Inhibition of pyruvate decarboxylase from Z. mobilis by novel analogues of thiamine pyrophosphate: Investigating pyrophosphate mimics

Article abstract of DOI:10.1039/b615861g

Replacement of the thiazolium ring of thiamine pyrophosphate with a triazole gives extremely potent inhibitors of pyruvate decarboxylase from Z. mobilis, with K_I values down to 20 pM; this system was used to explore pyrophosphate mimics and several effective analogues were discovered. The Royal Society of Chemistry.

Full text of DOI:10.1039/b615861g

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Inhibition of pyruvate decarboxylase from Z. mobilis by novel
analogues of thiamine pyrophosphate: investigating pyrophosphate
mimics{
Karl M. Erixon, Chester L. Dabalos and Finian J. Leeper*
Received (in Cambridge, UK) 1st November 2006, Accepted 27th November 2006
First published as an Advance Article on the web 8th December 2006
DOI: 10.1039/b615861g
communication, we report the synthesis of eight novel triazole-
based analogues of TPP, 9–16, and inhibition studies with
ZmPDC (Scheme 1).
Replacement of the thiazolium ring of thiamine pyrophosphate
with a triazole gives extremely potent inhibitors of pyruvate
decarboxylase from Z. mobilis, with KI values down to 20 pM;
this system was used to explore pyrophosphate mimics and
several effective analogues were discovered.
The analogues we planned to make contained a triazole ring in
place of the thiazolium ring of TPP. Thiamine itself is a convenient
starting material and nucleophilic substitution by azide, catalysed
by sodium sulfite,3 gave the azide 4. Cu(I)-catalysed cycloaddition
of 4 with 3-butynol or its tosylate (‘click’ chemistry4) gave the
triazole analogues 5 and 7 in 47 to 68% yield over the two steps
from thiamine. Conversion of tosylate 7 to its pyrophosphate 9
The coenzyme thiamine pyrophosphate (TPP) 2 is the active form
of vitamin B1 1 (Fig. 1). TPP cooperates in a number of enzymes,
which all catalyse the cleavage and formation of bonds adjacent to
the carbon of a carbonyl group. Examples include a-keto acid (e.g.
pyruvate) decarboxylases, dehydrogenases and oxidases, transke-
tolase, benzaldehyde lyase and acetohydroxyacid, acetolactate
and deoxyxylulose 5-phosphate synthases.1 As TPP-dependent
enzymes catalyse key steps in many metabolic pathways they are
also interesting targets for inhibition in the pharmaceutical and
agrochemical industries.
We have previously reported the synthesis of the isoelectronic
thiophene analogue of TPP, 3-deazathiamine pyrophosphate 3,
which proved to be an extremely potent inhibitor of several TPP-
dependent enzymes with KI , 14 pM for Zymomonas mobilis
pyruvate decarboxylase (ZmPDC), compared to a KD value of
approximately 0.35 mM for TPP.2
The pyrophosphate is of great importance in the binding of TPP
to enzymes, where it is coordinated to a magnesium ion in the
active site.1 However, in terms of drug properties, this group is
unsuitable, since it is highly charged and analogues possessing it
will suffer from poor bioavailability and cellular uptake. We,
therefore wanted to explore whether the pyrophosphate group
could be replaced by other analogues without losing too much
binding affinity for the enzyme. The synthesis of deazaTPP
involves 12 steps, however, so we also sought to use an analogue of
TPP which was easier to make for this investigation. In this
Scheme 1 Synthesis of TPP analogues. Reagents and conditions: (i)
NaN3 (2.5 eq.), Na2SO3 (0.1 eq.), H2O, 60–65 uC, 6 h, 83%; (ii) for 5:
3-butynol (1.0 eq.), sodium ascorbate (0.1 eq.), CuSO4?5H2O (0.01 eq.),
t-BuOH–H2O (2 : 1), 25 uC, 16 h, 81%; for 6: 3-pentynol (1.0 eq.),
n-BuOH, reflux at 130 uC, 96 h, 80% (isomeric mixture); (iii) 3-butynyl
tosylate (1.0 eq.), sodium ascorbate (0.1 eq.), CuSO4?5H2O (0.01 eq.),
t-BuOH–H2O (2 : 1), 25 uC, 24 h, 56%; (iv) TsCl (2.5 eq.), pyridine, 25
to 25 uC, 3 h, 30% of desired isomer; (v) (Bu4N)3HP2O7 (2.0 eq.), MeCN,
25 to 25 uC, 50–60%.
Fig. 1 Thiamine, TPP and deaza-TPP analogue.
Department of Chemistry, University of Cambridge, Lensfield Road,
Cambridge, UK CB2 1EW. E-mail: fjl1@cam.ac.uk;
Fax: +44 1233 336362; Tel: +44 1223 336493
{ Electronic supplementary information (ESI) available: Assay details,
double exponential curves and reactivation graphs. See DOI: 10.1039/
b615861g
960 | Chem. Commun., 2007, 960–962
This journal is ß The Royal Society of Chemistry 2007

View Article Online
was effected by displacement with pyrophosphate trianion as in the
synthesis of deazaTPP 3.2
Triazole 5 lacks the 4-methyl group of TPP but unfortunately
cycloaddition of azide 4 with 3-pentynol gives an approximately
1 : 1 mixture of the two possible regioisomers. We were only able
to separate this mixture, with difficulty, after tosylation to give 8.
Nevertheless this permitted access to the methyl triazole analogue
of TPP, 10.
Triazole-based analogues of TPP 9–16, in which the pyrophos-
phate has been replaced by other groups, were all made via similar
‘click’ chemistry starting from azide 4; details will be given in our
full paper. Most of these analogues are designed to be ligands to
the Mg2+ ion in the active site but 16 contains an amine group
which would be protonated at neutral pH and it was hoped this
would replace the Mg2+ ion.
Pyruvate decarboxylase (PDC) catalyses the conversion of
pyruvate to acetaldehyde and carbon dioxide. In the inhibition
studies, the activity of PDC was measured by a coupled enzyme
assay using the NADH-dependent reduction of acetaldehyde by
alcohol dehydrogenase (ADH) (Scheme 2). Because the TPP
analogues can only bind to the holo-enzyme once TPP has
dissociated and the dissociation of TPP is known to be very slow,5
we instead studied binding of the analogues to the apo-enzyme,
prepared as described previously.2 In the assay, the analogues
Fig. 2 Inactivation of ZmPDC by various concentrations of TPP
analogue 10; from top to bottom curve: ($) 2, (#) 4, (.) 6,
(n) 8, (&)10 mM. The curves are the best fits of the equation y =
(1 2 x)exp(2k1t) + x exp(2k2t), with x = 0.2 for each curve. Inset:
Apparent first order rate constants, k1, plotted vs. inhibitor concentrations;
the slope of the graph gives kon. the second order rate constant for the
initial binding.
(typically 2–10 mM) were incubated with apo-PDC in a Mg2+
-
Table 1 Inhibition parameters for analogues 9–16
containing buffer and small samples were taken out at timed
intervals (1–15 min) and added to the assay solution containing
ADH, Mg2+, NADH and an excess of TPP (100 mM). The assay
was then started by the addition of pyruvate. Under these
conditions any apo-PDC that does not already have the TPP
analogue bound will bind TPP when it is added and thus show
activity. However any TPP analogues that are only loosely bound
will be displaced by the excess TPP and so will also show activity.
Incubation of apo-PDC with analogues 9 to 13 gave a decrease
of activity to close to zero over a 15 min time-scale (see Fig. 2 and
supplementary material) but no decrease was observed for
analogues 14–16 under these conditions. The time-points obtained
did not fit well onto simple exponential curves, the initial decay
being too fast and the subsequent decay too slow. This suggests a
two-step process and indeed a much better fit was obtained using
a double exponential curve. When the faster of the two rate
constants obtained in each case was plotted against inhibitor
concentrations a straight line was obtained, which gave a value of
the second-order rate constant for the initial binding, kon (Fig. 2,
inset). The values of kon for the different analogues are collected
in Table 1.
Analogue
kon/mM21 min21
KI
30 ¡ 3 pM
20 ¡ 2 pM
1.2 ¡ 0.1 nM
0.95 ¡ 0.1 nM
0.14 ¡ 0.02 mM
NIa
9
0.05 ¡ 0.02
0.21 ¡ 0.03
0.17 ¡ 0.03
0.09 ¡ 0.01
0.02 ¡ 0.004
NIa
10
11
12
13
14–16
a
NI, no inhibition observed.
initial reversible binding is followed by a slower step, presumably a
conformational change of the enzyme, which effectively makes the
binding irreversible on the time-scale of the assay. A second
possible explanation arises from the fact that ZmPDC is a
homotetramer consisting of a dimer of dimers and there is
considerable evidence for communication between the two active
sites of each dimer in TPP-dependent enzymes.7 Thus binding of
the TPP analogue in the first active site of each dimer could be
faster than in the second active site. TPP itself has been reported to
show negative cooperativity in its binding to transketolase.8 In
support of the second explanation, we note that the apparent rate
constant for the slower of the two stages also seems to be
dependent on concentration of inhibitor.
The apparent two-step inhibition of TPP-dependent enzymes
has been seen before, most recently in our studies of a benzene
analogue of TPP2 but also with thiamine thiazolone pyrophos-
phate with wheat germ PDC.6 There are two possible explanations:
one is the normal explanation for slow-binding inhibition, that
In order to determine whether binding of the TPP analogues is
reversible and, if so, to obtain a KI value, a large excess of TPP
(1 or 10 mM) was added to the enzyme which had been fully
inactivated (,1% activity) by inhibitor (10 or 20 mM). The
recovery of activity was followed over a period of 8 days. In each
case recovery of some activity was observed and the activity
reached a plateau after up to ca. 2 days. In a control experiment,
uninhibited enzyme retained full activity, so this plateau of activity
is presumably due to establishment of the equilibrium between
binding of TPP and binding of the analogue. Using the reported
KD value for TPP (0.35 mM)5 and the relative concentrations of
Scheme 2 Coupled assay of PDC.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 960–962 | 961

View Article Online
TPP and analogue, the KI values for the analogues can be
calculated and are listed in Table 1.
One could try to displace the Mg2+, which was the idea of the
iminodiacetic acid analogue 16, but unfortunately this did not turn
out to be an effective inhibitor.
Not surprisingly, the two analogues possessing the important
pyrophosphate group (9 and 10) were the most potent inhibitors,
with 8.0 and 5.5%, respectively, of the original activity recovered
upon reactivation with 10 mM TPP, corresponding to KI values of
30 and 20 pM, which are in the same range as we previously
reported for deazaTPP.2 The 5-methyl group of 10, which TPP
possesses also, improves the strength of binding, as might be
expected, but interestingly it also speeds up the binding, perhaps
because of the greater hydrophobicity. The reason why analogues
such as 9, 10 and deazaTPP 3, bind so much more tightly than
TPP itself is thought to be because the enzyme stabilises neutral
rings at this position better than the positively charged thiazolium
ring.2,9 In this way the enzyme would promote both the initial
formation of the TPP ylide and the later decarboxylation step.
Some of the analogues containing pyrophosphate mimics also
show strong inhibitory activity. In 11 and 12, the bridging oxygen
in the pyrophosphate has been replaced by a CH2 or CF2 group.
These have recently been reported to be effective non-hydrolysable
mimics of pyrophosphate, used in e.g. antiviral nucleotide
triphosphate mimics.10 We expected these to be effective replace-
ments for the pyrophosphate because the crystal structures of PDC
complexed with TPP show no hydrogen bonds to the bridging
oxygen of the pyrophosphate.11 However, in our case, it appears
that the bridging oxygen is actually of some importance for the
binding. Although they bind faster than the corresponding
pyrophosphate 9, the KI values for both these analogues (11 and
12) were estimated to be 40-fold greater. Methylenediphosphonate
esters such as 11 are known to have higher pKa values than
pyrophosphates,12 which might explain why 11 binds faster, being
less hydrophilic, but less tightly than 9. Difluoromethylene-
diphosphonate esters such as 12, however, have a similar pKa to
pyrophosphates.13 Presumably the greater size of the CF2 group
compared to O reduces the binding affinity in this instance.
The results with the other pyrophosphate mimics suggest that
the charge on the group is of vital importance. Thus a marked
decrease of inhibition is seen in going from analogues 9–12
(charge = 23) to phosphoramidic acid 13 (22), to carbamate 14
and malonate 15 (21). Hence, if the magnesium is present in the
active site, the more negatively charged analogues will bind best.
In summary, eight new analogues of TPP based on a triazole
scaffold have been synthesised. These analogues were prepared in
just a few steps using ‘click’ chemistry and are shown to be very
effective at binding in the TPP binding site. Six of the analogues
contain mimics of the pyrophosphate group and several of these,
particularly the methylenediphosphonates 11 and 12, still bind
with high affinity. These results are likely to be relevant not only to
studies of TPP-dependent enzymic reactions but also to the wide
range of other proteins that bind pyrophosphate esters.
Notes and references
1 (a) F. Jordan, Nat. Prod. Rep., 2003, 20, 184; (b) A. Schellenberger,
Biochim. Biophys. Acta, 1998, 1385, 177.
2 S. Mann, C. Perez Melero, D. Hawksley and F. J. Leeper, Org. Biomol.
Chem., 2004, 2, 1732.
3 (a) J. A. Zoltewicz and G. M. Kauffman, J. Am. Chem. Soc., 1977, 99,
3134; (b) G. Uray, I. Kriessmann and J. A. Zoltewicz, Bioorg. Chem.,
1993, 21, 294; (c) J. A. Zoltewicz, G. Uray and G. M. Kauffman, J. Am.
Chem. Soc., 1980, 102, 3653.
4 F. Himo, T. Lovell, R. Hilgraf, V. V. Rostovtsev, L. Noodleman,
K. B. Sharpless and V. V. Fokin, J. Am. Chem. Soc., 2005, 127, 210.
5 R. J. Diefenbach and R. G. Duggleby, Biochem. J., 1991, 276, 3439.
6 R. Kluger, G. Gish and G. Kauffman, J. Biol. Chem., 1984, 259, 8960.
7 (a) E. M. Ciszak, L. G. Korotchkina, P. M. Dominiak, S. Sidhu and
M. S. Patel, J. Biol. Chem., 2003, 278, 21240; (b) F. Jordan,
N. S. Nemeria and E. Sergienko, Acc. Chem. Res., 2005, 38, 755; (c)
R. A. W. Frank, C. M. Titman, J. V. Pratap, B. F. Luisi and
R. N. Perham, Science, 2004, 306, 872; (d) F. Jordan, Science, 2004, 306,
818.
8 R. M. Egan and H. Z. Sable, J. Biol. Chem., 1981, 256, 4877.
9 F. Jordan, H. J. Li and A. Brown, Biochemistry, 1999, 38, 6369.
10 (a) N. A. Boyle, Org. Lett., 2006, 8, 187; (b) G. Wang, N. A. Boyle,
F. Chen, V. Rajappan, P. Fagan, J. L. Brooks, T. Hurd, J. M. Leeds,
V. K. Rajwanshi, Y. Jin, M. Prhavc, T. W. Bruice and P. D. Cook,
J. Med. Chem., 2004, 47, 6902; (c) L. J. Jennings, M. Macchia and
A. Parkin, J. Chem. Soc., Perkin Trans. 1, 1992, 2197.
11 (a) D. Dobritzsch, S. Ko¨nig, G. Schneider and G. Lu, J. Biol. Chem.,
1998, 273, 20196; (b) F. Dyda, W. Furey, S. Swaminathan, M. Sax,
B. Farrenkopf and F. Jordan, Biochemistry, 1993, 32, 6165.
12 G. M. Blackburn, D. E. Kent and F. Kolkmann, J. Chem. Soc., Chem.
Commun., 1981, 1188.
13 (a) G. R. J. Thatcher and A. S. Campbell, J. Org. Chem., 1993, 58, 2272;
(b) G. M. Blackburn, D. A. England and F. Kolkmann, J. Chem. Soc.,
Chem. Commun., 1981, 3523.
962 | Chem. Commun., 2007, 960–962
This journal is ß The Royal Society of Chemistry 2007