Yeast Hif1 [Hat1 (histone acetyltransferase 1)-interacting factor], a homologue of human NASP (nuclear autoantigenic sperm protein), is a histone chaperone that is involved in various protein complexes which modify histones during telomeric silencing and chromatin reassembly. For elucidating the structural basis of Hif1, in the present paper we demonstrate the crystal structure of Hif1 consisting of a superhelixed TPR (tetratricopeptide repeat) domain and an extended acid loop covering the rear of TPR domain, which represent typical characteristics of SHNi-TPR [Sim3 (start independent of mitosis 3)-Hif1-NASP interrupted TPR] proteins. Our binding assay indicates that Hif1 could bind to the histone octamer via histones H3 and H4. The acid loop is shown to be crucial for the binding of histones and may also change the conformation of the TPR groove. By binding to the core histone complex Hif1 may recruit functional protein complexes to modify histones during chromatin reassembly.

INTRODUCTION

Hif1 [Hat1 (histone acetyltransferase 1)-interacting factor] is a yeast homologue of human NASP (nuclear autoantigenic sperm protein) [1,2]. Mammalian NASP is found as a longer form known as tNASP (testicular NASP), expressed in the testis, embryonic tissues and some transformed cells [3], and as a shorter form known as sNASP (somatic NASP), present in all mitotic cells [4]. sNASP can bind both linker histone H1 and core histones H3 and H4 by means of its distinct domains [3,57]. Along with fusion yeast Sim3 (start independent of mitosis 3), a yeast protein that associates with the CENP-A (centromere protein-A) homologue Cnp1 (centromere protein 1) and also binds histone H3, NASP and Hif1 form the SHNi-TPR [Sim3-Hif1-NASP interrupted TPR (tetratricopeptide repeat)] family [1,2]. The SHNi-TPR family is predicted to share common structural features such as several tandem TPRs with an interrupted insertion, which may play an important role in mediating protein–protein interactions [1].

Hif1 was first found to be a component of yeast nuclear HAT-B (histone acetyltransferase type B) complex, which specifically acetylates Lys5 and Lys12 of nuclear histone H4 [8,9]. The nuclear HAT-B complex consists of, at least, Hat1, Hat2 and Hif1. Hif1 interacts with Hat1, with Hat2 bridging Hat1 and Hif1 [8,9], while it binds specifically to histones H3 and H4 [9]. Thus Hif1 functions in the chromatin reassembly process. It seems that Hif1 has no significant contribution to the enzyme activity of the HAT-B complex [10]; however, deletion of Hif1 results in a defect in telomeric silencing and DNA DSB (double-strand break) repair as observed for Hat1 [9]. Hif1 has been found to be recruited to an endonuclease HO-induced DNA DSB in the form of HAT-B complex after the phosphorylation of Ser129 of histone H2A and concomitant with the recombination repair factor Rad52 (radiation sensitive 52), irrespective of whether the repair was possible [11]. In agreement with a previous report that Hif1 could promote core histones depositing on to a relaxed plasmid in the presence of yeast cytosolic extract [9], it is assumed that Hif1 mediates the interaction between the HAT-B complex and nuclear histones to facilitate a mechanism to modify the core histones and then change the chromatin structure. However, the independent recruitment of Hif1 to a DNA DSB point during the recombinational repair process indicates that Hif1 may have functions in the post-repair chromatin reassembly process other than as part of the HAT-B complex [11]. A previous study reported that Hif1 deletion combined with H3 K14R and K23R mutation exhibited a greater loss in chromatin reassembly than following Hat1 or Hat2 deletion [12]. It was also reported that Hif1 is involved in a distinct complex in the nucleus that influences a histone H3-specific histone acetyltransferase activity [12].

Considering the above-mentioned studies together, Hif1 could bind to histones and may recruit multiple complexes to participate in histone modification and chromatin reassembly. In order to understand the functionality of Hif1 and how it binds to histones in chromatin reassembly, we took to solve the structure of Hif1 using X-ray crystallography and to assay the binding properties of Hif1 with histones. The results of the present study indicate that Hif1 binds to core histones purified from bovine thymus and yeast-reconstituted histone octamers via interactions between the inserted acid patch and histones H3 and H4. In addition, a docking mode is also described for predicting the interaction details between Hif1 and core histones. Our studies provide not only insights into the structural characters of SHNi-TPR family, but a clue for investigating the role of Hif1 and NASP in chromatin reassembly.

MATERIALS AND METHODS

Protein production

The ORF of yeast full-length hif1 was amplified by PCR using Saccharomyces cerevisiae S288c genomic DNA as the template, and cloned into pET22b (Novagen) and pET-GST (glutathione transferase) [modified from pET28a (Novagen)] vectors. The hif1ΔN7 (truncated with the corresponding region of N-terminal seven amino acids of Hif1) and various truncations of gst-hif1 fusion constructs were produced by PCR-based deletion mutation of the indicated region. A stop codon and RBS (ribosomal-binding site) sequence were introduced into all gst-hif1 fusion constructs during the PCR-based deletion mutation for producing discrete parts of Hif1, except for the GST–Hif1Δ(85–198) constructs where the whole acid patch (residues 85–198) of Hif1 was deleted. All constructions were confirmed by sequencing. Supplementary Figure S1 (http://www.biochemj.org/bj/462/bj4620465add.htm) shows the Hif1 variant truncations used in the pull-down assays. The recombinant wild-type Hif1 and truncated Hif1ΔN7 were expressed in Escherichia coli BL21(DE3) cells (Stratagene) and purified for crystallization with a C-terminal His6 tag. GST–Hif1 and variant GST–Hif1 truncations were also expressed in E. coli BL21(DE3) cells and purified for the pull-down assays with the His6 tag on both the N-terminus of GST and C-terminus of Hif1 retained.

All expression constructs were transformed into bacterial cells, and LB medium containing 100 μg/ml ampicillin for Hif1 expression or 50 μg/ml kanamycin for expression of GST–Hif1 and its variant truncations was innoculated with overnight-cultured bacterial cells at a ratio of 1:100. These cultures were grown at 37°C until the D600 reached 0.8. After cooling to 30°C the cultures were then induced by 0.25 mM IPTG and incubated for a further 8 h for protein expression. For expression of the selenomethionine-labelled Hif1ΔN7 used for the experimental phasing, the hif1ΔN7 construct was transformed into E. coli cells, but cultured in M9 medium with 100 mg/l lysine, phenylalanine and threonine and 50 mg/l isolucine, leucine and valine. Selenomethionine (60 mg/l) was added after cooling the culture to 16°C, but before target protein expression was induced by 0.25 mM IPTG for 30 h. E. coli cells were harvested by centrifugation at 5000 g for 8 min at 8°C, resuspended in lysis buffer [50 mM Tris/HCl (pH 8.0) and 500 mM NaCl] and sonicated on ice for 20 min. After centrifugation of the lysate the resulting supernatant was loaded on to a binding buffer (identical to the lysis buffer) -equilibrated Ni-chelating Sepharose column (GE Healthcare). Target proteins were washed and eluted by a gradient concentration of imidazole and further purified by gel filtration using a Superdex 200 column (GE Healthcare) pre-equilibrated with the binding buffer. Peak fractions of target protein were pooled and desalted with ion-exchange buffer [50 mM Tris/HCl (pH 8.0) and 100 mM NaCl] for ion-exchange chromatography. Pure protein fractions eluted by ion exchange using a HiTrap™ Q FF anion-exchange column (GE Healthcare) were pooled and concentrated. For the crystallization assays, Hif1 and Hif1ΔN7 then had their buffer changed to 20 mM Mes (pH 6.5) and 50 mM NaCl. The purity of target proteins was examined using SDS/PAGE. LC-MS was also used to assess the integrity of purified Hif1, which found that recombinant wild-type Hif1 is a full-length protein (UniProt entry Q12373).

The ORF of yeast histones H2A, H2B, H3 and H4 (UniProt entries P04911, P02293, P61830 and P02309 respectively) were also cloned from S. cerevisiae S288c genomic DNA and expressed in the E. coli Rosseta(DE3) strain. The yeast-reconstituted histones were prepared as described previously [1316]. Soluble H2A and H2B were purified using an Ni-affinity column, whereas denatured H3 and H4 were separated from the inclusion body by ion-exchange purification. Careful mixing of histones H2A, H2B, H3 and H4 in equimolar ratios in unfolding buffer [20 mM Tris/HCl (pH 7.5), 7 M guanidine hydrochloride, 1 mM EDTA and 5 mM 2-mercaptoethanol] and then a step-wise dialysis with refolding buffer [10 mM Tris/HCl (pH 7.5), 2 M NaCl, 1 mM EDTA and 5 mM 2-mercaptoethanol] was performed to reconstitute the yeast histone octamers. The final octamer product was purified using gel-filtration chromatography to exclude the misfolding oligomers. H2A–H2B dimers and H3–H4 tetramers were produced in the similar way. The gel-filtration chromatograms produced during purification of the reconstituted histones were plotted to clarify the oligomer state of the histones used in our research (Supplementary Figure S2 at http://www.biochemj.org/bj/462/bj4620465add.htm).

Crystallization and data collection

The purified Hif1 buffered in 20 mM MES (pH 6.5) and 60 mM NaCl was diluted to 25 mg/ml for crystallization. Protein (0.15 μl) was mixed with an equal volume reservoir solution using a Mosquito crystallization robot (TTP LabTech) under conditions for hanging-drop vapour-diffusion screening of protein complexes [17]. Initial crystals of native Hif1 were obtained after approximately 11 months in various wells. These crystals were hard to repeat, but initial crystals acquired in 0.1 M sodium cacodylate (pH 6.0) and 15% PEG4000 could diffract at a high-enough resolution to allow protein structure determination. Seleno-crystals were acquired after approximately 7 months under conditions of 0.1 M Tris/HCl (pH 8.0), 20% PEG8000 and 200 mM LiCl using the same crystallization method, but with a concentration of 40 mg/ml selenoprotein. All crystals were cryo-protected with cryoprotectant buffer containing 25% glycerol in the reservoir solution and then diffracted with 100 K nitrogen gas stream for data collection. Datasets of native and selenomethionine-labelled protein crystals were collected on beamline 17U at the SSRF (Shanghai Synchrotron Radiation Facility) at wavelengths of 0.97915 and 0.97917 Å respectively, using a Quantum 315r detector (Area Detector Systems Corporation). All the image data were processed, integrated and scaled using the HKL2000 software package [18] in orthorhombic lattice P222. Systematic absorption indicated the space group to be P212121.

Structure determination and analysis

The initial phase was solved using the AutoSol [19,20] module of PHENIX [21]. Using the Matthews probability calculator [22,23] a Vm of 1.76 Å3/Da was found with an extremely low probability of crystal packing and AutoSol failed to build a primary model with the provided sequence of Hif1ΔN7. Considering that there may be protein degradation before crystallization, we therefore manually built the primary model iteratively using Coot [24] and the Refmac5 module of the CCP4 program suit [25]. The primary model was therefore used for the search phase of native crystal data using the Phaser program [26] of CCP4. The structural model of the native crystal data was then completed by iterative building using Coot and refining with Refmac5. The model refinement was monitored by R-factor and Rfree and a TLS (Translation–Libration–Screw-rotation) refinement was introduced during the final step of refinement. The final model was built confidently and converged to reasonable R-factor and Rfree values of 21.1% and 24.3% respectively. The data collection and refinement statistics are listed in Table 1. The Dali server [27] was used to search for architecture with similarity to Hif1. Hif1 was docked to the yeast histone octamer using ClusPro [28]. All Figures were prepared using PyMOL (http://www.pymol.org).

Table 1
Crystallographic data collection and structure determination statistics

Values in parentheses are for the outermost resolution shell. Rmerge=∑hkli|Ii(hkl)−<I(hkl)>|/∑hkliIi(hkl), where Ii(hkl) is the intensity of ith observation and <I(hkl)> is the mean value for reflection hkl. R factor=∑hkl ||Fobs| − |Fcalc||/∑hkl|Fobs|, where Fobs and Fcalc are the observed and calculated structure-factor amplitudes respectively. Rfree is equivalent to R factor, but calculated with reflections excluded from the refinement process (5% of all reflections). Ramachandran plot statistics were defined using MolProbity.

Parameter SeMet–Hif1 Native Hif1 
Data collection   
 Wavelength (Å) 0.97917 0.97915 
 Space group P212121 P212121 
 Unit cell parameters a=48.2, b=78.4, c=81.0 a=48.3, b=78.3, c=82.7 
 Resolution range (Å) 50.00–3.25 (3.31–3.25) 50.00–2.10 (2.14–2.10) 
 Unique reflections 5180 (235) 18861 (908) 
 Completeness (%) 99.7 (99.6) 99.6 (99.0) 
 Average redundancy 7.2 (7.4) 4.5 (4.5) 
 Mean I/σ(I20.19 (9.84) 13.8 (2.1) 
Rmerge (%) 12.1 (30.6) 8.3 (65.6) 
Refinement statistics   
 Resolution range (Å)  41.30–2.10 
R factor (%)  21.11 
Rfree (%)  24.30 
 RMSD bond lengths (Å)  0.0079 
 RMSD bond angles (°)  1.0006 
 Mean B-factors of main/side chain (Å231.1/32.5 
 Ramachandran plot (percentage of residues)  
  Favoured (%)  98.0 
  Allowed (%)  2.0 
  Outlier (%)  
 PDB code  4NQ0 
Parameter SeMet–Hif1 Native Hif1 
Data collection   
 Wavelength (Å) 0.97917 0.97915 
 Space group P212121 P212121 
 Unit cell parameters a=48.2, b=78.4, c=81.0 a=48.3, b=78.3, c=82.7 
 Resolution range (Å) 50.00–3.25 (3.31–3.25) 50.00–2.10 (2.14–2.10) 
 Unique reflections 5180 (235) 18861 (908) 
 Completeness (%) 99.7 (99.6) 99.6 (99.0) 
 Average redundancy 7.2 (7.4) 4.5 (4.5) 
 Mean I/σ(I20.19 (9.84) 13.8 (2.1) 
Rmerge (%) 12.1 (30.6) 8.3 (65.6) 
Refinement statistics   
 Resolution range (Å)  41.30–2.10 
R factor (%)  21.11 
Rfree (%)  24.30 
 RMSD bond lengths (Å)  0.0079 
 RMSD bond angles (°)  1.0006 
 Mean B-factors of main/side chain (Å231.1/32.5 
 Ramachandran plot (percentage of residues)  
  Favoured (%)  98.0 
  Allowed (%)  2.0 
  Outlier (%)  
 PDB code  4NQ0 

GST pull-down assay

As most of the protein–protein interactions and post-translational modifications of core histones from eukaryotic cells are highly conserved [29], and rather the post-translational modifications, we first performed the GST pull-down assay using commercially available core histones purified from the bovine thymus (Sangon) instead of yeast histones. GST–Hif1 fusion protein and its variant truncations were expressed and purified and could bind to the glutathione–Sepharose resin (GE Healthcare). The mixtures containing GST–Hif1 or its truncations were incubated with target proteins in a low ionic strength binding buffer [25 mM Hepes, 200 mM KCl, 13 mM MgCl2, 10% glycerol, 0.1% NP40 (Nonidet P40) and 0.3% 2-mercaptoethanol] [9] and mixed with pre-equilibrated glutathione–Sepharose resin by rotating for 1 h. After removal of unbound proteins by low-ionic-strength washing buffer (25 mM Hepes, 200 mM KCl, 13 mM MgCl2, 10% glycerol, 0.1% Triton X-100 and 0.3% 2-mercaptoethanol), SDS loading buffer was employed to elute the binding proteins which were then analysed by SDS/PAGE. Yeast histone H2A–H2B dimers, H3–H4 tetramers and histone octamers were also reconstituted as described previously for validating the binding property of Hif1 with core histones. Pull-down experiments of yeast-reconstituted histones were the same as for the bovine core histones except a high-ionic-strength binding buffer and wash buffer (increased KCl concentration to 1 M) were applied to the binding experiments for assaying the binding ability with yeast histone octamers.

RESULTS AND DISCUSSION

Hif1 consists of TPR domain with an inserted acid patch

Hif1 was predicted to consist of an acid patch and structurally conserved tandem TPR repeats. However, the structure of the TPR domain and how it is inserted into the acid patch are not known. We used X-ray crystallography to determine the structure of Hif1 using the SAD (single-wavelength anomalous diffraction) method. Hif1 consists of nine helices, designated α1–α9, a long inserted acid patch and a C-terminal domain containing the NLS (nuclear localization sequence) (Figure 1). However, part of the acid patch (patch A) and the NLS domain are missing from the crystal structure due to unexpected proteolysis during the long crystallization process. Each of TPR motifs consists of two antiparallel α-helices. Helix α9 neighbours the preceding helix of TPR4 in a manner that resembles the interaction of α8 with the adjacent α7 repeats (Figure 2A). Together, the nine helices exhibit a right-handed superhelical twist which generates a groove surface shaped by the side chains of helices α1, α3, α5, α7 and α9 (Figure 2A). The acid patch of Hif1 surrounds the outer surface of the TPR domain as shown in Figure 2(A). Curiously, TPR2 is not stable because deletion of the whole acid patch causes the dissociation of the TPR domain between the two helixes of TPR2. GST–Hif1ΔM(80–199), constructed as described above and consisting of His6–GST–Hif11–79 and His6–Hif1200–385 co-expressed by polycistronic mRNA, could not self-reconstitute the TPR domain of Hif1 as shown by only the GST-containing part being retained in the glutathione resin (Figure 3A). Although keeping part of the acid patch, such as in the GST–Hif1ΔM(135–158) construct, could restore the TPR domain of Hif1 as shown by His6–GST–Hif11–134 and His6–Hif1159–385 both being retained in the GST-affinity resin (Figure 3A). Furthermore, GST–Hif1ΔM(80–158), with deletion of patch A of Hif1, could also restore the TPR domain as shown in the crystal structure of Hif1 (Figure 3B). In comparison, patch B, which was not subject to proteolysis during crystallization, is observed in the crystal structure of Hif1 to provide additional interactions via hydrogen bonds and hydrophobic interactions between helices α3 and α4 of TPR2 to maintain the conformation of the TPR domain (Figure 3B). This evidence suggests that patch B has an essential role for maintaining the conformation of the TPR domain.

Global alignment of yeast Hif1 across the SHNi-TPR family

Figure 1
Global alignment of yeast Hif1 across the SHNi-TPR family

Dots represent gaps in the sequence. Secondary structural elements of Hif1 are displayed above the sequence. The acid patch, which has a distinct conserved boundary from the TPR region, has been divided into patch A and patch B according the structural information for yeast Hif1. Patch A and patch B are indicated by the straight lines above and below the sequences respectively. The sequences are from UniProt entries Q12373 (Hif1), P49321 (sNASP) and Q9USQ4 (Sim3).

Figure 1
Global alignment of yeast Hif1 across the SHNi-TPR family

Dots represent gaps in the sequence. Secondary structural elements of Hif1 are displayed above the sequence. The acid patch, which has a distinct conserved boundary from the TPR region, has been divided into patch A and patch B according the structural information for yeast Hif1. Patch A and patch B are indicated by the straight lines above and below the sequences respectively. The sequences are from UniProt entries Q12373 (Hif1), P49321 (sNASP) and Q9USQ4 (Sim3).

Overall structure of yeast Hif1

Figure 2
Overall structure of yeast Hif1

(A) Cartoon representation of the Hif1 structure. Hif1 consists of nine superhelixed α-helices, designated α1 to α9, the TPR domain and an acid patch concatenating at least four 310 helixes that bind to the rear surface of TPR domain. The purple line represents patch A which is missing from the structure. TPR1–TPR4 are shown in red, green, blue and brown respectively, whereas the extended α-helix of TPR4 is turquoise. The TPR domain of Hif1 is shaped by these superhelixed helices. (B) Comparison of the groove surface of Hif1 with 14-3-3ζ/δ and the TPR domain of p67phox. Grooves are shaped by superhelixed α-helices. The upper panel shows the groove surface of Hif1. The lower left-hand panel shows the groove surface of 14-3-3ζ/δ with an inserted phosphoacetylated peptide of histone H3, whereas the lower right-hand panel shows the groove surface of the p67phox TPR domain with its C-terminus inserted. The Rac protein complexed with GTP is present at the rear of TPR domain of p67phox. The colour scheme of the TPR domains is the same for (A).

Figure 2
Overall structure of yeast Hif1

(A) Cartoon representation of the Hif1 structure. Hif1 consists of nine superhelixed α-helices, designated α1 to α9, the TPR domain and an acid patch concatenating at least four 310 helixes that bind to the rear surface of TPR domain. The purple line represents patch A which is missing from the structure. TPR1–TPR4 are shown in red, green, blue and brown respectively, whereas the extended α-helix of TPR4 is turquoise. The TPR domain of Hif1 is shaped by these superhelixed helices. (B) Comparison of the groove surface of Hif1 with 14-3-3ζ/δ and the TPR domain of p67phox. Grooves are shaped by superhelixed α-helices. The upper panel shows the groove surface of Hif1. The lower left-hand panel shows the groove surface of 14-3-3ζ/δ with an inserted phosphoacetylated peptide of histone H3, whereas the lower right-hand panel shows the groove surface of the p67phox TPR domain with its C-terminus inserted. The Rac protein complexed with GTP is present at the rear of TPR domain of p67phox. The colour scheme of the TPR domains is the same for (A).

The acid patch provides additional interactions to maintain the conformation of the TPR domain

Figure 3
The acid patch provides additional interactions to maintain the conformation of the TPR domain

(A) The dividing line indicates non-adjacent lanes spliced together. In GST–Hif1∆M(135–158) the TPR domain self-reconstitutes, but in GST–Hif1∆M(80–199) it does not. GST–Hif1∆M(80–199) was expressed by a polycistronic mRNA as two discrete parts of Hif1 that are separated by the deleted region of the acid patch. The same procedure applied to GST–Hif1∆M(135–158). Both constructs were subjected to Ni-NTA (Ni2+-nitrilotriacetate) and glutathione resin assays to assess their self-reconstitution of the TPR domain. The restoration of the TPR domain of GST–Hif1∆M(80–199) was shown by both fragments of His6–GST–Hif11–79 and His6–Hif1200–385 being retained by glutathione resin. The restoration of the TPR domain of GST–Hif1∆M(135–158) was shown by both fragments of His6–GST–Hif11–135 and His6–Hif1159–385 being retained by glutathione resin. (B) Patch B mediates interactions within TPR2. Interactions between patch B and the TPR domain were calculated using the PISA program [34,35]. Residues of the acid patch involved in interaction with helices α3 and α4 of TPR2 via hydrogen bonding and hydrophobic stacks are shown as a stick model and highlighted purple. Residues of TPR1 and TPR2 involved in the interaction are also shown as a stick model and highlighted cyan.

Figure 3
The acid patch provides additional interactions to maintain the conformation of the TPR domain

(A) The dividing line indicates non-adjacent lanes spliced together. In GST–Hif1∆M(135–158) the TPR domain self-reconstitutes, but in GST–Hif1∆M(80–199) it does not. GST–Hif1∆M(80–199) was expressed by a polycistronic mRNA as two discrete parts of Hif1 that are separated by the deleted region of the acid patch. The same procedure applied to GST–Hif1∆M(135–158). Both constructs were subjected to Ni-NTA (Ni2+-nitrilotriacetate) and glutathione resin assays to assess their self-reconstitution of the TPR domain. The restoration of the TPR domain of GST–Hif1∆M(80–199) was shown by both fragments of His6–GST–Hif11–79 and His6–Hif1200–385 being retained by glutathione resin. The restoration of the TPR domain of GST–Hif1∆M(135–158) was shown by both fragments of His6–GST–Hif11–135 and His6–Hif1159–385 being retained by glutathione resin. (B) Patch B mediates interactions within TPR2. Interactions between patch B and the TPR domain were calculated using the PISA program [34,35]. Residues of the acid patch involved in interaction with helices α3 and α4 of TPR2 via hydrogen bonding and hydrophobic stacks are shown as a stick model and highlighted purple. Residues of TPR1 and TPR2 involved in the interaction are also shown as a stick model and highlighted cyan.

In addition, structure-based comparison of Hif1 using the Dali server [27] returned a variety of similar structures of proteins containing tandem TPRs and 14-3-3ζ/δ proteins or protein complexes. These structures indicate Hif1 is also a scaffold protein that mediates protein–protein interactions. However, none of these proteins are found in the acid patch, which highlights it as a distinctive characteristic of the SHNi-TPR family. In particular two of the proteins found by the Dali server caught our attention. One is the model of 14-3-3ζ/δ protein with binding of a phosphoacetylated peptide of the histone H3 tail (PDB code 2C1J). The phosphoacetylated peptide spans the 14-3-3ζ/δ-binding cleft in an extended conformation [30]. The binding cleft resembles the TPR groove of Hif1 (Figure 2B), indicating that Hif1 may bind to a histone tail in a similar fashion. The other results of interest from the Dali server is the complex of the TPR domain of p67phox with the Rac-GTP (PDB code 1E96) [31]. The C-terminus of the p67phox protein inserts into the TPR groove, whereas the Rac protein in complex with GTP binds to the rear of TPR domain, which mediates interactions between Rac and p67phox [31]. Hif1 may bind other proteins or protein complexes in a similar way. In short, because of the high similarity of the structural architecture of the TPR domain, Hif1 may have a binding pattern similar to 14-3-3ζ/δ and the TPR domain of p67phox.

The acid patch is responsible for the binding of core histones

Hif1 is known as a nuclear protein involved in telomeric silencing and chromatin reassembly [8,9,12,32], but little is known regarding how Hif1 behaves in the nucleus. We first tested the binding ability of Hif1 with several candidate protein targets under the condition of low-ionic-strength buffer. Previous researchers have reported that Hif1 is involved in the nuclear HAT-B complex [8,9,11,12]. We therefore performed pull-down assays to analyse the binding ability of Hif1 with yeast Hat1 and bovine core histones. As shown in Figure 4(A), Hif1 binds strongly with the core histones H2A, H2B, H3 and H4, but could not bind to Hat1 even in the presence of core histones. This is in agreement with previous reports that the interaction between Hat1 and Hif1 is bridged by Hat2 [8].

Hif1 binding with bovine core histones

Figure 4
Hif1 binding with bovine core histones

The binding experiments were performed in of low-ionic-strength buffer. (A) Hif1-binding assay with Hat1 and bovine core histones. GST–Hif1 fusion protein was incubated with target proteins for pull-down assay. Input samples are in the left-hand three lanes. The protein marker lane (Marker) contained proteins of molecular mass (from top to bottom) 116, 66.2, 45, 35, 25, 18.4 and 14.4 kDa. (B) Mapping of the histone-binding region of Hif1. Various truncations of GST–Hif1 fusion proteins were incubated with target proteins for a pull-down assay.

Figure 4
Hif1 binding with bovine core histones

The binding experiments were performed in of low-ionic-strength buffer. (A) Hif1-binding assay with Hat1 and bovine core histones. GST–Hif1 fusion protein was incubated with target proteins for pull-down assay. Input samples are in the left-hand three lanes. The protein marker lane (Marker) contained proteins of molecular mass (from top to bottom) 116, 66.2, 45, 35, 25, 18.4 and 14.4 kDa. (B) Mapping of the histone-binding region of Hif1. Various truncations of GST–Hif1 fusion proteins were incubated with target proteins for a pull-down assay.

To understand which region of Hif1 is responsible for the binding of bovine core histones, we performed binding assays to test the affinity between core histones and the truncated GST–Hif1 fusion protein. As shown in Figure 4(B), GST–Hif1Δ(85–198), a construct with deletion of the whole acid patch of Hif1, exhibits no prominent binding ability with core histones. However, when the acid patch was present, at least the patch B region, Hif1 retained its histone-binding ability. GST–Hif1ΔM(80–158), a construct with deletion of residues 80–158 that retains patch B of Hif1, could bind to core histones as could GST–Hif1ΔM(135–158), a construct with deletion of part of patch A of Hif1 (Figure 4B). This evidence suggests that the acid patch of Hif1, especially patch B, has a crucial role in the binding of core histones. Accordingly, as discussed above, the acid patch could provide additional interactions to maintain the conformation of the TPR domain of Hif1. These interactions could probably induce conformational changes of the TPR domain during the histone-binding process to facilitate histone binding or recognition of Hif1.

Hif1 interaction with core histones as a scaffold protein

Attempting to co-crystallize Hif1 with core histones was unsuccessful and thus a docking model of Hif1 binding to core histones was calculated to predict their potential interaction. A model of the core histone octamer of yeast nucleosome (PDB code 1ID3) was used as the receptor protein during the docking process of the ClusPro server [28], whereas Hif1 was used as the ligand protein. The docking results are consistent with those of our binding assay, showing that the acid patch is responsible for histone binding and that Hif1 binds to the core histone octamer complex via histones H3 and H4. The dimer of histone H3 and H4 clamps Hif1 with both of them interacting with the acid surface of Hif1 (Figure 5). Histone H3 binds to the acid surface shaped by the TPR domain and patch B. In particular, the N-tail of histone H3 binds into the groove formed by the nine superhelixed TPR α-helices (Figure 5A). However, histone H4 binds to the rear surface of Hif1. The N-tail of histone H4 binds to the rear surface generated by the acid patch and α-helices of the TPR domain (Figure 5B). Structural analysis using the PISA program [33,34] reveals that there are 37 or 30 polar contacts between Hif1 and H3 or H4 respectively. Patch B in combination with helices α2, α3 and α4 forms the major interface for the binding of the global regions of histones H3 and H4. Although less flexibility of the histone tails was introduced during the process of docking, it could be supposed that N-tail of histone H3 may actually fall into the groove of Hif1 in a fashion resembling how the H3 peptide falls into the groove of 14-3-3ζ/δ and that the C-terminal tail of p67phox binds into the groove of its TPR domain (Figure 2B). Furthermore, the acid patch that concatenates at least four 310 helixes may exert a conformational change during protein–protein interactions. If this is correct, the groove of Hif1 could be induced into an easily accessible shape for recognizing the H3 tail.

A model of Hif1 docking into H3–H4 in a histone octamer

Figure 5
A model of Hif1 docking into H3–H4 in a histone octamer

Histones H3 and H4 are shown as green and brown cartoons, whereas Hif1 is shown as an electrostatic potential surface calculated by the APBS program [35]. The red and blue surfaces represent electrostatic potentials as indicated in the bar. The dimer of histone H3–H4 rides on the acid ‘ridge’ of Hif1. (A) Histone H3 binds to Hif1. (B) Histone H4 flanks Hif1.

Figure 5
A model of Hif1 docking into H3–H4 in a histone octamer

Histones H3 and H4 are shown as green and brown cartoons, whereas Hif1 is shown as an electrostatic potential surface calculated by the APBS program [35]. The red and blue surfaces represent electrostatic potentials as indicated in the bar. The dimer of histone H3–H4 rides on the acid ‘ridge’ of Hif1. (A) Histone H3 binds to Hif1. (B) Histone H4 flanks Hif1.

To validate whether Hif1 interacts with the histone octamer via H3 and H4 as the docking experiment indicated and to understand the capability of Hif1 binding with bovine H2A and H2B in a form of bovine core histones, as examined in the present study and differing from a previous report [9], we further examined the pattern of Hif1 binding with core histones. Yeast-reconstituted histones were manipulated as described previously for our binding assay (Supplementary Figure S2). Both the yeast-reconstituted H3–H4 tetramer and H2A–H2B dimer could bind to Hif1 in the low-salt binding buffer (Figure 6A), but Hif1 could not bind the H2A–H2B dimers in the presence of the H3–H4 tetramer (Figure 6A). This indicates that Hif1 binds to the bovine H2A–H2B–H3–H4 complex rather than binding the H3–H4 tetramer and repulsing the H2A–H2B dimer if they are both present in the sample of bovine core histones. Furthermore, the H2A–H2B dimer could not bind to Hif1 in a high-ionic-strength binding buffer, but could bind to Hif1 in the form of yeast-reconstituted core histone octamers or bovine core histones (Figure 6B). These results indicate that Hif1 binds to core histone octamers specifically, probably via interactions between Hif1 and H3–H4, which agrees with the predictions of our docking experiment.

Hif1 may bind to histone octamers via binding with H3–H4

Figure 6
Hif1 may bind to histone octamers via binding with H3–H4

The dividing lines indicate non-adjacent lanes spliced together. (A) Hif1 binding with yeast-reconstituted core histones. GST–Hif1 fusion protein was incubated with yeast-reconstituted histones of variant oligomer states for pull-down assay. The binding experiment was performed in low-ionic-strength buffer. Input samples are shown from the left-hand four lanes. Histone samples in the first, sixth and eighth lanes from the left-hand side are mixtures of yeast H2A/H2B dimers and H3/H4 tetramers. (B) Hif1 binding with core histones under high-ionic-strength conditions. GST-Hif1 fusion protein was incubated with bovine core histones, yeast histone octamers and yeast H2A–H2B dimers for pull-down assays.

Figure 6
Hif1 may bind to histone octamers via binding with H3–H4

The dividing lines indicate non-adjacent lanes spliced together. (A) Hif1 binding with yeast-reconstituted core histones. GST–Hif1 fusion protein was incubated with yeast-reconstituted histones of variant oligomer states for pull-down assay. The binding experiment was performed in low-ionic-strength buffer. Input samples are shown from the left-hand four lanes. Histone samples in the first, sixth and eighth lanes from the left-hand side are mixtures of yeast H2A/H2B dimers and H3/H4 tetramers. (B) Hif1 binding with core histones under high-ionic-strength conditions. GST-Hif1 fusion protein was incubated with bovine core histones, yeast histone octamers and yeast H2A–H2B dimers for pull-down assays.

Conclusions

We have solved the crystal structure of yeast Hif1. It consists of nine α-helices, which form a superhelixed TPR groove domain, and a long acid patch called patch B. This acid patch binds to the rear surface of the TPR domain that is crucial for the binding of core histones. In addition, Hif1 may bind to the histone octamer via interaction with H3–H4. Although the details of Hif1 binding to core histones remain to be addressed, we could imagine that the binding of Hif1 with core histones during chromatin reassembly could facilitate the binding of the Hat1–Hat2 complex with core histones in the nucleus, which is distinct from the Hat1–Hat2 complex binding with free histone H4 in the cytoplasm. It is probable that Hif1 binds to the histone octamer and thus recruits the Hat1–Hat2 or other protein complexes to change the epigenetic state of chromatin via histone modification during chromatin reassembly. The crystal structure of Hif1 not only offers insights into SHNi-TPR protein family, such as human NASP, but also provides clues to understand how the Hat1–Hat2–Hif1 complex functions in the nucleus. Detailed studies of the interactions between Hif1 and core histones or the Hat1–Hat2 complex are required for further understanding of the role of Hif1 in telomeric silencing and chromatin reassembly.

Abbreviations

     
  • DSB

    double-strand break

  •  
  • GST

    glutathione transferase

  •  
  • Hat

    histone acetyltransferase

  •  
  • HAT-B

    histone acetyltransferase type B

  •  
  • Hif1

    Hat1-interacting factor 1

  •  
  • NASP

    nuclear autoantigenic sperm protein

  •  
  • NLS

    nuclear localization sequence

  •  
  • RBS

    ribosomal-binding site

  •  
  • SHNi-TPR

    Sim3-Hif1-NASP interrupted TPR

  •  
  • Sim3

    start independent of mitosis 3

  •  
  • sNASP

    somatic NASP

  •  
  • TPR

    tetratricopeptide repeat

AUTHOR CONTRIBUTION

Hejun Liu conceived the project, performed and interpreted the experiments, and wrote the paper. Mengying Zhang performed and interpreted the experiments and wrote the paper. Wei He performed the experiments. Zhongliang Zhu interpreted the experiments. Maikun Teng supervised the project and interpreted the experiments. Yongxiang Gao supervised the project, interpreted the experiments and wrote the paper. Liwen Niu supervised the project, interpreted the experiments and wrote the paper.

We would like to thank the staff of SSRF for their help in data collection.

FUNDING

This work was supported by the Chinese Ministry of Science and Technology [grant numbers 2012CB917200 and 2011CBA00800] and the Chinese National Natural Science Foundation [grant numbers 31130018 and 31170726].

References

References
1
Dunleavy
E. M.
Pidoux
A. L.
Monet
M.
Bonilla
C.
Richardson
W.
Hamilton
G. L.
Ekwall
K.
McLaughlin
P. J.
Allshire
R. C.
A NASP (N1/N2)-related protein, Sim3, binds CENP-A and is required for its deposition at fission yeast Centromeres
Mol. Cell
2007
, vol. 
28
 (pg. 
1029
-
1044
)
[PubMed]
2
Cook
A. J. L.
Gurard-Levin
Z. A.
Vassias
I.
Almouzni
G.
A specific function for the histone chaperone NASP to fine-tune a reservoir of soluble H3–H4 in the histone supply chain
Mol. Cell
2011
, vol. 
44
 (pg. 
918
-
927
)
[PubMed]
3
Wang
H. Y.
Walsh
S. T. R.
Parthun
M. R.
Expanded binding specificity of the human histone chaperone NASP
Nucleic Acids Res.
2008
, vol. 
36
 (pg. 
5763
-
5772
)
[PubMed]
4
Richardson
R. T.
Batova
I. N.
Widgren
E. E.
Zheng
L. X.
Whitfield
M.
Marzluff
W. F.
O’Rand
M. G.
Characterization of the histone H1-binding protein, NASP, as a cell cycle-regulated somatic protein
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
30378
-
30386
)
[PubMed]
5
Wang
H. Y.
Ge
Z. Q.
Walsh
S. T. R.
Parthun
M. R.
The human histone chaperone sNASP interacts with linker and core histones through distinct mechanisms
Nucleic Acids Res.
2012
, vol. 
40
 (pg. 
660
-
669
)
[PubMed]
6
Osakabe
A.
Tachiwana
H.
Matsunaga
T.
Shiga
T.
Nozawa
R. S.
Obuse
C.
Kurumizaka
H.
Nucleosome formation activity of human somatic nuclear autoantigenic sperm protein (sNASP)
J. Biol. Chem.
2010
, vol. 
285
 (pg. 
11913
-
11921
)
[PubMed]
7
Finn
R. M.
Browne
K.
Hodgson
K. C.
Ausio
J.
sNASP, a histone H1-specific eukaryotic chaperone dimer that facilitates chromatin assembly
Biophys. J.
2008
, vol. 
95
 (pg. 
1314
-
1325
)
[PubMed]
8
Poveda
A.
Pamblanco
M.
Tafrov
S.
Tordera
V.
Sternglanz
R.
Sendra
R.
Hif1 is a component of yeast histone acetyltransferase B, a complex mainly localized in the nucleus
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
16033
-
16043
)
[PubMed]
9
Ai
X.
Parthun
M. R.
The nuclear Hat1p/Hat2p complex: a molecular link between type B histone acetyltransferases and chromatin assembly
Mol. Cell
2004
, vol. 
14
 (pg. 
195
-
205
)
[PubMed]
10
Poveda
A.
Sendra
R.
Site specificity of yeast histone acetyltransferase B complex in vivo
FEBS J.
2008
, vol. 
275
 (pg. 
2122
-
2136
)
[PubMed]
11
Qin
S.
Parthun
M. R.
Recruitment of the type B histone acetyltransferase Hat1p to chromatin is linked to DNA double-strand breaks
Mol. Cell. Biol.
2006
, vol. 
26
 (pg. 
3649
-
3658
)
[PubMed]
12
Ge
Z. Q.
Wang
H. Y.
Parthun
M. R.
Nuclear Hat1p complex (NuB4) components participate in DNA repair-linked chromatin reassembly
J. Biol. Chem.
2011
, vol. 
286
 (pg. 
16790
-
16799
)
[PubMed]
13
Luger
K.
Rechsteiner
T.
Richmond
T.
Becker
P.
Expression and purification of recombinant histones and nucleosome reconstitution
Chromatin Protocols
1999
Totowa
Humana Press
(pg. 
1
-
16
)
14
Dyer
P. N.
Edayathumangalam
R. S.
White
C. L.
Bao
Y.
Chakravarthy
S.
Muthurajan
U. M.
Luger
K.
Allis
C. D.
Carl
W.
Reconstitution of nucleosome core particles from recombinant histones and DNA
Methods in Enzymology
2003
San Diego
Academic Press
(pg. 
23
-
44
)
15
Hanson
B. L.
Alexander
C.
Harp
J. M.
Bunick
G. J.
Allis
C. D.
Carl
W.
Preparation and crystallization of nucleosome core particle
Methods in Enzymology
2003
San Diego
Academic Press
(pg. 
44
-
62
)
16
Tanaka
Y.
Tawaramoto-Sasanuma
M.
Kawaguchi
S.
Ohta
T.
Yoda
K.
Kurumizaka
H.
Yokoyama
S.
Expression and purification of recombinant human histones
Methods
2004
, vol. 
33
 (pg. 
3
-
11
)
[PubMed]
17
Radaev
S.
Li
S.
Sun
P. D.
A survey of protein–protein complex crystallizations
Acta Crystallogr. Sect. D. Biol. Crystallogr.
2006
, vol. 
62
 (pg. 
605
-
612
)
18
Otwinowski
Z.
Minor
W.
Carter
C. W.
Jr
Processing of X-ray diffraction data collected in oscillation mode
Methods in Enzymology
1997
San Diego
Academic Press
(pg. 
307
-
326
)
19
McCoy
A. J.
Storoni
L. C.
Read
R. J.
Simple algorithm for a maximum-likelihood SAD function
Acta Crystallogr. Sect. D Biol. Crystallogr.
2004
, vol. 
60
 (pg. 
1220
-
1228
)
20
Terwilliger
T. C.
Adams
P. D.
Read
R. J.
Mccoy
A. J.
Moriarty
N. W.
Grosse-Kunstleve
R. W.
Afonine
P. V.
Zwart
P. H.
Hung
L. W.
Decision-making in structure solution using Bayesian estimates of map quality: the PHENIX AutoSol wizard
Acta Crystallogr. Sect. D Biol.
2009
, vol. 
65
 (pg. 
582
-
601
)
21
Adams
P. D.
Afonine
P. V.
Bunkoczi
G.
Chen
V. B.
Davis
I. W.
Echols
N.
Headd
J. J.
Hung
L.-W.
Kapral
G. J.
Grosse-Kunstleve
R. W.
, et al. 
PHENIX: a comprehensive Python-based system for macromolecular structure solution
Acta Crystallogr. Sect. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
213
-
221
)
22
Matthews
B. W.
Solvent content of protein crystals
J. Mol. Biol.
1968
, vol. 
33
 (pg. 
491
-
497
)
[PubMed]
23
Kantardjieff
K. A.
Rupp
B.
Matthews coefficient probabilities: improved estimates for unit cell contents of proteins, DNA, and protein–nucleic acid complex crystals
Protein Sci.
2003
, vol. 
12
 (pg. 
1865
-
1871
)
[PubMed]
24
Emsley
P.
Lohkamp
B.
Scott
W. G.
Cowtan
K.
Features and development of Coot
Acta Crystallogr. Sect. D Biol. Crystallogr.
2010
, vol. 
66
 (pg. 
486
-
501
)
25
Winn
M. D.
Ballard
C. C.
Cowtan
K. D.
Dodson
E. J.
Emsley
P.
Evans
P. R.
Keegan
R. M.
Krissinel
E. B.
Leslie
A. G. W.
McCoy
A.
, et al. 
Overview of the CCP4 suite and current developments
Acta Crystallogr. Sect. D Biol. Crystallogr.
2011
, vol. 
67
 (pg. 
235
-
242
)
26
McCoy
A. J.
Grosse-Kunstleve
R. W.
Adams
P. D.
Winn
M. D.
Storoni
L. C.
Read
R. J.
Phaser crystallographic software
J. Appl. Crystallogr.
2007
, vol. 
40
 (pg. 
658
-
674
)
[PubMed]
27
Holm
L.
Rosenstrom
P.
Dali server: conservation mapping in 3D
Nucleic Acids Res.
2010
, vol. 
38
 (pg. 
W545
-
W549
)
[PubMed]
28
Kozakov
D.
Hall
D. R.
Beglov
D.
Brenke
R.
Comeau
S. R.
Shen
Y.
Li
K. Y.
Zheng
J. F.
Vakili
P.
Paschalidis
I. C.
Vajda
S.
Achieving reliability and high accuracy in automated protein docking: ClusPro, PIPER, SOU, and stability analysis in CAPRI rounds 13–19
Proteins
2010
, vol. 
78
 (pg. 
3124
-
3130
)
[PubMed]
29
Mardian
J. K. W.
Isenberg
I.
Yeast inner histones and the evolutionary conservation of histone-histone interactions
Biochemistry
1978
, vol. 
17
 (pg. 
3825
-
3833
)
[PubMed]
30
MacDonald
N.
Welburn
J. P. I.
Noble
M. E. M.
Nguyen
A.
Yaffe
M. B.
Clynes
D.
Moggs
J. G.
Orphanides
G.
Thomson
S.
Edmunds
J. W.
, et al. 
Molecular basis for the recognition of phosphorylated and phosphoacetylated histone H3 by 14-3-3
Mol. Cell
2005
, vol. 
20
 (pg. 
199
-
211
)
[PubMed]
31
Lapouge
K.
Smith
S. J. M.
Walker
P. A.
Gamblin
S. J.
Smerdon
S. J.
Rittinger
K.
Structure of the TPR domain of p67phox in complex with Rac·GTP
Mol. Cell
2000
, vol. 
6
 (pg. 
899
-
907
)
[PubMed]
32
Parthun
M. R.
Histone acetyltransferase 1: more than just an enzyme?
Biochim. Biophys. Acta
2012
, vol. 
1819
 (pg. 
256
-
263
)
[PubMed]
33
Krissinel
E.
Henrick
K.
Inference of macromolecular assemblies from crystalline state
J. Mol. Biol.
2007
, vol. 
372
 (pg. 
774
-
797
)
[PubMed]
34
Krissinel
E.
Crystal contacts as Nature's docking solutions
J. Comput. Chem.
2010
, vol. 
31
 (pg. 
133
-
143
)
[PubMed]
35
Baker
N. A.
Sept
D.
Joseph
S.
Holst
M. J.
McCammon
J. A.
Electrostatics of nanosystems: application to microtubules and the ribosome
Proc. Natl. Acad. Sci. U.S.A.
2001
, vol. 
98
 (pg. 
10037
-
10041
)
[PubMed]

Author notes

1

These authors contributed equally to this work.

The atomic co-ordinates and structure factors of Hif1 have been deposited in the PDB and will appear under accession code 4NQ0.