Characterization of a Linked Jumonji Domain of the KDM5/JARID1 Family of Histone H3 Lysine 4 Demethylases

Abstract

The KDM5/JARID1 family of Fe(II)- and α-ketoglutarate-dependent demethylases remove methyl groups from tri- and dimethylated lysine 4 of histone H3. Accumulating evidence from primary tumors and model systems supports a role for KDM5A (JARID1A/RBP2) and KDM5B (JARID1B/PLU1) as oncogenic drivers. The KDM5 family is unique among the Jumonji domain-containing histone demethylases in that there is an atypical insertion of a DNA-binding ARID domain and a histone-binding PHD domain into the Jumonji domain, which separates the catalytic domain into two fragments (JmjN and JmjC). Here we demonstrate that internal deletion of the ARID and PHD1 domains has a negligible effect on in vitro enzymatic kinetics of the KDM5 family of enzymes. We present a crystal structure of the linked JmjN-JmjC domain from KDM5A, which reveals that the linked domain fully reconstitutes the cofactor (metal ion and α-ketoglutarate) binding characteristics of other structurally characterized Jumonji domain demethylases. Docking studies with GSK-J1, a selective inhibitor of the KDM6/KDM5 subfamilies, identify critical residues for binding of the inhibitor to the reconstituted KDM5 Jumonji domain. Further, we found that GSK-J1 inhibited the demethylase activity of KDM5C with 8.5-fold increased potency compared with that of KDM5B at 1 mm α-ketoglutarate. In contrast, JIB-04 (a pan-inhibitor of the Jumonji demethylase superfamily) had the opposite effect and was ~8-fold more potent against KDM5B than against KDM5C. Interestingly, the relative selectivity of JIB-04 toward KDM5B over KDM5C in vitro translates to a ~10-50-fold greater growth-inhibitory activity against breast cancer cell lines. These data define the minimal requirements for enzymatic activity of the KDM5 family to be the linked JmjN-JmjC domain coupled with the immediate C-terminal helical zinc-binding domain and provide structural characterization of the linked JmjN-JmjC domain for the KDM5 family, which should prove useful in the design of KDM5 demethylase inhibitors with improved potency and selectivity.

Keywords: dioxygenase; histone demethylase; histone modification; metal ion-protein interaction; molecular modeling.

FIGURE 1.

The KDM5/JARID1 family. a, schematic representation of KDM5 family members. JmjN and JmjC represent the two parts of the Jumonji domain; ARID (AT-rich interactive domain) is a DNA binding domain, shown to bind to CCGCCC and GCAC(A/C) sequences for KDM5A and KDM5B, respectively (23, 24). In addition, KDM5 family members contain two or three PHD (plant homeodomain) domains, some of which have been shown to bind to unmethylated histone H3 Lys-4 (PHD1 in KDM5A (26, 45)), di-/trimethylated histone H3 Lys-4 (PHD3 in KDM5A (26)), and trimethylated histone H3 Lys-9 (PHD1 in KDM5C (25)). b, KDM5 family expression constructs made in this study. c, a manually assembled hypothetical three-dimensional model of the N-terminal half of KDM5B, based on PHYRE2-generated homology models. d, a closer view showing that the two joints (JmjN-ARID and PHD1-JmjC) are adjacent to each other. e, a 12% SDS-PAGE showing examples of the purified proteins used in this study. Lane 1, KDM5A(1–588)ΔAP; lane 2, KDM5B(1–755)ΔAP; lane 3, KDM5B(1–604)ΔAP; lane 4, KDM5C(1–839); lane 5, KDM5C(1–789)ΔAP; lane 6, KDM5C(1–618)ΔAP.

FIGURE 2.

Overview of demethylation reactions catalyzed by KDM5 family. a, for protein lysine demethylation, the Fe(II)- and α-ketoglutarate-dependent Jumonji dioxygenases generate a hydroxymethyl intermediate (-N-CH₂OH) for each reaction that subsequently decomposes to release a formaldehyde spontaneously (without additional enzymatic activities) and the demethylated lysine (with one methyl group removed). b, the reaction product formaldehyde can be converted by FDH to formate. This process is coupled with the reduction of NAD⁺ to NADH, and the fluorescence generated by NADH can be monitored to reflect the rate of the coupled reactions, by converting fluorescence intensities to formaldehyde concentrations using the calibration plot shown in c. c, maximum fluorescence intensities obtained after saturation of the fluorophore were plotted against the corresponding formaldehyde standard solutions of known concentrations to convert fluorescence intensity into formaldehyde concentration. d, low concentrations of DMSO (<10%) have no effect on activity of KDM5C(1–789)ΔAP. The velocity of reactions was determined with [S] = 15 μm H3K4(1–24)me3 and [E] = 0.5 μm JARID1C(1–789)ΔAP. The DMSO concentrations tested were 2-fold dilutions of 10% to 5, 2.5, 1.25, and 0.63%. e, traces of MALDI-TOF mass spectrometry of sample reactions of KDM5B(1–755)ΔAP at room temperature (left) and 37 °C (right). f and g, optimal pH for the demethylase activities of KDM5C(1–789)ΔAP activity (f), which is the same as our previously determined optimal conditions established for KDM5C(1–789) without the internal deletion ΔAP (43), and KDM5B(1–755)ΔAP (g), which has an optimum pH of 7.0. The reactions were carried out under the KDM5 reaction buffer conditions with varying pH as indicated (50 mm MES for KDM5C or 50 mm HEPES for KDM5B). h, FDH-coupled assays of demethylase activities of KDM5C(1–789)ΔAP and KDM5C(1–618)ΔAP under the conditions of [E] = 0.5 μm and [S] = 50 μm H3(1–24)K4me3. An assay with no enzyme was used to determine background fluorescence. i, the demethylase activity of KDM5B(1–755)ΔAP is greater at low ionic strength. j, KDM5B(1–755)ΔAP has no activity on trimethylated Lys-9- and trimethylated Lys-27-containing peptides.

FIGURE 3.

Kinetic parameters of KDM5C(1–789)ΔAP and KDM5B(1–755)ΔAP. a–d, Michaelis-Menten kinetic plots of KDM5C(1–789) (a and b) and KDM5B(1–755)ΔAP (c and d) for the cofactors αKG (a and c) and substrate H3 peptide (b and d), measured by an FDH-coupled demethylase assay. Initial velocities were plotted against αKG or peptide concentrations and fit with the Michaelis-Menten equation using GraphPad Prism version 5.0. Error bars, S.E. from two independent experiments. e and f, kinetic progression of demethylation reaction catalyzed by KDM5C(1–789) (e) and KDM5B(1–755)ΔAP (f), measured by MALDI-TOF mass spectrometry. g, summary of the kinetic constants for each demethylation step derived from the global fitting of the experimental data. The apparent k_cat value for each step was derived from the data in e and f using the formula ^appk_cat = k_i × [S]/[E], where i = 1, 2, and 3, and [S]/[E] = 20. H3K4me0, -me1, -me2, and -me3, un-, mono-, bi-, and trimethylated histone H3 Lys-4, respectively.

FIGURE 4.

Inhibition of GSK-J1 and JIB-04. a–d, inhibition of demethylation activity of KDM5B(1–755)ΔAP and KDM5C(1–789)ΔAP on H3(1–24)K4me3 substrate by GSK-J1 (a and b) and JIB-04 (c and d). e and f, cell growth-inhibitory effects of GSK-J4, -J5 (e), and JIB-04 (f) in human breast cancer cell lines. MDA-MB231 (red) or MCF7 (blue) cells were treated with the indicated concentrations of JIB-04 or GSK-J4 (solid line) or GSK-J5 (dashed line) or vehicle equivalent (DMSO; dotted line), and the percentage of untreated control cell growth was determined after 72 h by the sulforhodamine B assay. Data represent the means ± S.E. (error bars) of three or four independent experiments performed in triplicate.

FIGURE 5.

Structure of linked JmjN-JmjC of KDM5A(1–588)ΔAP and αKG binding. a, the KDM5A catalytic domain has two twisted β-sheets (a four-stranded minor sheet in cyan and an eight-stranded major sheet in dark blue and gray) with helices packed on the outer surfaces of the major and minor sheets. The yellow circle indicates the bound metal Mn(II). b and c, two views of Mn(II)-αKG (in yellow) binding in the active site of KDM5A. Note the octahedral coordination of Mn(II) by the His-483, Glu-485, His-571, αKG (two ligands), and a water molecule. d, omit electron densities, contoured at 5σ and 3σ above the mean, are shown for Mn(II) (magenta mesh) and αKG (green mesh), respectively. e, the corresponding helix αC (magenta) in the independently determined structure of KDM5C ARID domain (PDB entry 2JRZ). f and g, ITC measurements of binding of αKG (f) or N-oxalylglycine (g) to KDM5A(1–588)ΔAP and binding of αKG to KDM5B(1–604)ΔAP (h) or KDM5B(1–755)ΔAP (i). The dissociation constant (K_D) and the one-site binding model (n = 1) were calculated by fitting the ITC data.

FIGURE 6.

Structural comparison of KDM5A and KDM6 family demethylases. a, superimposition of KDM5A (colored) and KDM6A (in gray; PDB entry 3AVR). Note that the two terminal amino groups, Lys-501 of KDM5A and Lys-1137 of KDM6A, occupy the same space, and both are in close contact with αKG. b, a model of trimethylated lysine (Kme3) in the active site of KDM5A, adopted from the structure of KDM6A in complex with histone H3 peptide (residues 17–33) trimethylated at Lys-27 (PDB entry 3AVR). c, a model of GSK-J1 in the active site of KDM5A, adopted from superimposition of the structure of KDM6B in complex with GSK-J1 (PDB entry 4ASK). Note that GSK-J1 partially overlaps with αKG via the propanoic acid moiety, and the pyridyl-pyrimidine biaryl of GSK-J1 makes a bidentate interaction with the metal ion (silver ball). d, superimposition of KDM5A (colored) and KDM6B (in gray) in complex with GSK-J1. Note that Cys-481 of KDM5A (an invariant residue among the KDM5 family) replaces Pro-1388 of KDM6B, which packs directly against the aromatic ring of tetrahydrobenzazepine.

FIGURE 7.

Sequence alignment of KDM5 family members. White-on-black residues are invariant between the four sequences examined, whereas gray-highlighted positions are conserved (R and K; E and D; T and S; F and Y, V, I, L, and M). Positions highlighted with red arrows indicate junction points or termination sites in the various constructs indicated in Fig. 1, a and b. The Rice Jmj703 (another trimethylated histone H3 Lys-4 demethylase) was included for comparison (47).