If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Arabidopsis LHP1 (LIKE HETEROCHROMATIN PROTEIN 1), a unique homolog of HP1 in Drosophila, plays important roles in plant development, growth, and architecture. In contrast to specific binding of the HP1 chromodomain to methylated H3K9 histone tails, the chromodomain of LHP1 has been shown to bind to both methylated H3K9 and H3K27 histone tails, and LHP1 carries out its function mainly via its interaction with these two epigenetic marks. However, the molecular mechanism for the recognition of methylated histone H3K9/27 by the LHP1 chromodomain is still unknown. In this study, we characterized the binding ability of LHP1 to histone H3K9 and H3K27 peptides and found that the chromodomain of LHP1 binds to histone H3K9me2/3 and H3K27me2/3 peptides with comparable affinities, although it exhibited no binding or weak binding to unmodified or monomethylated H3K9/K27 peptides. Our crystal structures of the LHP1 chromodomain in peptide-free and peptide-bound forms coupled with mutagenesis studies reveal that the chromodomain of LHP1 bears a slightly different chromodomain architecture and recognizes methylated H3K9 and H3K27 peptides via a hydrophobic clasp, similar to the chromodomains of human Polycomb proteins, which could not be explained only based on primary structure analysis. Our binding and structural studies of the LHP1 chromodomain illuminate a conserved ligand interaction mode between chromodomains of both animals and plants, and shed light on further functional study of the LHP1 protein.
). The allelic mutations of LHP1 show pleiotropic phenotypes, such as small plant size, curled leaves, terminal flowers, early flowering, reduced fertility, low glucosinolate content, and low roots growth rate (
). Other studies demonstrate that LHP1 binds to trimethylated histone H3K9 (H3K9me3) and H3K27 (H3K27me3) with similar affinities in vitro, however, LHP1 specifically colocalizes with H3K27me3 in euchromatin in vivo (
), the function of LHP1 in heterochromatin needs to be further elucidated.
As multiple phenotypes shown by lhp1 mutants, LHP1 plays diverse roles during plant growth and development. LHP1 participates in flowering process by repressing two central flowering regulators, FT (FLOWERING LOCUS T, an activator gene of flowering) (
). Because LHP1 holds Pc-like functions, binding to methylated H3K27 and interacting with plant RING-finger proteins, LHP1 is taken as a component of the plant PRC1-like complex, catalyzing histone H2A ubiquitylation (
). Indeed, several studies illustrate that LHP1 works as the extra sex combs counterpart of PRC2 in animals (H3K27me3 binding and self-recruiting PRC2) and interacts with PRC2 and participates in establishing, spreading, and maintaining trimethylation of histone H3K27 (
Taken together, LHP1 plays important roles in plant development, and the interaction between LHP1 chromodomain and methylated histone H3K9 or H3K27 is critical for LHP1's functions. To better understand the mechanism of this interaction, in this study, we conducted quantitative binding assays and structural analyses of the LHP1 chromodomain with histone H3K9/K27 peptides. Our isothermal titration calorimetry (ITC) binding assays indicated that the LHP1 chromodomain binds to dimethylated or trimethylated H3K9/K27 peptides with comparable affinities, whereas no binding or weak binding to unmodified or monomethylated H3K9/K27 peptides. Our crystal structural studies provide molecular insights into lack of the selective binding of LHP1 to histone H3K9/K27 peptides.
Results and discussion
Chromodomain of Arabidopsis LHP1 binds to histone H3K9me2/3 and H3K27me2/3 peptides with comparable affinities
The chromodomains of HP1 and Pc in Drosophila are highly conserved and specifically recognize two sequence-related repressive histone markers, H3K9me3 and H3K27me3, respectively (
). The chromodomains of HP1 homologs in human, namely CBX1/HP1β, CXB3/HP1γ, and CBX5/HP1α, discriminate H3K9me3 from H3K27me3 effectively, whereas most Pc homologs in human (CBX2, 4, 6, 7, 8) bind to both H3K9me3 and H3K27me3 (
). To determine the binding ability of the LHP1 chromodomain in Arabidopsis, we measured the binding affinities of recombinant LHP1 chromodomain for both H3K9me3 (residues 1–15, with K9 trimethylated) and H3K27me3 (residues 19–33, with K27 trimethylated) peptides by ITC assays. Our data show that the LHP1 chromodomain binds to these two peptides similarly (Fig. 1), consistent with previous florescence polarization binding data (
). We further analyzed the effect of methylation states on the interaction and found that the LHP1 chromodomain bears similar binding affinities to both dimethylated H3K9 and H3K27 peptides, but no binding or weak binding to unmethylated or monomethylated cognates (Fig. 1). These data indicate that the chromodomain of LHP1 behaves differently from the HP1 and Pc proteins in Drosophila, but is similar to human Pc proteins.
Structure of the apo-form chromodomain of Arabidopsis LHP1 indicates an uncanonical chromodomain conformation
To illustrate the molecular mechanism of the interaction between the LHP1 chromodomain and methylated H3K9/K27 peptides, we tried to crystalize LHP1 chromodomain with or without methylated H3K9/K27 peptides of different lengths and solved the peptide-free and H3K9me3 (residues 5–10) peptide-bound crystal structures (Table 1 and Figs. 2 and 3). As shown in Figure 2, the structure of the apo LHP1 chromodomain is similar to the structures of previously reported HP1 or Pc chromodomains (
), containing three β strands and a 310 helix between the second and the third β strands followed by an α helix. The three β strands form a twisted antiparallel β-sheet harboring three potential methyllysine-binding, cage-forming aromatic residues Y108, W129, and W132 at one end of the β-sheet (Fig. 2A). However, a careful examination of the structure uncovers that the rings of the three aromatic residues are not mutually orthogonal, with the side chain of Y108 being parallel with that of W132, intruding into the potential methyllysine-binding pocket and making the pocket too shallow to fit for the methylated lysine residue (Fig. 2, A and B). This structural character indicates that there should be a conformational adjustment for the LHP1 chromodomain to interact with the methylated histone peptides.
Table 1Data collection and refinement statistics
a, b, c (Å)
31.7, 31.7, 106.4
26.1, 28.5, 70.5
α, β, γ (°)
90, 90, 120
90, 90, 90
35.26 -1.60 (1.63 -1.60)
No. of reflections work/free
No. atoms/average B-factor [Å2]
RMSD bonds (Å)/angles (º)
Ramachandran plot residues in favored regions (%)
Values in parentheses are for the highest resolution shell.
Comparison of the structure of the LHP1 chromodomain to those of the HP1/Pc chromodomains reveals one dramatical difference in LHP1, which has an extra-long C-terminal α-helix approximately perpendicular to the β-sheet of LHP1 and forms an L-shape architecture (Fig. 2, C and D). In the chromodomains of HP1 and Pc, the α-helix is much shorter and packed against the β-sheet to form a sphere shape (
). This extra-long α-helix is reminiscent of the second PWWP domain (another member of Royal family) of NSD3, which also bears an extra-long α-helix; however, the PWWP’s long helix shows a different orientation, packing against the β-sheet (
). Taken together, our peptide-free structure dictates that the chromodomain of LHP1 is an uncanonical chromodomain and needs a conformational change for ligand binding.
Structural basis for selective binding of ARKme3S motif by the chromodomain of Arabidopsis LHP1
In the LHP1–H3K9me3 complex structure, the chromodomain of LHP1 undergoes tremendous conformational changes to accommodate the methyllysine of the H3K9me3 peptide, which will be discussed in details latter. Briefly, the N-terminal residues of the LHP1 chromodomain and the H3K9me3 peptide form two antiparallel β strands by using an induced-fit mechanism (Fig. 3A), which is also observed in the previous HP1-H3K9me3 complex structures (
), and several intermolecular hydrogen bonds between H3 peptide and LHP1 further stabilize the complex (Fig. 3, B and C). For example, H3Q5, H3A7, H3R8, and H3S10 form main-chain hydrogen bonds with I100, Y108, and N144 from LHP1, and H3R8 and H3S10 form side-chain hydrogen bonds with N144, S147, and E140, respectively (Fig. 3B). The aromatic cage-forming residues are essential for the ligand recognition and LHP1 stability, because the mutations of Y108 or W132 to alanine disrupt the binding, whereas the W129A mutant protein is insoluble (Fig. 3D). These cage-forming residues are also vital for the function of LHP1 in vivo, which has been confirmed by previous cage residue mutant studies (
). In addition to the cation-π interaction, the side-chain hydrogen bonds between H3S10 and E140 are also indispensable, because the E140A mutation disrupts the peptide binding. This also suggests that the hydroxyl group of serine is important for the specific interaction and only serine or threonine can be tolerated at this site, consistent with the previous SPOT-blot assay (
). Furthermore, the interaction between the side chain of E140 and the hydroxyl group of serine/threonine are crucial for the chromodomain-ligand binding, because the glutamic acid residue is absolutely conserved in the chromodomains (Fig. 4A). On the other hand, the N144A mutant weakens, but does not abrogate binding. This also suggests that the arginine in the alanine-arginine-lysine-serine (ARKS) motif is replaceable, which is also in accordance with the previous SPOT-blot assay (
). Furthermore, all the mutants show similar binding behavior to both H3K9me3 and H3K27me3 peptides (Fig. 3D), indicating that the LHP1 chromodomain recognizes these two sites in a similar way.
In addition, the side chain of H3Q5 points away from the interaction surface, which indicates that the residue at this position is not critical for the specific interaction (Fig. 3B). H3T6 and the hydrophobic side chain but not the guanidino group of H3R8 are packed against two hydrophobic residues, F107 and A149, through hydrophobic interactions (Fig. 3E). Because the stabilization of H3T6 is mainly conferred by hydrophobic interactions, the corresponding residue H3A24 in the H3K27me3 peptide is also tolerated at this position. These two hydrophobic residues, namely F107 and A149, play an important role in binding to methylated H3K9 and H3K27 peptides, because the double mutations of these two residues to glycine significantly weaken the binding (Fig. 3D). H3A7 is anchored in a small hydrophobic pocket formed by residues I110, I127, W129, and L145 and the side chain of alanine fits in the small hydrophobic pocket perfectly (Fig. 3F), which also determines the protein-ligand binding specificity. Taken together, our binding and complex structure analyses demonstrate that the methylated lysine at position 0 and alanine at position -2 of the ARKS motif are absolutely required, serine or threonine is tolerated at position +1, whereas replacement of arginine at position -1 by hydrophobic residues should be acceptable (Fig. 4B), which is consistent with previous Pc protein study (
). This also suggests that additional lysine-methylated protein ligands of LHP1 may remain to be discovered.
Comparison of peptide-free and peptide-bound LHP1 chromodomain structures uncovers two conformational changes: the disordered N-terminal to additional β strand and the C-terminal rigid long α helix to a flexible short loop (Fig. 3G). These conformational adjustments facilitate the protein-peptide interaction by forming methyllysine-binding cage for K9me3 and for the F107-A149 hydrophobic pair to hold H3T6, N144 and S147 moving close to the H3K9me3 peptide to enhance the interactions. Overall, our structural studies indicated that the chromodomain of LHP1 adopts several conformational changes to bind to the histone H3 peptide.
Comparison of the chromodomain of LHP1 to other chromodomains supports a conserved protein-ligand recognition mechanism
Structural and sequence comparison of the LHP1 chromodomain with other chromodomains indicated that the methyllysine cage-forming residues are highly conserved (Fig. 4A) and they bind to their ligands by similar binding modes (Fig. 5). In addition to the highly conserved cage-forming residues, the interacting residue for serine of ARKS, E140 in LHP1, is absolutely conserved too, which is consistent with its indispensable role in the interaction (Fig. 3D). Although the domain structures of LHP1 and HP1 are the same (
), our structure study shows that their chromodomains bear different properties: the polar clasp in HP1 proteins decides their preference of binding to a methylated H3K9 histone tail, whereas the hydrophobic clasp in LHP1 and other human Pc proteins confers their binding to the two sequence-related methylated H3K9 and H3K27 histone tails more promiscuous (Figs. 4 and 5) (
). Our structural observation indicates that H3T6 is packed against a hydrophobic clasp formed by F107-A149, and H3T6-H3R8 also serves as a clasp to hold the “F107-A149 clasp” in turn (Fig. 3E), hence the hydrophobic clasp is important for the protein-ligand interaction in LHP1 as well (Fig. 3D). To explore if swapping the clasp residues of LHP1 with HP1 would change the binding selectivity of the LHP1 chromodomain, we made double mutations of the two clasp residues and found that the clasp residue swapping mutant does prefer to binding the H3K9me3 peptide over the H3K27me3 peptide (Fig. 3D), further validating the importance of the clasp residues in ligand selectivity of the LHP1 chromodomain. Based on the sequence alignment, Q146 of LHP1 is corresponding to the second residue of the “clasp residues” (Fig. 4A), which will not explain why LHP1 could bind to both H3K9me3 and H3K27me3 without structural information. Taken together, both our binding and structural studies indicate that LHP1 functions similar to human Pc proteins, and there may be other proteins resembling HP1 proteins in plants. Indeed, ADCP1 (Agenet Domain Containing Protein 1), a plant unique poly-Agenet domain (another member of Royal family in plant) containing protein has been verified as an HP1-like protein in plants (
The overall peptide-bound structures of LHP1 and other HP1/Pc homologs are similar, except the C-terminal conformational change to form an F107-A149 hydrophobic pair (Fig. 5, A–E). Structural comparison with another solved plant chromodomain, ZMET2 in Zea mays, suggests a similar ligand-binding induced C-terminal conformation change to form a polar clasp by E440 and D484 of ZMET2 to selectively bind to methylated H3K9 histone tail (Fig. 5F) (
). However, the C terminus of the ZMET2 chromodomain has the same orientation as the HP1/Pc chromodomains, which indicates that the new C-terminal direction of LHP1 chromodomain is a unique property.
In conclusion, our quantitative ITC-binding experiments illuminate that the chromodomain of LHP1 binds to dimethylated or trimethylated H3K9/K27 peptides without obvious sequence and methylation status preference, whereas it does not or weakly bind to unmethylated or monomethylated cognates. Our structural analyses reveal that although LHP1 is a unique homolog of HP1 protein in plants, its chromodomain behaves similar to those of human Pc proteins and interacts with methylated H3K9/K27 histone tails by a hydrophobic clasp.
Protein expression and purification
The chromodomain of Arabidopsis LHP1 protein (residues 106–160) was subcloned into a modified pET28-MHL vector. The encoded N-terminally His-tagged fusion protein was overexpressed in Escherichia coli BL21 (DE3) Codon plus RIL (Stratagene) at 15 °C and purified by affinity chromatography on Ni-nitrilotriacetate resin (Qiagen, or Nanjing Qingning Bio-Technology Co., Ltd) followed by tobacco etch virus protease treatment to remove the tag. The tag-removed protein was purified by a second round of Ni-nitrilotriacetate resin to remove the His-tag and tobacco etch virus protease and then further purified by a Superdex75 gel-filtration column (GE Healthcare). The final purified protein was concentrated to 8∼30 mg/ml in a buffer containing 20 mM Tris–HCl, pH 7.5, 150 mM NaCl, and 1 mM DTT. The mutant constructs were generated by using QuickChange Site-Directed Mutagenesis Kit (Stratagene, 200,518), and all the mutant proteins were purified similar to the wild-type one. All the constructs used in this article were sequenced and confirmed.
Isothermal titration calorimetry
For the ITC measurement, the concentrated proteins were diluted into 20 mM Tris–HCl, pH 7.5, 150 mM NaCl; the lyophilized peptides for H3K9 (residues 1–15, with K9 at different methylation states) and H3K27 (residues 19–33, with K27 at different methylation states) (Wuhan Bioyeargene Biotechnology Co., Ltd or Shanghai Apeptide CO., Ltd) were dissolved in the same buffer, and the pH value was adjusted by adding NaOH. Peptides concentrations were estimated from the mass and the volume of the solvent. All the measurements were performed in duplicate at 25 °C, using a VP-ITC microcalorimeter (MicroCal, Inc.) or an iTC-200 microcalorimeter (MicroCal, Inc.). The protein with a concentration of 50 μM was placed in the cell chamber, and the peptides with a concentration of 0.75 to 2 mM in syringe was injected in 25 (20 for iTC-200) successive injections with a spacing of 180 s (150 s for iTC-200) and a reference power of 13 μcal/s (5 μcal/s for iTC-200). Control experiments were performed under identical conditions to determine the heat signals that arise from injection of the peptides into the buffer. The data were fitted using the single-site binding model within the Origin software package (MicroCal, Inc.). iTC-200 data should be consistent with those from VP-ITC instrument, based on ITC results of same ITC assay using the two instruments.
For crystallization, the purified proteins were mixed with or without different length histone H3K9me3 or H3K27me3 peptides (encompassing residues 1–15, 19–33 as long form, and 5–10, 23–28 as short form), respectively. The mixture of protein and peptide with a molar ratio of 1:3 was crystallized using the sitting drop vapor diffusion method at 18 ºC by mixing 0.5 μl of the protein with 0.5 μl of the reservoir solution. The peptide-free crystals were obtained in a buffer containing 20 mM Tris–HCl, pH 7.5, 150 mM NaCl, and 1 mM DTT. The complex crystals of chromodomain of LHP1 and histone H3K9me3 peptide (residues 5–10) were obtained in a buffer containing 0.1 M Hepes, pH 7.5, 2.0 M ammonium sulfate. Before flash-freezing crystals in liquid nitrogen, the crystals were soaked in a cryoprotectant consisting of 85% reservoir solution and 15% glycerol.
Data collection and structure determination
The diffraction data of peptide-free crystals and complex crystals were collected at a rotating copper anode x-ray generator and beamline BL17U1 of the Shanghai Synchrotron Radiation Facility (SSRF) at 100K, respectively. Diffraction images were processed using XDS (
) was used for interactive model building. For the structure of peptide-free LHP1 was solved by molecular replacement with coordinates from PDB entry 1Q3L (deposited by Jacobs & Khorasanizadeh in 2003). After initial restrained model refinement with REFMAC (
) were used for further model refinement. The structure of LHP1–H3K9me3 complex was solved by molecular replacement with coordinates from the LHP1 crystal structure. REFMAC was used for restrained model refinement. Crystal diffraction data and refinement statistics for the structure were displayed in Table 1.
Coordinates and structure factor amplitudes of LHP1 chromodomain and LHP1 chromodomain-H3K9me3 were deposited in the PDB with accession numbers 7VZ2 and 7VYW, respectively.
Conflict of interest
The authors declare that they have no conflicts of interest with the contents of this article.
We would like to thank Dr Wolfram Tempel of University of Toronto, Zeyuan Guan of Huazhong Agricultural University, and Huan Zhou of SSRF beamline BL17U1 for assistance in data collection, Wolfram Tempel for structure determination, and Aiping Dong for reviewing the crystallographic models. We also thank the research associates at Center for Protein Research (CPR), Huazhong Agricultural University, for technical support. This work was carried out with the support of Shanghai Synchrotron Radiation Facility, beamline 17U1 (proposal 2019-SSRF-PT-010749). This work was supported by the National Natural Science Foundation of China grants ( 31500615 ), the Priority Academic Program Development of the Jiangsu Higher Education Institutes, China (PAPD), and Six-talent Professorship of Jiangsu province, China ( SWYY-104 ).
Y. L. conceptualization; Y. L., X. Y., M. Z., Y. Y., F. L., X. Y., M. Z., Z. W., and S. Q. methodology; Y. L., X. Y., M. Z., Y. Y., F. L., X. Y., M. Z., and Z. W. investigation; Y. L., S. Q., and J. M. writing–original draft; Y. L., X. Y., M. Z., Y. Y., F. L., X. Y., M. Z., Z. W., S. Q., and J. M. writing–review and editing; Y. L., Y. Y., and J. M. visualization; Y. L. and J. M. supervision; Y. L. funding acquisition; Y. L., X. Y., M. Z., Y. Y., F. L., X. Y., M. Z., Z. W., S. Q., and J. M. formal analysis.
Biochemical genetics of plant secondary metabolites in Arabidopsis thaliana: The glucosinolates.