Structural insights into the histone H1-nucleosome complex Bing-Rui Zhoua, Hanqiao Fenga, Hidenori Katoa, Liang Daib,c, Yuedong Yangb,c,1, Yaoqi Zhoub,c,1, and Yawen Baia,2 a Laboratory of Biochemistry and Molecular Biology, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892; and bSchool of Informatics and cCenter for Computational Biology and Bioinformatics, Indiana University School of Medicine, Indiana University-Purdue University Indianapolis, Indianapolis, IN 46202

Linker H1 histones facilitate formation of higher-order chromatin structures and play important roles in various cell functions. Despite several decades of effort, the structural basis of how H1 interacts with the nucleosome remains elusive. Here, we investigated Drosophila H1 in complex with the nucleosome, using solution nuclear magnetic resonance spectroscopy and other biophysical methods. We found that the globular domain of H1 bridges the nucleosome core and one 10-base pair linker DNA asymmetrically, with its α3 helix facing the nucleosomal DNA near the dyad axis. Two short regions in the C-terminal tail of H1 and the C-terminal tail of one of the two H2A histones are also involved in the formation of the H1–nucleosome complex. Our results lead to a residue-specific structural model for the globular domain of the Drosophila H1 in complex with the nucleosome, which is different from all previous experiment-based models and has implications for chromatin dynamics in vivo.

E

ukaryotic genomic DNA is packaged into chromatin through association with positively charged histones to form the nucleosome, the structural unit of chromatin (1–3). The nucleosome core consists of an octamer of histones with two copies of H2A, H2B, H3, and H4, around which ∼146 bp of DNA winds in ∼1.65 left-handed superhelical turns (4). At this level of the DNA packaging, chromatin resembles a beads-on-a-string structure, with the nucleosome core as the beads and the linker DNA between them as the strings (5). At the next level of DNA packaging, H1 histones bind to the linker DNA and the nucleosome to further condense the chromatin structure (6, 7). H1mediated chromatin condensation plays important roles in cellular functions such as mitotic chromosome architecture and segregation (8), muscle differentiation (9), and regulation of gene expression (10, 11). Linker H1 histones typically are ∼200 amino acid residues in length, with a short N-terminal region, followed by a ∼70–80amino acid structured globular domain (gH1) and a ∼100-amino acid unstructured C-terminal domain that is highly enriched in Lys residues. H1 stabilizes the nucleosome and facilitates folding of nucleosome arrays into higher-order structures (12–15). gH1 alone confers the same protection from micrococcal nuclease digestion to the nucleosome as the full-length H1 does (16). The N-terminal region of H1 is not important for nucleosome binding (16, 17), whereas the C terminus is required for H1 binding to chromatin in vivo (18, 19) and for the formation of a stem structure of linker DNA in vitro (17, 20, 21). The globular domain structures of avian H5 (22) and budding yeast Hho1 (23), which are both H1 homologs, have been determined at atomic resolution and show similar structures. In addition, numerous studies have indicated that gH1/gH5 binds around the dyad region of the nucleosome (14, 24), leading to many conflicting structural models for how the globular domain of H1/H5 binds to the nucleosome (SI Appendix, Fig. S1) (24– 26). These models are divided into two major classes, symmetric and asymmetric, on the basis of the location of gH1/gH5 in the nucleosome. In the symmetric class, gH1/gH5 binds to the nucleosomal DNA at the dyad and interacts with both linker DNAs www.pnas.org/cgi/doi/10.1073/pnas.1314905110

(16, 17, 27, 28). In the asymmetric class, gH1/gH5 binds to the nucleosomal DNA in the vicinity of the dyad axis and to 10 bp (27, 29–32) or 20 bp (19, 29, 33, 34) of one linker DNA, or is located inside the DNA gyres, where it interacts with histone H2A (35). In addition, Zhou and colleagues also characterized the orientation of gH5 in the gH5-nucleosome complex (29). The use of nonuniquely positioned nucleosomes and indirect methods may have contributed to the differences in these models (SI Appendix, Fig. S1). Multidimensional nuclear magnetic resonance (NMR), and in particular methyl-based NMR, provides a direct approach to the structural characterization of macromolecular complexes (36, 37). We have previously assigned chemical shifts of the methyl groups of the side chains of residues Ile, Leu, and Val in the core histones (38) and the backbone amides in the disordered histone tails (39), which provide the fingerprints for investigating the interactions between H1 and the nucleosome. Here, we used NMR, along with several other methods, to determine the location and orientation of the globular domain of a stable mutant of Drosophila H1 on a well-positioned nucleosome. Results Histone Regions Involved in the Formation of the H1–Nucleosome Complex. To identify the nucleosome-binding regions of H1, we

first used a gel shift assay to examine the binding of several H1 fragments, which contain gH1 and various lengths of C-terminal regions, to the nucleosome centered on 167 bp DNA with the Widom “601” sequence (40), which can uniquely position the Significance Linker H1 histones control the accessibility of linker DNA between two neighbor nucleosomes to DNA-binding proteins and regulate chromatin folding. We investigated the structure of the H1–nucleosome complex through a combination of multidimensional nuclear magnetic resonance spectroscopy, site-directed mutagenesis-isothermal-titration calorimetry and computational design/modeling. The results lead to a unique structural model for the globular domain of H1 in complex with the nucleosome that contains residue-level information and have implications for the dynamics of chromatin in vivo. In addition, our approach will be useful for testing the hypothesis that the globular domain of H1 variants might have distinct binding geometries within the nucleosome, and thereby contribute to the heterogeneity of chromatin structure.

Author contributions: B.-R.Z. and Y.B. designed research; B.-R.Z., H.F., L.D., Y.Y., and Y.Z. performed research; H.K. contributed new reagents/analytic tools; B.-R.Z., H.F., L.D., Y.Y., Y.Z., and Y.B. analyzed data; and B.-R.Z. and Y.B. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1

Present address: Institute for Glycomics and School of Informatics and Communication Technology, Griffith University, Southport, QLD 4222, Australia.

2

To whom correspondence may be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1314905110/-/DCSupplemental.

PNAS Early Edition | 1 of 6

BIOPHYSICS AND COMPUTATIONAL BIOLOGY

Edited by S. Walter Englander, The University of Pennsylvania, Philadelphia, PA, and approved October 22, 2013 (received for review August 6, 2013)

nucleosome. To be consistent with later NMR experiments, we used a stable mutant of H1 in which four residues in the hydrophobic core of gH1 are replaced with the corresponding residues in gH5 (22) (Fig. 1 A and B and SI Appendix, Fig. S2). This stable H1 mutant is necessary to allow NMR experiments to be conducted at the high temperature (35 °C) required to observe the methyl groups in the H1-nucleosome complex (38). In this study, H1 refers to this stable mutant unless specified otherwise. We found that all fragments of H1 shift the position of the nucleosome in the agarose gel (Fig. 1C). In particular, the fragment containing residues 37–211 shifted the nucleosome as efficiently as H137–256. Using isothermal titration calorimetry (ITC), we further showed that H11–256, H137–211, and the wild type full-length H11–256 have similar binding affinities for the nucleosome (SI Appendix, Fig. S3 and Table S1). Therefore, to avoid signal overlaps in NMR spectra (SI Appendix, Fig. S4A), we chose H137–211 for subsequent NMR studies. Using 15 N/13C-labeled H137–211, we assigned the observable residues of H137–211 in complex with the nucleosome (SI Appendix, Fig. S4B). The unobserved residues of H137–211, which are presumably folded into the nucleosome, include gH1 (45–118), residues 119ASAKKEK125 immediately following the globular domain, and the Lys-rich region 164KKPKAKKAVAT174 in the middle of the C-terminal tail (Fig. 1 D and E). We further demonstrated that H137–211 and the wild-type full-length H11–256 used the same regions to bind to the nucleosome (SI Appendix, Fig. S4C). Similar results were observed when the nucleosome with 208 bp DNA was used (SI Appendix, Fig. S4D), indicating that H1 binds within the 10 bp of the linker DNAs that enter/exit the nucleosome core particle. To determine whether the histone core of the nucleosome interacts with H137–211, we compared the methyl spectra of the nucleosome histones in the absence or presence of H137–211. We found that H137–211 did not perturb the methyl spectra (SI Appendix, Fig. S5A), indicating that H137–211 and the histone core are not in direct contact. We then tested whether the C-terminal tails of H2A and the N-terminal tails of H3 interact with H137–211, as they are close to the linker DNA (Fig. 2A). The peak intensities/

volumes of the C-terminal residues (119–122) of the 15N-labeled H2A decreased by about half in the 1H-15N heteronuclear single quantum correlation (HSQC) spectra on binding of H137–211 or the wild-type full-length H1 1–256, with little changes in their chemical shifts (Fig. 2 B–E), whereas those in 15N-labeled H3 and the N-terminal tails of 15N-labeled H2A remain unchanged in both peak intensities and chemical shifts (Fig. 2 B–E and SI Appendix, Fig. S5 B and C). Similar results were observed for H137–211 in complex with the nucleosome array, which has a longer linker DNA (SI Appendix, Fig. S6). These results indicate that one of the two disordered H2A C-terminal tails in the nucleosome and two short regions in the disordered C-terminal tails of the wild-type full-length H11–256 became folded upon binding of H1 to the nucleosome. Determination of the Orientation of gH1 on the Nucleosome by PRE.

We next performed paramagnetic relaxation enhancement (PRE) experiments to determine the orientation of gH1 in the H137–211nucleosome complex. In the PRE experiments, residues immediately outside the two ends of gH1 (P44 and A119) and on the surface (L60, A83, and K109) were chosen to minimize the perturbation to the gH1 structure (Fig. 1B). These residues were each individually mutated to Cys and linked to the paramagnetic compound [(S-(2,2,5,5-tetramethyl-2,5-dihydro-1Hpyrrol-3-yl) methyl methanethiosulfonate) (MTSL) or ((S-methanethiosulfonylcysteaminyl)ethylenediamine-N,N,N′,N′-tetraacetic acid) (MTS-EDTA-Mn2+)] through a disulfide bond. Binding of paramagnetic spin-labeled H137–211 reduced the peak intensities of the observable backbone amide 1H-15N signals in the nucleosome in a manner dependent on distance from the paramagnetic center (42). The peak intensities of residues in the C-terminal tails of H2A were only weakly affected by MTSL but were strongly affected by MTS-EDTA-Mn2+ (Fig. 3 A–D and SI Appendix, Fig. S7), consistent with the anticipated PRE effects from the two types of spin labels (42). The intensity changes in the observable residue T119 in H2A, which is near the folded region of H2A, indicate it is closer to H137–211 residues A83 and K109 than to H137–211 residues L60, A119, and P44 (Fig. 3 A–D, Fig. 1B, and Fig. 2A). These results indicate that in the H137–211–nucleosome

Fig. 1. The globular domain and two discrete regions of linker histone H1 bind to the nucleosome. (A) Sequence of H1, highlighting the globular domain (cyan), the quadruple mutations (green and bold), and the C-terminal regions (red) that are involved in the binding of the nucleosome, and secondary structures. (B) The structural model of the globular domain, which is modeled using the B chain of the gH5 globular domain structure as the template (22). The mutated residues are shown as green sticks. The balls indicate the positions for spin labeling, except that positions 44 and 119 outside the gH1 are indicated with residues 45 and 118 in gH1, respectively. (C) Gel shift assay results for different fragments of H1. The nucleosome contains 167 bp DNA centered with the 601 sequence. (D) Overlay of 1H-15N HSQC spectra of H1 in free form (blue) and in complex with the nucleosome (black). In this experiment, the ratio of H137–211 to the nucleosome in the complex is ∼0.7. The disappearance of peaks indicates that many residues in H137–211 become folded on binding of the nucleosome. (E) Deviation of Cα chemical shifts of observable residues of H1 in complex with the nucleosome from random coil values.

2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1314905110

Zhou et al.

Fig. 2. Asymmetric binding of H137–211 to the nucleosome. (A) Nucleosome structure, highlighting residues H2A T119 and H3 K37. The DNA structure is taken from the tetra-nucleosome structure with 10 bp DNA at both entry and exit regions (41): H2A is in orange, H3 is in blue, and H4 is in green. (B) Overlay of 1H-15N HSQC spectra of the nucleosome with 15N-labeled H2A in the absence (black) or the presence (orange) of mutant H137–211. (C) Same as in B except the wild-type full-length H11–256 (magenta) was used instead of mutant H137–211. Note that only strong peaks in the spectra are shown (SI Appendix, Fig. S7). In this experiment, the ratio of H137–211 to the nucleosome is 1.0. The dashed lines indicate that the volumes of the peaks in the boxes are integrated together. (D) Ratios of peak volumes of T119, K121, and the region that include K118, E120, and K122 of the H2A tails in the nucleosome in complex with mutant H137–211 to those in the free nucleosome. (E) Same as in D except that the wild-type full-length H11–256 was used instead of mutant H137–211. The asterisks in B–E indicate that the peaks in the dashed box were integrated together when the peak volumes were measured.

complex, the α3 helix of gH1, which is within the region 83– 109, is closer to the nucleosome core than the gH1 N- and Cterminal border residues (P44 and A119) and loop 1 residue L60 (Fig. 1B). Determination of the Location of gH1 on the Nucleosome by PRE. We further used the methyl groups of Ile, Val, and Leu residues in gH1 as probes to determine the location of gH1 on the nucleosome. We first assigned most of the methyl groups in gH1 by combining three approaches (Fig. 3E and SI Appendix, Fig. S8): mutation of specific residues, comparison of the peak positions of the methyl groups of gH1 in the free form and in complex with Zhou et al.

Important H1 Residues for Nucleosome Binding. We used site-directed mutagenesis and ITC to identify gH1 residues that are important for the binding of H137–211 to the nucleosome. Positively charged Lys residues on the surface of gH1 were each mutated to Ala, and the effects of the mutations on the dissociation equilibrium constant (KD) were measured by ITC (Fig. 4A and SI Appendix, Table S1). Mutation of each of the six residues (K58, K91, K95, K102, K107, and K116) showed larger effects (> a factor of 2.5) than others. These mutations are located on two distinct surfaces on nearly opposite sides of the gH1 structure (Fig. 4B). One surface includes K91 and K95 in the α3 helix. The other surface includes residues K58 at the C-terminal region of the α1 helix and residues K102, K107, and K116 in the two β strands. In addition, ITC experiments showed that gH1 binding to the nucleosome was ∼12 times weaker than H137–211 (SI Appendix, Fig. S9). These results are consistent with the NMR observation that residues 119ASAKKEK125 and 164KKPKAKKAVAT174 in the middle of the C-terminal tail of H11–256 are folded. A Structural Model for the gH1–Nucleosome Complex. Taking all the data together, our experimental results suggest that gH1 uses the two positively charged surfaces to bridge the nucleosome core and the linker DNA asymmetrically. GH1 is close to one of the two H2A C-terminal tails and one of the two H3 N-terminal tails, with its α3 helix facing the nucleosome core. Using these restraints and the HADDOCK program (43), we docked the gH1 onto the nucleosome by forcing the K91 and K95 residues in the α3 helix of gH1 to interact with the nucleosomal DNA near the dyad region and the K58, K102, K107, and K116 residues to interact with one of the nearby 10-bp linker DNAs (SI Appendix, Fig. S10). After initial rigid body docking and subsequent energy minimization that allows contacting residues to adopt different conformations, the calculated low-energy structures were clustered. We found that one of the clusters was most consistent with all of the experimental results (Fig. 5 A and B and SI Appendix, Fig. S10D), including the mutation effects on the binding affinity between H137–211 and the nucleosome (Fig. 5C), PRE effects on H2A T119 in the nucleosome by spin labels in gH1 (Fig. 5D), PNAS Early Edition | 3 of 6

BIOPHYSICS AND COMPUTATIONAL BIOLOGY

the nucleosome, and spin-labeling effects on the methyl groups in gH1 in the absence or presence of the nucleosome. Using the assigned methyl groups, we examined the PRE effects on the methyl groups of gH1 with MTSL labeled at position T119 of H2A or at position K37 of H3 in the nucleosome (Fig. 2A). The residues at these positions are disordered in the nucleosome but are close to the nucleosome core, with relatively fixed locations. The spin labels have little effects on the chemical shifts of the methyl groups in gH1, indicating that they do not perturb the structure of the gH1–nucleosome complex (Fig. 3 E–G). The spinlabel at position 119 of H2A has large PRE effects on the methyl groups of L97, V98, and V99 at the C-terminal region of the α3 helix of gH1, whereas the spin-label at position 37 of H3 strongly affected not only these methyl groups but also L60 in loop 1 and L103 and I104 in the C-terminal region of the β1 strand (Fig. 3 F–I). Because there are two H3 K37 residues in the nucleosome structure that are separated by ∼65 Å (Fig. 2A), the observation that methyl groups in one region of gH1 are strongly affected by the MTSL labels at the H3 K37 sites further indicates that gH1 binds to the nucleosome asymmetrically. This is also consistent with the folding of only one of the two H2A C-terminal tails in the nucleosome on H137–211 binding (Fig. 2 B and C). Furthermore, as MTSL spin labels in gH1 have only small effects on the backbone amide NMR signals from the disordered H2A T119 (Fig. 3A), gH1 must be distant from this disordered H2A Cterminal tail in the H137–211–nucleosome complex. Therefore, the large PRE effects on some methyl groups of gH1 from MTSL labels at H2A T119 (Fig. 3H) indicate that gH1 is close to the folded H2A C-terminal tail in the H1–nucleosome complex.

Fig. 3. Location and orientation of gH1 on the nucleosome determined by PRE effects. (A–D) PRE effects of spin labeling of H137–211 at positions 44, 60, 83, 109, and 119 on the amide protons of H2A C-terminal tail residues T119 and K122 and N-terminal R3 and K5 residues in the nucleosome. (E–G) Methyl-spectra of gH1 in complex with the nucleosome (red) and the nucleosomes with MTSL labels at H2A T119 (orange) and H3 K37 (blue), respectively. (H and I) Bar graphs showing PRE effects on the methyl groups of gH1 with MTSL label at H2A T119 (orange) and H3 K37 (blue), respectively.

and PRE effects on the methyl groups of gH1 by spin labels at H2A T119 (Fig. 5E) and H3 K37 (Fig. 5F). In the model, although gH1 directly interacts with only one linker DNA, it is very close to the other linker DNA (Fig. 5B and SI Appendix, Fig. S10E). A slight shift of this linker DNA toward the nucleosome core could also lead to its interaction with gH1. Therefore, it is possible that gH1 may interact with both DNA linkers, one strongly and the other weakly (SI Appendix, Fig. S6A).

(29). However, the orientation of gH5 in their model is opposite to that of gH1 in ours: The α3 helix of gH5 in their model binds to the linker DNA instead of the nucleosomal DNA, as in our model (SI Appendix, Fig. S12 B and C). One possibility for the differences among various models of the gH1/H5–nucleosome complex might be that there are multiple gH1/H5 binding modes for the nucleosome. Another possibility is that earlier models are

The C-Terminal Tail of Histone Variant H2A.Z Disfavors H1 Binding.

Earlier studies have shown that the nucleosome containing histone variant H2A.Z affects H1 binding (44), and the C-terminal tail of H2A.Z may play a role in the reduced binding affinity (45). The amino acid sequence of the very C-terminal region (119TEKKA123) of H2A is different from corresponding residues (123EETVQ127) of H2A.Z. To test the role of the C-terminal tail of H2A.Z in H1 binding, we measured the binding affinity of H137–211 to the nucleosome containing H2A with the last six residues replaced by the corresponding residues of H2A.Z. We found that the binding was reduced to an undetectable level by ITC (SI Appendix, Fig. S11), whereas deletion of the last five residues of H2A increased KD by a factor of ∼10. These results further confirm the earlier NMR observation that one of the C-terminal tails of H2A is involved in the formation of the H1–nucleosome complex. Discussion Our structural model of the gH1–nucleosome complex differs from all previous experiment-based models in either the location or orientation of gH1/H5 or the length of linker DNA involved. The observation that MTSL labels at H2A T119 or H3 K37 have large PRE effects on the methyl groups in only one region of gH1 clearly shows that the gH1 binds to the nucleosome asymmetrically. For a symmetric binding, large PRE effects on methyl groups of gH1 from spin labels at H3 K37 are not expected because they are separated by more than 20 Å (SI Appendix, Fig. S12A). It is interesting to note that Zhou and coworkers have also found that gH5 binds to the nucleosome complex asymmetrically, involving 10-bp linker DNA (SI Appendix, Fig. S1) 4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1314905110

Fig. 4. Effects of mutations in gH1 on the binding affinity of H137–211 to the nucleosome. (A) Effects of mutation of surface residues in gH1 on the binding affinity between H137–211 and the nucleosome. (B) Structural illustration of the distribution of the gH1 residues whose Ala mutations lead to a large decrease in binding affinity.

Zhou et al.

derived from less-controlled systems, using indirect methods (SI Appendix, Fig. S1). Among several computation-based models of the gH1/H5– nucleosome complex (30–32, 46), two of them show similarities to our model in terms of both location and orientation of gH1 in the complex (30, 31). In addition, our model is consistent with several other experimental observations. For example, the two C-terminal tails of H2A in the Xenopus nucleosome cross-link to different positions of the DNA in the nucleosome on addition of H1 (47). Calf H1 is cross-linked to the C-terminal region of H2A (48). The location of gH1 in our model is very close to the disordered negatively charged C-terminal tail of HMGN2 in complex with the nucleosome, consistent with its role in inhibiting the binding of H1 to the nucleosome (38). Binding of H1 to the 10-bp linker DNA that neighbors the nucleosome core particle has been used to explain the preferential “out-of-phase” population of AT bases in the region (31, 49). The identification of two major discrete DNA-binding surfaces of gH1 in our mutation studies is in excellent agreement with earlier mutation studies of the histone H5 globular domain (50) and Mouse H10, using fluorescence recovery after photo-bleaching methods (19) (SI Appendix, Fig. S13). In addition, within the limit of the resolution, the tetra-nucleosome crystal structure can accommodate gH1 in our model (41) with only small clashes with the linker DNA that connects the neighboring nucleosomes (SI Appendix, Fig. S14), suggesting that a small structural rearrangement of the linker DNA could allow H1 to condense chromatin to the twostart zigzag nucleosome high-order structure (7, 51). Our finding that two discrete regions in the C-terminal tail of H1, 119ASAKKEK125 and 164KKPKAKKAVAT174, are involved in nucleosome binding is consistent with the earlier experimental observations that the C-terminal tail of H1 plays an important role in nucleosome binding and chromatin structure condensation (18, 20, 21). In particular, our results support the conclusions that specific regions in the C-terminal tail of H1, rather than the distribution of positively charged residues, are responsible for interaction with linker DNA (17, 18, 52, 53). In the Zhou et al.

case of human H1.5, it has been shown that the seven residues following the globular domain, 121PKAKKAG127, are necessary for the formation of the linker DNA stem structure (17). Six of these seven residues are conserved in the 164KKPKAKKAVAT174 region of the C-terminal tail of Drosophila H1. Our results show that NMR provides a powerful tool for identifying specific nucleosome-binding regions in the C-terminal tail of H1. It has been shown that the H2A.Z-containing nucleosome is typically present at regions flanking the nucleosome-free region (54), near the DNA double-strand break (55), and at the boundary of euchromatin and heterochromatin, which prevents spread of heterochromatin regions (56). In general, these chromatin regions are less condensed and more dynamic. Our finding that the very last several residues at the H2A.Z C-terminal tail inhibit the binding of H1 to the H2A.Z nucleosome provides a possible mechanistic explanation for the dynamic features of the H2A.Z-enriched chromatin regions. Finally, our approach provides an experimental tool for testing the hypothesis that the globular domains of individual H1 variants might have distinct binding geometries within the nucleosome that contribute to the heterogeneity of chromatin structure (9, 19, 57). Materials and Methods Design of a Stable Mutant of gH1 and Preparation of Samples. The gH1 structure model was built using homology modeling and the structure of gH5. Protein design programs were used to select the mutation (V53I, S56A, C81V, and A97L) that stabilizes gH1 structure. The Drosophila H1 gene was synthesized and inserted into the pET42b vector. Mutations were made using the QuikChange kit. Proteins were expressed in Escherichia coli and purified using chromatography. MTSL or MTS-EDTA was linked to the protein, with the target site mutated to Cys. Nucleosomes were reconstituted by stepwise salt dialysis in the absence of reducing agent, followed by HPLC to remove free DNA and immature nucleosomes (SI Appendix, Materials and Methods and SI Appendix, Fig. S15). NMR, PRE, and ITC Experiments. Isotope-labeled H1 mixed with the nucleosome, or vice versa. The spin-labeled H1 was mixed with nucleosomes containing 15Nlabeled histones or vice versa. 1H-15N HSQC or 1H-13C heteronuclear multiple quantum correlation (HMQC) spectra were recorded. The NMR peak

PNAS Early Edition | 5 of 6

BIOPHYSICS AND COMPUTATIONAL BIOLOGY

Fig. 5. An asymmetrical structural model of the gH1-nucleosome complex. (A and B) Structural model of the gH1-nucleosome complex in surface and ribbon representations. The rectangular frame approximately shows the enlarged regions in C–F. (C) Illustration of the Lys residues that show a large increase in KD (> a factor of 2.5) when mutated to Ala. (D) Illustration of the disordered residue H2A_T119 and residues P44 (H45), L60, A83, K109, and A119 (S118) in gH1 in the gH1-nucleosome complex. (E) Illustration of the ordered residue H2A_T119 and the methyl groups in gH1. The methyl groups in red have an Ipara/Idia [Intensity (paramagnetic)/Intensity (diamagnetic)] of less than 0.3, whereas other methyl groups in cyan have larger values. (F) Illustration of one of the two H3_K37 residues and the methyl groups in gH1. The methyl groups in red have Ipara/Idia of less than 0.3, whereas methyl groups in cyan have larger values.

intensities or volumes were measured. ITC experiments were performed on a VP–ITC microcalorimeter (Microcal). The dissociation constant (KD ) and the stoichiometry of binding (n) were determined by fitting the observed binding curves to the model with independent n and KD values (SI Appendix, Materials and Methods).

dyad; the K58, K102, K107, and K106 residues were forced to interact with the nearby linker DNA. A cluster of structures that is most consistent with the experimental data were selected as the final model (SI Appendix, Materials and Methods).

Docking Calculation. We docked gH1 to the 167-bp DNA obtained from the tetra-nucleosome structure using HADDOCK program. Residues Lys91 and Lys95 in gH1 were forced to interact with the nucleosomal DNA near the

ACKNOWLEDGMENTS. We thank Dr. Jemima Barrowman for editing the manuscript. This work is supported by the intramural research program of National Cancer Institute and National Institutes of Health (NIH) (to Y.B.) and by the National Institute of General Medical Sciences of the NIH (R01GM085003, to Y.Z.).

1. Kornberg RD (1974) Chromatin structure: A repeating unit of histones and DNA. Science 184(4139):868–871. 2. Kornberg RD, Thomas JO (1974) Chromatin structure; oligomers of the histones. Science 184(4139):865–868. 3. Khorasanizadeh S (2004) The nucleosome: From genomic organization to genomic regulation. Cell 116(2):259–272. 4. Luger K, Mäder AW, Richmond RK, Sargent DF, Richmond TJ (1997) Crystal structure of the nucleosome core particle at 2.8 A resolution. Nature 389(6648):251–260. 5. Felsenfeld G, Groudine M (2003) Controlling the double helix. Nature 421(6921): 448–453. 6. Thoma F, Koller T (1977) Influence of histone H1 on chromatin structure. Cell 12(1): 101–107. 7. Robinson PJ, Rhodes D (2006) Structure of the ‘30 nm’ chromatin fibre: A key role for the linker histone. Curr Opin Struct Biol 16(3):336–343. 8. Maresca TJ, Freedman BS, Heald R (2005) Histone H1 is essential for mitotic chromosome architecture and segregation in Xenopus laevis egg extracts. J Cell Biol 169(6):859–869. 9. Lee H, Habas R, Abate-Shen C (2004) MSX1 cooperates with histone H1b for inhibition of transcription and myogenesis. Science 304(5677):1675–1678. 10. Fan Y, et al. (2005) Histone H1 depletion in mammals alters global chromatin structure but causes specific changes in gene regulation. Cell 123(7):1199–1212. 11. Shen X, Gorovsky MA (1996) Linker histone H1 regulates specific gene expression but not global transcription in vivo. Cell 86(3):475–483. 12. Thoma F, Koller T, Klug A (1979) Involvement of histone H1 in the organization of the nucleosome and of the salt-dependent superstructures of chromatin. J Cell Biol 83(2 Pt 1):403–427. 13. Singer DS, Singer MF (1976) Studies on the interaction of H1 histone with superhelical DNA: Characterization of the recognition and binding regions of H1 histones. Nucleic Acids Res 3(10):2531–2547. 14. Simpson RT (1978) Structure of the chromatosome, a chromatin particle containing 160 base pairs of DNA and all the histones. Biochemistry 17(25):5524–5531. 15. Noll M, Kornberg RD (1977) Action of micrococcal nuclease on chromatin and the location of histone H1. J Mol Biol 109(3):393–404. 16. Allan J, Hartman PG, Crane-Robinson C, Aviles FX (1980) The structure of histone H1 and its location in chromatin. Nature 288(5792):675–679. 17. Syed SH, et al. (2010) Single-base resolution mapping of H1-nucleosome interactions and 3D organization of the nucleosome. Proc Natl Acad Sci USA 107(21):9620–9625. 18. Hendzel MJ, Lever MA, Crawford E, Th’ng JP (2004) The C-terminal domain is the primary determinant of histone H1 binding to chromatin in vivo. J Biol Chem 279(19): 20028–20034. 19. Brown DT, Izard T, Misteli T (2006) Mapping the interaction surface of linker histone H1(0) with the nucleosome of native chromatin in vivo. Nat Struct Mol Biol 13(3): 250–255. 20. Hamiche A, Schultz P, Ramakrishnan V, Oudet P, Prunell A (1996) Linker histonedependent DNA structure in linear mononucleosomes. J Mol Biol 257(1):30–42. 21. Bednar J, et al. (1998) Nucleosomes, linker DNA, and linker histone form a unique structural motif that directs the higher-order folding and compaction of chromatin. Proc Natl Acad Sci USA 95(24):14173–14178. 22. Ramakrishnan V, Finch JT, Graziano V, Lee PL, Sweet RM (1993) Crystal structure of globular domain of histone H5 and its implications for nucleosome binding. Nature 362(6417):219–223. 23. Ali T, Coles P, Stevens TJ, Stott K, Thomas JO (2004) Two homologous domains of similar structure but different stability in the yeast linker histone, Hho1p. J Mol Biol 338(1):139–148. 24. Thomas JO (1999) Histone H1: Location and role. Curr Opin Cell Biol 11(3):312–317. 25. Vignali M, Workman JL (1998) Location and function of linker histones. Nat Struct Biol 5(12):1025–1028. 26. Crane-Robinson C (1997) Where is the globular domain of linker histone located on the nucleosome? Trends Biochem Sci 22(3):75–77. 27. Fan L, Roberts VA (2006) Complex of linker histone H5 with the nucleosome and its implications for chromatin packing. Proc Natl Acad Sci USA 103(22):8384–8389. 28. Meyer S, et al. (2011) From crystal and NMR structures, footprints and cryo-electronmicrographs to large and soft structures: Nanoscale modeling of the nucleosomal stem. Nucleic Acids Res 39(21):9139–9154. 29. Zhou YB, Gerchman SE, Ramakrishnan V, Travers A, Muyldermans S (1998) Position and orientation of the globular domain of linker histone H5 on the nucleosome. Nature 395(6700):402–405.

30. Pachov GV, Gabdoulline RR, Wade RC (2011) On the structure and dynamics of the complex of the nucleosome and the linker histone. Nucleic Acids Res 39(12): 5255–5263. 31. Cui F, Zhurkin VB (2009) Distinctive sequence patterns in metazoan and yeast nucleosomes: Implications for linker histone binding to AT-rich and methylated DNA. Nucleic Acids Res 37(9):2818–2829. 32. Bharath MM, Chandra NR, Rao MR (2003) Molecular modeling of the chromatosome particle. Nucleic Acids Res 31(14):4264–4274. 33. Wong J, Li Q, Levi BZ, Shi YB, Wolffe AP (1997) Structural and functional features of a specific nucleosome containing a recognition element for the thyroid hormone receptor. EMBO J 16(23):7130–7145. 34. An W, Leuba SH, van Holde K, Zlatanova J (1998) Linker histone protects linker DNA on only one side of the core particle and in a sequence-dependent manner. Proc Natl Acad Sci USA 95(7):3396–3401. 35. Pruss D, et al. (1996) An asymmetric model for the nucleosome: A binding site for linker histones inside the DNA gyres. Science 274(5287):614–617. 36. Bax A (2011) Triple resonance three-dimensional protein NMR: Before it became a black box. J Magn Reson 213(2):442–445. 37. Tugarinov V, Hwang PM, Kay LE (2004) Nuclear magnetic resonance spectroscopy of high-molecular-weight proteins. Annu Rev Biochem 73:107–146. 38. Kato H, et al. (2011) Architecture of the high mobility group nucleosomal protein 2-nucleosome complex as revealed by methyl-based NMR. Proc Natl Acad Sci USA 108(30):12283–12288. 39. Zhou BR, et al. (2012) Histone H4 K16Q mutation, an acetylation mimic, causes structural disorder of its N-terminal basic patch in the nucleosome. J Mol Biol 421(1): 30–37. 40. Thåström A, Bingham LM, Widom J (2004) Nucleosomal locations of dominant DNA sequence motifs for histone-DNA interactions and nucleosome positioning. J Mol Biol 338(4):695–709. 41. Clore GM, Tang C, Iwahara J (2007) Elucidating transient macromolecular interactions using paramagnetic relaxation enhancement. Curr Opin Struct Biol 17(5):603–616. 42. Dominguez C, Boelens R, Bonvin AM (2003) HADDOCK: A protein-protein docking approach based on biochemical or biophysical information. J Am Chem Soc 125(7): 1731–1737. 43. Thakar A, et al. (2009) H2A.Z and H3.3 histone variants affect nucleosome structure: Biochemical and biophysical studies. Biochemistry 48(46):10852–10857. 44. Vogler C, et al. (2010) Histone H2A C-terminus regulates chromatin dynamics, remodeling, and histone H1 binding. PLoS Genet 6(12):e1001234. 45. Wong H, Victor JM, Mozziconacci J (2007) An all-atom model of the chromatin fiber containing linker histones reveals a versatile structure tuned by the nucleosomal repeat length. PLoS ONE 2(9):e877. 46. Lee KM, Hayes JJ (1998) Linker DNA and H1-dependent reorganization of histoneDNA interactions within the nucleosome. Biochemistry 37(24):8622–8628. 47. Boulikas T, Wiseman JM, Garrard WT (1980) Points of contact between histone H1 and the histone octamer. Proc Natl Acad Sci USA 77(1):127–131. 48. Travers AA, Muyldermans SV (1996) A DNA sequence for positioning chromatosomes. J Mol Biol 257(3):486–491. 49. Goytisolo FA, et al. (1996) Identification of two DNA-binding sites on the globular domain of histone H5. EMBO J 15(13):3421–3429. 50. Schalch T, Duda S, Sargent DF, Richmond TJ (2005) X-ray structure of a tetranucleosome and its implications for the chromatin fibre. Nature 436(7047):138–141. 51. Dorigo B, et al. (2004) Nucleosome arrays reveal the two-start organization of the chromatin fiber. Science 306(5701):1571–1573. 52. Lu X, Hansen JC (2004) Identification of specific functional subdomains within the linker histone H10 C-terminal domain. J Biol Chem 279(10):8701–8707. 53. Fang H, Clark DJ, Hayes JJ (2012) DNA and nucleosomes direct distinct folding of a linker histone H1 C-terminal domain. Nucleic Acids Res 40(4):1475–1484. 54. Cairns BR (2009) The logic of chromatin architecture and remodelling at promoters. Nature 461(7261):193–198. 55. Morrison AJ, Shen X (2009) Chromatin remodelling beyond transcription: The INO80 and SWR1 complexes. Nat Rev Mol Cell Biol 10(6):373–384. 56. Zofall M, et al. (2009) Histone H2A.Z cooperates with RNAi and heterochromatin factors to suppress antisense RNAs. Nature 461(7262):419–422. 57. Alami R, et al. (2003) Mammalian linker-histone subtypes differentially affect gene expression in vivo. Proc Natl Acad Sci USA 100(10):5920–5925.

6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1314905110

Zhou et al.

SI Appendix Structural insights into the histone H1-nucleosome complex Bing-Rui Zhou, Hanqiao Feng, Hidenori Kato, Liang Dai, Yuedong Yang, Yaoqi Zhou, Yawen Bai

 

1  

SI MATERIALS AND METHODS Computational design of a stable mutant of H1 In order to observe the methyl groups in the H1-nucleosome complex with reasonable sensitivity by NMR, a temperature of 35oC or higher is needed. At such a high temperature, Drosophila gH1 and the H1-nucleosome complex are unstable. To design a stable mutant of H1 without perturbing its structure and surface charge distribution, we first used SWISS-MODEL (http://swissmodel.expasy.org/) to build a structural model of gH1 using the crystal structure of Chicken H5 globular domain as the template. We then substituted each residue in the hydrophobic core of gH1 with every one of the other 19 amino acids and evaluated the stability of each mutant while the rest were fixed using three protein design programs (RosettaDesign (1), EGAD (2), and RosettaDesignSR (3)). The consensus mutations made by all three programs were further examined by position-specific mutation profiles (PSSM) from PSI-Blast (4), sequence-derived structure entropy (SDSE (5) and structurebased prediction of stability change (DMUTANT (6)). A combination of four mutations (V53I, S56A, C81V, and A97L) leads to a higher probability in PSSM than the wild type, lower sequence-derived structural entropy, and more stable than or nearly equivalent to wild type in protein stability according to DMUTANT. Protein and DNA purification The Drosophila H1 gene was synthesized (Bio Basic Inc., Canada) and inserted into the pET42b vector using Nde I and Bam HI cloning sites. A hex-histidine tag was added at the N-terminus. Mutations of H1 were made using the QuikChange kit and verified by DNA sequencing. H1 and its mutants (including gH1 and H1 fragments) were expressed in Escherichia coli BL21(DE3) RIPL cells and purified using Ni-NTA chromatography, followed by RP HPLC. DNA and histones were obtained as described in our early studies (7). Nucleosome and H1-nucleosome reconstitution Nucleosomes were reconstituted and purified according to protocols described in our earlier studies (7). For nucleosome labeled with MTSL (Toronto Research Chemicals), histone octomer with H3K37C or H2AT119C (H3 Cys110 was mutated to Ala in all cases) were refolded in 10 mM Tris-HCl, 2M NaCl, 1 mM EDTA and 5 mM β-Mecaptoethanol (pH 7.4). The histone octamer was purified by gel filtration using the same buffer but in the absence of β-Mecaptoethanol. The concentration of the octamer was determined with a UV spectrometer. A 10fold excess of MTSL was added to the octamer and the reaction was allowed to continue overnight in the dark at 4oC. Nucleosomes were reconstituted by stepwise salt dialysis in the absence of reducing agent, followed by HPLC to remove free DNA and immature nucleosomes. H1 and nucleosome were mixed at a ratio of 0.75 at a concentration below 10 µM in H1 chemical shift perturbation studies but at a ratio 1 for histone core (methyl-labeled), histone tails (15Nlabeled), and PRE experiments.  

2  

NMR and PRE experiments In order to obtain PRE-derived distance restraints between histone H1 and the disordered H2A C-terminus, each of the selected position (Pro44, Leu60, Ala83, Lys109, and Ala119) in H1 were mutated to Cys. The lyophilized protein powder of H1 bearing those single mutations was dissolved in 20 mM Tris-HCl and 10 mM DTT (pH 8.0) and incubated at room temperature for 2 hours. The sample was then passed trough a G-25 Sephadex column equilibrated with 50 mM sodium phosphate (pH 7.8) and 0.1 M NaCl. A 10-fold excess of spin label (MTSL, MTS-EDTA-Ca2+ or MTS-EDTA-Mn2+, Toronto Research Chemicals.) was immediately added to the H1 samples. The ligation of spin label to the cysteine occurred overnight in the dark at room temperature. The completion of the reaction was confirmed by mass spectrometry and analytical reverse phase HPLC (Fig. S15). Free spin labels were removed by buffer exchange on a G-25 Sephadex column equilibrated with NMR buffer (20 mM sodium phosphate, pH 6.0, 10% D2O). The spin labeled H1 was then mixed with 15N-labeled H2A nucleosomes at 1:1 ratio in NMR buffer. 1H-15N HSQC spectra of the nucleosome were recorded. For MTSL labeled H1-nucleosome complex, after reduction of the MTSL using a 5-fold ascorbic acid, a second set of HSQC spectra was recorded. To obtain PRE-derived distance restraints from the nucleosome to the globular domain of H1, H2A Thr119 or H3 Lys37 was mutated to Cys (the H3 Cys110 was mutated to Ala simultaneously) and linked to MTSL as described above. The completion of the labeling was confirmed by mass spectrometry. Nucleosomes bearing the MTSL spin labels were reconstituted in the buffer without reducing agent. Leu, Val, and Ile methyl-labeled H1 was mixed with spin labeled nucleosomes at a ratio of 0.75 in the NMR buffer (20 mM Tris-D11-DCl, pD 7.4). 1 H-13C HMQC spectra of the H1-nucleosome complex were recorded before or after the reduction of the MTSL. NMR spectra were collected on a Bruker 700 MHz NMR machine equipped with a cyro-probe. The concentrations of the NMR samples are ~100-150 µM. The data was analyzed using NMRPipe. The peak intensities of the well-separated peaks or the volume of the merged peaks were measured using NMRViewJ (One Moon Scientific Inc.) or UCSF Sparky software. ITC experiments All ITC experiments were performed on a VP-­‐ITC microcalorimeter (Microcal) at 25 °C. H1 and nucleosome samples were extensively dialyzed against the ITC buffer (20 mM Tris-HCl, pH7.4, 1 mM EDTA and 1mM DTT) and degassed before use. In a typical titration experiment, 10 µM of nucleosome containing 167 bp “601” DNA was titrated with 100 µM H1 in the ITC buffer. The ITC data was analyzed using Origin 7.0 software (Microcal). Binding curves were generated by plotting the heat change of the binding reaction against the ratio of the total concentration of H1 to the total concentration of the nucleosomes. The dissociation constant (KD) and the stoichiometry of binding (n) were determined by fitting the observed binding curves to the model with independent n and KD  

3  

values. Assignment of ILV methyl groups in gH1 in complex with the nucleosome The methyl groups of gH1 in complex with the nucleosome were assigned using combination of three approaches. Firstly, we used specific mutations to assign the methyl groups of V98, L103, L94, and L60 using single mutations (V98I, L103I, L94V, and L60V) and methyl groups of L69 and I71 using a double mutation (L69V/I71V) (Fig. S8A-D), which are based on the observation that mutations lead to the disappearance of the methyl peaks from their positions in the 1H-13C-TROSY spectra of the WT protein. Due to peak position changes of neighbor residues upon L60V mutation, L60 cannot be assigned unambiguously by mutation alone. It is assigned later after other changed peaks (L69, L86, and L97) are assigned. In addition, the methyl peaks arising from a natural abundance of non-labeled histones in the nucleosome were identified by comparing 13CH3 and 13CHD2 labeled Ile, Leu and Val in gH1 in complex with the nucleosome, which show systematic shifts in peak positions. Those natural-abundance peaks do not show chemical shift changes in those two spectra (Fig. S8F). Secondly, we assigned the methyl groups of the free gH1 using backbone-based three-dimensional methods (Fig. S8G) and then overlaid the 1H-13C TROSY spectra of gH1 in its free form and in complex with the nucleosome to assign the methyl groups of V81 and I104 (Fig. S8H). With the assignment of V81 and V98, we are also able to assign V99 since only three Val residues exist in gH1. Thirdly, a MTSL label at the N-terminal position 44 allowed us to assign the methyl groups in residues L86 and I57 and to confirm the assignment of V81 methyl groups (Fig. S8I, J). These assignments are based on the structural model of gH1 in which L86 and V81 are the only Leu and Val residues that are close to the spin label site whereas I57 and I104 are the two Ile residues that are most remote from the spin-label site. Finally, L117 is assigned based on the fact that it is the last unassigned Leu residue. L60 and L97 peaks will be distinguished based on the spin label at H2A T119. H2A T119 is close to V98 and therefore L97 in the final model whereas L60 is far away. I53, I75, and I90 are not distinguishable. Docking calculation Guided by the results of the PRE and mutation-ITC analyses, we docked gH1 to the 167 bp DNA obtained from the tetra-nucleosome structure (pdb ID: 1ZBB) using HADDOCK version 2.1, installed on the National Institutes of Health Biowulf server. HADDOCK (High Ambiguity Driven biomolecular DOCKing) makes use of data resulting from NMR titration experiments, mutagenesis or bioinformatic predictions as restraints for molecular docking, which is introduced as Ambiguous Interaction Restraints (AIRs) to drive the docking process. An AIR is defined as an ambiguous distance between all residues shown to be involved in  

4  

the interaction (for details, see http://www.nmr.chem.uu.nl/haddock/). In our docking calculation, the six gH1 residues whose Ala mutations show large effects on binding affinity and the residues in the DNA region near the dyad and in the nearby linker DNA are chosen as active residues to derive AIRs for docking calculations. Lys91 and Lys95 were forced to interact with the nucleosomal DNA near the dyad while the K58, K102, K107 and K106 residues were forced to interact with the nearby linker DNA. There is no direct contact between gH1 and the nucleosome in the starting structures (Fig. S10C). A total of 1,000 structures were generated during the rigid-body energy minimization step (it0), and were subsequently subjected to simulated annealing (it1) during which residues of gH1 and DNA in contact were allowed to move freely. The resulting structures were sorted according to intermolecular energy (sum of the van der Waals, electrostatic, and AIRs energy), and the 200 structures with the lowest energies were selected and clustered using the standard protocol. The cluster that is most consistent with the experimental data was selected as the final model for the gH1/nucleosome complex.

 

5  

References 1. 2.

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

Rohl CA, Strauss CEM, Misura KMS, & Baker D (2004) Protein structure prediction using Rosetta. Methods Enzymol 383:66-93. Pokala N & Handel TM (2005) Energy functions for protein design: adjustment with protein-protein complex affinities, models for the unfolded state, and negative design of solubility and specificity. J Mol Biol 347:203227. Dai L, Yang Y, Kim HR, & Zhou Y (2010) Improving computational protein design by using structure-derived sequence profile. Proteins 78:2338 2348. Altschul SF, et al. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25:3389-3402. Chan CH, et al. (2004) Relationship between local structural entropy and protein thermostability. Proteins 57:684-691. Zhou HY & Zhou YQ (2002) Distance-scaled, finite ideal-gas reference state improves structure-derived potentials of mean force for structure selection and stability prediction. Protein Sci 11:2714-2726. Kato H, et al. (2011) Architecture of the high mobility group nucleosomal protein 2-nucleosome complex as revealed by methyl-based NMR. Proc Natl Acad Sci USA 108(30):12283-12288. Travers, A. (1999) The location of the linker histone on the nucleosome. Trends Biochem Sci 24(1): 4-7. Cui F & Zhurkin VB (2009) Distinctive sequence patterns in metazoan and yeast nucleosomes: implications for linker histone binding to AT-rich and methylated DNA. Nucleic Acids Res 37(9):2818-2829. Zhou BR, et al. (2012) Histone H4 K16Q mutation, an acetylation mimic, causes structural disorder of its N-terminal basic patch in the nucleosome. J Mol Biol 421(1):30-37. Goytisolo FA, et al. (1996) Identification of two DNA-binding sites on the globular domain of histone H5. EMBO J 15(13):3421-3429. Pachov GV, Gabdoulline RR, & Wade RC (2011) On the structure and dynamics of the complex of the nucleosome and the linker histone. Nucleic Acids Res 39(12):5255-5263. Zhou YB, Gerchman SE, Ramakrishnan V, Travers A, & Muyldermans S (1998) Position and orientation of the globular domain of linker histone H5 on the nucleosome. Nature 395(6700):402-405. Brown DT, Izard T, & Misteli T (2006) Mapping the interaction surface of linker histone H1(0) with the nucleosome of native chromatin in vivo. Nat Struct & Mol Biol 13(3):250-255. Stasevich TJ, Mueller F, Brown DT, & McNally JG (2010) Dissecting the binding mechanism of the linker histone in live cells: an integrated FRAP analysis. EMBO J. 29(7):1225-1234.

 

 

6  

Fig. S1. Summary of major experiment-based models of the gH1/H5 complex in the literature. The diagrams are modified from references (8, 9).

 

7  

A

B

Fig. S2. Design of stable mutant of gH1. (A) Sequence alignment of the globular domains of Chicken H5 and Drosophila H1 and its mutant. (B) Circular dichroism melting curves of Drosophila gH1 (blue) and its quadruple mutant (red).

 

8  

A

B

C

Fig. S3. ITC results of H1 binding to the nucleosome with 167 bp DNA. (A) The fullwt length H11-256. (B) The full-length quadruple mutant H11-256. (C) The truncated quadruple mutant H137-211. These H1 proteins have essentially the same binding affinity to the nucleosome, 0.29 µM and 0.28 µM and 0.21 µM, respectively (See Table S1).

 

9  

B A

C

D

Fig. S4. 1H-15N HSQC spectra of H1. (A), Free H137-256. (B) H137-211 in complex with the nucleosome with 167 bp of DNA. (C) Overlay of 1H-15N HSQC spectra of H137-211 (red) and fulllength H11-256 (black) in complex with the nucleosome. The extra number of peaks in black is 70, which is close to 73 expected from H1 regions 1-36 and 212-256 (excluding 7 prolines), indicating that these two regions are disordered in the H11-256-nucleosome complex. Note that peaks in the overlapped regions are counted based on 3D experiments (HNCO and HNCA) using 15N- and 13C-labeled samples. (D) Overlay of H137-211 in complex with ‘601’-centered 167 bp (black) and 208 bp DNA nucleosome (red). No additional peaks of H137-211 become undetectable with the increasing linker DNA length, suggesting that no additional stable binding of the histone tail to the extended DNA region. The shifts of some peaks are likely due to transient interactions between the extended linker DNA and the H1 tails.

 

10  

A

B

C

Fig. S5. The N-terminal tails of H3 and the core histones in the nucleosome do not interact with H137-211. (A) Overlay of methyl-TROSY NMR spectra of the core histones (black) in the absence of H137-211 and in the presence of H137-211 for H2A (orange), H2B (red), H3 (blue) and H4 (green). (B) 1H-15N HSQC spectra of 15Nlabeled H3 in the nucleosome (black) alone and in complex with H137-211 (blue). (C) Bar graph showing the ratio of peak intensities of the corresponding residues of H3 in the presence and absence of H137-211.

 

11  

A

B

C

Fig. S6. H137-211 preferentially binds to the nucleosome asymmetrically. (A) MNase digestion of 167 bp DNA nucleosome in the presence or absence of H1371 H-15N spectra of 211, showing preferential protection of 10 bp DNA. (B) nucleosome array (10 x 208 bp) with 15N-labeled H2A in the absence (black) and presence (orange) of wild type full-length H11-256. We note that the disappearance of the peaks from residues 118-120 of H2A in the spectra is likely caused by the large size of the array (> 2 x 106 Dalton). (C) Peak intensity changes of the residues on H2A tail upon binding of H11-256 to the nucleosome.

 

12  

Fig. S7. Illustration of PRE effects on T119 and K121 of H2A from spin-labels at the five sites: 44, 60, 83, 109 and 119 in H137-211. Full assignment of the peaks in the spectra can be found in reference (10).

 

13  

 

14  

Fig. S8. Assignment of the methyl groups in gH1 in complex with the nucleosome using H137-132. (A) V98a is assigned with V98I mutation since only one peak disappeared in the Val region. (B and C) Assignment of L103 and L94 using L103I and L94V mutations. Note that L103 and L94 are neighbors in the model structure of gH1. (D) L60 is assigned after other shifted peaks caused by L60V mutation are assigned. (E) Assignment of L69 and I71 using double mutation L69V/I71V. (F) Identification of natural abundance peaks from histones, shown with asterisks. (G) Assignment of methyl groups in the free gH1. (H), Overlay of the methyl spectra of free gH1 and gH1 in complex with the nucleosome, which allows assignment of V81 and I104 methyl groups, and V99 since only three Val residues exist in gH1. (I and J) Overlap of spectra of reduced (blue) and oxidized (red) of gH1 in free form and in complex with the nucleosome with spin-label at position 44. (K) Cα atom of N-terminal residue 45 and methyl groups in the gH1 structural model.

 

15  

A

B Time (min)

0,9

0

20

40

60

80

0,8

µcal/sec

0,7 0,6 0,1

0,0

kcal/mole of injectant

12

10

8 0,0

0,5

1,0

1,5

2,0

Molar Ratio

Fig. S9. ITC results of gH1 binding to the nucleosome with 167 bp DNA and 167 bp DNA alone. (A) gH1 binding to the nucleosome. Molar ratio is gH1 over the nucleosome. The Kd of 1.3 µM indicates that the binding affinity of gH1 is ~12 time weaker than that of H137-211. The enthalpy change associated with the formation of the complex is reversed when compared with that of H137-211 binding. (B) gH1 binding to DNA. Molar ratio is gH1 to free DNA. These results indicate that binding of H137-211 C-terminal tail leads to large negative enthalpy, whereas binding of gH1 leads to large positive change of enthalpy. We note that the large heats of the posttransition baselines come from non-specific binding of gH1 to the nucleosome after the specific binding site is occupied.

 

16  

A

C

B

D

E

F

Fig. S10. Two conformations of the gH5 crystal structure and modeling of the gH1-nucleosome complex. (A and B) The two conformations are different in the loop region between the two beta-strands. In addition, the side chain of K85 and R42 orient differently and may be classified into two different charged surfaces as indicated by the dashed circles. The residues in blue spheres are shown to be important for DNA binding by Goytisolo et al. (11). Earlier models of H1/H5 in complex with the nucleosome are built using the conformation in (A) with the exception of the model by Pachov et al. (12). Our spin-labeling data is more  

17  

consistent with the conformation in (B). Note that the amino acid sequence numbers of gH5 are different from the corresponding residues in Drosophila gH1. (C) Initial structures of gH1 and DNA for HADDOCK calculations. (D) 10 structures with lowest energy in the cluster. (E) The gH1-nucleosome complex, showing that the loop 2 residues (blue) in gH1 are close to the other linker DNA. (F), The shortest distance of the two atoms between gH1 and the linker DNA that are not in direct contact with gH1 is 5.1 Å.

 

18  

A

B

 

Fig. S11. The C-terminal region of histone variant H2A.Z disfavors H137-211 binding to the nucleosome. ITC profiles of H137-211 binding by the nucleosome with modified C-terminal residues of H2A. (A) The nucleosome includes the H2A with the last 6 residues deleted, which increased Kd by a factor of ~10 when compared with the WT H2A. (B) The nucleosome contains the H2A with the last 5 residues replaced by the corresponding residues of H2A.Z. Again the large heat values observed here are due to non-specific binding of H137-211 to nucleosomal DNA that occurs after the specific binding site is saturated.

 

19  

A

B

C

Fig. S12. Our gH1-nucleosome model is different from previous models of the gH1/H5 complexes. (A) In a symmetric model, the distance between any residues in gH1 and H3 K37 in the nucleosome is more than 20 Å, which is inconsistent with the observation that MTSL labels at H3 K37 have large PRE effects on only one region of gH1 (Fig. 3I). (B) The gH5-nucleosome model modified from Fig. 2b in Zhou et al. (13), in which the α3 helix of gH5 binds to the linker DNA. (C) Our model, in which the α3 helix of gH1 binds to the nucleosomal DNA. The three helices are colored in cyan (α1), purple (α2) and magenta (α3).

 

20  

Fig. S13. Correlation of mutation effects on in vitro nucleosome binding affinity of Drosophila H137-211 and in vivo chromatin association kinetics of Mouse H10 (14). (A) Correlation between the change in dissociation activation free energy (ΔΔG‡D) of Mouse H10 and the change in dissociation equilibrium binding free energy (ΔΔGD) of H137-211, where ΔΔG‡D = -RTln( t50%(mut)/t50%(wt)) and ΔΔGD = RTln(KD(mut)/KD(wt)). The numbers in the circles represent the mutation pairs in Drosophila H1 and Mouse H1o. 1. dWT/mWT; 2. dK73A/mK52A; 3. dK80A/mK59A; 4. dR63A/mR42A; 5. dH46A/mH25G; 6. dK61A/mK40A; 7. dK107A/mK85A; 8. dK116A/mR94A; 9. dK91A/mK69A; 10. dK95A/mK73A. (B) Free energy diagram illustration for the dissociation/association of H1 and the nucleosome. Drosophila H1 and Mouse H10 molecules have high sequence homology. The correlation would be expected if they bind to the nucleosome in a similar mode in vitro and in vivo, and implies that the binding mode is conserved and robust. The simplest interpretation of the correlation in (A), which is referred to as the Brønsted plot in chemical reaction and protein folding studies, is that the structure of the rate-limiting transition state of the association/dissociation process for the H1-nucleosome/chromatin complex is close to the dissociated state. Therefore, the destabilization of the H1-nucleosome/chromatin complex by the mutation in gH1 dominantly affects both the dissociation rate constant (1/t50) and equilibrium constant (KD). In this diagram, the effective association reaction should include the diffusion process and binding in the model that consists early diffusion and late binding (15).

 

21  

A

B

Fig. S14. Possible position of gH1 in the tetranucleoosme structure and dynamic binding of gH1 to each of the two linker DNA. (A) The gH1 in the model is largely compatible with the tertra-nucleosome structure. (B) Dynamic binding of H1 to the two linker DNAs.

 

22  

Fig. S15. Illustration of spin-label efficiency monitored by mass spectrometry and reverse phase HPLC. (A) MTS-EDTA-Mn2+ label at position A83 in H137-211. Upper panel, MALDI-TOF measurement of H137-211(A83C). Lower panel, MALDITOF measurement of H137-211 (A83C-MTS-EDTA-Mn2+). (B) MTSL label at position H3 K37. The reconstituted nucleosome is analyzed using reverse phase HPLC. The upper panel shows the profile of the nucleosome with H2A, H2B, H3(K37C/C110A), and H4. The lower panel shows the profile of the nucleosome with H2A, H2B, H3 (K37C-MTSL/C110A), and H4.

 

23  

Table S1. Effects of mutations in the globular domains of Drosophila H1 on nucleosome binding affinity and FRAP life times of corresponding mutations in Mouse H1o (14). ________________________________________________________________ H1

KD(µM)

H1o

t50%(s)

________________________________________________________________ H11-256 (WT)

0.29 ± 0.02

H11-256

0.28 ± 0.02

H137-211

0.21 ± 0.02

WT

52

H46A

0.38 ± 0.07

H25G

12

K58A

0.83 ± 0.08

K61A

0.46 ± 0.05

K40A

39

R63A

0.37 ± 0.05

R42A

16

K72A

0.36 ± 0.03

K73A

0.21 ± 0.03

K52A

28

K80A

0.37 ± 0.04

K59A

37

K85A

0.30 ± 0.03

K91A

0.7 ± 0.2

K69A

4

K92A

0.20 ± 0.02

K95A

0.83 ± 0.09

K73A

8

K102A

0.56 ± 0.08

K107A

0.55 ± 0.05

K85A

10

K109

0.39 ± 0.04

K116A

0.58 ± 0.09

R94A

17

________________________________________________________________ Except the WT H11-256, all others include quadruple mutations. Additional single mutations are based on H137-211. D_gH1_mut

45

SHPPTQQMIDAAIKNLKERGGSSLLAIKKYITATYKVDAQKLAPFIKKYLKSLVVNGKLIQTKGKGASGSFKLS118

M_gH10

24

DHPKYSDMIVAAIQAEKNRAGSSRQSIQKYIKSHYKVGENADSQ-IKLSIKRLVTTGVLKQTKGVGASGSFRLA96

 

24  

Table S2. Summary of major NMR samples ________________________________________________________________ 1. H11-256 (15N/13C) and nucleosome 2. H137-211 (15N/13C) and nucleosome 3. H137-211 (15N/13C) 4. H137-211 and nucleosome (H2A{15N/13C/2H}/H2B/H3/H4 and 167 bp DNA) complex 5. H11-256 (WT) and nucleosome (H2A{15N/13C/2H}/H2B/H3/H4 and 167 bp DNA) complex 6. H137-211 and nucleosome (H2A/H2B/H3{15N/13C/2H}/H4 and 167 bp DNA) complex 7. H137-211 and nucleosome (H2A{Ile, Leu, Val(13CH3) and 15N}/H2B/H3/H4 and 167 bp DNA) complex 8. H137-211 and nucleosome (H2A/H2B{Ile, Leu, Val(13CH3) and 15N}/H3/H4 and 167 bp DNA) complex 9. H137-211 and nucleosome (H2A/H2B/H3{Ile, Leu, Val(13CH3) and 15N}/H4 and 167 bp DNA) complex 10. H137-211 and nucleosome (H2A/H2B/H3/H4{Ile, Leu, Val(13CH3) and 15N} and 167 bp DNA) 11. H137-132(15N and Ile, Leu, Val-(13CH3 or -13CHD2)) 12. H137-132(15N and Ile, Leu, Val-(13CH3 or -13CHD2)) and nucleosome 13. H137-211(P44C/L60/A83C/K109/A119C) individually labeled with MTSL/MTS-EDTA-Mn2+ or Ca2+ and nucleosome (15N/13C/2H H2A) 14. H137-132(15N and Ile, Leu, Val-(13CH3 or -13CHD2)) and nucleosome (H2AT119C or H3K37C or H3A114C) with MTSL spin label ________________________________________________________________

 

25  

Table S3. Summary of ITC samples ________________________________________________________________ 1. H11-256 (WT)/Nucleosome (H2A/H2B/H3/ H4 and 167 bp DNA) 2. H11-256/Nucleosome (H2A/H2B/H3/ H4 and 167 bp DNA) 3. H137-211/Nucleosome (H2A/H2B/H3/ H4 and 167 bp DNA) 4. H137-211 with individual mutation (H46A, K58A, K61A, R63A, K72A, K73A, K80A, K85A, K91A, K92A, K95A, K102A, K107A, K109A, K116A)/ Nucleosome (H2A/H2B/H3/ H4 and 167 bp DNA) 5. H137-211/Nucleosome (H2A(T119E/K121T/K122V/A123Q)/H2B/H3/H4 and 167 bp DNA) 6. H137-211/Nucleosome (H2AΔC6/H2B/H3/H4 and 167 bp DNA) 7. H145-119/Nucleosome (H2A/H2B/H3/ H4 and 167 bp DNA) 8. H145-119/167 bp DNA ________________________________________________________________

 

26