The perils of rational design-unexpected irreversible elimination of fluoride from 3-fluoro-2-methylacyl-CoA esters catalysed by α-methylacyl-CoA racemase (AMACR; P504S)

Article abstract of DOI:10.1039/c4cc06127f

α-Methylacyl-CoA racemase (AMACR; P504S) catalyses 'racemization' of 2-methylacyl-CoAs, the activation of R-ibuprofen and is a promising cancer drug target. Human recombinant AMACR 1A catalyses elimination of 3-fluoro-2-methyldecanoyl-CoAs to give E-2-methyldec-2-enoyl-CoA and fluoride anion, a previously unknown reaction. 'Racemization' of 2-methyldec-3-enoyl-CoAs was also catalysed, without double bond migration. This journal is

Full text of DOI:10.1039/c4cc06127f

ChemComm
View Article Online
View Journal
COMMUNICATION
The perils of rational design – unexpected
irreversible elimination of fluoride from
Cite this:DOI: 10.1039/c4cc06127f
3-fluoro-2-methylacyl-CoA esters catalysed by
a-methylacyl-CoA racemase (AMACR; P504S)†
Received 5th August 2014,
Accepted 18th September 2014
Maksims Yevglevskis, Guat L. Lee, Michael D. Threadgill, Timothy J. Woodman and
Matthew D. Lloyd*
DOI: 10.1039/c4cc06127f
www.rsc.org/chemcomm
a-Methylacyl-CoA racemase (AMACR; P504S) catalyses ‘racemization’ of these reactions with the appropriate substrates. This communication
2-methylacyl-CoAs, the activation of R-ibuprofen and is a promising reports that human recombinant AMACR 1A4 is able to catalyse
cancer drug target. Human recombinant AMACR 1A catalyses elimination an elimination reaction with 3-fluoro-2-methylacyl-CoA sub-
of 3-fluoro-2-methyldecanoyl-CoAs to give E-2-methyldec-2-enoyl-CoA strates to give the corresponding unsaturated 2-methylacyl-CoA
and fluoride anion, a previously unknown reaction. ‘Racemization’ ester and a fluoride anion.22 It is also able to catalyse ‘racemization’
of 2-methyldec-3-enoyl-CoAs was also catalysed, without double of unsaturated 2-methylacyl-CoA esters but does not catalyse
bond migration.
migration of the double bond.
The known substrate 2-methyldecanoyl-CoA 1S allows the
a-Methylacyl-CoA racemase (AMACR,‡ P504S; E.C. 5.1.99.4) catalyses course of the enzyme reaction to be followed using proton/
the key ‘racemization’ step in the degradation of branched-chain deuterium exchange in 2H2O. However this assay only yields
fatty acids and is also important in the pharmacological activa- information on exchange of the a-proton; to obtain information
tion of R-ibuprofen and related drugs.1–3 The enzyme catalyses on the stereochemistry a time consuming and scale-limited
the conversion of either epimer of a 2-methylacyl-CoA ester to a derivatization of products is needed. It was hoped that using
ca. 1 : 1 mixture of 2R- and 2S-epimers.4 AMACR has also been fluorinated analogues of 2-methylacyl-CoA esters would overcome
proposed to be involved in the uni-directional chiral inversion of this problem. Specifically syn- and anti-3-fluoro-2-methyldecanoyl-
mandelic acid in mammals5 but this was recently shown to CoA, 2S and 2R, were chosen as it was anticipated that chiral
proceed by a distinct pathway.6
inversion and a-proton exchange of this epimeric pair of substrates
AMACR protein levels and enzyme activity are increased in could be directly and simultaneously observed by changes in the
prostate cancers7,8 a subset of colon cancers9 and various other 1H and 19F NMR spectra. S- and R-E-2-Methyldec-3-enoyl-CoAs 3S
cancers1 and it is widely recognised as a promising new drug and 3R were chosen for testing in order to determine whether
target.1,2,10,11 However, relatively few chemical inhibitors of AMACR AMACR could catalyse double bond migration into conjugation with
have been reported,12–15 largely due to the lack of a convenient, high- the carbonyl group, whilst E-2-methyldec-2-enoyl-CoA 4 was selected
throughput assay.
as the proposed product of this reaction (Fig. 1).
No X-ray crystal structure for a mammalian AMACR has been
The required substrates were synthesised by extension of
reported but the enzyme is proposed to catalyse its reaction reported methods.3,4,12 anti-3-Fluoro-2-methyldecanoic acid 5 was
by removal of the a-proton by Asp-152 or the His-122/Glu-237 synthesised by the method of Carnell et al.12 using octanal (ESI,†
pair1,2,4,16,17 to form an enolate intermediate.18 Non-stereoselective Scheme S2). syn-3-Fluoro-2-methyldecanoic acid 6 was synthesised
reprotonation of the enolate gives a B1 : 1 mixture of 2-methylacyl- by aldol reaction of N-propanoyl-Evans’ auxiliary with 2-octenal,
CoAs with the original and epimeric configuration at the a-carbon.2,4 followed by hydrogenation, conversion to the methyl ester, treat-
Enolates are common intermediates in a number of enzymatic ment with DAST12 and deprotection (ESI,† Scheme S3). S- and
reactions, including condensations,19 double-bond migrations20 R-2-Methyldec-3-enoic acids 7 were synthesised by reaction of
and elimination reactions,21 and AMACR could potentially perform the Grignard reagent derived from E-1-crotyl chloride 8 with
CO2, followed by chiral resolution of the resulting acids as
Medicinal Chemistry, Department of Pharmacy & Pharmacology, University of Bath,
the N-acylated R-Evans’ auxiliary derivatives. Metathesis with
1-octene and deprotection gave the required unsaturated acids
(ESI,† Scheme S4). 2-Methyldec-2-enoic acid 9 was synthesised
Claverton Down, Bath BA2 7AY, UK. E-mail: M.D.Lloyd@bath.ac.uk;
Fax: +44-(0)1225-386114; Tel: +44-(0)1225-386786
† Electronic supplementary information (ESI) available: Assignment of stereo-
by a Wittig reaction between octanal and the ylide derived from
ethyl 2-bromopropanoate and Ph3P, followed by deprotection
chemical configurations and description of reactions; synthetic methods; enzyme
kinetic plots. See DOI: 10.1039/c4cc06127f
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.

View Article Online
Communication
ChemComm
This shows that the geometry of the double bond of the
precursor acid can be assigned and confirms that the product
of the enzymatic reaction is the E-isomer.
Both 2S and 2R gave the same E-acid product 4, consistent
with an E1cb mechanism21 in which an enolate intermediate18 is
used to expel the fluoride. The resulting E-double bond suggests that
the reaction occurs with the substrate in an anti-conformation with
respect to the a-proton and fluoride. This contrasts with enoyl-CoA
hydratase,21,24 which catalyses its E1cb reaction with syn-elimination
because the two catalytic glutamate residues are on the same face of
the substrate.24 anti-elimination by AMACR probably results from
the combination of a number of factors:21 firstly, the substrate side-
chain is bound by a hydrophobic surface,17 allowing adoption of the
more favourable anti-conformation; secondly, fluorine is a relatively
small substituent;25 and third, fluoride is likely to be highly solvated
in aqueous solution. Work in similar chemical systems suggests that
fluoride is eliminated by an E1cb-like E2 mechanism26,27 with anti-
stereochemistry. It is also notable that other enzymes with enolate or
enediol intermediates also eliminate HF from substrate analogues,
including butyryl-CoA dehydrogenase,28 uronate isomerase29 and
glyoxalase I.30 Some of these enzymes28,29 also possess active-site
aspartate or glutamate residues acting as bases.
Fig. 1 Structures of the acyl-CoA esters incubated with human recombinant
AMACR 1A.
(ESI,† Scheme S5). These acids were converted into their corre-
sponding acyl-CoA esters by activation with N,N0-carbonyldiimidazole
followed by reaction with CoA-SH.3
Initially, acyl-CoA esters were incubated with human recombi-
nant AMACR 1A4 in the presence of 2H2O. Incubation of 2R and 2S
with active AMACR was expected to result in formation of a mixture
of epimers at carbon-2 with exchange of the a-proton for deuterium
resulting in formation of a broad single peak (2-bond coupling to the
2H is not normally observed). Unexpectedly, the peak at ca. 1.1 ppm
diminished with time (Fig. 2) and a new singlet at ca. 1.75 ppm
simultaneously arose and increased in intensity. In the 19F spectrum
the signal for 2R and 2S slowly disappeared and a new signal
appeared at d ꢀ122 ppm, which is characteristic of inorganic
fluoride. Taken together these observations suggested that an
elimination reaction had occurred. Product levels increased
over time when 2S or 2R were incubated with active AMACR
(ESI,† Fig. S1). For 2R, this reaction was not observed in
negative controls containing heat-inactivated enzyme.
Kinetic analysis of the AMACR-catalysed elimination reaction
showed that Michaelis–Menten behaviour was observed. The follow-
ing kinetic parameters were determined for 2R by the Direct Linear
Plot:31,32 Km = 21 mM; Vmax = 96.5 nmol minꢀ1 mgꢀ1 protein;
kcat = 0.0758 sꢀ1; kcat/Km = 3612 Mꢀ1 ꢀ1. This compares to Km
s =
277 mM; Vmax = 39.3 nmol minꢀ1 mgꢀ1 protein; kcat = 0.0310 sꢀ1
;
kcat/Km = 112 Mꢀ1 sꢀ1 for S-2-methyldecanoyl-CoA 1S,3 implying
that the elimination reaction is B32ꢁ more efficient than the
racemization reaction (as judged by kcat/Km). This is probably
due to the electron-withdrawing effect of the fluorine atom
increasing the a-proton acidity.12 For 2S, significant background
conversion was observed in negative controls, probably due to
the anti-arrangement of the a-proton and fluorine atom in a
favourable staggered conformation. The following approximate
values for kinetic parameters were obtained for 2S: Km = 40 mM;
Comparison of the 1H NMR spectrum of the unsaturated
acyl-CoA product from the enzymatic elimination reaction showed
that it was identical to 4. Compound 4 had been synthesised from
the corresponding E-acid, as assigned by the alkene proton signal
at d 6.92 ppm (The Z-acid alkene proton appears at 6.09 ppm).23
Vmax = 50.6 nmol minꢀ1 mgꢀ1 protein; kcat = 0.0397 sꢀ1
;
kcat/Km = 993 Mꢀ1 sꢀ1. 2R needs to adopt a gauche conformation
for anti-elimination, and hence the background non-enzymatic
reaction is less favoured.
Incubation of 4 with active AMACR in the presence of fluoride
did not show any conversion to 2, showing that the elimination
reaction is irreversible. Control experiments with ꢂ-fenoprofenoyl-
CoA showed that AMACR was equally active in the presence and
absence of fluoride anions, showing the enzyme was not
inactivated. The irreversibility of the elimination of 2 probably
results from the high levels of hydration of the fluoride anion.
Fluoride is also a hard nucleophile making it less likely to react
with the soft conjugate electrophile.
Unsaturated 2-methylacyl-CoA esters were also investigated
as substrates for AMACR. Incubation of 3S and 3R with active
AMACR in the presence of 2H2O resulted in a-proton exchange,
as judged by conversions of the doublet at d ca. 1.1 (methyl) into a
broad single peak and of the doublet of doublets at 5.29 ppm into
a doublet in the 1H NMR spectrum (ESI,† Fig. S2). Exchange of
Fig. 2 1H NMR spectrum of 2R when incubated with AMACR over time.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
4 D. J. Darley, D. S. Butler, S. J. Prideaux, T. W. Thornton, A. D. Wilson,
T. J. Woodman, M. D. Threadgill and M. D. Lloyd, Org. Biomol.
Chem., 2009, 7, 543–552.
5 L.-B. Gao, J.-Z. Wang, T.-W. Yao and S. Zeng, Chirality, 2012, 24,
86–95.
6 M. Yevglevskis, C. R. Bowskill, C. C. Y. Chan, J. H.-J. Heng,
M. D. Threadgill, T. J. Woodman and M. D. Lloyd, Org. Biomol.
Chem., 2014, 12, 6737–6744.
7 J. Luo, S. Zha, W. R. Gage, T. A. Dunn, J. L. Hicks, C. J. Bennett,
C. N. Ewing, E. A. Platz, S. Ferdinandusse, R. J. Wanders, J. M. Trent,
W. B. Isaacs and A. M. De Marzo, Cancer Res., 2002, 62, 2220–2226.
8 C. Kumar-Sinha, R. B. Shah, B. Laxman, S. A. Tomlins, J. Harwood,
W. Schmitz, E. Conzelmann, M. G. Sanda, J. T. Wei, M. A. Rubin and
A. M. Chinnaiyan, Am. J. Pathol., 2004, 164, 787–793.
9 Z. Jiang, G. R. Fanger, B. F. Banner, B. A. Woda, P. Algate, K. Dresser,
J. C. Xu, S. G. Reed, K. L. Rock and P. G. Chu, Cancer Detect. Prev.,
2003, 27, 422–426.
the a-proton is required for the chiral inversion of substrates by
AMACR,3,4 and therefore it is highly likely that chiral inversion
has also taken place. Substrate incubated with heat-inactivated
enzyme under the same conditions showed no changes, showing
that AMACR catalyses ‘racemization’ of unsaturated substrates.
Native human and rat enzymes have been previously reported33,34
not to bind similar unsaturated acyl-CoA esters (based on a competi-
tion assay), suggesting they bind relatively weakly compared to
saturated substrates.
Rearrangement of the double bond of 3S or 3R was not
catalysed by AMACR, with no formation of 4 as shown by the
absence of the characteristic methyl group singlet at ca. d 1.8
in the 1H NMR spectrum. These results imply that either no
proton donor is in close proximity to the distal end of the double
bond to facilitate migration or that reprotonation of the enolate
intermediate18 to give ‘racemization’ is much more efficient.
Incubation of the 2-unsaturated acyl-CoA ester 4 with active
AMACR showed, by 1H NMR analysis, that it was not converted
to 3 or any other product. It is not clear whether this is due to 4
failing to bind to AMACR or if it binds but does not undergo
a reaction. It is known that 2-methylene acyl-CoA esters, which
also possess a sp2-hybridised carbon-2, behave as reversible tight-
binding dead-end inhibitors of human AMACR 1A in vivo,2,13
suggesting that 4 may also be bound.
10 S. Zha, S. Ferdinandusse, S. Denis, R. J. Wanders, C. M. Ewing,
J. Luo, A. M. De Marzo and W. B. Isaacs, Cancer Res., 2003, 63,
7365–7376.
11 K. Takahara, H. Azuma, T. Sakamoto, S. Kiyama, T. Inamoto,
N. Ibuki, T. Nishida, H. Nomi, T. Ubai, N. Segawa and
Y. Katsuoka, Anticancer Res., 2009, 29, 2497–2505.
12 A. J. Carnell, I. Hale, S. Denis, R. J. A. Wanders, W. B. Isaacs, B. A. Wilson
and S. Ferdinandusse, J. Med. Chem., 2007, 50, 2700–2707.
13 A. Morgenroth, E. A. Urusova, C. Dinger, E. Al-Momani, T. Kull,
G. Glatting, H. Frauendorf, O. Jahn, F. M. Mottaghy, S. N. Reske and
B. D. Zlatopolskiy, Chem. – Eur. J., 2011, 17, 10144–10150.
14 B. A. P. Wilson, H. Wang, B. A. Nacev, R. C. Mease, J. O. Liu,
M. G. Pomper and W. B. Isaacs, Mol. Cancer Ther., 2011, 10, 825–838.
15 A. J. Carnell, R. Kirk, M. Smith, S. McKenna, L.-Y. Lian and
R. Gibson, ChemMedChem, 2013, 8, 1643–1647.
The results in this Communication demonstrate that human
AMACR 1A is able to catalyse irreversible elimination of substrates,
probably by an E1cb or E1cb-like E2 mechanism. The reaction is of
potential utility for measuring the AMACR activity, since quantifica-
tion of both the enoyl-CoA and fluoride products is possible. It is
also notable that several AMACR inhibitors with similar structures to
2R and 2S have been reported.12 Given that the only difference
between these compounds and 2R is the length of the side-chain, it
is quite possible that these compounds also undergo an elimination
reaction. Fluorine atoms are often used in drug molecules
(with 420% of all drugs containing at least one fluorine atom35),
but it is important to consider that they may be reactive under
certain circumstances. These results also extend the range
of substrates for the AMACR racemization reaction to include
2-methyl-3-enoyl-CoA esters.
16 F. A. Sattar, D. J. Darley, F. Politano, T. J. Woodman,
M. D. Threadgill and M. D. Lloyd, Chem. Commun., 2010, 46,
3348–3350.
17 P. Bhaumik, W. Schmitz, A. Hassinen, J. K. Hiltunen, E. Conzelmann
and R. K. Wierenga, J. Mol. Biol., 2007, 367, 1145–1161.
18 S. Sharma, P. Bhaumik, W. Schmitz, R. Venkatesan, J. K. Hiltunen,
E. Conzelmann, A. H. Juffer and R. K. Wierenga, J. Phys. Chem. B,
2012, 116, 3619–3629.
19 R. B. Hamed, J. R. Gomez-Castellanos, A. Thalhammer, D. Harding,
C. Ducho, T. D. W. Claridge and C. J. Schofield, Nat. Chem., 2011, 3,
365–371.
20 J. M. Schwab and D. C. T. Lin, J. Am. Chem. Soc., 1985, 107,
6046–6052.
21 J. R. Mohrig, Acc. Chem. Res., 2013, 46, 1407–1416.
22 M. D. Lloyd, T. J. Woodman, M. Jevglevskis and M. D. Threadgill,
UK Pat., 1401817.0, 2014.
23 F. Bargiggia and O. Piva, Tetrahedron: Asymmetry, 2003, 14,
1819–1827.
24 P. Kasaragod, W. Schmitz, J. K. Hiltunen and R. K. Wierenga, FEBS
J., 2013, 280, 3160–3175.
We thank Dr Simon E. Lewis (Department of Chemistry,
University of Bath, U.K.) for helpful discussions. This work was
supported by Prostate Cancer UK. The authors are members of
the Cancer Research @ Bath (CR@B) network.
25 K. L. Kirk, J. Fluorine Chem., 2006, 127, 1013–1029.
26 P. Ryberg and O. Matsson, J. Am. Chem. Soc., 2001, 123, 2712–2718.
27 P. Ryberg and O. Matsson, J. Org. Chem., 2002, 67, 811–814.
28 R. H. Abeles and G. Fendrich, J. Am. Chem. Soc., 1981, 21, 6685–6695.
29 L. Williams, T. Nguyen, Y. Li, T. N. Porter and F. M. Raushel,
Biochemistry, 2006, 45, 7453–7462.
30 J. W. Kozarich, R. V. J. Chari, J. C. Wu and T. L. Lawrence,
J. Am. Chem. Soc., 1981, 103, 4593–4595.
Notes and references
31 A. Cornish-Bowden and R. Eisenthal, Biochem. J., 1974, 139,
721–730.
32 R. Eisenthal and A. Cornish-Bowden, Biochem. J., 1974, 139,
715–720.
33 W. Schmitz, R. Fingerhut and E. Conzelmann, Eur. J. Biochem., 1994,
222, 313–323.
34 W. Schmitz, C. Albers, R. Fingerhut and E. Conzelmann,
Eur. J. Biochem., 1995, 231, 815–822.
‡ Abbreviations used: AMACR, a-methylacyl-CoA racemase; CoA, coenzyme
A; DAST, diethylaminosulfur trifluoride; ppm, parts per million.
1 M. D. Lloyd, D. J. Darley, A. S. Wierzbicki and M. D. Threadgill,
FEBS J., 2008, 275, 1089–1102.
2 M. D. Lloyd, M. Yevglevskis, G. L. Lee, P. J. Wood, M. D. Threadgill
and T. J. Woodman, Prog. Lipid Res., 2013, 52, 220–230.
3 T. J. Woodman, P. J. Wood, A. S. Thompson, T. J. Hutchings,
G. R. Steel, P. Jiao, M. D. Threadgill and M. D. Lloyd, Chem. Commun., 35 H. Chen, S. Viel, F. Ziarelli and L. Peng, Chem. Soc. Rev., 2013, 42,
2011, 47, 7332–7334.
7971–7982.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.