Structur al analysis of human KDM5B guide s histone demethylase inhibitor development

Members of the KDM5 (also known as JARID1) family are 2-oxoglutarate and Fe 2+ dependent oxygenases acting as histone H3K4 demethylases, thereby regulating proliferation, stem cell self-renewal and differentiation. Here we report crystal structures of the catalytic core of the human KDM5B enzyme in complex with three inhibitor chemotypes. These scaffolds exploit several aspects of the KDM5 active site, while their selectivity profiles reflect the hybrid features with the KDM4 and KDM6 families. Whereas GSK-J1, a previously identified KDM6 inhibitor shows about 7-fold less inhibitory activity towards KDM5B, KDM5-C49, displays >25-100 -fold selectivity between KDM5B and KDM6B. The cell-permeable derivative KDM5-C70 provides an anti-proliferative effect in myeloma cells, leading to genome-wide elevated H3K4me3 levels. The selective inhibitor GSK467 exploits unique binding modes, however lacks cellular potency in the myeloma system. Taken together, these structural leads deliver multiple starting points for further rational and selective inhibitor design.


INTRODUCTION
Methylation of lysine residues on histone tails is a reversible epigenetic modification that plays a key role in gene regulation. The transcriptional response is dependent on the specific residue methylated, the number of methyl groups added or removed, as well as on a complex chromatin cross-talk with other chromatin regulators and transcription factors, suggested to fine-tune transcription [1][2][3] . Removal of methyl groups from lysine residues is catalysed by two classes of histone lysine (K)-demethylases (KDMs): the flavine adenine dinucleotide (FAD)dependent amino oxidases 4 (grouped as KDM1 subfamily) 5 and the Fe(II) and 2-oxoglutarate (2OG) dependent oxygenases that contain a conserved catalytic Jumonji C (JmjC) -domain (belonging to subfamilies KDM2-KDM7) 2,3 . In contrast to the KDM1 demethylases which are limited to demethylate mono-and di-methyl lysine residues, the JmjC-type KDMs are able to catalyse demethylation of all 3 possible methylation states of methyl-lysyl residues.
Distinctive features between the KDM subfamilies include defining domain organisation patterns as well as site-and methylation state specificity for the different methyl marks 2,3 .
The reaction mechanism for the JmjC-demethylases is dependent on molecular oxygen and proceeds through the oxidation of 2OG and Fe(II), yielding CO 2 , succinate and a highly reactive Fe(IV)-oxoferryl intermediate 6 .
Members of the KDM5 subfamily of JmjC-KDMs are transcriptional co-repressors by specifically catalysing the removal of all possible methylation states from lysine 4 of histone H3 (H3K4me3/me2/me1), a histone modification that occurs largely around transcriptional start sites of actively transcribed genes 7 . KDM5 enzymes are often found as components of transcriptional complexes with repressors such as REST, histone deacetylases or histone methyl transferases (HTMs) 8 . In mammals the KDM5 subfamily encompasses four proteins, KDM5A (known as JARID1A or RBP2), KDM5B (known as JARID1B or PLU1), and the KDM5C (JARID1C or SMCX) and KDM5D (JARID1D or SMCY) members, the latter two encoded on the X and Y chromosomes, respectively.
The KDM5 family is conserved from yeast to humans, displaying a similar domain architecture with an N-terminal Jumonji (JmjN) domain, a DNA binding ARID domain (ATrich interactive domain), a catalytic JmjC-domain, a C5HC2 zinc finger motif located Cterminally to the JmjC-domain, a PLU-1 motif, as well as two to three methyl-lysine or methyl-arginine binding plant homeodomain (PHD) domains denoted PHD1, PHD2 and PHD3 (Fig. 1). These additional domains contribute critically to genomic KDM5 target gene occupation, for example PHD domains bind to modified lysine residues in a sequencespecific manner 9 . However, the complex role of the particular KDM5 PHD domains is not fully understood at present. Whereas in KDM5C the PHD1 domain binds the H3K9me2/me3 methyl mark, the same domain in KDM5B, and to some extent also KDM5A, recognises preferentially unmethylated H3K4 marks, the product of the KDM5 mediated demethylation reaction. Thus, at least in KDM5B and KDM5A 10 the PHD1 domain, sandwiched between the JmjN and C domains, may protect H3K4 from re-methylation 11 and moreover could allosterically regulate the catalytic activity of the Jmj domain 12 .
The KDM5 enzymes play pivotal roles both during normal development and in pathological conditions 2,13 , with KDM5A and KDM5B linked to control of cell proliferation, differentiation, and to several cancer types. KDM5C plays a role in neuronal development, whereas the Y-chromosome encoded KDM5D is widely expressed and involved in spermatogenesis.
Development of chemical tools to interrogate KDM5 biology is progressing slowly 13,14 -this impasse prompted us to determine crystal structures of the catalytic core of human KDM5 enzymes, compare them to related KDM4 and KDM6 enzymes, and to characterise KDM inhibitors, which were used to investigate H3K4me3 epigenomic signatures. The complex structures with distinct chemotypes reveal several distinguishing features and plasticity of the KDM5 cofactor site, which could be utilized in further inhibitor development.

RESULTS
ARID and PHD domains are dispensable for activity -Full-length human KDM5B is a 1544-residue protein that requires the JmjN, ARID, PHD1, JmjC domain as well as the C5HC2 zinc finger for in vivo demethylase activity 15 . To further understand the effects of domain structure on catalytic activity and to facilitate structural studies of the Jmj catalytic core we first engineered deletion constructs and analysed their enzymatic properties.
Assuming that the C-terminal PLU and PHD2 and PHD3 domains were dispensable for enzymatic activity, a KDM5B-c1 construct encompassing the JmjN, ARID, PHD1, JmjC and the zinc finger motif was cloned and found to be active ( Fig. 1a- Supplementary Fig. 1-2). Based on structural analysis 16,17 of other KDMs that reveal close contacts between the JmjN and JmjC domains, we next designed deletion constructs of KDM5B-c1 where the PHD1 domain and parts of the ARID domain were deleted (KDM5Bc-2-KDM5Bc4; Fig. 1a). All proteins were expressed as N-terminally His-tagged proteins in baculovirus infected SF9 insect cells and purified using Ni-affinity resin followed by size-exclusion chromatography. To probe the effect of the ARID and PHD1 deletion, proteins were analysed for H3K4me3/me2 demethylase activity using either a formaldehyde dehydrogenase (FDH) coupled assay or a direct Rapid-Fire mass spectrometry (RF-MS) assay 18,19 (Fig. 1b, Supplementary Fig. 1, Supplementary Tables 1 and 2).
We note that cleavage of the His-tag, re-purification and freezing makes the KDM5B-c4 construct amenable to gradual loss of activity, however provides material that can be used for crystallisation. Using the FDH-coupled assay and a histone H3(1-21)K4me2 peptide as substrate, KDM5B-c2 revealed approximately 40% of the catalytic efficiency (k cat /K m ) compared to KDM5B-c1 (Fig. 1b, Table 3). The initial structure of KDM5B-c4, crystallised with inhibitor KDM5-C49 (1), was solved by molecular replacement using KDM4A (PDB ID 2bp5) as a search model. Later on, diffracting crystals were also obtained for construct KDM5B-c2, confirming the ligand binding modes observed for KDM5B-c4.
The overall fold of this catalytic core reveals three conserved domains: the JmjN (residues 31-72) and the JmjC domain (residues 375-602) associate tightly and form together with seven residues of the ARID domain (residues 94-100) an extended double-stranded β-helix (DSBH), characteristic of the 2OG oxygenase superfamily (labelled in pink, blue and yellow, Fig. 2a), a C-terminal helical domain (residues 604-671, 737-753, green), and a beta-sheet composed of 3 beta-strands (residues 673-734, cyan) (Supplementary Fig. 3) that harbours a C5HC2-zinc finger motif. A Mn 2+ ion, used in the experiment to replace the active site Fe 2+ , has an octahedral coordination through His 499 , Glu 501 and His 587 , all part of the conserved HxD/E..H metal chelation motif found in 2-OG oxygenases 20 and adjacent to the cofactorbinding site (Fig. 2b). In most structures a HEPES buffer molecule is bound in a cavity at the interface between the extended DSBH-fold and a helical domain (Supplementary Fig. 4).
The C-terminal helical domain is composed of 4 helices, and a Zn-finger C5HC2 -motif is found ( Fig. 2a and Fig. 2c), similar to the GATA-like motif in the KDM6 subfamily 21,22 . In all structures a Zn 2+ molecule is coordinated in a tetrahedral geometry by Cys 692 , Cys 695 , Cys 715 and His 718 (Fig. 2c, Supplementary Fig. 5). Interestingly, this domain appears to be sensitive to the redox environment. In crystals grown and mounted within a 3-6 day period a second Zn 2+ molecule is observed and coordinated by Cys 706 , Cys 708 , Cys 723 and Cys 725 as noted in the ligand structures of KDM5B with 2,4-PDCA (2), N-oxalylglycine (NOG, 3), GSK-J1 (4) and GSK467 (5). In the complex structure of KDM5B with the inhibitor KDM5-C49, determined from a crystal that appeared after three weeks, two intra-molecular disulfide bonds are observed (Supplementary Fig. 5). In KDM5 proteins the C5HC2 zinc finger motif is required for its in vivo catalytic activity 15 and in vitro data show that demethylase activity of KDM5B-c1 is increased in the presence of the reducing agent TCEP (Supplementary Fig.   6). This potentially redox-sensitive motif is similar in packing to the Jmj core as observed in the KDM6 proteins with a GATA-like domain 21,22 , which is involved in substrate binding with a putative induced fit mechanism for substrate recognition. At present it is unknown if this oxidation/reduction mechanism has an impact upon enzymatic activity of KDM5B and further studies will reveal its possible physiological implications.
An overlay of the KDM5B-c4 complex structures provides an overall C α rmsd value of 0.36-0.98Å, with largest differences found in the end of beta strand 5 and the loop connecting β5 and β6 (Supplementary Fig. 7). Alternative conformations are observed for Tyr 425 , and Asp 428 is found on either side of the DSBH-fold, (Supplementary Fig. 7), resulting in a cofactor pocket of different size and shape.
Comparison to KDM4 and KDM6 subfamilies -Superimposition of the KDM5B structures with KDM6A (PDB ID 3AVR) or KDM4A (PDB ID 2P5B) shows that the domain architecture and the overall fold is similar to the KDM6 enzymes, however major differences are found in the loop structures and the JmjC domain (overall backbone rmsd value of C α =2.3 Å). Instead the KDM5B JmjC domain is more related to its counterpart in KDM4A ( Supplementary Fig. 8), although the Zn-binding substructure found in KDM4A 17 is missing in KDM5B. The surface charge distribution between the proteins is dissimilar, likely reflecting different substrate specificities and substrate binding sites (Fig. 3a-c). Interestingly, β-strand 5 in KDM5B, that is located close to the ARID and PHD1 deletion site, is found at a slightly different position in both KDM6A and KDM4A. In the latter proteins, this β-strand projects 7-8 Å away from the 2OG-binding pocket compared to KDM5B, rendering interpretations and comparisons difficult since the corresponding strands in KDM4 and KDM6 harbour arginine or asparagine residues (Arg 1027 and Asn 86 ) in homologous sequence positions to Arg 98 . To explore whether the location of Arg 98 in the KDM5B-c4 deletion construct affects activity we mutated this residue to a glycine or a lysine residue, which occupies in the structure of KDM4A the same position as Arg 98 , but is located on a distinct secondary structure element. Using the FDH coupled assay and a histone H3(1-21)K4me2 peptide as substrate, both mutants were found to be active, with lower K m and higher catalytic efficiencies compared to KDM5B-c4 ( Fig. 1b and Supplementary Table 1) indicating indeed a critical role for Arg 98 . These data suggest that despite measurable activity, the engineered KDM5B-c4 construct has limitations for correlating full-length construct activities, however the data also highlight the utility of this construct for analysis of inhibitor binding modes. The complex behaviour of this structural segment around Arg 98 is further emphasized by the fact that the crystal structure of construct KDM5B-c2, (PDB ID 5f3V) revealed that Arg 98 is found in an identical position as in construct KDM5-c4 (PDB ID 5a3w). This construct, in terms of kinetic constants more similar to construct KDM5B-c1, shows that Lys 100 has moved approximately 7.8 Å away from the 2OG-binding pocket ( Supplementary Fig. 9) indicating that this part of the construct indeed contains a flexible segment that has an impact on catalytic behaviour. Taken together, these data suggest that substrate binding is mediated in part by elements between the PHD and the JmjC domain, however significant interpretations of inhibitor binding can be derived from either KDM5B-c2 or KDM5B-c4 ligand structures.
Comparison of the 2OG-binding pocket between KDM5B (PDB ID 5A1F) and KDM6A reveals significant differences (Fig. 3d), whereas superimposition of the same structure of KDM5B (PDB ID5A1F) with KDM4A (PDB ID 2P5B) reveals a more similar shape of the 2OG-binding pocket (Fig. 3e) where the most significant differences are Trp 486 and Ala 599 .
In KDM4A (PDB ID 2OQ6) Asn 86 , Asp 311 , Asp 135 , Lys 421 and Glu 169 interact with the H3K9me3 peptide and the different residues found in KDM5B could possibly reflect the different substrate specificity i.e. methylated H3K4 versus H3K9 as in KDM4A 23 .
Domain arrangement of KDM5B determined by SAXS -Building on the high-resolution models of the catalytic core, we next characterised the KDM5B-c1 construct, carrying the additional PHD1 and ARID domains by small-angle X-ray scattering (SAXS) and rigid-body modelling. To study the domain arrangement in this construct we determined the SAXS solution structure of the KDM5B-c1 and KDM5B-c4 constructs (Fig. 3f Table 5). In contrast to the KDM5B structures however, we do not observe in the KDM5C structure electron density for Arg 98 or the beta strands that correspond to β5, β12, or the loop connecting β5 and β6, again suggesting that this region is highly flexible.
The structural relationships between KDM5B and the KDM4 and KDM6 subfamilies are also reflected by the selectivity profile of compound KDM5-C49. Calculation of Ki values from the IC50 determinations 28,29 reveals 25-150 -fold differences between KDM5 and KDM6 members, whereas the close relationship to the KDM4 family is indicated by 2-30 fold differences in Ki (Supplementary Table 4). To investigate this aspect further, the inhibitor was co-crystallised with the KDM6-related member UTY 22 as well as with KDM4A 17 (Supplementary Table 5). The structures reveal several decisive similarities and differences across these Jmj subfamilies (Supplementary Fig. 13 Fig. 13). Comparison of UTY·KDM5-C49 with the structure of KDM6A (PDB ID 3AVR), in complex with NOG and H3K27me3 peptide (Supplementary Fig. 13), reveals that the pyridine ring overlaps as expected with the NOG position, whereas the dimethyl amino ethyl portion in KDM5-C49 is located within the methylated H3K27me3 binding pocket. This is further supported by an overlay of the KDM5B·KDM5-C49 complex structure with the related plant demethylase JMJ703 with a substrate H3K4me3 peptide (PDB ID 4IGQ) (Supplementary Fig. 13). Assuming similar interactions are formed with the highly conserved KDM6B and KDM4C enzymes, we hypothesize that the somewhat fewer interactions detected in the KDM5-C49 UTY complex may explain the lower inhibitory activity, whereas the similar interactions observed with KDM4A could account for the low selectivity between KDM4C and KDM5B. However, in the absence of structural data for KDM5-c1 constructs that contain the ARID and PHD1 domains, which were used for the inhibitor assays, we cannot rule out that further stabilising interactions may be formed between the ethylamide portion and the additional elements present in the KDM5B-c1 and other KDM5 protein constructs. With this caveat further interpretations about structurefunction relationships appear difficult.
In order to further map the relationships between KDM4, 5 and 6 subfamilies we used the previously identified inhibitor GSK-J1 21 , a KDM6 and KDM5 subfamily specific inhibitor ( Fig. 4b and Supplementary Fig. 11). In KDM5B-c4 GSK-J1 is bound to the 2OG-binding site in an almost identical mode as previously observed with UTY·GSK-J1 22 and JmjD3/KDM6B·GSK-J1 21 . The heterocyclic biaryl ring system of GSK-J1 makes bidentate interactions with the catalytic metal, facilitated by a side-chain movement of Arg 98 to accommodate the inhibitor molecule. As also observed in the KDM6 enzymes, the Mn 2+ ion translocates ∼2.0 Å away from the HXE …H metal chelation triad with GSK-J1 as compared to other structures, and this movement results in an indirect water-bridge interaction between His 587 and the Mn 2+ ion, confirming the previously observed elasticity of the active site metal position in 2OG oxygenases 21,22 . The tetrahydrobenzazepine part of GSK-J1 has a bent conformation and is located in a cavity between β-strands 5 and 9, and this part of GSK-J1 has previously been shown to occupy the peptide-binding site in KDM6B 21 and UTY (KDM6C) 22 potentially explaining the higher selectivity towards the KDM6 enzymes.
Whereas the ligand structures of KDM5B with KDM5-C49 and GSK-J1 highlight the overlapping selectivity of the inhibitors with the KDM4 and KDM6 families, the structure of KDM5B with GSK467 (compound 5) provides a possible template for selective KDM5B inhibitor development. GSK467 has been disclosed by GlaxoSmithKline (GSK) as part of a sub-micromolar inhibitor series for the KDM4 family and KDM5C with cellular activity (IC-50 <10uM) in cellular imaging assays 30 . In our AlphaScreen assay GSK467 shows a calculated KI value of 10 nM for KDM5B (Supplementary Table 4) with an apparent 180fold selectivity to KDM4C and no measurable inhibitory effects towards KDM6 or other JmJ members (Supplementary Fig. 14). GSK467A was co-crystallised with KDM5B revealing that GSK467 ( Fig. 4c and Supplementary Fig. 11) is found in the 2OG cofactor-binding pocket, where the inhibitor makes a mono-dentate interaction with the catalytic metal via its pyrido-nitrogen, with the two remaining coordination sites occupied by water molecules. The pyrido [3,4-d]pyrimidine-4(3H)-one forms hydrophobic interactions with Trp 519 , Phe 496 and Tyr 488 , and the pyrimidin-4(3H)-one oxygen forms a polar interaction with Lys 517 (Fig. 4c).
The benzyl ring is located in a cavity formed by Tyr 425 , Gln 88 , Ala 426 and Arg 98 , and induces a movement of Arg 98 and the loop containing the Leu 101 -Ala 373 fusion site away from the DSBH-fold (Supplementary Fig. 7). Thus, it is possible that further interactions are formed between this part of the molecule and the additional domains present in the KDM5B-c1 construct.

The inhibitor KDM5-C70 increases global H3K4me3 levels -Having identified KDM5-
C49 and GSK467 as potent inhibitor scaffolds for KDM5 enzymes, we next set out to explore their utility in interrogating KDM5 biology. Given the role of KDM5 in cancer biology 13  shows anti-proliferative effects after 7 days of treatment ( Fig. 5a-b; Supplementary Fig. 15) at elevated concentrations (estimated 50% reduction of viability/proliferation for KDM5-C70 ca. 20 µM; for GSK467 >50 µM). Neither KDM5-C70 nor GSK467 increased the rate of apoptosis in myeloma cells (Supplementary Fig. 15) but treatment with KDM5-C70 decreased the level of phosphorylation of retinoblastoma protein (pRb) compared with both the vehicle control and GSK467 ( Fig. 5c and Supplementary Fig. 16) whilst leaving the total level of pRb unchanged, indicating impairment of cell cycle progression, and in line with the anti-proliferative effects observed. Chromatin immunoprecipitation followed by next generation sequencing (CHIPseq) showed an increase in H3K4me3 levels around transcription start sites with KDM5-C70 but little change with GSK467A (Fig. 6a) at 50 µM inhibitor concentrations. When all H3K4me3 peaks are considered, which includes promoters and enhancers distant from the transcription start site, the increase seen with KDM5-C70 is more pronounced (Fig. 6b, Supplementary Fig. 17). Analysis between KDM5-C70 and the DMSO control reveals 2872 differentially bound peaks at a false discovery rate of 0.1 ( Supplementary Fig. 18, Supplementary Dataset 1), of which 2728 show increased H3K4me3 with KDM5-C70. Examination of the genomic location of differentially bound peaks shows a modest number of peaks close to transcription start sites with a larger number being intergenic (Fig. 6c). Gene ontology (GO) analysis of the differential bound peaks increased with KDM5-C70 shows enrichment of terms associated with B cell malignancies whereas the much smaller number of peaks associated with decreased binding on addition of KDM5-C70 are associated with cell cycle progression (Supplementary Table 6). Examples of the transcription start sites of three genes, associated with myeloma biology, are given ( Fig. 6d-f).

DISCUSSION
Although their precise physiological roles are still incompletely understood, the KDM5 family members have attracted attention due to their important roles in stem cell biology, development and oncology 13 12,13 . Although the apparent lack of increasing H3K4me3 levels using GSK467 cannot be explained at present, it is however correlated to a weak anti-proliferative effect. The specific perturbation of the ethyl-ester pro-drug KDM5-C70 leading to the observed cell cycle arrest in the multiple myeloma cell system is likely to be related to the observed dysregulation of cell cycle and metabolic genes.
Although the cellular potency of KDM5-C70 is weak and about one to two orders of magnitude beyond a desired cellular potency characterising high-quality chemical probes 36       For display of electron density and stereoviews for ligands see Supplementary Fig.11). Sidechains are displayed in cyan, water molecules as red spheres, and the metal centre as green sphere.   protease cleavable C-terminal 10x-histidine tag (KDM5C). The KDM5C deletion constructs were generated as described for KDM5B-c4 (residue Phe 8 -Thr 772 , Leu 83 -Gly 384 deleted with insertion of a 4x-glycine linker). All constructs were confirmed by Sanger DNA sequencing.

Protein expression and purification -Recombinant KDM5B constructs were expressed in
Sf9 cells and generation of recombinant baculo viruses, insect cell culture, and infections were performed according to the manufacturer's instructions (Invitrogen). The cells were collected 72 hours post infection and suspended in a buffer containing 50 mM HEPES pH 7.5, 500 mM NaCl, 10 mM imidazole, 5% glycerol, 0.5 mM TCEP, and a protease inhibitor mix (Calbiochem). All KDM5B variants were purified using nickel affinity chromatography using a stepwise gradient of imidazole. The eluted protein was then incubated with TEV protease at 4 °C overnight followed by size-exclusion chromatography (Superdex 200). The TEV protease and uncleaved protein were removed using nickel affinity chromatography and the mass was verified by electrospray ionization time-of-flight mass spectrometry (ESI-TOF, Agilent LC/MSD). KDM5 proteins for biochemical assays were purified as above but without cleavage of the histidine tag. UTY/KDM6C and JMJD2/KDM4 proteins were expressed and purified as described previously 17,22 whereas KDM2, 3, and 6B are described in detail elsewhere 37 .  Table 2c). All IC50 determinations were performed using 2-OG concentrations at or near the Km value for the respective enzyme and incubation times with substrate were determined from the linear range of enzyme progress curves for respective enzymes (Supplementary   Fig. 6). All reagents were from Sigma Aldrich unless otherwise stated and were of the  Table 2a) and antibodies to product methyl marks were purchased from Abcam, Cell Signalling Technology or Millipore (Supplementary Table   2B).

Activity assays-reagents and conditions:
Demethylase AlphaScreen -The demethylase AlphaScreen assay was performed in 384-well plate format using white proxiplates (Perkin Elmer) and transfer of compound (100 nl) was performed using an ECHO 550 Acoustic Dispenser (Labcyte). After establishment of suitable purification conditions (see above), enzyme samples showed normal distribution of their activities. All subsequent steps were carried out in assay buffer (50 mM HEPES pH 7.5, 0.1% (w/v) bovine serum albumin and 0.01 % (v/v) Tween-20). In brief 5 µl of assay buffer containing demethylase enzyme at 2X final assay concentration (see Supplementary Table   3c for assay specifics) was pre-incubated for 15 minutes with dilutions of compound. The  Table 2c) and was stopped by addition of 5 µl assay buffer containing EDTA (30 mM) and NaCl (800 mM). The final concentration of DMSO was 1%.
Streptavidin Donor beads (0.08 mg/ml) and Protein-A conjugated acceptor beads (0.08 mg/ml) were pre-incubated for 1 hour with anti-methyl mark antibody (4X final assay concentration) and the presence of histone H3 product methyl mark was detected by addition of the pre-incubated AlphaScreen beads (5 µl). Detection was allowed to proceed for 2 hour at room temperature and the assay plates were read in a BMG Pherastar FS plate reader (Excitation 680 nM / Emission 570 nM). Data were normalized to the (no enzyme) control and the IC50 values were determined from the nonlinear regression curve fit using GraphPad Prism 5. Using the kinetic parameters for substrate and cofactor, apparent Ki values were extrapolated from IC50 values using the described relationships between IC50 and Ki values assuming competitive inhibition 28,29 .

Formaldehyde Dehydrogenase (FDH) Coupled Enzyme Assay
The resulting data was used to fit the Morrison equation 29 in GraphPad Prism. All crystals were cryo-protected with mother liquor supplemented with 25% ethylene glycol before they were flash frozen in liquid nitrogen. Data sets were collected on beamlines, I02, I03 or I04-1 at the Diamond Light source UK. Metal contents of crystals were investigated from a KDM5B crystal using X-ray fluorescence scanning on beamline I02 or I03 at Diamond Light Source, UK, on a Vortex-EX fluorescence detector (Hitachi High-Technologies Science). A peak for Zn (observed peak at 9667.14 eV, expected peak at 9658.6 eV) and a peak for Mn (observed peak at 6548.69 eV, expected peak at 6539.0eV) were observed as expected based on the structural work. Small-Molecule KDM5 Inhibitors -KDM5-C49 and KDM-C70 were identified from a patent application (WO 2014053491 A1) and were purchased from Xcessbio or synthesized as described in Supplementary Notes. A sample of GSK467 was kindly provided by GSK and details of its synthesis and profiling are described elsewhere 30 . Signalling. After primary antibody incubation, the membrane was washed three times and incubated with the appropriate HRP conjugated secondary antibody (Cell Signalling) for 1h.

Assessment of cellular activity for GSK467 and KDM5-C70 -
Excess secondary antibody was then rinsed off by washing the membrane three times. The membrane was incubated with SuperSignal West Dura Extended Duration substrate (ThermoScientific) for 5 minutes and luminescence from the membrane was detected using Fuji Medical X-ray film (Fujifilm).