Cancer-associated 2-oxoglutarate analogues modify histone methylation by inhibiting histone lysine demethylases

Article abstract of DOI:10.1016/j.jmb.2018.06.048

Histone lysine demethylases (KDMs) are 2-oxoglutarate-dependent dioxygenases (2-OGDDs) that regulate gene expression by altering chromatin structure. Their dysregulation has been associated with many cancers. We set out to study the catalytic and inhibito

Full text of DOI:10.1016/j.jmb.2018.06.048

Accepted Manuscript
Cancer-associated 2-oxoglutarate analogues modify histone
methylation by inhibiting histone lysine demethylases
Tuomas Laukka, Matti Myllykoski, Ryan E. Looper, Peppi
Koivunen
PII:
S0022-2836(18)30431-5
doi:10.1016/j.jmb.2018.06.048
YJMBI 65803
DOI:
Reference:
To appear in:
Journal of Molecular Biology
Received date:
Revised date:
Accepted date:
14 May 2018
26 June 2018
27 June 2018
Please cite this article as: Tuomas Laukka, Matti Myllykoski, Ryan E. Looper, Peppi
Koivunen , Cancer-associated 2-oxoglutarate analogues modify histone methylation by
inhibiting histone lysine demethylases. Yjmbi (2018), doi:10.1016/j.jmb.2018.06.048
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.

ACCEPTED MANUSCRIPT
Cancer-associated 2-oxoglutarate analogues modify histone methylation by
inhibiting histone lysine demethylases
1
1
2
1
Tuomas Laukka , Matti Myllykoski , Ryan E. Looper and Peppi Koivunen
1
Biocenter Oulu, Faculty of Biochemistry and Molecular Medicine, Oulu Center for Cell-Matrix
Research, University of Oulu, FIN-90014 Oulu, Finland
2
Department of Chemistry, University of Utah, Salt Lake City, UT 84112, USA
To whom correspondence should be addressed: Peppi Koivunen, University of Oulu, Faculty of
Biochemistry and Molecular Medicine, Tel: +358 48 294 5822, Fax: +358 8 531 5037, E-mail:
peppi.koivunen@oulu.fi
1

ACCEPTED MANUSCRIPT
ABSTRACT
Histone lysine demethylases (KDMs) are 2-oxoglutarate-dependent dioxygenases (2-OGDDs) that
regulate gene expression by altering chromatin structure. Their dysregulation has been associated with
many cancers. We set out to study the catalytic and inhibitory properties of human KDM4A, KDM4B,
KDM5B, KDM6A and KDM6B, aiming in particular to reveal which of these enzymes are targeted by
cancer-associated 2-oxoglutarate (2-OG) analogues. We used affinity-purified insect cell-produced
enzymes and synthetic peptides with trimethylated lysines as substrates for the in vitro enzyme activity
assays. In addition, we treated breast cancer cell-lines with cell-permeable forms of 2-OG analogues and
studied their effects on the global histone methylation state. Our data show that KDMs have substrate
specificity. Among the enzymes studied, KDM5B had the highest affinity for the peptide substrate but the
2+
lowest affinity for the 2-OG and the Fe cosubstrate/cofactors. R-2-hydroxyglutarate (R-2HG) was the
most efficient inhibitor of KDM6A, KDM4A and KDM4B, followed by S-2HG. This finding was
supported by accumulations of the histone H3K9me3 and H3K27me3 marks in cells treated with the cell-
permeable forms of these compounds. KDM5B was especially resistant to inhibition by R-2HG, while
citrate was the most efficient inhibitor of KDM6B. We conclude that KDM catalytic activity is
susceptible to inhibition by tumorigenic 2-OG analogues and suggest that the inhibition of KDMs is
involved in the disease mechanism of cancers in which these compounds accumulate, such as the
isocitrate dehydrogenase mutations.
Keywords: KDM, fumarate, succinate, R-2HG, S-2HG
Abbreviations: KDM: Histone lysine demethylase; 2-OGDD: 2-oxoglutarate-dependent dioxygenase; 2-
OG: 2-oxoglutarate; R/S-2HG; R/S-2-hydroxyglutarate; TET: Ten-eleven translocation; HIF-P4H:
hypoxia-inducible factor prolyl 4-hydroxylase; DEF: diethyl fumarate; DMS: dimethyl succinate; TFMB-
R/S-2HG: Trifluoromethyl benzyl-esterified R/S-2-hydroxyglutarate; IDH: isocitrate dehydrogenase; FH:
fumarate hydratase; SDH: succinate dehydrogenase.
2

ACCEPTED MANUSCRIPT
INTRODUCTION
Epigenetic modifications of the nucleic acid and protein components of chromatin are central regulators
of eukaryotic transcription. These modifications include the reversible methylation of DNA and histones
1
and further histone modifications such as acetylation, phosphorylation and ubiquitination The chromatin
modifications are written or erased by enzymes, with several of the erasers belonging to the family of 2-
1
-4
oxoglutarate-dependent dioxygenases (2-OGDDs) . The family contains ~70 enzymes in humans,
including a large number of histone lysine demethylases (KDMs), the DNA demethylases, ten-eleven
translocation (TET) enzymes, the hypoxia-inducible factor prolyl 4-hydroxylases (HIF-P4Hs, also known
1
, 2, 4
as EglNs and PHDs) and several enzymes required for collagen biosynthesis
.
The 2-OGDDs share a common reaction mechanism, although the substrate can vary from DNA to
1
, 2, 4
2+
RNA, protein, fatty acid or other small molecules
. 2-OGDDs require Fe as a cofactor and molecular
oxygen and 2-oxoglutarate (2-OG, -ketoglutarate) as cosubstrates, and they yield CO and succinate.
2
Many also need a reducing agent (such as vitamin C) to retain iron in the divalent state. By coupling the
sensing of cellular oxygen and 2-OG levels with that of the iron redox status, the 2-OGDDs act as central
cellular metabolic sensors.
2
-OG is a central intermediate of the Krebs cycle, where it is produced by isocitrate dehydrogenase
(
IDH) isoenzymes 2 and 3. A 2-OG carrier mediates its efflux from mitochondria to the cytosol, where it
can be used for amino acid synthesis or converted via reductive metabolism to acetyl-CoA and used for
fatty acid synthesis. 2-OG can also be generated from several amino acids which can be converted to
glutamate and transaminated to 2-OG. In the cytosol IDH1 provides an additional source of non-
mitochondrial 2-OG.
1
-4
2
-OGDDs can be competitively antagonized by 2-OG analogues . These include the Krebs cycle
intermediates fumarate, succinate, oxaloacetate and citrate, of which fumarate and succinate accumulate
5-8
in fumarate hydratase (FH) and succinate dehydrogenase (SDH) mutant tumors, respectively . IDH
mutants that occur in certain cancers convert 2-OG to R-2-hydroxyglutarate (R-2HG), which is a 2-OG
9-13
analogue oncometabolite capable of targeting 2-OGDDs . Moreover, hypoxia and deficiency in the
3

ACCEPTED MANUSCRIPT
metabolizing enzyme has been shown to cause the accumulation of S-2-hydroxyglutarate (S-2HG), which
14, 15
can also target 2-OGDDs
.
Histone methylation, carried out by histone methyl transferases, occurs principally on the lysyl and
1
arginyl side chains of histones H3 and H4 . Up to three methyl groups can be added to a lysyl residue.
2
, 3
Histone demethylation is carried out by histone demethylases, of which more than 20 are KDMs . These
are classified into two subfamilies one of which is the JmjC domain-containing KDM family that forms
the largest subfamily within the 2-OGDDs. The KDMs are highly specific for the sequence surrounding
the substrate lysine and the methylation status. Histone methylation can enable context-dependent
activation or repression of transcription. In general, the methylation of histone H3 lysine 9 (H3K9) or
H3K27, or histone H4 lysine 20 (H4K20) correlates with silent heterochromatin and transcriptional
repression, whereas methylation of H3K4, H3K36 and H3K79 correlate with active euchromatin and
enhanced transcription. The outcome of the methylation nevertheless depends on its site and extent, the
presence of additional chromatin modifications and many other regulatory factors. KDMs play an
important role in the epigenetic regulation of gene expression involved in development, cellular
1
proliferation and differentiation, senescence and genomic stability . Aberrant patterns of modification to
chromatin have been associated with the onset and progression of many diseases, and recent evidence
1
, 16, 17
links several KDMs to the pathogenesis of a number of cancers
. Therefore many writers and erasers
3
are pursued as small molecule targets for therapeutic benefit .
Despite being high in number, the KDMs remain less well studied than the other 2-OGDDs. We
therefore set out to determine the catalytic properties of several H3K4me3, H3K9me3 and H3K27me3
demethylating KDMs. We also studied the inhibitory properties of the physiologically or pathologically
occurring 2-OG analogues on their activity. Finally, we used the cell-permeable forms of 2-OG analogues
to study their effects on chromatin marks in cellulo.
4

ACCEPTED MANUSCRIPT
RESULTS
Kinetic Analyses of KDMs Reveal Substrate Specificity
In order to determine the catalytic properties of the enzymes, human full-length KDM4A, KDM4B,
KDM5B and KDM6A were expressed in insect cells by means of baculoviruses as FLAG or His-tagged
recombinant proteins and affinity-purified using anti-FLAG or Ni-NTA chromatography. The eluted
fractions were run on SDS-PAGE followed by Coomassie Blue staining (Fig. 1a). Since a baculovirus
coding for the full-length human KDM6B did not yield any soluble recombinant enzyme (data not
shown), a baculovirus coding for the catalytic JmjC domain of KDM6B (aa 1164-1682) with a FLAG-tag
was used instead to obtain this enzyme (Fig. 1a).
The catalytic activities of the KDMs were analysed by a method based on measurement of the
1
4
14
lysine demethylation-coupled stoichiometric release of CO from 2-oxo-[1- C]glutarate (Fig. 1d).
2
Synthetic peptides of length 19-28 residues representing histone H3 and containing a trimethylated lysine
were used as substrates (Table 1). All the purified KDMs were active under these conditions and their
activity was dependent on time (Fig. 1b, c, Fig. S1A-E) and the substrate concentration (Fig. 2a-c,
Fig.S2A-I), suggesting that the method was suitable for determining the catalytic properties of these
enzymes.
Substrate affinity was determined first. KDM4A and KDM4B used H3K9me3 and H3K27me3 as
substrates but not H3K4me3 (Fig. 2a, Table 2, Fig.S2C-F), the K values for these substrates being ~20-60
m

M (Table 2). KDM5B used H3K4me3 as a substrate with a K value of 0.8 M, but it did not
m
demethylate the H3K9me3 or H3K27me3 substrates (Fig. 2b, Table 2). KDM5B also demethylated a
dimethylated form of the H3K4 peptide with a K value of 0.5 M, which was highly similar to that for
m
the trimethylated form (Table 2, Fig. S2B). KDM6A and KDM6B did not use H3K4me3 or H3K9me3 as
substrates but only H3K27me3 (Fig. 2c, Table 2, Fig. S2G-I). Interestingly, KDM6A had a markedly
higher K value for the H3(21-44)K27(me3) substrate than did KDM6B (170 vs. 6 µM, Table 2, Fig.
m
5

ACCEPTED MANUSCRIPT
S2G, I). We therefore determined the K value of KDM6A for a peptide H3(17-44)K27(me3) that was
m
four residues longer and found it to be 130 M (Table 2, Fig. S2H).
To evaluate the turnover rate and substrate specificity, we calculated the k and k /K values for
cat
cat
m
all enzymes and substrates (Table 2). The k values were highly similar for all enzymes ranging from
cat
-1
-1
18
0
.004 to 0.04 s (Table 2) and in the same range than the 0.015 s reported earlier for KDM5B . For
KDM4A and KDM5B there was no difference in the k values between H3K9me3 and H3K27me3 and
cat
H3K4me3 and H3K4me2 substrates, respectively (Table 2). For KDM4B the k value for H3K9me3 was
cat
about double of that for the H3K27me3 substrate (Table 2). For KDM6A the k for the longer
cat
H3K27me3 substrate was about six-fold higher than that for the four residues shorter peptide (Table 2).
-1
-1
The k /K values for the enzymes varied from 0.016 to 0.00004 s M (Table 2). For KDM4A no
cat
m
difference in the specificity between H3K9me3 and H3K27me3 substrates were detected whereas for
KDM4B H3K9me3 was a more specific substrate (Table 2). No major differences were detected in the
specificity of KDM5B between the tri- and dimethylated H3K4 substrates whereas for KDM6A the
longer H3K27me3 substrate was more specific than the shorter one (Table 2). Among H3K27me3
demethylating enzymes KDM6A had the highest specificity for this substrate (Table 2).
The K values of the enzymes studied for the 2-OG cosubstrate were 6-50 µM, KDM4B having the
m
highest affinity and KDM6B the lowest (Fig. 2d, e, Table 3, Fig. S3). These values were similar to those
8
, 19, 20
of HIF-P4H-1, FIH and TET2, which are other 2-OGDDs operating in the nucleus (Table 3)
. The Km
values of the other enzymes studied here for iron (<0.1 µM) were lower than that of KDM6B, which was
2
1
6
µM (Fig. 2f, Table 3) and thus similar to that of TET2 .
KDMs Are Susceptible to Inhibition by 2-OG Analogues
We next studied the inhibition of KDMs by the cancer-associated 2-OG analogues fumarate, succinate, R-
2HG and S-2HG and the metabolic regulators citrate and oxaloacetate. All the recombinant KDM
enzymes studied were susceptible to competitive inhibition by succinate, R-2HG, S-2HG, citrate and
6

ACCEPTED MANUSCRIPT
oxaloacetate (Fig. 3a-d, Table 3, Figs. S5-9). Fumarate was only a weak inhibitor of KDM4A and
KDM6A, with IC values of 2300 and 3000 µM (Fig. S4A, B), respectively, and did not inhibit KDM4B,
50
KDM5B or KDM6B (Table 3). The oncometabolite R-2HG appeared to be the most potent inhibitor of
KDM4A, KDM4B and KDM6A among the 2-OG analogues studied, with IC values of <200 µM (Fig.
5
0
3a, Table 3, Fig. S5A, D), and R-2HG was also the second most potent inhibitor of KDM6B, with an IC50
value of 350 µM (Table 3, Fig. S5E). Unlike the other KDMs studied, R-2HG was only a very weak
inhibitor of KDM5B, with an IC value of 3600 µM, which was also the case with its S-enantiomer
50
(
Table 3, Fig. S5C, Fig. S6C). On the other hand, S-2HG was the most efficient inhibitor of KDM6A
among the 2-OG analogues studied, with an IC value of 180 µM, and the second and third most efficient
5
0
inhibitor of KDM4A and KDM4B, respectively, with IC values of 290 and 450 µM (Fig. 3b, Table 3,
5
0
Fig. S6B). Succinate was the third and fourth most potent inhibitor of KDM6A and KDM6B,
respectively, with IC values of 270 and 550 µM (Fig. 3c, Table 3, Fig. S7E), while citrate was the most
50
potent inhibitor of KDM6B, with an IC value of 130 µM (Fig. 3d, Table 3), this being lower than the
50
concentration of 2-OG in the assay. For the other KDMs studied, the IC values of citrate were 290-
5
0
1600M (Table 3, Fig. S8). It was also the most potent inhibitor of KDM5B among the 2-OG analogues
studied, with an IC value of 800 µM (Table 3, Fig. S8C). The IC values of all the KDMs studied for
50
50
oxaloacetate were 380-900 M (Table 3, Fig. S9), which was the smallest variation among the IC values
5
0
of the KDMs studied here for any of the 2-OG analogues.
Treatment of Breast Cancer Cells with Cell-Permeable 2-OG Analogues Increases Global H3K27
and H3K9 Methylation Levels
To evaluate the potential of the cancer-associated 2-OG analogues to inhibit KDMs in cellulo, we treated
three subtypes of breast cancer cells, BT474 (ER+, HER2+), SK-BR-3 (ER-, HER2+) and MCF7 (ER+,
HER2-) with cell-permeable forms of fumarate (DEF), succinate (DMS), R-2HG (TFMB-R-2HG) and S-
2HG (TFMB-S-2HG), the cells being exposed to increasing concentrations of these for 72 h. DMSO-
7

ACCEPTED MANUSCRIPT
treated cells were used as controls. These compounds have been earlier reported to increase intracellular
8
,20,22
concentrations of fumarate, succinate, R-2HG and S-2HG
. Changes in global histone methylation
with respect to H3K9 and H3K27, the marks demethylated by KDM4A, KDM4B, KDM6A and KDM6B,
which were the most susceptible to inhibition by 2-OG analogues (Table 3), were determined by
immunoblotting. HIF-1 stabilization was also assessed in SK-BR-3 and MCF7 cells. Fumarate, R-2HG
and S-2HG significantly increased both H3K27me3 and H3K9me3 levels in all three cell lines (Fig. 4),
but the increases were more modest in the SK-BR-3 cells than in the BT474 and MCF7 cells (Fig. 4).
Succinate slightly increased H3K27me3 and H3K9me3 levels in the BT474 and SK-BR-3 cells but these
changes were modest by comparison with the other compounds (Fig. 4a, b). The most prominent effect of
succinate was seen on H3K9me3 levels in the BT474 cells (Fig. 4a). R-2HG and S-2HG treatment
showed dose-dependent increases in H3K27me3 and H3K9me3 levels in all the cell lines studied (Fig. 4).
Fumarate, succinate, R-2HG and S-2HG stabilized HIF1 in SK-BR-3 and MCF7 cells (Fig. S10).
Altogether, these data show that increased intracellular 2-OG analogue concentrations can lead to
increased H3K27 and H3K9 methylation. The differences in sensitivity observed in the different cells
likely stem from differences in the endogenous expression levels of histone demethylases and
methyltransferases, and the differences in the genomic background of the cells.
DISCUSSION
We report here on the catalytic and inhibitory properties of the high molecular weight full-length insect
cell affinity-purified human enzymes KDM4A, KDM4B, KDM5B and KDM6A and the JmjC domain of
KDM6B using the same assay detecting demethylation-coupled 2-OG decarboxylation in all cases. This
setting provides a basis for comparing the results between the enzymes and for gaining an understanding
of which enzymes are actually inhibited by the 2-OG-analogue oncometabolites and are causally linked to
the transformed phenotype. Previous reports have concentrated on kinetic and inhibitory data for the
catalytic domains of human KDM4A and KDM5B and full-length KDM5B, these data having been
8

ACCEPTED MANUSCRIPT
generated by methods that detect substrate demethylation by MALDI-TOF-MS or antibody-based
AlphaScreen or formaldehyde dehydrogenase (FDH) assay, which detects formaldehyde generated by
2, 18, 23-29
substrate hydroxylation and subsequent demethylation
. Of the available methods MALDI-TOF-
MS is the only direct method available to detect the demethylated product. The AlphaScreen technology
relies on the product-recognizing antibodies. Therefore any unspecificity in them is a weakness of this
method. Although being indirect the formaldehyde detected in the FDH assay stems directly from the
hydoxylated product. This assay measures continuous formation of formaldehyde, unlike the others that
measure product at end point. The substrate demethylation-coupled 2-OG decarboxylation assay used
14
14
here measures C-labeled CO derived from the C-labeled 2-OG providing high specificity and
2
sensitivity. As the 2-OGDDs can catalyse to a small extent uncoupled decarboxylation of 2-OG negative
control reactions where the peptide substrate is omitted must be performed in parallel with ones
containing the substrate. The benefit of the indirect assays compared with direct ones is that the one assay
can be used for all enzymes that use the same reaction method. For example, the AlphaScreen requires a
specific antibody for all products.
Our data show that the catalytic properties among the studies KDMs are highly similar, yet there
are specificities among the enzymes. KDM4A and KDM4B demethylated peptides with trimethylated
lysines in positions 9 and 27 in histone H3, while the other enzymes studied were specific for a single
substrate. The former is in agreement with a previous observation, but our data do not suggest any
difference in the affinity of the peptides representing the two demethylation sites in histone H3 for
KDM4A, unlike the 20-fold higher K value for H3K27me3 compared with H3K9me3 reported by others
m
2
.
Our K value of 20 M for KDM4A for H3K9me3 is in agreement with those of 23-71 M reported
m
2
, 24, 26
elsewhere for the catalytic domain of the enzyme
. The K values for the two substrates studied for
m
KDM4A and KDM4B were highly similar, KDM4B having a slightly lower affinity and higher
specificity for the H3K27me3 peptide than for H3K9me3. The KDM4 paralogues A, B and C have also
3
0
been reported to accept H3K36me3 and H3K36me2 as substrates . Among the enzymes studied here,
KDM5B showed the highest affinity and specificity for its substrates. The K value of 0.8 M for the
m
9

ACCEPTED MANUSCRIPT
trimethylated H3K4 was very similar to the values of 0.6 and 4.0 M reported by others using the FDH-
assay and peptides representing the same demethylation site as a substrate, but with only the catalytic
27, 28
domain of the enzyme as a catalyst
. The K value for the dimethylated H3K4 substrate, 0.5 M, was
m
identical to that reported for the catalytic domain of KDM5B when detected by substrate demethylation
18
with MALDI-TOF-MS . However, the K value reported for a 17-residue H3K4me3 peptide in the case
m
2
9
of KDM5B when using substrate demethylation with AlphaScreen was 15 nM . Among the present
enzymes, KDM6A had the lowest affinity for its H3(21-44)K27me3 substrate, the figure of 170 M being
significantly different from the 6 M obtained for KDM6B. Since it has been reported on the basis of
structural studies that amino acids 17-21 of histone H3 contribute to the substrate binding of KDM6A, we
3
1
tested a longer peptide that included these amino acids as a substrate and found that its K value was
m
lower, 130 M. It was nevertheless still the highest among the enzymes studied here, however, its kcat
value for the longer peptide was highly similar if not identical with KDM6B. Altogether, these data
suggest that the assay used, the substrate length and the catalyst length (full-length vs. catalytic domain)
may affect the kinetic values obtained for KDMs. As the KDMs can also demethylate dimethylated
lysines, and in some cases even monomethylated ones, the product of an assay using a trimethylated
2
, 30
peptide as a substrate may serve as a substrate for a second catalytic reaction . Also, the affinities for
differently methylated substrates may vary, adding further complexity to the interpretation of the results.
The affinities for 2-OG and iron were highly similar for all the KDMs studied except for KDM6B,
2
+
which had a >3-fold higher K value for 2-OG and a >60-fold higher value for Fe than any other
m
enzyme. Such findings would suggest that when these reactants become limiting, KDM6B activity is the
first to decline. Among the nuclear 2-OGDDs, the kinetics of KDM6B resembled those of TET2, whereas
HIF-P4H-1 and FIH possess values that are more similar to those for the other KDMs (Table 3). While
affinities of the KDMs for iron have not been reported before, our 2-OG K values for KDM4A and
m
KDM5B are in agreement with the figures of 3-15 M reported by others for the catalytic domains of the
18, 24, 27
corresponding enzymes obtained by different methods
. Intracellular 2-OG levels are estimated to
10

ACCEPTED MANUSCRIPT
23
be in the low millimolar range , i.e. higher than the reported K values of the KDMs. The concentration
m
of 2-OG in the nucleus is not known, however, and the simultaneous presence of 2-OG and 2-OG
analogues in cellulo probably alters the in vivo K values of enzymes relative to the in vitro values,
m
rendering them higher, and therefore 2-OG levels may under certain conditions be limiting for the full
3
2
activity of KDMs. Iron deficiency has been shown to stabilize HIF1 , and as the KDM6B K value of 6
m
1
9

M for iron was much larger than that of the HIF-P4Hs that regulate HIF stability, this could be
expected to be of some physiological relevance.
Somatic mutations in IDH1 and IDH2 identified in gliomas, acute myeloid leukaemia,
chondrosarcoma and cholangiosarcoma, and less commonly in bladder cancer, for example, lead to the
9
-11, 33-35
accumulation of R-2HG at millimolar levels
. Among the 2-OG analogues studied here, R-2HG
was the most potent antagonist of KDM4A, KDM4B, KDM6A and KDM6B, suggesting that the
oncogenesis promoted by R-2HG involved inhibition of histone demethylation and especially
accumulation of H3K27me3 and H3K9me3. This is in agreement with the reported accumulation of these
36
marks in IDH1/2 mutant glial tumors . Treatment of breast cancer cells with cell-permeable R-2HG
resulted in concentration dependent-accumulation of the H3K27me3 and H3K9me3 marks, suggesting
that R-2HG can also efficiently target the enzymes modifying these markers in cellulo. R-2HG was a
more potent or equally potent inhibitor of the KDMs studied here than was S-2HG for which levels up to
3
7
3
00 M haven been reported in hypoxic mammalian cells and 750 M in S-2HG dehydrogenase mutant
3
8
patients . This is contrary to our earlier report for other 2-OGDDs such as FIH, TET1/2, collagen prolyl
1
3
4
-hydroxylase and HIF-P4Hs, for which S-2HG was always a more efficient inhibitor than R-2HG . For
the HIF-P4Hs, R-2HG is actually not an inhibitor at all but can act as a weak alternative cosubstrate
13
instead of 2-OG . These data suggest potential structural differences in the binding pocket for 2-OG or its
competitive analogues between KDMs and other 2-OGDDs.
5
, 7
Succinate, which can accumulate up to millimolar levels in SDH mutant paragangliomas , was a
moderately potent inhibitor of KDM6A and KDM6B, which is in agreement with the reported
11

ACCEPTED MANUSCRIPT
3
9
accumulation of the H3K27me3 mark in SDH mutant paragangliomas and pheochromocytomas .
Treatment of breast cancer cells with cell-permeable succinate nevertheless only resulted in minor
increases in the H3K27me3 marks, a higher accumulation of H3K9me3 being detected in one cell line.
These observations may suggest cell-type specific differences in the epigenetic marks mediated by
succinate. Among the 2-OG analogues studied here, fumarate, which accumulates up to millimolar levels
6, 7
in FH-mutant renal cancers in particular , was the poorest inhibitor of the KDMs with IC values of ≥
5
0
2
.3 mM. Despite the high IC values, cell-permeable fumarate increased the accumulation of the
50
25
H3K27me3 and H3K9me3 marks in the breast cancer cells, in agreement with data reported elsewhere .
Altogether, these data suggest that fumarate may mediate its effects in these histone marks indirectly and
in a manner independent of KDM inhibition.
Surprisingly, citrate, a central regulator of lipid and glucose metabolism, appeared to be a very
potent inhibitor of KDM6B and a moderate inhibitor of KDM6A. Elevated plasma citrate levels up to 160
4
0

M have been reported associated e.g. with non-alcohol fatty liver disease . Interestingly, KDM6A
activation and H3K27me3 demethylation have been associated with brown adipocyte thermogenetic
4
1
programming, which is a mechanism by which rodents combat obesity . Moreover, ATP-citrate lyase, an
4
2
enzyme that converts citrate to acetyl-CoA, is required for another modification, histone acetylation . It
has also been reported recently that chronic exposure of the heart to R-2HG, which causes contractile
dysfunction, contributed to increased intracellular citrate concentration and modifications to the cardiac
4
3
epigenome . Oxaloacetate, a central regulator of several key biosynthetic pathways, such as
gluconeogenesis and the synthesis of several essential amino acids, was a moderate inhibitor of KDM4A,
KDM6A and KDM6B and its IC values did not exceed the millimolar level for any of the KDMs
50
studied. Altogether, these data may provide an understanding of the mechanisms by which environmental
factors such as diet can influence epigenetic regulation.
Although genetic alterations in KDMs have been identified in various human cancers, the precise
1
, 16, 17
role of KDMs in tumorigenesis remains unclear
. In some cancers the alterations lead to down-
regulation of the corresponding enzyme and in others to upregulation. Of the KDMs studied here,
12

ACCEPTED MANUSCRIPT
inactivating mutations have mainly been reported for KDM6A in lung, liver, esophageal, renal and
bladder cancer, multiple myeloma and leukaemia, while KDM5B expression has been reported to be
4
4
amplified in prostate, breast and bladder cancer . Our inhibitory data support these disease mechanisms
in that the catalytic activity of KDM6A, unlike that of KDM5B, was significantly more susceptible to
inhibition by 2-OG-analogue oncometabolites. The tumor suppressor role of KDM6A has been further
4
5
supported by its ability to prevent epithelial to mesenchymal transition in breast cancer stem cells . For
KDM4A both activating and inactivating mutations have been reported, in breast cancer and bladder
4
4
cancer, respectively, while for KDM4B and KDM6B only activating mutations have been found . The
association of upregulation of KDM4A, KDM4B, KDM5B and KDM6B with certain cancers has made
3
them targets for developing inhibitors to treat these diseases . Our present data shed light on the
mechanism behind the increased histone methylation levels observed in many cancers that involve FH,
SDH and IDH mutations, but they also support the concept that it should be possible to develop specific
inhibitors for the various KDMs, as they possess catalytic properties that vary from one another and are
distinct from those of other 2-OGDD.
MATERIALS AND METHODS
Expression and Purification of Recombinant Enzymes
The baculoviruses for human full-length KDM5B and KDM6A with C-terminal FLAG tags were a gift
from Dr. Qin Yan. cDNA clones for human full-length KDM4A and KDM4B in the pReceiver-I01
plasmid were a gift from Dr. Susanne Schlisio, and the corresponding baculoviruses with N-terminal His-
tags were generated. The baculoviral construct for human KDM6B 1164-1682 (coding for the catalytic
JmjC domain) was amplified by PCR using the corresponding cDNA as a template. The C-terminal
FLAG tag was generated by PCR and the constructs were cloned into the pVL1393 backbone. The
corresponding baculovirus was generated using the BacMagic-3 DNA kit (Novagen). The baculoviruses
were used to express recombinant proteins in Sf9 insect cells in TNM-FH media supplemented with 10%
fetal bovine serum. The cells were infected with the respective baculoviruses and harvested 72 h after
13

ACCEPTED MANUSCRIPT
infection, washed with PBS, and homogenized in a buffer containing 10 mM Tris-HCl (pH 7.8), 150 mM
NaCl, µmM glycine, 5 µM FeSO , 0.1% Triton X-100 and EDTA-free protease inhibitor (Roche).
4
Soluble fractions of the FLAG-tagged enzymes were purified with an anti-FLAG M2 affinity gel (Sigma).
For His-tagged KDM4A and KDM4B, the soluble fractions were affinity purified using Ni-NTA agarose
(
Qiagen) and eluted with imidazole, which was then removed with PD-10 columns (Sigma). The fractions
collected were analysed using 8% SDS-PAGE under reducing conditions followed by Coomassie Blue
staining.
Kinetic Activity Assays
The catalytic activities of the KDM enzymes were determined using an assay in which the stoichiometric
lysine demethylation-coupled decarboxylation of 2-OG to CO was detected. The assay was modified
2
2
1, 46-48
from one used to determine the catalytic activities of several 2-OGDDs
out in a final volume of 50 µl and contained 50 mM Tris-HCl (pH 7.8), 2 mg/ml bovine serum albumin
Roche), 60 g/ml catalase (Sigma), 0.1 mM dithiothreitol, 2 mM sodium ascorbate, 50 µM iron(II)
. The reactions were carried
(
1
4
sulphate, 10% (v/v) dimethyl sulphoxide (DMSO), 2-oxo[1- C]glutarate (Perkin-Elmer) and peptide
substrates representing histone H3(1-21)K4me3 (Anaspec), H3(1-21)K4me2 (Anaspec), H3(1-19)K9me3,
H3(21-44)K27(me3) (Innovagen) and H3(17-44)K27(me3) (Innovagen) and 0.2-1.5 M of the respective
2
+
enzyme. K values for the substrates, 2-OG and Fe were determined by varying the concentration of the
m
component in question while keeping the concentrations of the others saturating and constant. In the IC50
determinations a four-fold 2-OG concentration relative to the 2-OG K value of the respective enzyme
m
was used. These were 50 M for KDM4A, 20 M for KDM4B, 40 M for KDM5B, 37.5 M for
KDM6A and 200 M for the KDM6B JmjC domain. Saturating concentrations of the other
cofactors/cosubstrates were used in the presence of increasing concentrations of the respective 2-OG
analogue in the reaction mixture. The reaction time was 15 or 20 min at 37°C in an ambient O2
concentration, after which the reactions were stopped by adding 100 µl of 1M potassium phosphate pH 5
14

ACCEPTED MANUSCRIPT
14
and the amount of C-labelled CO generated was scintillated. K and V values were determined from
2
m
max
the Michaelis-Menten saturation curves and Lineweaver-Burk plots using Excel (Microsoft). The kcat
values were calculated from the Vmax values taking in account the enzyme concentration and purity. IC50
values were determined from the saturation curves using Excel (Microsoft).
Cell Culture Experiments with Cell-Permeable Fumarate, Succinate, R-2HG and S-2HG
Diethyl fumarate (DEF) (Sigma) and dimethyl succinate (DMS) (Sigma) were purchased. Trifluoromethyl
benzyl (TFMB)-esterified R-2-hydroxyglutarate (TFMB-R-2HG) and TFMB-S-2HG were synthesized as
22
described before . BT474 and SK-BR-3 cells were maintained in RPMI 1640 medium (Sigma). MCF7
cells were maintained in Dulbecco’s minimal essential medium (Biochrom) supplemented with 10 g/ml
insulin. All media were supplemented with 10% fetal bovine serum (Biowest). Cells were seeded 5-6 h
prior to the initiation of the treatment, at 35-45% confluence and treated with increasing concentrations of
DEF, DMS, TFMB-R-2HG or TFMB-S-2HG. The compounds were diluted in DMSO and control cells
were treated with an equivalent volume of DMSO. The treatments lasted for 72 h and equal volumes and
concentrations of the compounds were added to the cells every 24 h.
Immunoblotting
Histone fractions of the cells were obtained by lysing them in a buffer containing 250 mM sucrose, 60
mM KCl, 15 mM NaCl, 15 mM Tris (pH 7.5), 5 mM MgCl , 1 mM CaCl , 1 mM DTT, 10 mM butyric
2
2
acid, 0.3 % (v/v) NP-40 and protease inhibitor mixture without EDTA (Roche Applied Science). The
lysed cells were centrifuged, the supernatant was discarded and the cell pellets were resuspended and
incubated in 0.2 M HCl overnight. The samples were again centrifuged and the supernatant containing the
histones was collected. The resulting histone fractions were analysed in SDS-PAGE under reducing
concentrations followed by Western blotting with polyclonal rabbit antibodies against H3K27me3
(
Millipore, 07-449), H3K9me3 (Abcam, ab8898) and total histone H3 (Abcam, ab1791). Polyclonal goat
anti-rabbit immunoglobulin (Agilent) was used as a secondary antibody, followed by ECL
15

ACCEPTED MANUSCRIPT
immunodetection. For HIF1 detection whole cell lysates were prepared by lysing the cells in 50 mM
Tris (pH 8.0), 150 mM NaCl and 0.5% NP40. The extracts were analyzed in SDS-PAGE under reducing
conditions followed by Western blotting with monoclonal antibody against human HIF1 (BD
Biosciences), followed by ECL immunodetection. Western blotting for β-actin (Novus, NB600-501) was
used as a loading control.
Acknowledgments:
This work was supported by Academy of Finland Grants 266719 and 308009 (PK), and by grants from
the S. Jusélius Foundation (PK), the Emil Aaltonen Foundation (PK and TL), the Finnish Medical
Foundation (TL), the Jane and Aatos Erkko Foundation (PK) and the Finnish Cancer Organizations (PK).
We thank T. Aatsinki and E. Lehtimäki for their excellent technical assistance. The authors declare that
they have no conflicts of interest regarding the contents of this paper.
References
1
2
.
.
Kaelin W. G.,Jr, McKnight S. L. (2013) Influence of metabolism on epigenetics and disease. Cell
53, 56-69.
1
Williams S. T., Walport L. J., Hopkinson R. J., Madden S. K., Chowdhury R., Schofield C. J. &
Kawamura A. (2014) Studies on the catalytic domains of multiple JmjC oxygenases using peptide
substrates. Epigenetics 9, 1596-1603.
3.
McAllister T. E., England K. S., Hopkinson R. J., Brennan P. E., Kawamura A. & Schofield C. J.
(
2016) Recent Progress in Histone Demethylase Inhibitors. J. Med. Chem. 59, 1308-1329.
4
5
.
.
Koivunen P., Laukka T. (2018) The TET enzymes. Cell Mol. Life Sci. 75, 1339-1348.
Selak M. A., Armour S. M., MacKenzie E. D., Boulahbel H., Watson D. G., Mansfield K. D., Pan
Y., Simon M. C., Thompson C. B. & Gottlieb E. (2005) Succinate links TCA cycle dysfunction to
oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer. Cell. 7, 77-85.
6.
Isaacs J. S., Jung Y. J., Mole D. R., Lee S., Torres-Cabala C., Chung Y. L., Merino M., Trepel J.,
Zbar B., Toro J., Ratcliffe P. J., Linehan W. M. & Neckers L. (2005) HIF overexpression correlates
with biallelic loss of fumarate hydratase in renal cancer: novel role of fumarate in regulation of HIF
stability. Cancer. Cell. 8, 143-153.
16

ACCEPTED MANUSCRIPT
7.
Pollard P. J., Briere J. J., Alam N. A., Barwell J., Barclay E., Wortham N. C., Hunt T., Mitchell M.,
Olpin S., Moat S. J., Hargreaves I. P., Heales S. J., Chung Y. L., Griffiths J. R., Dalgleish A.,
McGrath J. A., Gleeson M. J., Hodgson S. V., Poulsom R., Rustin P. & Tomlinson I. P. (2005)
Accumulation of Krebs cycle intermediates and over-expression of HIF1alpha in tumours which
result from germline FH and SDH mutations. Hum. Mol. Genet. 14, 2231-2239.
8
9
.
.
Koivunen P., Hirsilä M., Remes A. M., Hassinen I. E., Kivirikko K. I. & Myllyharju J. (2007)
Inhibition of hypoxia-inducible factor (HIF) hydroxylases by citric acid cycle intermediates: possible
links between cell metabolism and stabilization of HIF. J. Biol. Chem. 282, 4524-4532.
Yan H., Parsons D. W., Jin G., McLendon R., Rasheed B. A., Yuan W., Kos I., Batinic-Haberle I.,
Jones S., Riggins G. J., Friedman H., Friedman A., Reardon D., Herndon J., Kinzler K. W.,
Velculescu V. E., Vogelstein B. & Bigner D. D. (2009) IDH1 and IDH2 mutations in gliomas. N.
Engl. J. Med. 360, 765-773.
1
0. Parsons D. W., Jones S., Zhang X., Lin J. C., Leary R. J., Angenendt P., Mankoo P., Carter H., Siu I.
M., Gallia G. L., Olivi A., McLendon R., Rasheed B. A., Keir S., Nikolskaya T., Nikolsky Y.,
Busam D. A., Tekleab H., Diaz L. A.,Jr, Hartigan J., Smith D. R., Strausberg R. L., Marie S. K.,
Shinjo S. M., Yan H., Riggins G. J., Bigner D. D., Karchin R., Papadopoulos N., Parmigiani G.,
Vogelstein B., Velculescu V. E. & Kinzler K. W. (2008) An integrated genomic analysis of human
glioblastoma multiforme. Science 321, 1807-1812.
1
1
1. Dang L., White D. W., Gross S., Bennett B. D., Bittinger M. A., Driggers E. M., Fantin V. R., Jang
H. G., Jin S., Keenan M. C., Marks K. M., Prins R. M., Ward P. S., Yen K. E., Liau L. M.,
Rabinowitz J. D., Cantley L. C., Thompson C. B., Vander Heiden M. G. & Su S. M. (2009) Cancer-
associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739-744.
2. Ward P. S., Patel J., Wise D. R., Abdel-Wahab O., Bennett B. D., Coller H. A., Cross J. R., Fantin
V. R., Hedvat C. V., Perl A. E., Rabinowitz J. D., Carroll M., Su S. M., Sharp K. A., Levine R. L. &
Thompson C. B. (2010) The common feature of leukemia-associated IDH1 and IDH2 mutations is a
neomorphic enzyme activity converting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer. Cell. 17,
225-234.
1
1
3. Koivunen P., Lee S., Duncan C. G., Lopez G., Lu G., Ramkissoon S., Losman J. A., Joensuu P.,
Bergmann U., Gross S., Travins J., Weiss S., Looper R., Ligon K. L., Verhaak R. G., Yan H. &
Kaelin W. G.,Jr. (2012) Transformation by the (R)-enantiomer of 2-hydroxyglutarate linked to
EGLN activation. Nature 483, 484-488.
4. Intlekofer A. M., Wang B., Liu H., Shah H., Carmona-Fontaine C., Rustenburg A. S., Salah S.,
Gunner M. R., Chodera J. D., Cross J. R. & Thompson C. B. (2017) L-2-Hydroxyglutarate
production arises from noncanonical enzyme function at acidic pH. Nat. Chem. Biol. 13, 494-500.
1
1
1
5. Van Schaftingen E., Rzem R. & Veiga-da-Cunha M. (2009) L: -2-Hydroxyglutaric aciduria, a
disorder of metabolite repair. J. Inherit. Metab. Dis. 32, 135-142.
6. Cloos P. A., Christensen J., Agger K. & Helin K. (2008) Erasing the methyl mark: histone
demethylases at the center of cellular differentiation and disease. Genes Dev. 22, 1115-1140.
7. Varier R. A., Timmers H. T. (2011) Histone lysine methylation and demethylation pathways in
cancer. Biochim. Biophys. Acta 1815, 75-89.
17

ACCEPTED MANUSCRIPT
1
8. Tarhonskaya H., Nowak R. P., Johansson C., Szykowska A., Tumber A., Hancock R. L., Lang P.,
Flashman E., Oppermann U., Schofield C. J. & Kawamura A. (2017) Studies on the Interaction of
the Histone Demethylase KDM5B with Tricarboxylic Acid Cycle Intermediates. J. Mol. Biol. 429,
2895-2906.
1
2
2
2
2
9. Hirsilä M., Koivunen P., Xu L., Seeley T., Kivirikko K. I. & Myllyharju J. (2005) Effect of
desferrioxamine and metals on the hydroxylases in the oxygen sensing pathway. Faseb J. 19, 1308-
1310.
0. Koivunen P., Hirsilä M., Kivirikko K. I. & Myllyharju J. (2006) The length of peptide substrates has
a marked effect on hydroxylation by the hypoxia-inducible factor prolyl 4-hydroxylases. J. Biol.
Chem. 281, 28712-28720.
1. Laukka T., Mariani C. J., Ihantola T., Cao J. Z., Hokkanen J., Kaelin W. G.,Jr, Godley L. A. &
Koivunen P. (2016) Fumarate and Succinate Regulate Expression of Hypoxia-inducible Genes via
TET Enzymes. J. Biol. Chem. 291, 4256-4265.
2. Losman J. A., Looper R. E., Koivunen P., Lee S., Schneider R. K., McMahon C., Cowley G. S.,
Root D. E., Ebert B. L. & Kaelin W. G.,Jr. (2013) (R)-2-hydroxyglutarate is sufficient to promote
leukemogenesis and its effects are reversible. Science 339, 1621-1625.
3. Chowdhury R., Yeoh K. K., Tian Y. M., Hillringhaus L., Bagg E. A., Rose N. R., Leung I. K., Li X.
S., Woon E. C., Yang M., McDonough M. A., King O. N., Clifton I. J., Klose R. J., Claridge T. D.,
Ratcliffe P. J., Schofield C. J. & Kawamura A. (2011) The oncometabolite 2-hydroxyglutarate
inhibits histone lysine demethylases. EMBO Rep. 12, 463-469.
2
2
4. Hancock R. L., Masson N., Dunne K., Flashman E. & Kawamura A. (2017) The Activity of JmjC
Histone Lysine Demethylase KDM4A is Highly Sensitive to Oxygen Concentrations. ACS Chem.
Biol. 12, 1011-1019.
5. Xiao M., Yang H., Xu W., Ma S., Lin H., Zhu H., Liu L., Liu Y., Yang C., Xu Y., Zhao S., Ye D.,
Xiong Y. & Guan K. L. (2012) Inhibition of alpha-KG-dependent histone and DNA demethylases by
fumarate and succinate that are accumulated in mutations of FH and SDH tumor suppressors. Genes
Dev. 26, 1326-1338.
2
2
6. Couture J. F., Collazo E., Ortiz-Tello P. A., Brunzelle J. S. & Trievel R. C. (2007) Specificity and
mechanism of JMJD2A, a trimethyllysine-specific histone demethylase. Nat. Struct. Mol. Biol. 14,
689-695.
7. Horton J. R., Liu X., Gale M., Wu L., Shanks J. R., Zhang X., Webber P. J., Bell J. S., Kales S. C.,
Mott B. T., Rai G., Jansen D. J., Henderson M. J., Urban D. J., Hall M. D., Simeonov A., Maloney
D. J., Johns M. A., Fu H., Jadhav A., Vertino P. M., Yan Q. & Cheng X. (2016) Structural Basis for
KDM5A Histone Lysine Demethylase Inhibition by Diverse Compounds. Cell. Chem. Biol. 23, 769-
781.
28. Kristensen L. H., Nielsen A. L., Helgstrand C., Lees M., Cloos P., Kastrup J. S., Helin K., Olsen L.
&
Gajhede M. (2012) Studies of H3K4me3 demethylation by KDM5B/Jarid1B/PLU1 reveals strong
substrate recognition in vitro and identifies 2,4-pyridine-dicarboxylic acid as an in vitro and in cell
inhibitor. Febs J. 279, 1905-1914.
18

ACCEPTED MANUSCRIPT
2
3
9. Sayegh J., Cao J., Zou M. R., Morales A., Blair L. P., Norcia M., Hoyer D., Tackett A. J., Merkel J.
S. & Yan Q. (2013) Identification of small molecule inhibitors of Jumonji AT-rich interactive
domain 1B (JARID1B) histone demethylase by a sensitive high throughput screen. J. Biol. Chem.
288, 9408-9417.
0. Whetstine J. R., Nottke A., Lan F., Huarte M., Smolikov S., Chen Z., Spooner E., Li E., Zhang G.,
Colaiacovo M. & Shi Y. (2006) Reversal of histone lysine trimethylation by the JMJD2 family of
histone demethylases. Cell 125, 467-481.
3
3
1. Sengoku T., Yokoyama S. (2011) Structural basis for histone H3 Lys 27 demethylation by
UTX/KDM6A. Genes Dev. 25, 2266-2277.
2. Peyssonnaux C., Zinkernagel A. S., Schuepbach R. A., Rankin E., Vaulont S., Haase V. H., Nizet V.
&
Johnson R. S. (2007) Regulation of iron homeostasis by the hypoxia-inducible transcription
factors (HIFs). J. Clin. Invest. 117, 1926-1932.
3
3
3. Ward P. S., Patel J., Wise D. R., Abdel-Wahab O., Bennett B. D., Coller H. A., Cross J. R., Fantin
V. R., Hedvat C. V., Perl A. E., Rabinowitz J. D., Carroll M., Su S. M., Sharp K. A., Levine R. L. &
Thompson C. B. (2010) The common feature of leukemia-associated IDH1 and IDH2 mutations is a
neomorphic enzyme activity converting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer. Cell. 17,
225-234.
4. Amary M. F., Bacsi K., Maggiani F., Damato S., Halai D., Berisha F., Pollock R., O'Donnell P.,
Grigoriadis A., Diss T., Eskandarpour M., Presneau N., Hogendoorn P. C., Futreal A., Tirabosco R.
&
Flanagan A. M. (2011) IDH1 and IDH2 mutations are frequent events in central chondrosarcoma
and central and periosteal chondromas but not in other mesenchymal tumours. J. Pathol. 224, 334-
43.
3
3
3
3
5. Cerami E., Gao J., Dogrusoz U., Gross B. E., Sumer S. O., Aksoy B. A., Jacobsen A., Byrne C. J.,
Heuer M. L., Larsson E., Antipin Y., Reva B., Goldberg A. P., Sander C. & Schultz N. (2012) The
cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics
data. Cancer. Discov. 2, 401-404.
6. Lu C., Ward P. S., Kapoor G. S., Rohle D., Turcan S., Abdel-Wahab O., Edwards C. R., Khanin R.,
Figueroa M. E., Melnick A., Wellen K. E., O'Rourke D. M., Berger S. L., Chan T. A., Levine R. L.,
Mellinghoff I. K. & Thompson C. B. (2012) IDH mutation impairs histone demethylation and results
in a block to cell differentiation. Nature 483, 474-478.
7. Intlekofer A. M., Dematteo R. G., Venneti S., Finley L. W., Lu C., Judkins A. R., Rustenburg A. S.,
Grinaway P. B., Chodera J. D., Cross J. R. & Thompson C. B. (2015) Hypoxia Induces Production
of L-2-Hydroxyglutarate. Cell. Metab. 22, 304-311.
3
3
8. Kranendijk M., Struys E. A., Salomons G. S., Van der Knaap M. S. & Jakobs C. (2012) Progress in
understanding 2-hydroxyglutaric acidurias. J. Inherit. Metab. Dis. 35, 571-587.
9. Letouze E., Martinelli C., Loriot C., Burnichon N., Abermil N., Ottolenghi C., Janin M., Menara M.,
Nguyen A. T., Benit P., Buffet A., Marcaillou C., Bertherat J., Amar L., Rustin P., De Reynies A.,
Gimenez-Roqueplo A. P. & Favier J. (2013) SDH mutations establish a hypermethylator phenotype
in paraganglioma. Cancer. Cell. 23, 739-752.
19

ACCEPTED MANUSCRIPT
4
4
0. van de Wier B., Balk J. M., Haenen G. R., Giamouridis D., Bakker J. A., Bast B. C., den Hartog G.
J., Koek G. H. & Bast A. (2013) Elevated citrate levels in non-alcoholic fatty liver disease: the
potential of citrate to promote radical production. FEBS Lett. 587, 2461-2466.
1. Zha L., Li F., Wu R., Artinian L., Rehder V., Yu L., Liang H., Xue B. & Shi H. (2015) The Histone
Demethylase UTX Promotes Brown Adipocyte Thermogenic Program Via Coordinated Regulation
of H3K27 Demethylation and Acetylation. J. Biol. Chem. 290, 25151-25163.
4
4
2. Wellen K. E., Hatzivassiliou G., Sachdeva U. M., Bui T. V., Cross J. R. & Thompson C. B. (2009)
ATP-citrate lyase links cellular metabolism to histone acetylation. Science 324, 1076-1080.
3. Karlstaedt A., Zhang X., Vitrac H., Harmancey R., Vasquez H., Wang J. H., Goodell M. A. &
Taegtmeyer H. (2016) Oncometabolite d-2-hydroxyglutarate impairs alpha-ketoglutarate
dehydrogenase and contractile function in rodent heart. Proc. Natl. Acad. Sci. U. S. A. 113, 10436-
10441.
4
4
4. Park S. Y., Park J. W. & Chun Y. S. (2016) Jumonji histone demethylases as emerging therapeutic
targets. Pharmacol. Res. 105, 146-151.
5. Choi H. J., Park J. H., Park M., Won H. Y., Joo H. S., Lee C. H., Lee J. Y. & Kong G. (2015) UTX
inhibits EMT-induced breast CSC properties by epigenetic repression of EMT genes in cooperation
with LSD1 and HDAC1. EMBO Rep. 16, 1288-1298.
4
4
6. Kivirikko K. I., Myllylä R. (1982) Posttranslational enzymes in the biosynthesis of collagen:
intracellular enzymes. Methods Enzymol. 82 Pt A, 245-304.
7. Hirsilä M., Koivunen P., Gunzler V., Kivirikko K. I. & Myllyharju J. (2003) Characterization of the
human prolyl 4-hydroxylases that modify the hypoxia-inducible factor. J. Biol. Chem. 278, 30772-
30780.
48. Koivunen P., Hirsilä M., Gunzler V., Kivirikko K. I. & Myllyharju J. (2004) Catalytic properties of
the asparaginyl hydroxylase (FIH) in the oxygen sensing pathway are distinct from those of its prolyl
4-hydroxylases. J. Biol. Chem. 279, 9899-9904.
20

ACCEPTED MANUSCRIPT
FIGURE LEGENDS
Fig. 1. Expression, purification and catalytical activity measurements of KDMs. (a) SDS-PAGE and
Coomassie Blue analysis of affinity purified human recombinant KDM4A (4A), KDM4B (4B), KDM5B
(
5B), KDM6A (6A) and KDM6B (6B*). For KDM6B the JmjC catalytic domain was expressed. (b-c)
Time course analysis of KDM4B and KDM6A. KDM4B (b) and KDM6A (c) retained a linear increase in
velocity as a function of time up to 20 min. 100 µM H3K9me3 was used as a substrate in (b) and 200 µM
H3(21-44)K27me3 in (c). (d) Schematic drawing of KDM catalytic reaction and analysis catalytical
14
activity by determination of released CO from the stoichiometric lysine demethylation-coupled
2
decarboxylation of 2-OG to CO2.
Fig. 2. Kinetic analyses of KDMs. Michaelis-Menten curves and Lineweaver-Burk plots (inset) of
KDM4A for H3(21-44)K27me3 substrate (a), KDM5B for H3K4(1-21)me3 substrate (b), KDM6A for
H3(21-44)K27me3 substrate (c), KDM6A for 2-oxoglutarate (d), KDM6B* for 2-oxoglutarate (e) and
KDM6B* for iron (f). *JmjC domain.
Fig. 3. KDMs are susceptible to competitive inhibition by 2-OG analogues. (a-b) IC curves of KDM4B
5
0
and KDM4A for R-2HG and S-2HG, respectively. (c-d) IC curves of KDM6A and KDM6B for
5
0
succinate and citrate, respectively.
Fig. 4. Treatment of cells with cell-permeable forms of fumarate, succinate, R-2HG and S-2HG results in
increased global histone methylation. H3K27me3 and H3K9me3 levels determined by immunoblotting of
BT474 (a), SK-BR-3 (b) and MCF7 (c) cells incubated with increasing concentrations of diethyl fumarate
(
DEF), dimethyl succinate (DMS), TFMB-R-2HG and TFMB-S-2HG for 72 h.
21

ACCEPTED MANUSCRIPT
Table 1. Synthetic peptides representing histone H3 demethylation sites used as substrates for
KDMs.
Substrate
Sequence
Manufacturer
Anaspec
H3K4me3(1-21)
H3K4me2(1-21)
H3K9me3(1-19)
H3K27me3(21-44)
H3K27me3(17-44)
ART-K(Me3)-QTARKSTGGKAPRKQLA-GGK(Biotin)
ART-K(Me2)-QTARKSTGGKAPRKQLA-GGK(Biotin)
MARTKQTAR-K(Me3)-STGGKAPRKQ
ATKAAR-K(Me3)-SAPATGGVKKPHRYRPG-GK(Biotin)
Anaspec
Innovagen
Innovagen,
Anaspec
RKQLATKAAR-(KMe3)-SAPATGGVKKPHRYRPG-GK(Biotin) Innovagen
22

ACCEPTED MANUSCRIPT
Table 2. K , k and k /K values of KDM4A, KDM4B, KDM5B, KDM6A and KDM6B* for
m
cat
cat
m
methylated histone substrates.
H3(21-44)
K27me3
17 ± 4
H3(17-44)
K27me3
-
H3K4me3
H3K4me2
-
H3K9me3
20 ± 10
KDM4A
KDM4B
KDM5B
KDM6A
K (µM)
>1000/inactive
>2000/inactive
0.8 ± 0.3
0.009 ± 0.003 0.008 ± 0.003
0.011
>2000/inactive
m
-1
k (s )
0.005 ± 0.001 0.004 ± 0.0007
cat
-
1
-1
k /K (s M )
0.00025
30 ± 17
0.009 ± 0.003
0.0003
0.00024
60 ± 20
0.004 ± 0.002
0.00007
cat
m
K (µM)
-
-
-
m
-
1
k (s )
cat
-1
-1
k /K (s M )
cat
m
K (µM)
0.5 ±0.3
>2000/inactive >2000/inactive
m
-
1
k (s )
cat
-1
-1
k /K (s M )
0.016
cat
m
K (µM)
-
>2000/inactive
>2000/inactive
170 ± 40
0.007 ± 0.001
0.00004
6 ± 3
0.015 ± 0.0005
130 ± 30
0.04 ± 0.01
0.003
m
-
1
k (s )
cat
-1
-1
k /K (s M )
cat
m
KDM6B* K (µM)
>2000/inactive
-
-
m
-
1
k (s )
cat
-1
-1
k /K (s M )
0.0025
cat
m
Values are means ± S.D. of at least 3 independent assays.
JmjC domain
*
23

ACCEPTED MANUSCRIPT
Table 3. K values of KDM4A, KDM4B, KDM5B, KDM6A and KDM6B* for 2-oxoglutarate and
m
2
+
Fe by comparison with values of selected nuclear resident 2-oxoglutarate-dependent dioxygenases.
References indicate values not determined here.
2+
2
-OG
µM
Fe
KDM4A
KDM4B
KDM5B
KDM6A
KDM6B*
HIF-P4H-1
FIH
15 ± 9
6 ± 3
10 ± 2
8 ± 4
<0.1
<0.1
<0.1
<0.1
6 ± 2
50 ± 208
18
2 ± 0.44
0.03 ± 0.002
0.5 ± 0.2
6
46
25 ± 3
55 ± 20
2
0
20
TET2
4.8 ± 4
Values are means ± S.D. of at least 3 independent assays.
JmjC domain
*
24

ACCEPTED MANUSCRIPT
Table 3. IC values of succinate, fumarate, (R)-, (S)-2-hydroxyglutarate, citrate and oxaloacetate
50
for KDM4A, KDM4B, KDM5B, KDM6A and KDM6B* by comparison with values of selected
nuclear resident 2-oxoglutarate-dependent dioxygenases. References indicate values not determined
here.
Fumarate
2300 ± 200
Succinate
800 ± 150
R-2HG
S-2HG
290 ± 20
450 ± 130
1600 ± 200
180 ± 60
Citrate
800 ± 250
1600 ± 200
800 ± 300
290 ± 40
Oxaloacetate
µM
160 ± 10
150 ± 30
3600 ± 1400
180 ± 30
350 ± 100
not an inhibitor
1100-1500
KDM4A
KDM4B
KDM5B
KDM6A
KDM6B*
HIF-P4H-1
FIH
300 ± 50
700 ± 50
900 ± 100
440 ± 50
>5000
>5000
2300 ± 900
1400 ± 500
270 ± 40
3000 ± 1000
>5000 8
550 ± 808
750 ± 250
130 ± 10
380 ± 80 8
1000 ± 4208
1400 ± 370
not determined
13
13,#
8
120 ± 108
830 ± 540
630
6300 ± 1300
8
13,21
13,21
8
>10000
400 ± 70
>10000
190-300
850 ± 190
>5000
20
20
13
13
20
TET2
570 ± 190
5000
1600
Values are means ± S.D. of at least 3 independent assays.
JmjC domain.
*
#
K value
i
25

ACCEPTED MANUSCRIPT
Highlights




Disturbed histone methylation levels have been found in various types of cancers.
Many tumors contain mutations that lead to accumulation of 2-OG analogues.
2-OG analogues can inhibit several KDMs and increase histone methylation in cells.
Inhibition of KDM activity contributes to disease mechanism in such cancers.
26

Graphics Abstract

Figure 1

Figure 2