Figure 1.
Random mutagenesis identifies residues in the unstructured regions of Htz1 that are required for function.
[A] Overview of random mutagenesis screen. Yeast strain CMY307 containing PGAL1-regulated HA-HTZ1 and asf1∆ was transformed with a library of randomly mutated htz1 alleles. Individual clones were picked into 96-well plates and pre-grown in glucose for 72H to repress PGAL1-HA-HTZ1 before plating in duplicate onto SG-Ura and SD-Ura to assess growth in the presence and absence of WT Htz1. CMY307 transformed with plasmids carrying WT Htz1 (black box) or empty vector (blue box) were included as controls on each plate. An example of a glucose-sensitive clone is shown boxed in red. [B] Western blot of protein lysates prepared from CMY307 grown in galactose and at various time-points after glucose addition show that WT HA-Htz1 protein is undetectable by 48H. [C] Frequency of mutations at each Htz1 residue recovered from non-functional alleles identified by screening as described in [A]. The coloured bars in the graph represent residues within known structured regions. A schematic of H2A.Z protein structure, where boxes depict α-helices, the black bars indicate unstructured regions and the M6 region is boxed in pink, is shown. The N- and C-terminal tails and the inter-helical loops (LN, L1, L2, LC) are also labelled and residues I109 and S111 are indicated with red arrows. [D] The frequency of mutations at each residue in a sample of clones from the random mutagenesis library that were not screened for Htz1 function in yeast. [E] Summary of mutation frequencies per residue in each structural region of Htz1. The protein sequence was divided into regions including the N- and C-termini, the various alpha helices and the inter-helical loops, as indicated in [C]. The mutation frequency for each region was normalised to the number of residues in that region to allow comparison. S1 (serine 1) was treated separately as most of the N-terminal mutations were at this residue.
Figure 2.
I109T and S111P point mutations disrupt normal Htz1 function and chromatin association.
[A] Htz1 C-terminal mutants are sensitive to cytotoxic stress. To compare the growth of strains carrying WT or mutant htz1 alleles in the presence of the indicated drugs, cells were serially diluted 1:5 and spotted onto plates, from left to right in each panel. The identities of the htz1 alleles are indicated at the left of the panel. [B] C-terminal mutants have a chromatin association defect. Sub-cellular fractionation was used to isolate cytosolic (Cy), nuclear (N) and insoluble chromatin fractions (C), from total cell lysate (T). Fractions were analysed by SDS-PAGE and immunoblotting, with anti-G6PDH used as a cytoplasmic loading control, anti-histone H4 as a chromatin loading control, and anti-HA used to detect HA-tagged Htz1. The identity of the Htz1 protein is indicated above each panel. [C] Quantification of HA-Htz1 protein levels normalised to H4 (Htz1/H4) from Western blots. Graphs show the averages of mutant HA-Htz1 levels normalised to the WT level in chromatin (left) and total protein (right) fractions from four independent experiments. Error bars correspond to the standard error of the mean and asterisks indicate the results of two-tailed paired t-tests between WT and the corresponding mutant, where * = P < 0.01, ** = P < 0.005, *** = P < 0.001. [D] ChIP analysis of WT and mutant Htz1 proteins reveals reduced mutant Htz1 occupancy. Htz1 enrichment at the heterochromatin (HMR) flanking genes GIT1 and OCA4 and the euchromatic genes COY1 and GAL10 were calculated relative to a negative control region within the silent mating type locus HMR. Average mutant enrichments are shown relative to WT; error bars represent standard deviations from two replicates.
Figure 3.
C-terminal mutants have reduced affinity for the SWR deposition complex.
[A] Anti-HA antibodies were used to immunoprecipitate cell lysates from strains expressing FLAG-tagged Swc2 and either HA-tagged WT, mutant, or no Htz1 (htz1∆). Input and anti-HA IP samples were analysed by anti-HA and anti-flag immunoblotting, with an example blot shown on the left. The position of the antibody light chain is indicated (*). Levels of co-immunoprecipitated Swc2-FLAG for each strain were normalised to the amount of immunoprecipitated HA-Htz1, expressed relative to WT, and averages are depicted in the graph (right; n= 4). [B]-[D] Effects of deleting Swr1 on HA-Htz1 levels in chromatin. [B] Sub-cellular fractionation performed and labelled as described in Figure 2B for WT Htz1 in SWR1 and swr1∆ backgrounds. A representative blot is shown on the left, and chromatin levels of HA-Htz1 normalised to H4, expressed relative to WT and averaged are shown on the right (n = 3). [C] Quantification of chromatin HA-Htz1 protein levels as in [B] but where each double mutant is normalised to the corresponding single HA-Htz1 mutant (n = 3). [D] Quantification of chromatin HA-Htz1 protein levels as in [B] but where each double mutant is normalised to the WT swr1∆ level (n = 3). Error bars indicate standard error of the mean. Asterisks indicate the results of two-tailed paired t-tests between the indicated strains, where * = P < 0.05, ** = P < 0.005, *** = P < 0.001.
Figure 4.
Loss of INO80 complex activity increases the association of C-terminal mutants with chromatin.
[A] Western blots of sub-cellular fractions generated from WT and arp8∆cells, labelled as in Figure 2B. A representative example is shown and chromatin HA-Htz1 protein levels normalised to H4, expressed relative to WT and averaged from 3 biological replicates, are shown on the right. [B] Quantification of chromatin HA-Htz1 protein levels as in [A] but where each double mutant is compared to the corresponding single HA-Htz1 mutant (n = 3 for bars 1-4 from the left and n = 4 for bars 5 & 6). [C] Quantification of chromatin HA-Htz1 protein levels as in [A] but where each double mutant is normalised to the WT arp8∆ level (n = 3). Error bars indicate standard error of the mean. Asterisks indicate the results of two-tailed paired t-tests between the indicated strains, where * = P < 0.05, ** = P < 0.01, *** = P < 0.005.
Figure 5.
Point mutations in the nucleosome docking domain reduce the physical stability of Htz1’s association with chromatin.
[A]-[B] Mutants are less resistant to washing with increased ionic-strength buffers. [A] HA-Htz1 levels in total cell lysates (T), and in insoluble chromatin fractions after washing with buffers containing 100, 200, 300 or 400 mM NaCl were determined by immunoblotting with anti-HA and anti-H4. [B] For each strain, levels of HA-Htz1 normalised to H4 for each washed chromatin sample were compared to the 100 mM wash sample and averaged (n = 3). Error bars indicate standard error of the mean. [C]-[F] Computational modelling of point mutations at equivalent positions in the mouse H2A.Z crystal structure predicts changes in local intermolecular interactions. KiNG software was used to model the effect of mutating residues in the mouse H2A.Z nucleosome crystal structure [18], at the positions equivalent to yeast serine 111 and isoleucine 109 (glycine 106 and isoleucine 104 respectively). In each panel, blue, grey and cyan lines represent the main chains, hydrogen atoms and side chains respectively, apart from G106, I104 and the residues substituted at these sites, where the main chains are coloured red and the side chains are pale pink. Green and blue dots indicate stabilising van der Waals interactions; yellow, orange and red spikes indicate small intermolecular clashes and pink spikes indicate large clashes. Local interactions are modelled for G106 [C], but when substituted with proline [D] substantial van der Waals overlaps are seen. I104 makes several stabilising van der Waal’s contacts [E], which are lost upon substitution with threonine [F] (right hand side of panel). In addition the hydrogen-bonding potential of the threonine hydroxyl group is unsatisfied.