Homodimeric structure of sNASP

We first wished to determine the structures of sNASP in monomeric and dimeric forms in order to gain insights into how they bind histones. Crystallization of both forms of sNASP TPR domain was attempted; however, only the dimeric protein crystallized, and a 2.9 Å structure was solved (Fig. 2A; Table 1). There is one sNASP molecule per asymmetric unit, and an apparent homodimer results from the packing of a twofold crystallographic symmetry-related molecule via an ~60-residue C-terminal helix. Surprisingly, this long helix is formed by the predicted second helix of the TPR4 motif and the entire coiled-coil domain. Apart from this long C-terminal helix, each of the TPR units from 1 to 3 consists of two antiparallel α helices, while TPR4 has a separate N-terminal helix (Fig. 2A). For ease of description, we denote these helices α1 to α7 sequentially from the N terminus, and the long C-terminal helix is termed αC, which is further separated into an N-terminal α8 segment belonging to the TPR4 motif, and a C-terminal α9 segment, which corresponds to the predicted coiled-coil domain (Fig. 2A). Homodimerization of sNASP entails the formation of an extended antiparallel coiled-coil between the two αC helices, and the intermolecular interaction buries a surface area of 1877 Ų. The coiled-coil interaction involves six pairs of charged residues and a number of hydrophobic ones (Supplemental Fig. S1B). Not surprisingly, removal of αC impairs the dimerization ability of sNASP in solution (Supplemental Fig. S1C).

Open in a separate window

Figure 2.

Homodimeric structure of sNASP. (A) A cartoon representation of the sNASP TPR domain dimer. One molecule is colored according to the TPR repeat motifs (α1–α8) as indicated, and the CC domain (α9) is in magenta. The symmetry-related molecule is colored gray. The two sNASP TPR molecules dimerize via antiparallel packing of the long αC helix, which is composed of α8 and α9. (B) Electrostatic potential of homodimer surface viewed from two directions related by 180° rotation. Uncharged and negatively and positively charged surface regions are colored white, red, and blue, respectively. (C) Pull-downs of histone H3.1–H4 with GST-tagged sNASP fragments (TPR, amino acids 40–320; TPRΔαC, amino acids 40–259; and αC, amino acids 259–320) analyzed by Coomassie-stained SDS-PAGE. (D) Supercoiling assay for in vitro nucleosome assembly by sNASP. (Left) ϕX174 DNA (lane 1) was treated with DNA topoisomerase I (lane 2) and incubated with histone octamers (lane 3, negative control), and increasing amounts (0.25 and 0.5 μg) of yNap1 (lanes 4,5, positive control), full-length sNASP (lanes 6,7), sNASP TPR monomer (lanes 8,9), dimer (lanes 10,11), and the ΔαC variant (lanes 12,13), together with 0.2 μg of histone octamer, were incubated with 0.1 μg of relaxed ϕX174 DNA. Arrowheads labeled R and S at the left indicate positions of supercoiled and relaxed DNA. (Right) Input protein samples for the supercoiling assay analyzed by Coomassie-stained SDS-PAGE.

Table 1.

Data collection and refinement statistics

Open in a separate window

The TPR motifs pack via stacking of the N-terminal helix of the next TPR motif onto the groove between the helix pair of the preceding TPR motif, principally via hydrophobic interactions, and the packing results in a right-handed superhelical twist (Fig. 2A). While charge distribution on the inner (concave) surface of the TPR domain is mainly neutral, the exposed surface of α1, which forms one rim of the concave surface, is positively charged (Fig. 2B). In contrast, the loop connecting α3 and α4 in TPR2 and that connecting α5 and α6 in TPR3, both of which are part of the convex side of the sNASP surface, are negatively charged. Additionally, a 63-residue segment of the DE-rich region, which is inserted between α3 and α4, is disordered in the structure. It is interesting that the approximately seven C-terminal turns of the long αC are highly charged, but oppositely charged residues are segregated on the opposing surface sides along the C-terminal segment of αC, and the negatively charged surface region from the symmetry-related sNASP molecule forms the other rim of the concave surface sNASP (Fig. 2B). Previous studies showed that the TPR domain of sNASP is capable of binding histone H3–H4 and possesses the nucleosome assembly activity in vitro (Osakabe et al. 2010; Bowman et al. 2016, 2017), while the coiled-coil domain may provide a structural scaffold mediating protein–protein interactions (Rose and Meier 2004). Our GST pull-down assay indicates that both the intact sNASP TPR domain (Fig. 2C, lane 3), which is mixture of monomeric and dimeric forms, and its homodimerization-defective truncation variant encompassing α1–α7 (Fig. 2C, lane 4) can interact with histone H3.1–H4, whereas the C-terminal long helix αC (amino acids 259–320), which contains TPR4 α8 and the CC domain, cannot (Fig. 2C, lane 5). Interestingly, we found that, although not needed for H3–H4 binding, the coiled-coil domain of sNASP is important for its nucleosome assembly activity, as judged by an in vitro supercoiling assay (Fig. 2D). In this assay, both monomeric and dimeric forms of the sNASP TPR domain appear to have levels of nucleosome assembly activity comparable with that of the full-length protein.

Structure of monomeric sNASP in complex with H3–H4 and ASF1

sNASP has been shown to form a quaternary complex with histones H3 and H4 and their chaperone ASF1 (Bowman et al. 2017). To investigate how sNASP binds histone H3–H4 together with ASF1, we assembled the quaternary complex using the sNASP TPR domain, the globular domain of ASF1A (amino acids 1–172), and full-length histones H3.1 and H4 under a medium salt concentration condition (0.5 M NaCl). Gel filtration results showed that both dimeric and monomeric forms of the sNASP TPR domain can form stable complexes with the preassembled ASF1A–H3.1–H4 complex. They eluted from a Superdex 200 10/300GL sizing column at ~11- and 13-mL elution volume, respectively, while the ASF1A–H3.1–H4 complex eluted at ~14.7 mL (Fig. 3A). However, crystallization trials of the two sNASP-containing complexes with a large number of conditions were unfruitful.

Open in a separate window

Figure 3.

Structure of the monomeric sNASP bound to H3–H4 and ASF1A. (A, left panels) Sizing column elution profiles of dimeric (top panel) and monomeric (middle panel) sNASP TPR domain with the ASF1-H3–H4 complex, and the ASF1A–H3–H4 complex alone (bottom panel). The profiles of sNASP dimers and monomers, shown in blue and magenta dashed lines, are superimposed for reference in respective panels. (Right panels) SDS-PAGE analysis of 0.5-mL fractions spanning the 8.0- to 18.5-mL elution volume corresponding to the three sizing column runs displayed in the left panels, respectively. Each gel panel was assembled from two gels with a thin vertical gray line separating them. (B) Overall structure of the sNASP–ASF1A–H3–H4 quaternary complex. sNASP is colored yellow, ASF1A is colored green, H3 is colored cyan, and H4 is colored magenta. Orange dashed line denotes the disordered DE-rich region within TPR2, and the gray dashed line indicates the disordered GS linker joining sNASP and ASF1A. (C) A top view of the structure in cartoon drawing superimposed with a semitransparent surface representation of sNASP (gray) and ASF1A (light green). sNASP regions involved in binding the N-terminal segment of H3 and the C-terminal tail of H4 are indicated with a dashed rectangle and circle, respectively. (D) Conformational differences of the N-terminal region of H3 in the structures with the sNASP–ASF1 complex (cyan), nucleosome core particle (NCP, orange) (PDB: 5AV9), the Rtt109-Asf1 complex (green) (PDB: 5ZBA), the MCM2-ASF1 complex (yellow) (PDB: 5C3I), ASF1 (light blue) (PDB: 2IO5), and DAXX (magenta) (PDB: 4HGA). (E) Superposition of the sNASP TPR monomer structure, colored the same as above, with that of the sNASP dimer, colored green and magenta, and yeast Hif1 (PDB: 4NQ0) colored in gray.

We reasoned that one possible reason for unsuccessful crystallization might be due to ASF1A–H3–H4's propensity to precipitate under low salt concentrations (<0.5 M NaCl). To overcome this problem, we engineered a fusion protein in which the globular domain of ASF1A (amino acids 1–156) is tethered to the C-terminal end of the sNASP TPR domain through a 20-residue GS linker, with the sequence of four repeats of GGGGS. The sNASP–ASF1A fusion protein was then coexpressed with full-length human H3.1 and H4 in E. coli, and the complexes with H3.1–H4 were also stable at 0.5 M NaCl. Furthermore, the fusion protein forms monomeric and dimeric complexes with H3.1–H4 much like the corresponding complexes assembled with separately expressed sNASP and ASF1A (Supplemental Fig. S2A), and now the fusion protein complexes can tolerate a low salt concentration, even at 100 mM NaCl. Importantly, we were able to achieve crystallization of the sNASP–ASF1A–H3–H4 complex using the monomeric fusion protein, and succeeded in solving a 3.0-Å resolution structure by molecular replacement (Fig. 3B; Table 1). There are two sNASP–ASF1A–H3–H4 complexes, both with a 1:1:1:1 subunit stoichiometry, in the crystallographic asymmetric unit (ASU). The two complexes within the ASU share a highly similar structure, manifested by an overall root-mean-square deviation (RMSD) of 0.16 Å. The contacts between the two complexes within the ASU are mediated through ASF1–ASF1 interactions and do not appear to have physiological implications. Thus, we will choose the complex (chains A–D) having a better quality electron density map for further analysis from now on.

In the structure, the electron density for the GS linker is completely absent, and the C-terminal end of sNASP and the N-terminal end of ASF1A to which the GS linker connects are spatially close (Fig. 3B). This observation indicates that the linker has a flexible conformation, and artificial tethering of sNASP and ASF1A via this linker did not constrain their positioning for histone binding. The structure shows that sNASP makes direct contacts with both histones H3 and H4, which bind ASF1A as a heterodimer in the same manner as seen in the structures of ASF1–H3–H4 complexes (Fig. 3B,C; English et al. 2006; Natsume et al. 2007; Zhang et al. 2018). sNASP interacts with histones H3 and H4 through the convex surface of the TPR domain: TPR1 through TPR3 with H3, and TPR3–TPR4 with H4 (Fig. 3B). Intriguingly, histone H3 interacts with sNASP principally via its N-terminal flexible region encompassing residues 40–57, which normally adopts a helical structure (αN) in NCP. The conformation of this H3 region is highly variable outside the nucleosomal context, and αN is mostly unwound in the present structure (Fig. 3D). Previous structural studies found that αN could adopt different orientations when binding histone chaperones, but the helical structure remains mostly intact. The drastic structural change of αN in histone H3 was first observed in the structure of Rtt109, a yeast H3K56 acetyltransferase, bound to the Asf1–H3–H4 complex (Zhang et al. 2018). Our observation here reinforces the notion that conformational dynamics of this H3 region are important for recognition by histone chaperones. The very N-terminal 39 residues of H3 are disordered, but based on the location of the ordered portion of the H3 N-terminal tail from residue 40 onward, the disordered portion of the H3 N-terminal tail is expected to be in the vicinity of the also disordered DE-rich acidic region between α3 and α4 of sNASP (Fig. 3B). The oppositely charged H3 tail and the acidic region of sNASP may interact via Coulomb's force. Additionally, histone H4 interacts with sNASP via its C-terminal β strand, which is stabilized by interaction with ASF1A, and the very C-terminal tail (Fig. 3B,C; Supplemental Fig. S2B).

The sNASP TPR domain in the complex structure is apparently in a monomeric form, and its structure is very similar to that of the yeast Hif1 monomer; the two can be superimposed with an RMSD of 2.18 Å. The structures of monomeric and dimeric forms of sNASP are nearly identical throughout the four TPR motifs, except that the very C-terminal helix α9, which joins with α8 to form the long, continuous αC in the dimeric structure of sNASP, folds back to form an antiparallel coiled coil with α8 in the present structure (Fig. 3E). Apart from this difference, the two TPR domain structures can be superimposed with an RMSD of 0.57 Å.

Interaction between sNASP and H3–H4

The structure shows that the N-terminal segment of histone H3 (amino acids 40–57) snakes through the convex surface of sNASP formed by TPR1, TPR2, and TPR3 (Fig. 4A), as evidenced by a clear, continuous electron density map (Supplemental Fig. S3A). This region of histone H3 interacts with sNASP extensively through a mixture of hydrophobic interactions, salt bridges, and hydrogen bonds (Fig. 4A). Outside the N-terminal region of H3, only Glu105 in the central α2 helix of H3 is found to interact with sNASP via hydrogen bonding to Gln221 and Glu224 located in α6 of TPR3. We deleted the entire N-terminal tail (amino acids 1–60) of histone H3.1 (H3.1Δ60) and tested the interaction between the sNASP TPR domain and H3.1Δ60–H4 in the presence and absence of ASF1A. GST pull-down results (Fig. 4B) show that the sNASP TPR domain dramatically reduced binding with ASF1A–H3.1Δ60–H4 (Fig. 4B, lane 13), whereas no obvious impact is observed with full-length H3.1–H4, alone (Fig. 4B, lane 10) or in complex with ASF1A (Fig. 4B, lane 12), and H3.1Δ60–H4 (Fig. 4B, lane 11). This observation validates the structural finding that the N-terminal tail of histone H3 is essential for sNASP binding when ASF1 is present, and also reveals that sNASP binds H3–H4 differently when ASF1 is absent.

Open in a separate window

Figure 4.

Histone H3 N-terminal tail binding sites in sNASP. (A) Detailed view of interactions between residues 40–57 of histone H3 (cyan) and sNASP TPR (yellow). The involved residues are shown in a stick representation, with sNASP and H3 residues labeled in black and blue, respectively, and black dashed lines indicate hydrogen bonds. (B, top panel) Pull-down of protein samples indicated with a plus sign by the GST-tagged sNASP TPR domain or GST. (Bottom panel) Input samples for GST pull-down experiments. Corresponding lane numbers are labeled at the bottom. (C) GST pull-down of the ASF1A–H3.1–H4 complex by sNASP or its indicated mutants. (D) Cellular colocalization test of WT or mutant sNASP with H3.1. Scale bar, 10 μm. (E) GST pull-down of H3.1 and its mutants by sNASP. (F) Colocalization test of sNASP with H3.1 and its indicated mutants.

To further delineate sNASP–H3 interactions, we carried out structure-guided mutagenesis of sNASP and tested the binding of these mutants to the ASF1A–H3.1–H4 complex. Prominently among sNASP–H3 interactions, Ile51 of H3 contacts Leu185 and Ile188 located in α4 of sNASP via hydrophobic interactions; and the guanidino group of Arg52 of H3 interacts with Asp184 in α4 and Glu225 in α6 of sNASP via hydrogen bonds and charge interactions, while the carbonyl of Arg52 makes a hydrogen bond with the nitrogen atom of the indole ring of Trp180 of sNASP. GST pull-downs with the D184R, E225R, D184R E225R, or L185A I188A mutants of sNASP show that these mutants effectively lost binding to the ASF1A–H3–H4 complex (Fig. 4C, lanes 4–7). These in vitro pull-down results are supported by LacO-LacI targeting experiments in mammalian A03_1 cells, where sNASP or its derivatives were fused with mCherry and LacI, while H3.1 was fused to EGFP (Fig. 4D; Supplemental Fig. S3B,C). Reciprocally, we made I51A, R52A, and I51A R52A mutants of H3, and GST pull-down experiments showed that the individual I51A or R52A mutations have minimal impact on sNASP–H3 binding, while the I51A R52A combination mutant of H3 greatly impaired the interaction with sNASP (Fig. 4E, lanes 9–11). The in vitro GST pull-down results are corroborated by cellular targeting experiments (Fig. 4F; Supplemental Fig. S3D). It is worth pointing out that Ile51 and Arg52 are strictly conserved among histone variants across species (Supplemental Fig. S3E), suggesting a common sNASP binding mode of H3 variants in the presence of ASF1.

In addition to the above-mentioned sNASP–H3 interactions, the hydroxyl and the carbonyl groups of Ser216 of sNASP make hydrogen bonds with Tyr54 and Ser57 of H3, respectively. Ser57 of H3 also contacts Asn218 of sNASP via a hydrogen bond. Tyr41 of H3 interacts with Glu79 and Tyr91 of sNASP, and Asn175 of sNASP forms a hydrogen bond with Arg40 of H3 (Fig. 4A). Interestingly, a S216Y change is among somatic mutations of sNASP in cancer tabulated in the COSMIC database (Bamford et al. 2004). A GST pull-down of the ASF1A–H3–H4 complex by the S216Y mutant of sNASP showed no obvious effect, while cellular targeting experiments showed noticeable reduction of colocalization of the sNASP mutant with histone H3 (Fig. 4D; Supplemental Figs. S3C, S4A). This result indicates a possible link between histone binding and human pathogenesis by sNASP alterations. As a contrast, certain sNASP or H3.1 mutations at the interface between the two showed minimal or no impact on sNASP–histone interaction (Supplemental Fig. S4), thus rendering support of the significance of our earlier analyses.

The sNASP interaction region of histone H4 is located in the C-terminal tail of histone H4 (amino acids 95–102), which is visible in its entirety in the structure. Residues 95–97 form an antiparallel β sheet with ASF1A, and the side chain of Phe100 inserts into a hydrophobic pocket between two β sheets of ASF1A (Fig. 5A; Supplemental Fig. S5A). In this area, we noted the only direct interaction between ASF1A and sNASP involved a hydrogen bond between Asn231 of sNASP and ASF1A Ser142's main chain carbonyl group. The sole function of ASF1A appears to be to stabilize the C-terminal residues of H4, thus enabling them to participate in intermolecular interactions. A similar mechanism of Asf1 was noted in the previous study of its function in regulating H3K56 acetylation of Rtt109 (Zhang et al. 2018).

Open in a separate window

Figure 5.

Interaction between sNASP and histone H4. (A) Depiction of interactions involving H4 C-terminal tail encompassing residues 95–102 (magenta), ASF1 (green), and sNASP (yellow). The involved residues are shown in a stick representation, and black dashed lines indicate hydrogen bonds. (B) GST pull-down of ASF1A–H3.1–H4 or ASF1A–H3.1–H4Δ95-C complexes by the WT or Y255R Q267R mutant of sNASP. (C, left panels) From top to bottom, sizing column elution profiles of the ASF1A–H3–H4Δ95-C complex with dimeric and monomeric sNASP and without sNASP, respectively. (Right panels) From top to bottom, SDS-PAGE analysis of 0.5-mL fractions spanning 8.0–19.5 mL corresponding to three column runs shown in the left panels. Each gel panel was assembled from two gels with their boundary indicated by a thin vertical gray line.

sNASP–H4 interactions involve hydrogen bonds between Tyr255 of sNASP and Arg95 and the C-terminal carboxylate of Gly102 of H4, Glu263 of sNASP and Arg95 of H4, and Gln267 of sNASP and the carbonyl group of Gly101 of H4. To test their significances, we created a sNASP Y255R Q267R double mutant and an H4 deletion mutant, H4Δ95-C, removing all residues from Arg95 to the C terminus. However, both cellular targeting and in vitro GST pull-down assays show that these mutations of sNASP and H4 do not compromise the binding between sNASP and histones H3–H4 (Fig. 5B; Supplemental Fig. S5B,C). Curiously, we noticed that in the GST pull-down experiments with the ASF1A–H3–H4Δ95-C complex, ASF1A is evicted from the sNASP–histone complex (Fig. 5B, lane 6). Gel filtration experiments also show that both monomeric and dimeric forms of the sNASP TPR domain can evict ASF1A from the ASF1A–H3–H4Δ95-C complex (Fig. 5C). These results indicate that while the interaction between sNASP and the C-terminal tail of H4 is of marginal effect, it is important for ASF1A to maintain a firm grip on H3–H4 in the presence of sNASP, which appears to be able to compete ASF1A off H3–H4 through binding to a yet unidentified site in H3–H4 either overlapping with the ASF1A binding sites or through an allosteric effect.

Protein expression and purification

sNASP fragments were expressed in E. coli BL21 CodonPlus (DE3) RIL cells cultured in LB medium at 37°C, followed by induction of protein production with the addition of 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) for 4 h. Selenomethionine (SeMet)-substituted sNASP TPR protein was prepared by inhibition of the E. coli methionine biosynthesis pathway. Briefly, the pRSFDuet-His₆-SUMOtag-sNASP TPR plasmid was transformed into E. coli BL21 CodonPlus (DE3) cells, and the bacteria were cultured in the LB medium at 37°C. The cells were collected and cultured in the M9 minimal medium at 37°C, followed by addition of 25 mg/L SeMet (Acros Organics); Lys, Thr, and Phe at 100 mg/L; and Leu, Ile, and Val at 50 mg/L to inhibit the endogenous synthesis of methionine when the OD₆₀₀ value of the culture reached 0.6. After incubation for 1 h, the temperature was reduced to 16°C and protein expression was induced by the addition of IPTG to a final concentration of 0.3 mM, followed by further shaking for 36 h.

All purification steps were carried out at 4°C. The following procedures were used for purification of sNASP TPR domain or its mutants. Cell pellets were resuspended in 5 vol of the binding buffer A with 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, 20 mM imidazole, and 1 mM phenylmethylsulphonyl fluoride (PMSF), before lysing by sonication. Cell debris were removed by centrifugation, and the supernatant was incubated for 1 h at 4°C with Ni²⁺-NTA agarose resins pre-equilibrated with buffer A, followed by washing the resins three times with buffer A. The bound proteins were eluted with buffer B (20 mM Tris-HCl at pH 7.5, 500 mM NaCl, 500 mM imidazole) and immediately supplemented with 1 mM each ethylenediaminetetraacetic acid (EDTA) and dithiothreitol (DTT). After removing the His₆-SUMO tag by using Ulp1 (SUMO protease), the protein was further purified on a HiLoad 16/60 Superdex 200 column (GE Healthcare) in buffer C (20 mM Tris-HCl at pH 7.5, 100 mM NaCl, 1 mM DTT). Elution fractions were analyzed by SDS-PAGE, and the fractions enriched with the desired proteins were pooled and concentrated to ~30 mg/mL by ultrafiltration and stored at −80°C before use.

Full-length sNASP was expressed in suspension culture of HEK293F cells in the SMM 293T-II medium (Sino Biological, Inc.). A total of 300 mL of HEK293F cells was transfected with 300 μg of pcDNA3-sNASP-strep plasmid plus 600 μg of linear polyethylenimine (PEI) (Polysciences), and cells were harvested at 48 h after transfection. The sNASP-strep protein was purified through successive steps of affinity purification with Strep-TactinXT beads (IBA Life Science), ion exchange with a heparin column (GE Healthcare), and size exclusion chromatography with a Superdex 200 Increase 10/300 GL column (GE Healthcare) in buffer D (20 mM Tris-HCl at pH 7.5, 500 mM NaCl, 5% glycerol, 1 mM DTT).

The complex of the sNASP TPR-ASF1A fusion protein with histones H3.1 and H4 was produced by coexpression of the pRSFDeut-His₆-SUMOtag-sNASP TPR-(G₄S)₄-ASF1A HBD plasmid and the pCDFDuet1–H3.1–H4 plasmid in E. coli BL21 CodonPlus (DE3) RIL cells. Production of the protein complex was induced with 0.5 mM IPTG for 4 h at 37°C. The protein complex was first purified on Ni²⁺-NTA agarose resins. After removing the His₆-SUMO tag through digestion with Ulp1, the protein complex was first purified on a heparin column (GE Healthcare), followed by a HiLoad 16/60 Superdex 200 column (GE Healthcare) in buffer C (20 mM Tris-HCl at pH7.5, 100 mM NaCl, 1 mM DTT). Highly purified samples were pooled and concentrated to ~25 mg/mL and stored at −80°C.

ASF1A proteins were expressed in bacterial cells using the pRSFDuet-His₆-ASF1A_1–172 or the pRSFDuet-His₆-SUMOtag-ASF1A_1–156 plasmids for 4 h at 37°C with 0.5 mM IPTG. After purification with Ni²⁺-NTA resins, the His₆-ASF1A_1–172 protein was directly eluted and further purified through a sizing column, whereas for His₆-SUMOtag-ASF1A_1–156, the His₆-SUMO tag was removed on the column, and the ASF1A_1–156 protein in the flowthrough was collected and further purified in the sizing column. Histones H3.1–H4 were coexpressed using the pCDFDuet1–H3.1–H4 plasmid in E. coli cells for 4 h at 37°C with 0.5 mM IPTG. Histone complexes were first purified by ion exchange column using a 5-mL SP column (GE Healthcare), and further purified through a gel filtration column at 2 M NaCl. H3.1–H4 mutant complexes were expressed and purified following the same procedure.

Assembly of wild-type and mutant ASF1A–H3.1–H4 complexes was carried out by mixing H3.1–H4 and His₆-ASF1A_1–172 or ASF1A_1–156 at a 1:2 molar ratio for 30 min on ice, and loading onto a HiLoad 16/60 Superdex 200 (GE Healthcare) gel filtration column in buffer D (20 mM Tris-HCl at pH 7.5, 500 mM NaCl, 5% glycerol, 1 mM DTT). Fractions enriched with the ternary complex were pooled and concentrated to ~10 mg/mL.

Crystallization, data collection, and structure determination

Native crystals of the dimeric sNASP TPR domain were obtained by mixing 1.0 μL of protein (15–18 mg/mL) with 1.0 μL of reservoir solution containing 0.1 M sodium cacodylate (pH 6.5), 32% (v/v) PEG 300, and 0.2 M calcium acetate. The SeMet-substituted sNASP TPR domain was crystallized in 0.1 M MES (pH 6.5), 30% (v/v) PEG300, and 0.18 M calcium acetate at a concentration of 12–15 mg/mL. The sNASP_40–320–(G₄S)₄–ASF1A_1–156–H3.1–H4 complex protein at a concentration of 10–12 mg/mL was crystallized in 18% (v/v) tacsimate (pH 5.0), and 11% (w/v) PEG2000MME. All crystals were grown by sitting-drop vapor diffusion at 4°C. Cryogenic data collection was performed with cryoprotectants prepared by supplementing the crystallization mother liquors with 15% glycerol before flash-freezing in liquid nitrogen.

For structural determination of the dimeric sNASP TPR domain, X-ray diffraction data were collected at the Shanghai Synchrotron Radiation Facility (SSRF) beamline BL17U equipped with a Quantum 315r CCD detector (ADSC) using a wavelength of 0.97915 Å. Data were processed using the HKL2000 software package (Otwinowski and Minor 1997). Four Se sites were identified with SHELXD (Collaborative Computational Project, Number 4 1994) using the 3.6 Å SeMet data set. Refinement of the heavy atom substructure, phasing, and density modification was performed using PHENIX (Adams et al. 2010). The experimentally determined phase was then extended to the 2.9 Å resolution native data set. The initial model was manually built with COOT (Emsley and Cowtan 2004), and the model was improved by iterative cycles of refinement and model adjustment. The final structure had R_work and R_free values of 0.187 and 0.223, respectively, and good stereochemical quality, with 97.2% and 2.8% of the residues in the favored and allowed regions of the Ramachandran plot, respectively. The ordered structure contained residues 40–103 and 167–320 of sNASP TPR. Residues 104–166 of TPR2 were rich in acidic residues and had no clear electron density, and thus were not included the final model.

Crystal diffraction data for the sNASP–ASF1–H3–H4 complex were also collected at SSRF beamline BL17U at a wavelength of 0.97913 Å using a Dectris Eiger 16 M detector, and processed with the HKL2000 package. The complex structure was determined by molecular replacement using PHASER with the structures of ASF1A–H3–H4 (PDB: 2IO5) and the structure of the sNASP TPR domain encompassing the α1–α7 region as the search models. The structure was then manually adjusted using COOT and refined in PHENIX. The final structure contained residues 40–103 and 167–320 of the sNASP TPR domain, residues 40–135 of histone H3.1, residues 23–102 of histone H4, and residues 1–154 of ASF1A. The refined model had satisfactory R-values (R_work/R_free = 0.165/0.203) and good Ramachandran statistics (97.4% favored and 2.6% allowed). Finally, the models were validated with MolProbity (Chen et al. 2010). Detailed statistics for data collection and structure refinement are shown in Table 1. Structural figures were prepared using PyMOL (Schrödinger LLC).

Atomic coordinates and structure factors have been deposited in PDB with accession numbers 7V6P and 7V6Q for the sNASP TPR domain and the sNASP–ASF1A–H3–H4 complex, respectively.

Molecular size analysis by gel filtration

The sNASP TPR domain was loaded onto a Superdex 200 Increase 10/300 GL column (GE Healthcare) equilibrated in a buffer containing 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, and 5% glycerol. Peak 1 contained fraction eluted between 11.5 and 13.0 mL, and peak 2 contained fractions eluted between 14.0 and 15.5 mL. Samples of both peak 1 and peak 2 were reloaded onto the sizing column for further analysis. Analyses of various protein complexes and individual components were carried out using the samples prepared as described before, and in the cases of testing binding with sNASP TPR domain, other components were added in roughly twofold excess. The loading samples were generally diluted to ~1–3 mg/mL, and a total volume of 0.5 mL in 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, and 5% glycerol. The samples were then injected at 0.4 mL/min onto a Superdex 200 Increase 10/300 GL column (GE Healthcare). Elution fractions between 8.0 and 19.5 mL were pooled and analyzed by SDS-PAGE and stained with Coomassie blue. The gel filtration chromatograms were plotted by OriginPro 8.0 (OriginLab Corporation).

Analytical ultracentrifugation (AUC)

Sedimentation velocity experiments were carried out in a Beckman Coulter ProteomeLab XL-I analytical ultracentrifuge with absorbance detection. Samples were prepared in a buffer containing 20 mM Tris-HCl (pH 7.5), 500 mM NaCl, and 1 mM EDTA. Samples of the sNASP TPR domain from peak 1 and peak 2, and peak 2 of full-length sNASP, all at ~1 mg/mL in a total volume of 400 μL, were loaded into the cell chamber against buffer blank and spun at 55,000 rpm for 7 h at 20°C. Data were analyzed using the c(M) model of molecular mass distribution in Sedfit (Schuck 2000).

Topological assay for nucleosome assembly

DNA supercoiling assays were performed as described previously(Osakabe et al. 2010) with the following modifications. Relaxed DNA template was prepared by combining ϕX174 RF I DNA (NEB N3021S) with excess DNA topoisomerase I (TaKaRa 2240A) in the assembly buffer (10 mM Tris-HCl at pH 7.5, 125 mM NaCl, 2 mM MgCl₂, 0.5 mM DTT, 0.1 mg/mL BSA) for 1 h at 37°C and kept at room temperature ready for use. Histone octamers (200 ng) were preincubated with 250/500 ng of yNAP1, full-length sNASP, sNASP TPR monomer, sNASP TPR dimer, or sNASP TPRΔαC (2.5-fold/fivefold to octamer), respectively, in the assembly buffer for 30 min at 37°C. To initiate the assembly reaction, 100 ng of relaxed plasmid DNA was added to the chaperone–histone mixture and incubated for 90 min at 37°C. The standard 25-μL chromatin assembly reaction was stopped by adding EDTA to 20 mM and SDS to 0.5%, with an additional 0.5 mg/mL proteinase K, and incubation for 15 min at 55°C followed by phenol-chloroform extraction. The DNA samples were analyzed on a 1% agarose gel in 1× TAE buffer (40 mM Tris-acetate, 1 mM EDTA) for 4 h at 90 V with ethidium bromide (EB) staining.

GST pull-downs

For pull-down experiments between the GST-sNASP TPR domain and ASF1A–H3.1–H4 proteins, 3 µg of GST, GST-sNASP TPR, or its mutants were immobilized on 5 µL of glutathione Sepharose 4B resin, which was then mixed with 5 µg of ASF1A–H3.1–H4 or its mutants in 500 µL of the binding buffer (20 mM Tris-HCl at pH 7.5, 300 mM NaCl, 5 mM 2-mercaptoethanol, 0.02% NP-40, 1 mM EDTA), and the mixture was rotated overnight at 4°C. The resins were then washed five times with 1 mL of binding buffer with 500 mM NaCl. Bound proteins were eluted in SDS sample buffer, separated on a 15% SDS-PAGE gel, and stained with Coomassie blue. GST pull-down experiments between GST-sNASP fragments and WT or mutant H3.1–H4 complexes were performed following the same procedure.

We thank staff scientists at the Shanghai Synchrotron Radiation Facility beamlines BL17U and BL19U for assistance for X-ray data collection, and Dr. Xiaoxia Yu for technical support in handling AUC. This work was supported in part by grants from the Natural Science Foundation of China (91853204, 31991162, and 31521002), the Ministry of Science and Technology of China (2017YFA0504202, 2018YFE0203300, 2017YFA0506600, and 2019YFA0508900), and the Strategic Priority Research Program of Chinese Academy of Sciences (XDB37010101). C.-P.L. is supported by the Youth Innovation Promotion Association of the Chinese Academy of Sciences (2018125).

Author contributions: R.-M.X. conceived the project. C.-P.L. performed protein expression, purification, crystallization, and structure determination. C.-P.L and M.W. collected X-ray data and refined models. C.-P.L. performed AUC experiments, GST pull-downs, and gel filtration experiments. W.J. performed fluorescence colocalization experiments and protein expression in 293F cells. J.C. assisted and provided support of all experiments. J.H. performed the supercoiling assays under the supervision of G.L. C.-P.L. and R.-M.X. wrote the manuscript, and all authors discussed the results and contributed to the final manuscript.