromatin modifications and histone variants on larger-scale chromosome organization has yet to be established in equivalent detail. In this study, we examined the influence of linker histone H1 and variant histone H2A.Z on chromosome condensation in yeast cells. In vitro, H2A.Z-containing nucleosome particles are less stable than H2A nucleosomes (Suto et al. 2000; Abbott et al. 2001; Lewis et al. 2021) while the inclusion of H2A.Z in nucleosome arrays have been reported to increase folding of the arrays (Fan et al. 2002; Lewis et al. 2021).

Experiments in yeast and mammalian systems identified a role for H2A.Z or the SWR-C complex in promoting chromosome stability (Krogan et al. 2004; Rangasamy et al. 2004; Keogh et al. 2006; Kim et al. 2009; Sharma et al. 2013, Thattikota et al. 2018, Sales-Gil et al 2021). Here, we demonstrate that loss of H2A.Z in budding yeast cells causes a defect in rDNA condensation. While deletion of SWR1, coding for a core component of the SWR-C complex that deposits H2A.Z in chromatin, does not significantly affect chromosome condensation in our assay, we observe that the condensation defects caused by H2A.Z loss can be rescued by also removing Swr1. Prior reports found that a subset of phenotypes caused by loss of H2A.Z in yeast can be suppressed by deleting the SWR1 gene or other components of the SWR-C complex, suggesting that some defects observed in strains lacking H2A.Z are caused by deleterious action of the SWR-C complex in H2A.Z's absence (Halley et al. 2010; Morillo-Huesca et al. 2010, 2019; Hang and Smith 2011). The specific effects of SWR-C on chromatin or other cellular processes in cells that lack H2A.Z have not been defined (Morillo-Huesca et al. 2010, 2019).

How might H2A.Z/SWR-C affect chromosome condensation? Our mRNA expression analysis suggests that altered transcription of the condensin genes is unlikely to be the cause of the condensation defect we observe (Table 2). Schizosaccharomyces pombe cells lacking H2A.Z exhibit aberrant chromosome segregation in mitosis, associated with premature dissociation of the condensin complex from chromatin (Kim et al. 2009), while human condensin subunit CAP-H (the yeast Brn1 homolog) binds to histone H2A and H2A.Z in vitro, and HeLa cells depleted for H2A.Z have reduced condensin association with chromatin (Tada et al. 2011). However, we fail to observe a change in the association of condensin with the rDNA locus in htz1Δ cells. Unlike in S. pombe, Saccharomyces cerevisiae cells lacking H2A.Z exhibit synthetic lethality with mutations in mitotic checkpoint genes (Krogan et al. 2004; Tong et al. 2004; Kim et al. 2009), suggesting H2A.Z makes distinct contributions to chromosome segregation in budding yeast.

Most prior studies examining H1's influence on chromosome condensation in eukaryotes indicated a broadly positive role (Ner and Travers 1994; Shen et al. 1995; Fan et al. 2003; Maresca et al. 2005; see Introduction). However, vertebrate linker histone H1.8, the predominant linker histone expressed in oocytes and early embryos, suppresses the binding of condensins and DNA topoisomerase II on mitotic chromosomes (Choppakatla et al. 2021). While we found that deletion of HHO1 in otherwise wild-type yeast strains did not affect rDNA condensation, we made the unexpected discovery that loss of H1 can alleviate the condensation defects observed in htz1Δ strains, suggesting that a primary cause of chromosome instability in strains lacking H2A.Z is a deleterious action of histone H1. We previously reported that a spore viability defect caused by the absence of histone H2A.Z can be partially suppressed if cells also lack histone H1, while the combined loss of both H1 and H2A.Z is associated with elevated gene conversion events in cells progressing through meiosis (Chigweshe et al. 2022). Our results indicate that chromosome stability in yeast cells relies on balancing the activities of histone H1 and H2A.Z.

Loss of Swr1 or other Swr complex subunits can suppress some of the phenotypes caused by the absence of H2A.Z; one proposal to explain this observation is that Swr destabilizes nucleosomes in cells lacking H2A.Z (Halley et al. 2010; Morillo-Huesca et al. 2010; Hang and Smith 2011). Loss of histone H1 leads to a similar suppression of H2A.Z loss. Might the modes of suppression in swr1Δ and hho1Δ cells be related? If H1 is required to recruit the Swr complex to chromatin, then hho1Δ cells might phenocopy swr1Δ cells. In support of this possibility, the Swr complex subunit Arp4 has been reported to physically interact with H1 in yeast cells (Georgieva et al. 2015). However, we do not observe a dramatic decrease in H2A.Z association with chromatin in hho1Δ cells (Fig. 7b), as would be predicted if H1 was a primary determinant of Swr complex recruitment.

Loop-extruding yeast condensins can readily bypass large molecules, including RNA polymerase; thus, inhibition of chromatin condensation by chromatin structure is unlikely to occur at the level of individual nucleosomes (Kong et al. 2020; Pradhan et al. 2021). However, particular chromatin states may attract or inhibit the association or action of SMC protein complexes (Muñoz et al. 2020), and modulating the stiffness of nucleosome arrays via posttranslational modifications might reduce the efficiency of loop extrusion (Kim et al. 2023). In vitro, histone H1 has been shown to promote higher-order chromatin packing that may decrease the efficiency of loop extrusion (Perišić et al. 2019). An inhibitory influence of H1 on condensin binding or action is suggested by our observation that loss of H1 can partially suppress mutations in the DNA binding domain of the Brn1 condensin. Loss of H2A.Z might create chromatin domains that are inefficiently manipulated by the condensin complex due to exacerbated action of histone H1, while elimination of histone H1 might relieve this structural impediment. Our observations open new avenues to examine the role of local chromatin changes to chromosome condensation.

Data availability

The authors affirm that all data necessary for confirming the conclusions of this article are represented fully within the article and its tables and figures.

Supplemental material available at GENETICS online.

Funding

S.G.H. gratefully acknowledges funding from National Institutes of Health grant (1R15GM102824-01).