Gene network analysis

Yeast genetic and protein interactions data (Krogan et al. 2006; Collins et al. 2007a, 2007b; Fiedler et al. 2009) were downloaded from the BIOGRID database (Stark et al. 2006 and references therein). A gene/protein interaction network was built using GeneMania algorithm (Warde-Farley et al. 2010) as implemented in cytoscape web application. Networks were constructed and visualized using cytoscape version 3.1.1 (Cline et al. 2007).

Strains used in the study

S. cerevisiae strains used in this study were generated through standard molecular genetic procedures and are listed in Supplementary Table S1. Primer sequences used for overlapping polymerase chain reactions are listed in Supplementary Table S2. Tetrad analysis was carried out using standard methods. A total of 9 tetrads were dissected for each cross as indicated in text.

Extraction of proteins and Western blots

Whole cell extracts were prepared using trichloroacetic acid (TCA) as described previously (Kao and Osley 2003). Proteins were separated using 10% SDS-PAGE, transferred onto nitrocellulose paper, and immunoblotted with antibodies mouse monoclonal IgG α-MYC (Santa Cruz Biotechnology) primary antibody, or rabbit monoclonal α-TAP (Thermo scientific) primary antibody at 1:7500 dilutions. Equivalent loading in each lane was assessed by Ponceau S.

Immunoprecipitations (IPs)

IPs were performed as previously described (Kobor et al. 2004). Histone IPs were carried out essentially as described above with the following modifications: 2 µg of anti-acetyl H4 antibody (Abcam: ab46983) was conjugated with 25 µl of protein G Dynabeads (Thermo Fisher Scientific: 10003D) for one hour at room temperature. The antibody-conjugated beads were then incubated with the cell lyses for overnight at 4°.

Indirect immunofluorescence

Indirect immunofluorescence was performed using a modified method of (Poveda et al. 2004). Briefly, cells grown to mid log phase (~2 × 10⁷) were fixed in 4% formaldehyde for 60 min with gentle agitation at room temperature. Cells were washed in 1x PBS twice and resuspended in 500 μL of Spheroplasting Buffer (2% Glucose, 1x Amino Acids, 1x Sorbitol, 20 mM Tris-HCl pH 7.5 and 1x YNB), and kept at 4° overnight. Next morning, 10 mg/ml zymolyase 100T and 1.42 M beta-mercaptoethanol were added to 200μL of cells in spheroplasting buffer, and incubated at 30° for 90 min. Cells were then washed twice with 1mL of 1xPBS +0.05% Tween20 and suspended in 1xPBS +0.05% Tween20. The cell suspension was added to poly-D-lysine coated cover slips and dried for 20 min followed by three washes with 1x PBS. 1xPBS +1 mg/mL BSA blocking buffer was added for 30 min, and were then washed three times with 1xPBS. Primary antibody (anti-Myc) was added in a 1:250-dilution in 1xPBS+1mg/mL BSA at room temperature for 1 hr in a humid chamber. After three washes with 1xPBS, fluorescent secondary antibodies (anti-mouse, Alexa488) in a 1:1000 dilution in 1xPBS+1mg/mL BSA were incubated in the dark for 45 min. Subsequently, the cover slips were washed three times with 1xPBS, and incubated with 1:5.000 diluted DAPI in 1xPBS for 10 min. Coverslips were then washed three times with 1xPBS prior to adhering the cover slips onto a clean slide with Fluorescent Mounting Media (Dako). Microscopy was performed using Olympus Spinning Disk Confocal Microscopy at the Sick Children Hospital Bio-Imaging Facility.

Spot assays for DNA-damage sensitivity and histone Over-expression

Yeast strains were inoculated in 5ml YPD medium and were grown overnight at 30° within shaking. The optical density of each culture was measured and adjusted to OD600nm = 0.2. Cells were then grown to OD600nm = 0.5, and then 1/4 serial dilutions of each cell culture were spotted onto plates containing 50 mM or 100 mM hydroxyurea. Plates were incubated at 30° for 3 days.

For histone over-expression assay, strains were transformed with tagged histone H3 under the control of a galactose-inducible GAL1-10 promoter or the empty vector control (pYES2). Cells were grown overnight in 5 ml minimal medium (YNB) either minus uracil or minus uracil/minus leucine and containing 2% raffinose. Expression of H3 was induced by addition of 2% galactose for about four hours, and cultures were then grown to approximately OD600nm = 0.8. Sixfold serial dilutions of each strain were performed and cell culture were spotted onto plates containing minimal media either lacking uracil or lacking both uracil and leucine, with either glucose (H3 = OFF) or galactose (H3 = ON) as carbon source. Plates were incubated at 30° for 3 days.

Hif1 interacts With Hat1/Hat2 via acidic region of TPR2

Several studies have previously established that S. cerevisiae Hif1 physically interacts with the Hat1/Hat2-complex presumably via Hat2 (for review see 42). However, the specific region on Hif1 that might be responsible for this interaction has not yet been identified. Our evolutionary analysis (refer to Figures) indicates that Hif1 has a conserved pattern of four TPR motifs, an acidic region which disrupts the TPR2, and a C-terminal putative NLS. These conserved regions provide candidate functional units that might be involved in an interaction with the HAT1-complex. Therefore, to map the Hat1 interaction region in S. cerevisiae Hif1 we generated a series of truncation mutants lacking either a specific TPR1-4 motif, or the first alpha helix of TPR2 (before start of the acidic region), or only the acidic region (Figure S3A). We also generated five C-terminal deletion mutants lacking either the NLS, or the last 35 C-terminal residues (including the NLS), or the whole C-terminus until the TPR4, or the C-terminus including TPR4, or the C-terminus until TPR2 (Figure S3A). Wild type and mutant Hif1 proteins were N-terminally tagged with 12×MYC and were transformed in hif1 yeast cells that endogenously expressed C-terminally tagged Hat1 TAP-tag (where TAP is tandem affinity purification). We maintained transformed plasmids (pRB415-12×MYC) using a LEU2 marker on media without leucine. Subsequently, the expression of 12MYC-HIF1 mutants as well as HAT1-TAP was assessed on Western blots using anti-MYC and anti-TAP antibodies, respectively, which showed even expression levels (Figure S3 B, C).

In order to identify the Hat1 interacting region of Hif1, we immunoprecipitated Hat1-TAP and probed for 12MYC-Hif1 to confirm that the full length (F.L.) Hif1 and Hat1 proteins interact (Figure 2A). We predicted that 12MYC-Hif1 truncation mutants without the TPR2 acidic region should not co-immunoprecipitate with Hat1-TAP. We began by assessing the ability of C-terminal truncation 12MYC-Hif1 mutants to immunoprecipitate with Hat1-TAP. We observed that none of the five C-terminal deletion mutations affected Hif1-Hat1 interaction, suggesting that the Hif1 C-terminus is not required for this interaction (Figure 2A).

Open in a separate window

Figure 2

Western blot analysis of Co-IP fractions of Hat1/2-TAP and Hif1 C-terminal (external) and internal deletions constructs. A: (Left) Input fractions of Co-IP experiments for various Hif1 C-terminal (external) deletions. (Right) Western blot analysis of Co-IP fractions of Hif1 C-terminal mutants to assess their ability to immunoprecipitate with Hat1-TAP B: (Left) Input fractions of Co-IP experiments for various Hif1 internal deletion mutants. (Right) Co-IP samples of Hif1 internal deletions. C: (Left) Input fractions of Co-IP experiments for various Hif1 C-terminal (external) deletions. (Right) Western blot analysis of Co-IP fractions of Hif1 C-terminal mutants to assess their ability to immunoprecipitate with Hat2-TAP D: (Left) Input fractions of Co-IP experiments for various Hif1 internal deletion mutants. (Right) Co-IP samples of Hif1 internal deletions. The red arrows represent the position of HAT2. Note: The size difference of various Hif1 truncated mutants represents various deletions. The top panels were probed with anti-MYC antibody whereas bottom panels were probed with anti-TAP antibody.

Next, we tested the Hat1-binding abilities of 12MYC-Hif1 mutants carrying internal deletions for a specific TPR motif or only the acidic region. As shown in Figure 2B, internal deletions for either of TPR1, TPR3 and TPR4 did not prevent 12MYC-Hif1 association with the Hat1-TAP. This suggests that none of the three above noted TPR motifs are required for the Hif1-Hat1 interaction. Interestingly, deletion of TPR2 resulted in loss of an interaction (see lane 3 from the right on Figure 2B) indicating that this region is critical for binding with the Hat1-complex. The TPR2 is composed of acidic interruption region that is flanked by two alpha helices (Liu et al. 2014). We decided to assess the role of various TPR2 segments in the Hat1 binding. Remarkably, the deletion of TPR2 alpha helix found prior to the start of acidic region did not affect the ability of 12MYC-Hif1 to co-immunoprecipitate with Hat1-TAP (Figure 2B). Conversely, the deletion of the acidic region completely abolished this interaction (see lane 4 from the right on Figure 2B) indicating that Hat1 interaction depends on the acidic region within TPR2.

Hif1 has previously been shown to physically interact with the Hat1/Hat2-complex via Hat2 (Haigney et al. 2015). To further characterize the Hif1 interaction with the Hat1/Hat2-complex, we extended our Co-IP analysis and mapped the Hat2 interaction region on Hif1 (Figure 2C, D). As expected, we observed that the deletion of the TPR2 completely abolished Hif1-Hat2 interaction (Figure 2D). Similar to what we observed for Hat1 (see above), the mutant lacking the TPR2 acidic region resulted in the loss of Hif1-Hat2 interaction. In contrast neither the C-terminal deletion mutations nor the mutations carrying internal deletions for TPR1, TPR3 and TPR4 abolished the Hif1-Hat2 interaction indicating that these regions are not required (Figure 2C, D). Taken together these results demonstrate that the acidic region that interrupts TPR2 is essential for Hif1 interaction with the Hat1/Hat2-complex.

Hif1 interacts With histones H3/H4 via acidic region of TPR2

Hif1 has been shown to bind with histones H3/H4 in vitro (Ai and Parthun 2004). Furthermore, as a member of the Hat1-complex, Hif1 participates in the acetylation of newly synthesized histone H4 (Campos et al. 2010). To investigate the mechanism through which Hif1 binds histones H3/H4 in vivo, we carried out Co-IP experiments using the Hif1 mutants described above (see Figure S3A). As in Figure 2, each of the Hif1 mutants were expressed in cells lacking the wild type HIF1 (Δhif1), and histones were pulled down using an antibody against acetylated H4 (Figure 3A, B). Consistent with what we observed for the Hat1/Hat2, the Hif1 mutants lacking the TPR2 and/or acidic-interruption region failed to immunoprecipitate with the acetylated histone H4 (Figure 3B). In contrast, neither the C-terminal deletion mutations nor the mutations carrying internal deletions for TPR1, TPR3 and TPR4 abolished the Hif1 association with H4 (Figure 3A, B). These results are consistent with a recent study suggesting that Hif1 TPR2 acidic region is important for H3/H4 binding in vitro (Liu et al. 2014; Zhang et al. 2016).

Open in a separate window

Figure 3

Co-IP analysis of Hif1 truncation mutants and histone H4/ Asf1. A: (Left) Input fractions for various Hif1 C-terminal (external) deletions. (Right) Western blot of Co-IP fractions of Hif1 C-terminal mutants to monitor their ability to co-immunoprecipitate with acetylated H4. B: (Left) Input fractions of Co-IP experiments for Hif1 internal deletion mutants. (Right) Co-IP samples of Hif1 internal deletions to examine their ability to come down with acetylated H4. C: (Left) Input fractions for various Hif1 C-terminal (external) deletions. (Right) Western blot analysis of Co-IP fractions of Hif1 C-terminal mutants to assess their ability to immunoprecipitate with Asf1 D: (Right) Input fractions of Co-IP experiments for Hif1 internal deletion mutants. (Left) Western blot analysis of Co-IP samples of Hif1 internal deletion mutants to assess their ability to immunoprecipitate with Asf1. E: One step co-IP using TEV protease elution of Asf1-TAP was performed in order to obtain better resolution of Asf1-TAP and the shorter Hif1 internal deletion clones (see methods). The Asf1 size difference between inputs and IPs is due to TEV cleavage. Note: The top panels were probed with anti-MYC antibody whereas middle panel was probed with either anti-acH4 or anti-TAP antibody to detect acetylated H4 or Asf1, respectively.

Hif1 interacts With Asf1 via acidic region of TPR2

Asf1 is a key generalized H3/H4 chaperone that functions in an array of chromatin-related processes including the transport of newly synthesized H3/H4 and chromatin assembly pathways (English et al. 2006; Campos et al. 2010; Alvarez et al. 2011) . Asf1 and Hif1 interaction has been extensively reported where Asf1 is thought to function downstream of Hif1 in the H3/H4 transport pathway (Campos et al. 2010). In vitro studies have suggested that the interaction between Hif1 and Asf1 is likely mediated by histones H3/H4 (Haigney et al. 2015; Bowman et al. 2017). To examine what region of Hif1 is required for an interaction with Asf1 in vivo, we employed our described Co-IP strategy. We expressed the Hif1 mutants (see Figure S3A) in hif1 cells expressing C-terminally TAP-tagged Asf1 from its endogenous locus. Our Co-IP experiments indicated the TPR2 as the region responsible for the Hif1-Asf1 interaction (Figure 3C-E). Remarkably, we observed that the Hif1 mutant lacking only the acidic region of TPR2 was defective in its ability to immunoprecipitate with Asf1 (Figure 3C-E). In contrast, the C-terminal deletion mutations and the mutations carrying the internal deletions for TPR1, TPR3 and TPR4 were found not to be required for Hif1-Asf1 interaction (Figure 3C-E). These results establish that the acidic region of TPR2 is functionally important for Hif1-Asf1 interaction. Because the TPR2 acidic region is also essential for H3/H4 binding (see above), these results, taken together, suggest that the in vivo Asf1-Hif1 interaction likely mediated by H3/H4 consistent with previous in vitro (Haigney et al. 2015) and in vivo studies (Campos et al. 2010).

The conserved basic patch is essential for Hif1 nuclear localization

Our molecular evolutionary analysis indicated that Hif1 carries a conserved basic patch at its extreme C-terminus which shares similarity to the consensus NLS (Figure S2). To examine the functionality of the putative NLS we carried out an indirect immunofluorescence (IF) analysis using 12MYC tagged Hif1 cells. In accordance with previous studies (Poveda et al. 2004), the full length 12MYC-Hif1 mainly localized to the nucleus (Figure 4). Strikingly, we observed that Hif1 mutant lacking the C-terminal basic patch was mostly located in the cytoplasm indicating that this region is required for the proper nuclear import of the protein (Figure 4). We extended our IF analysis using the Hif1 mutants described above. All Hif1 mutants carrying an internal deletion predominantly localized to the nucleus indicating that eliminating any of these regions does not abolish nuclear localization. Conversely, Hif1 mutants carrying C-terminal external deletions were found to be defective for proper nuclear localization and the signal was observed mostly in the cytoplasm (Figure S4). Together these observations establish that Hif1 C-terminal basic patch is essential for the protein’s nuclear localization consistent with function as an NLS.

Open in a separate window

Figure 4

Indirect immunofluorescence analysis of Hif1. Top panel was stained with DAPI to capture the nuclear orientation. Middle panel was probed with anti-MYC antibody to examine the localization of either full length Hif1 (F.L) or truncated mutant lacking basic patch (-B.P). Cells transfected with an empty vector were used as a control. Red arrows point toward the nuclei of representative cells. The bottom panel represents merge of DAPI and anti-MYC staining.

hif1Δ mutants are sensitive to histone Over-expression

Excessive soluble histones are known to have deleterious effects on genomic stability (Singh et al. 2010; Huddleston 2011). In S. cerevisiae excessive histones are immediately degraded via a pathway that includes Rad53 and the Proteasome (Gunjan and Verreault 2003; Singh et al. 2009). Previously, Lsm1 has been shown to function in cell cycle based regulation of histone mRNA decay (Herrero and Moreno 2011). Interestingly, NASP-family proteins among chordates have been reported to also be important for regulation of histone metabolism (Dilworth et al. 1987; Cook et al. 2011). We therefore asked if Hif1 has a role in regulating histone dynamics. We initiated our analysis by assessing the effect of histone over-expression in hif1Δ cells. If Hif1 is important for the regulation of histones, we expected hif1Δ cells to be sensitive to histone over-expression. We also included hat1Δ and hat2Δ strains in our analysis since these proteins along with Hif1 form a nuclear NuB4 complex (Parthun 2012). Furthermore, a lsm1Δ strain was included because cells lacking LSM1 have been reported to exhibit hypersensitivity to histone over-expression (Herrero and Moreno 2011). Wild-type, hif1Δ, hat1Δ, hat2Δ and lsm1Δ strains were transformed with plasmids encoding histone H3 gene under the control of GAL1 promoter or with an empty vector. Subsequently, raffinose-containing medium was used to grow the transformants and cell cultures in serial dilutions were plated on glucose (promoter OFF) or galactose (promoter ON)-containing media.

We found that hif1Δ and hat2Δ cells were mildly sensitive to H3 over-expression whereas hat1Δ cells were not affected (Figure 5A), suggesting a role of Hif1 and Hat2 proteins in histone homeostasis. We also observed that hif1Δ, hat1Δ, and hat2Δ cells were mildly sensitive to genotoxic agent hydroxyurea (HU) which stalls replication forks and can induce DNA DSBs after prolonged exposure (Petermann et al. 2010) (Figure S5). In accordance with a previous report (Herrero and Moreno 2011), cells lacking LSM1 exhibited a hypersensitivity to histone H3 over-expression (Figure 5A) as well as to HU (Figure S5). Since hif1Δ, hat2Δ and lsm1Δ cells exhibit some degree of phenotype in response to H3 over-expression, we reasoned that Hif1, Hat2 and Lsm1 might be functionally related. To answer this question, we carried out tetrad analysis of crosses between haploid lsm1Δ and hif1Δ, hat2Δ, and hat1Δ cells. For each cross we dissected nine tetrads after sporulation. In contrast to lsm1Δhat1Δ cells which grew normally, a severe phenotype of slow growth was observed for lsm1Δhat2Δ and lsm1Δhif1Δ segregants (Figure 5B). These observations suggest a synergistic functional link between Lsm1 and Hif1/Hat2 proteins.

Open in a separate window

Figure 5

Histone over-expression analysis. A: Effect of histone over-expression in hif1Δ, hat1Δ, hat2Δ and lsm1Δ mutant strains. Note: Growth on galactose turns GAL1 promoter ON resulting in histone over-expression (O/E) B: Tetrad analysis of yeast strains. Yeast haploid strains lsm1Δ and hif1Δ, hat2Δ, and hat1Δ were crossed. Representative tetrads that were dissected for each cross are shown. Double mutants are shown as circles.

To further establish the functional link between Hif1 and Hat2 proteins, we generated a double knockout lacking both the HIF1 and HAT2. Interestingly, hif1Δ hat2Δ double knockout cells were synthetic sick and exhibited severe growth defects on YPD media (Figure 6). We then expressed back the above described truncation mutants into hif1Δ hat2Δ cells to examine whether they could rescue the observed phenotype. Interestingly, the full length Hif1 successfully rescued the phenotype, however, none of the examined truncation mutants were able to overcome the growth defects of hif1Δ hat2Δ cells (Figure 6). These results indicate the requirement of various Hif1 functional domains for its proper functioning. Taken together, the overall pattern of sensitivity to H3 over-expression and genetic interaction of Hif1/Hat2 with Lsm1 suggest that these proteins presumably function synergistically to maintain appropriate histone levels.

Open in a separate window

Figure 6

Analysis of hif1Δ hat2Δ double knockout cells for growth defects on YPD media. A: Strains were grown to an OD at 600nm of 0.5 before being plated at fivefold serial dilutions on YPD. B: Strains were cloned into pRB4151-2MYC lacking the Leucine amino acid (-Leu) for selectivity. Hif1 F.L. was transformed into hif1Δ/hat2Δ and hif1Δ to rescue the phenotype. C, D: Various Hif1 truncations were expressed back into hif1Δ hat2Δ double knockout cells to examine their ability to rescue the phenotype. Note: C1 and D1 represent 3 days of growth. E: Western blot analysis of whole cell extracts to examine the expression of Hif1 truncation mutants in hif1Δ hat2Δ double knockout cells.

Conservation of Hif1 among fungi

We have examined Hif1 homologs with a conserved TPR motif architecture within all major fungal groups. This finding is consistent with our previous report (Nabeel-Shah et al. 2014) which established that NASP-family proteins are conserved throughout the major eukaryotic super-groups. NASP-family proteins share a common feature of being histone chaperones however a degree of functional diversity does exist across different species. A clear example of this functional diversity has been noted for closely related S. cerevisiae and S. pombe fungal lineages. S. cerevisiae Hif1 is a H3/H4 chaperone that binds with Hat1/Hat2 to mediate H4 acetylation (Parthun 2012) whereas S. pombe Sim3 is a CenpA specific chaperone and has been reported to functionally overlap with Asf1 (Dunleavy et al. 2007; Tanae et al. 2012). Consistent with these previous reports, our MSA analysis revealed lineage specific differences at key amino acid positions within various Hif1 homologs. For example, variable length insertions into TPR2 and TPR4 motifs argue in favor of functional diversity within closely related lineages.

The presence of large number of acidic residues interrupting TPR2 results in an overall net negative charge which appears necessary for an interaction with positively charged histones (Liu et al. 2014). The acidic region is critical for proper Hif1 function (Liu et al. 2014) and has been reported to be a subject of strong purifying selection (Nabeel-Shah et al. 2014). Several studies have experimentally established the functional importance of this TPR2 acidic interruption region. Considering its conservation during evolution, and taking earlier findings into account we suggest that the observed variations within the acidic interruption region likely accounts for the functional diversity previously reported within closely related fungal lineages (see above).

We thank Janusan Baskararajah and Oluwafikemi Omotoso for technical assistance. We also thank Dr. Costin Antonescu (Ryerson University) for excellent technical advice and helpful discussions. The plasmid pRB415A-12Myc was a generous gift from Dr. Roberto Botelho (Ryerson University). This work was supported by NSERC Discovery Grants to JF (grant number: 06448–2015) and SAS (grant number: 04405-2015). NSD was supported by King Abdullah Scholarship Program from the Saudi Arabia ministry of higher education. All authors have read and approved the final manuscript.