The 2-Oxoglutarate-Dependent Oxygenase JMJD6 Catalyses Oxidation of Lysine Residues to give 5S-Hydroxylysine Residues

Article abstract of DOI:10.1002/cbic.201000641

Amino acid analyses reveal that JMJD6-catalysed hydroxylation of RNA-splicing regulatory protein fragments occurs to give hydroxylysine products with 5S stereochemistry. This contrasts with collagen lysyl hydroxylases, which give 5R-hydroxylated products. The work suggests that more than one subfamily of lysyl hydroxylases has evolved and illustrates the importance of stereochemical assignments in proteomic analyses.

Full text of DOI:10.1002/cbic.201000641

DOI: 10.1002/cbic.201000641
The 2-Oxoglutarate-Dependent Oxygenase JMJD6 Catalyses Oxidation of
Lysine Residues to give 5S-Hydroxylysine Residues
Monica Mantri,[a] Nikita D. Loik,[a] Refaat B. Hamed,[a, b] Timothy D. W. Claridge,[a] James S. O. McCullagh,*[a] and
Christopher J. Schofield*[a]
This paper is dedicated to Anthony (Tony) C. Willis on the occasion of his retirement in gratitude for many years of assistance with
amino acid analysis.
Collagen prolyl- and lysylhydroxylases were the first enzymes
catalysing post-translational hydroxylations on the carbons of
proteins to be identified. Both collagen prolyl- and lysylhydrox-
ylases are FeII- and 2-oxoglutarate (2OG)-dependent oxygenas-
es. Subsequently, 2OG oxygenases have been shown to cata-
lyse other protein hydroxylation reactions.[1,2] The aspartyl and
asparaginyl residues of proteins containing epidermal growth-
factor-type domains undergo carbon hydroxylation.[1] Intracell-
ularly localised proteins also undergo hydroxylation on carbon.
Following the identification of the hypoxia-inducible-factor
(HIF) prolyl and asparaginyl hydroxylases,[3] the asparaginyl hy-
droxylase, factor inhibiting HIF (FIH), has been shown to cata-
lyse multiple hydroxylations in ankyrin repeat domains.[3] There
is also evidence for alternative substrates for the HIF prolylhy-
droxylases.[4,5] Further Ne-methyl lysine demethylation is cata-
lysed by a widespread family of histone demethylases[6] (JmjC
subfamily) in a mechanism that probably involves initial hy-
droxylation of the Ne-methyl group.[1,7,9]
dues in splicing regulatory proteins.[13] Here we report that
JMJD6-catalysed C-5 lysyl hydroxylation occurs to give the hy-
droxylysine product with 5S stereochemistry. This result con-
trasts with the collagen lysyl hydroxylases that give products
with the 5R stereochemistry.[14]
Webby et al. have reported that the splicing regulatory pro-
tein LUC7-like2 (LUC7L2) undergoes JMJD6-catalysed C-5 hy-
droxylation at Lys 269 and that a peptide fragment of
LUC7L2267–278 is a JMJD6 substrate.[13] To define the stereochem-
istry of the JMJD6-catalysed product, we began by synthesis-
ing a LUC7L2267–278 peptide with a 2S,5R-hydroxylysine at resi-
due 269, that is, with the 5R stereochemistry as observed in
collagen lysyl hydroxylase catalysis,[14] and comparing it by
means of NMR and MS analyses with
a hydroxylated
LUC7L2267–278 peptide produced by JMJD6 catalysis. To prepare
the synthetic standard, suitably protected 2S,5R-hydroxylysine
was synthesised as previously reported[15] and incorporated by
solid-phase synthesis to give the peptide NPK(OH)RSRSREHRR
(K(OH)=5R-hydroxylysine). For comparison, LUC7L2267–278 was
synthesised and incubated with recombinant human JMJD6
produced in Escherichia coli. The synthetic and enzymatically
produced peptides were then compared by 1D TOCSY
(Figure 1). These analyses, confirmed the regiochemistry of
JMJD6-catalysed hydroxylation as occurring at C-5. However, at
least with the small amount of enzymatically produced materi-
al available, the NMR analyses were unable to unambiguously
distinguish between 5R and 5S products because the differen-
ces in the coupling constants for the relevant positions are
small; for 2S,5R-hydroxylysine, the J values for H-6 and H-6’ are
13.3, 3.1 Hz (H-6, 2.97 ppm) and 13.3, 9.6 Hz (H-6’, 2.73 ppm),
respectively;[16] for 2S,5S-hydroxylysine, the values are 13.1,
1.7 Hz (H6, 2.96 ppm) and 13.1, 10.0 Hz (H-6’, 2.73 ppm).[16] We
therefore carried out amino acid analyses after hydrolysis of
the hydroxylated peptide to investigate the stereochemistry of
JMJD6-catalysed hydroxylation.
There are variations in the regio- and stereochemical specif-
icity of the studied 2OG-dependent hydroxylases that catalyse
post-translational modifications. In the case of collagen prolyl-
hydroxylation, different modifications to the same type of resi-
due are associated with different functions (Scheme 1). For ex-
ample, trans-4-prolyl hydroxylation stabilises the collagen triple
helix,[10] whereas trans-3-prolyl hydroxylation is proposed to
destabilise collagen to enable further modifications. In the
case of the epidermal growth-factor-type domain protein, hy-
droxylation of aspartyl and asparaginyl residues occurs to give
the 3R products,[11] whereas in HIF and ankyrin repeat domains,
hydroxylation of asparginyl residues occurs to give the 3S
products.[5,12] Thus, in addition to identifying the residue that is
oxidised, it is important to define the regio- and stereochemi-
cal outcome of hydroxylation reactions. JMJD6 has been re-
cently reported to catalyse hydroxylation at C-5 of lysyl resi-
Initially, we optimised a method[17] for amino acid analysis
that employs derivatisation of a mixture of all four stereoiso-
mers of racemic C-5 hydroxylysine with 6-aminoquinolyl-N-hy-
droxysuccinimidyl carbamate (AQC), followed by positive ion
electrospray ionisation (ESI) MS analysis. Separation of the two
sets of diastereomers (2S,5S/2R,5R and 2S,5R/2R,5S) was ach-
ieved by optimisation of HPLC conditions. The extracted ion
chromatogram corresponded to the mass of hydroxylysine de-
rivatised on both its amino groups ([M+H]+ 503.1 Da).
[a] M. Mantri,+ N. D. Loik,+ Dr. R. B. Hamed, Dr. T. D. W. Claridge,
Dr. J. S. O. McCullagh, Prof. Dr. C. J. Schofield
Department of Chemistry, Oxford University
Mansfield Road, Oxford OX1 3TA (UK)
Fax: (+44)1865-275674
E-mail: christopher.schofield@chem.ox.ac.uk
[b] Dr. R. B. Hamed
Department of Pharmacognosy, Faculty of Pharmacy, Assiut University
71256 Assiut (Egypt)
[+] These authors contributed equally to this work.
We then carried out acid hydrolysis of the LUC7L2267–278
,
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201000641.
product of JMJD6 catalysis for amino acid analysis. For compar-
ChemBioChem 2011, 12, 531 – 534
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
531

Scheme 1. Different observed regio- and stereochemistries of the hydroxylation products of 2OG-dependent oxygenases in animals. Products of A) collagen
and HIF prolylhydroxylation, B) epidermal growth factor-type domain and aspartyl and asparaginyl hydroxylations and C) collagen and splicing regulatory pro-
tein hydroxylation (this work).
Figure 1. 1H NMR spectra of A) JMJD6-catalysed hydroxylation of LUC7L2267–278 peptide and B) synthetic peptide LUC7L2267–278 containing a 2S,5R-hydroxylysine
residue at position 269. 1D TOCSY spectra of C) the JMJD6-catalysed LUC7L2267–278 product and D) synthetic peptide LUC7L2267–278 containing 2S,5R hydroxyly-
sine; E) 1H NMR spectrum of a 2S,5R-hydroxylysine standard.
ison, modern cow-bone collagen (Femur, type-1) extracted by
standard laboratory procedure[18] was also analysed. Initial
experiments in tris(hydroxymethyl)aminomethane (Tris) buffer
were hampered by its reaction with AQC to give derivatised
buffer. We therefore tested other buffers—(3-(N-morpholino)-
propanesulfonic acid (MOPS), 4-(2-hydroxyethyl)-1-piperazinee-
thanesulfonic acid (HEPES) and phosphate buffers)—MOPS
gave the highest yields and was used in subsequent analyses.
LC-MS analyses after hydrochloric acid-catalysed hydrolysis
and derivatisation of both the JMJD6 and collagen products
revealed peaks at 8.54 min (corresponding to 2S,5S- or 2R,5R-
hydroxylysine) and 8.74 min (corresponding to 2S,5R- or 2R,5S-
hydroxylysine, Figure 2). Although for the collagen sample, the
2S,5R/2R,5S peak was larger, these results could not be used
for stereochemical assignment. This is probably because, under
the acidic hydrolysis conditions, C-2 epimerisation of C-5 hy-
droxylysine occurs through the formation of a six-membered
lactone involving stereochemically favoured 6-exo-trig cyclisa-
tion of the C-5 hydroxy group onto the protected C-2 carbox-
ylic acid (Scheme 2); lactone formation lowers the pKa of the
C-2 hydrogen, thereby enabling epimerisation by enolisa-
tion.[19] To minimise C-2 epimerisation, we developed an
enzyme-catalysed hydrolysis procedure involving initial treat-
ment with trypsin to cleave the amide bond C-terminal to the
hydroxylysine residue, followed by treatment with carboxypep-
tidase-Y to release the C-5 hydroxylated lysine (Scheme 2).
When treated in this manner, collagen gave a single peak cor-
responding to the retention time of derivatised 2S,5R-hydroxy-
532
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 531 – 534

Figure 2. Amino acid analyses comparing the hydroxylysine products of JMJD6 catalysis with that in collagen. A) a) Hydroxylysine product of JMJD6-catalysed
LUC7L2267–278 after treatment with trypsin and carboxypeptidase Y; b) 2S,5R-hydroxylysine; c) mixture of 2R,5R/2S,5S- and 2R,5S/2S,5R-hydroxylysine (58:42, re-
spectively); B) a’) Hydroxylysine product after treatment of collagen with trypsin and carboxypeptidase Y; b’) Collagen standard after HCl-catalysed hydrolysis;
c’) 2S,5R-hydroxylysine; d’) mixture of 2R,5R/2S,5S- and 2R,5S/2S,5R-hydroxylysine (58:42, respectively).
es, which give products with the 2S,5R stereochemistry. Bioin-
formatic studies imply that JMJD6 is more closely related to
the JmjC family of Ne-methyl lysine demethylases and hydroxy-
lases (FIH) than it is to the collagen hydroxylases.[20] Thus,
whilst it seems likely that most, if not all, animal 2OG oxygen-
ases evolved from a common ancestor, at least two different
subfamilies of lysyl hydroxylases might have evolved with dif-
ferent stereochemical selectivities. Given the proximity of the
C-5 and Ne-amino group of lysine, it is possible that an Ne-
methyl demethylases have evolved from a C-5 hydroxylase, or
vice versa.
Experimental Section
Synthesis and purification of peptides: Peptides were synthesised
by using standard methodology.[21] In the case of the peptide con-
taining 5-hydroxylysine, Fmoc hydroxylysine (Boc-oxazolidine) was
synthesised as described by Spetzer.[15] Peptides were purified on a
Vydac C18 Peptide column and a Waters Quattro micro LC-MS
system.
Scheme 2. Outcomes of acid- and enzyme-catalysed hydrolyses of hydroxy-
Sample preparation for NMR analysis: Hydroxylation reactions
lysine-containing peptides. Note the formation of 2S- and 2R-hydroxylysine
were initiated by mixing LUC7L2267–278 peptide (1 mm final concen-
tration), 2OG (5 mm) and FeII (4 mm) in ammonium acetate
(10 mm, pH 7.5) and JMJD6 (200 mm), followed by incubation at
378C. For large-scale hydroxylation 10 mg of peptide was used.
After 2 h, hydroxylation of the peptide was analysed by MALDI-
TOF MS. The assay mixture was further incubated (at room temper-
ature) until MS analysis showed substantial hydroxylation (>90%).
Methanol (200 mL) was then added to each Eppendorf tube to pre-
cipitate protein. The mixture was centrifuged (10 min, 14000 rpm,
14462g), and the decanted solution was freeze-dried, dissolved in
0.1% aqueous formic acid and purified by LC-MS.
products through epimerisation at C-2 in HCl-catalysed hydrolysis[19] and for-
mation of the product with retained C-2S stereochemistry when using the
enzyme-catalysed hydrolysis procedure.
lysine (it is subsequently assumed that the 2S (l) stereochemis-
try is maintained). This result demonstrates that C-5 hydroxyl-
ation does not ablate trypsin activity. In contrast, and un-
expectedly, when the hydroxylated LUC7L2267–278 peptide was
subjected to the same enzyme-catalysed hydrolysis protocol, a
single peak corresponding to 2S,5S-hydroxylysine (1.4Æ
0.4 pmolmLÀ1; n=4) was observed (the limit of detection is
1 pmolmLÀ1; Figure 2).
Sample preparation for amino acid analysis: For amino acid anal-
ysis, the hydroxylation reaction was initiated by mixing LUC7L2267–278
(100 mm), 2OG (500 mm), FeII (400 mm) and ascorbic acid (100 mm) in
MOPS buffer (50 mm, pH 7.5) and JMJD6 (20 mm), followed by incu-
bation at 378C (1 h). Recombinant JMJD6 was purified as report-
ed.[13] After it had been ensured by MALDI-TOF MS that the reac-
In conclusion, we have found that JMJD6 catalyses hydroxyl-
ation of lysyl residues to give products with the 2S,5S-stereo-
chemistry, a clear difference from the collagen lysyl hydroxylas-
ChemBioChem 2011, 12, 531 – 534
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
533

[1] C. Loenarz, C. J. Schofield, Nat. Chem. Biol. 2008, 4, 152–156.
[2] E. Flashman, C. J. Schofield, Nat. Chem. Biol. 2007, 3, 86–87.
[3] a) J. Webb, M. Coleman, C. Pugh, Cell. Mol. Life Sci. 2009, 66, 3539–
3554; b) W. G. Kaelin, Jr., P. J. Ratcliffe, Mol. Cell 2008, 30, 393–402.
[4] a) M. E. Cockman, D. E. Lancaster, I. P. Stolze, K. S. Hewitson, M. A. McDo-
nough, M. L. Coleman, C. H. Coles, X. Yu, R. T. Hay, S. C. Ley, C. W. Pugh,
N. J. Oldham, N. Masson, C. J. Schofield, P. J. Ratcliffe, Proc. Natl. Acad.
Sci. USA 2006, 103, 14767–14772; b) M. E. Cockman, J. D. Webb, P. J.
Ratcliffe, Ann. N. Y. Acad. Sci. 2009, 1177, 9–18.
tion was complete, protein was precipitated by adding methanol
to the mixture. The sample was centrifuged (10 min, 14000 rpm,
14462g), then freeze dried. The assay mixtures were resuspended
in 0.1% aq. formic acid and purified by HPLC on a Phenomenex
Synergi Hydro RP C18 column. The purified peptide was then sus-
pended in MOPS (50 mm) and incubated with trypsin (2 mL,
100 mgmLÀ1; Promega) for 2 h at 378C. The trypsin was removed
by centrifugation (10 min, 14000 rpm, 14462g). Carboxypeptida-
se Y (1 mL, 5 mgmLÀ1, Roche) was then added to the supernatant,
and the mixture incubated (378C, overnight). Carboxypeptidase Y
was then removed by centrifugation (14000 rpm, 14462g, 10 min).
For the acid hydrolysis, an excess of HCl (6m) was added to each
sample, which was then incubated for 24 h at 1108C under an
inert atmosphere. All the samples were concentrated in a vacuum
centrifuge (Speed-Vac, Heto).
´
[5] M. L. Coleman, M. A. McDonough, K. S. Hewitson, C. Coles, J. Mecinovic,
M. Edelmann, K. M. Cook, M. E. Cockman, D. E. Lancaster, B. M. Kessler,
N. J. Oldham, P. J. Ratcliffe, C. J. Schofield, J. Biol. Chem. 2007, 282,
24027–24038.
[6] L. Xie, K. Xiao, E. J. Whalen, M. T. Forrester, R. S. Freeman, G. Fong, S. P.
Gygi, R. J. Lefkowitz, J. S. Stamler, Sci. Signaling 2009, 2, ra33.
[7] R. J. Klose, E. M. Kallin, Y. Zhang, Nat. Rev. Genet. 2006, 7, 715–727.
[8] K. Agger, J. Christensen, P. A. Cloos, K. Helin, Curr. Opin. Genet. Dev.
2008, 18, 159–168.
[9] R. J. Hopkinson, R. B. Hamed, N. R. Rose, T. D. W. Claridge, C. J. Schofield,
ChemBioChem 2010, 11, 506–510.
[10] J. Myllyharju, Matrix Biol. 2003, 22, 15–24.
[11] C. T. Przysiecki, J. E. Staggers, H. G. Ramjit, D. G. Musson, A. M. Stern,
C. D. Bennett, P. A. Friedman, Proc. Natl. Acad. Sci. USA 1987, 84, 7856–
7860.
[12] L. A. McNeill, K. S. Hewitson, T. D. Claridge, J. F. Seibel, L. E. Horsfall, C. J.
Schofield, Biochem. J. 2002, 367, 571–575.
[13] C. J. Webby, A. Wolf, N. Gromak, M. Dreger, H. Kramer, B. Kessler, M. L.
Nielsen, C. Schmitz, D. S. Butler, J. R. Yates III, C. M. Delahunty, P. Hahn,
A. Lengeling, M. Mann, N. J. Proudfoot, C. J. Schofield, A. Bçttger, Sci-
ence 2009, 325, 90–93.
[14] a) B. Witkop, Cell. Mol. Life Sci. 1956, 12, 372–374; b) B. Witkop, Chem.
Soc. 1955, 3, 60–82.
Cow-bone collagen was extracted from a femur according to the
demineralisation and gelatinisation method of Longin.[18] Modern
cow-bone collagen digests were carried out by using either hydro-
chloric acid or the trypsin-carboxypetidase Y method. For the acid-
catalysed hydrolysis, 400 mL aqueous collagen solution (3 gLÀ1
)
was dried in a Pyrex tube, which was then placed in an air-tight
vial with HCl (200 mL, 6n) at 100 mbar and 1108C for 24 h. For the
enzyme protocol, 50 mL of aqueous collagen solution (3 gLÀ1) was
used, and the enzyme-catalysed hydrolysis procedure was carried
out as above. Hydrolysed peptides were derivatised by 6-amino-
quinolyl-N-hydroxysuccinimidyl carbamate (AQC) zin acetonitrile
according to the AccQTag Solution Protocol (Waters) and analysed
by LC-MS as described in the Supporting Information.
[15] J. C. Spetzler, T. Hoeg-Jensen, Tetrahedron Lett. 2002, 43, 2303–2306.
[16] P. Allevi, M. Anastasia, Tetrahedron: Asymmetry 2004, 15, 2091–2096.
[17] a) S. A. Cohen, D. P. Michaud, Anal. Biochem. 1993, 211, 279–287;
b) S. A. Cohen, D. P. Michaud, K. D. Antonis in Techniques in Protein
Chemistry IV, Academic Press, San Diego, 1993, pp. 289–298.
[18] R. Longin, Nature 1971, 230, 241–242.
See the Supporting Information for other methods.
Acknowledgements
The authors thank Dr. Celia J. Webby, Dr. Alexander Wolf and
Prof. Angelika Bçttger for support and advice. This work was
funded by the Biotechnology and Biological Sciences Research
Council, the Wellcome Trust and a Scatcherd European scholar-
ship.
[19] P. B. Hamilton, R. A. Anderson, J. Biol. Chem. 1955, 213, 249–258.
[20] M. Mantri, T. Krojer, E. A. Bagg, C. J. Webby, D. S. Butler, G. Kochan, K. L.
Kavanagh, U. Oppermann, M. A. McDonough, C. J. Schofield, J. Mol. Biol.
2010, 401, 211–222.
´
[21] N. R. Rose, S. S. Ng, J. Mecinovic, B. M. R. Liꢁnard, S. H. Bello, Z. Sun,
M. A. McDonough, U. Oppermann, C. J. Schofield, J. Med. Chem. 2008,
51, 7053–7056.
Keywords: collagen
·
conformation analysis
·
enzyme
Received: October 24, 2010
catalysis · JMJD6 · lysyl hydroxylation
Published online on January 18, 2011
534
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 531 – 534