Gold(i)-catalysed synthesis of a furan analogue of thiamine pyrophosphate

Article abstract of DOI:10.3762/bjoc.10.270

An analogue of thiamine having a furan ring in place of the thiazolium ring has been synthesised by a short and efficient route, involving gold(I)-catalysed cyclisation of an alkynyl alcohol to form the furan ring. The furan analogue of thiamine diphospha

Full text of DOI:10.3762/bjoc.10.270

Gold(I)-catalysed synthesis of a furan analogue
of thiamine pyrophosphate
Amjid Iqbal, El-Habib Sahraoui and Finian J. Leeper*
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2014, 10, 2580–2585.
University of Cambridge, Department of Chemistry, Lensfield Road,
Cambridge, CB2 1EW, U.K.
doi:10.3762/bjoc.10.270
Received: 05 September 2014
Accepted: 16 October 2014
Published: 05 November 2014
Email:
Finian J. Leeper* - fjl1@cam.ac.uk
* Corresponding author
Associate Editor: J. S. Dickschat
Keywords:
© 2014 Iqbal et al; licensee Beilstein-Institut.
License and terms: see end of document.
furan synthesis; gold-catalysed cyclisation; pyruvate decarboxylase;
thiamine diphosphate
Abstract
An analogue of thiamine having a furan ring in place of the thiazolium ring has been synthesised by a short and efficient route,
involving gold(I)-catalysed cyclisation of an alkynyl alcohol to form the furan ring. The furan analogue of thiamine diphosphate
(ThDP) was also made and tested for binding to and inhibition of pyruvate decarboxylase (PDC) from Zymomonas mobilis (overex-
pressed in E. coli with a N-terminal His-tag). It is a very strong inhibitor, with a Ki value of 32.5 pM. It was also shown that the
furan analogue of thiamine can be functionalised at the C-2 position, which will allow access to mimics of reaction intermediates of
various ThDP-dependent enzymes.
Introduction
The biologically active form of vitamin B1 is thiamine diphos-
phate (ThDP, 1, Figure 1), which is an essential cofactor and
involved in a number of metabolic pathways, including oxida-
tive and non-oxidative decarboxylation of α-keto acids (e.g.,
pyruvate dehydrogenase, pyruvate decarboxylase), the forma-
tion of amino acid precursors (acetohydroxy acid synthase), and
ketol transfer between sugars (transketolase) [1]. One common
Figure 1: Structure of thiamine diphosphate (ThDP, 1).
feature of ThDP-dependent enzymes is to catalyse the cleavage
and formation of bonds adjacent to the carbon of a carbonyl
group with the thiazolium ring of ThDP acting as an electron The catalytic cycle for pyruvate decarboxylase (PDC, a
sink during catalysis in order to stabilise what would otherwise representative ThDP-dependent enzyme), first proposed by
be an acyl carbanion in the form of an enamine intermediate [2]. Breslow [3,4], is given in Scheme 1. After formation of the
2580

Beilstein J. Org. Chem. 2014, 10, 2580–2585.
ThDP ylid, the keto group of the substrate is attacked by the Although many crystal structures of ThDP dependent enzymes
thiazolium C-2 carbanion to form the tetrahedral pre-decar- exist, there still are some unanswered questions regarding
boxylation intermediate 2-(2-lactyl)-ThDP, LThDP. It has been mechanistic processes of these enzymes (e.g., how the substrate
proposed that the covalent attachment of pyruvate to form binds, what is the role of important catalytic groups, how are
LThDP introduces significant strain, the release of which is a individual steps accelerated). In order to answer these questions,
driving force for the decarboxylation [5]. The effects of this it would help to have snapshots of the enzyme bound to various
type of strain on bond angles and lengths have recently been reaction intermediates in the cycle. However, this is usually not
observed in a high resolution crystal structure of another ThDP- possible when using the natural cofactor because the enzyme
dependent enzyme, transketolase [6]. Another important factor turns over rapidly, which in most cases prevents the trapping of
of this decarboxylation is the presence of an electron sink in the reaction intermediates bound to the enzyme. One solution to
form of the positively charged nitrogen atom. Decarboxylation this problem is to use analogues of ThDP that closely mimic the
gives the C-2α-carbanion/enamine resonance forms of the post reaction intermediates of the catalytic cycle, but crucially are
decarboxylation intermediate. Protonation of this enamine inter- unable to continue the reaction mechanism. To this end we have
mediate at C-2α gives HEThDP and final elimination of previously synthesised inactive analogues of the ThDP cofactor,
acetaldehyde completes the catalytic cycle.
including 3-deazaThDP 2, in which the nitrogen atom of the
thiazolium ring is replaced by a carbon [9,10]. 3-DeazaThDP is
The 4’-aminopyrimidine ring of ThDP can interconvert among isoelectronic with ThDP, but the lack of positive charge means
three ionization/tautomeric states, the protonated form and two that deprotonation at C-2 is not possible.
neutral forms, the amino tautomer and the imino tautomer
(Scheme 1) and it is believed to be the imine nitrogen atom of 3-DeazaThDP 2 is an almost irreversible inhibitor, which binds
the latter that is the base that deprotonates C-2 to form the ylid 25,000 times more strongly (Ki = 14 pM) than ThDP to PDC
[7,8].
from Zymomonas mobilis (ZmPDC) and 500 times more
Scheme 1: Mechanism of pyruvate decarboxylase and structures of some previously synthesised ThDP analogues.
2581

Beilstein J. Org. Chem. 2014, 10, 2580–2585.
strongly (Ki = 5 nM) to the E1 component of α-ketoglutarate
dehydrogenase [10]. This extremely tight binding can be ratio-
nalised by the structural similarity between 2 and the ylid of
ThDP, with increased hydrophobic interactions between the
enzyme and the electrically neutral analogue compared to the
positively charged ThDP [11].
The synthesis of 2 was achieved in 12 steps in reasonable yield
but was lengthy and time-consuming. For this reason other
analogues have also been studied. Two triazole analogues 3
(Ki = 20 pM) and 4 (Ki = 30 pM against ZmPDC) were synthe-
sised in just four synthetic steps [12]. However, there is no
possibility of attaching a substituent at the 2-position of the
triazole to mimic enzymic reaction intermediates. Another
analogue of ThDP 5 with a benzene ring in place of the thia-
zolium ring was synthesised and also found to be a very strong
inhibitor of ZmPDC, binding at almost the same rate as 2 [10].
With this benzene analogue also, it would be difficult to func-
tionalise the 2-position selectively.
Scheme 2: Dehydrative cyclization catalysed by gold(I) and its
presumed mechanism [31].
In this paper, we report a short and efficient synthesis of a
furan-based analogue of thiamine. This analogue was converted
to its diphosphate, which is another extremely potent inhibitor
of ZmPDC, and could also be functionalised at the 2-position
using a Friedel–Crafts acylation.
Results and Discussion
Synthesis of ThDP analogue 17
The key step in our planned synthesis was the formation of the
furan ring. Homogeneous gold-catalysed reactions have been
used recently in the synthesis of furans from alkynes [13-20].
The ease with which alkynes, allenes and alkenes can be
activated by Au(I) catalysts to form carbon–carbon and
carbon–heteroatom bonds makes the method an appealing
strategy towards diverse chemical targets [21-30]. Furthermore,
the catalysts are stable to air and moisture and the reactions
Scheme 3: Synthesis of furan 12. Reagents and conditions:
(i) TBDMS-Cl, N-methylimidazole. (ii) n-BuLi, −78 °C, then 9 and to rt,
55% (iii) TBAF, THF, 79%. (iv) AuCl, THF, 85%.
generally proceed under mild conditions. Aponick and ethers, 7 and 9. The alkyne 7 was then coupled with ketone 9
co-workers reported a catalytic dehydrative cyclisation reaction using n-BuLi. TBAF-mediated deprotection gave unsaturated
of alkynyl alcohols catalysed by simple gold(I) salts [31]. The alcohol 11, the substrate for dehydrative cyclization with AuCl.
reaction proceeds rapidly under mild, open flask conditions to The order of addition, reaction time and temperature were all
provide aromatic heterocycles such as furans, pyrroles and thio- found to be important for this reaction: the best yield was
phenes in high yield with low catalyst loadings (Scheme 2).
obtained when a solution of 11 in THF was added at 0 °C to the
dry AuCl. The mixture was stirred for 10 min and then filtered
Following this synthetic strategy, it was decided to explore the through Celite to remove the catalyst. Using this method an
cyclisation of alkyne tetraol 11, which should give furan 12 85% yield of furan 12 could be obtained. Delay in filtration or
with hydroxyethyl and hydroxymethyl substituents at the 2- and increase in temperature caused the catalyst to decompose and
4-position respectively (Scheme 3). This could then be made the product difficult to purify. Subsequent to our syn-
converted into a furan analogue of ThDP.
thesis of 12 a very similar synthesis of the same compound was
published by Deslongchamps and coworkers [32], which only
The first step in the synthesis of tetraol 11 was the protection of differed in that the cyclisation of 11 to 12 in excellent yield was
3-butynol (6) and 1,3-dihydroxyacetone (8) as their TBDMS catalysed by Hg(II), instead of Au(I).
2582

Beilstein J. Org. Chem. 2014, 10, 2580–2585.
With furan 12 in hand, it was then transformed into the furan between the subunits in each dimer [33]. One of the main
analogue 17 of ThDP by the same route as previously used for reasons for using this enzyme for our study is that, unlike most
the synthesis of deazaThDP 2 [9,10], Scheme 4. Thus selective other PDCs (e.g., from yeast), it does not show allosteric
oxidation of the benzylic alcohol with MnO2, condensation with activation by its own substrate and thus gives normal
3-anilinopropionitrile, and then reaction with acetamidine gives Michaelis–Menten kinetics [34-38], and also it shows good
the thiamine analogue 15. Tosylation of the alcohol and dis- stability. Previously the native enzyme expressed in Escherichia
placement of tosylate by the diphosphate trianion worked well coli has been used in such studies but, in order to simplify the
to give ThDP analogue 17.
purification, in this study the gene was cloned into a pET28a
vector, to give the enzyme an N-terminal His6-tag. This form of
the protein was overexpressed in high yield in E. coli, and was
easily purified on a Ni2+-NTA column. The His6-tag does not
appear to affect the enzymic activity.
The normal coupled enzyme assay for PDC was used, in which
PDC catalyses the decarboxylation of pyruvate to acetaldehyde
and alcohol dehydrogenase (ADH) converts the acetaldehyde to
ethanol using NADH (Scheme 5). The oxidation of NADH to
NAD+ is monitored by UV spectroscopy at 340 nm.
Scheme 5: Coupled assay of PDC activity.
In order to follow the time-course of binding of furan 17,
various concentrations (1–8 µM) were incubated with apo-PDC
in a Mg2+-containing buffer and aliquots were taken out at
timed intervals (1–30 min) and added to the assay solution
containing excess ThDP (100 µM). Under these conditions any
apo-PDC that does not already have the furan analogue bound
Scheme 4: Synthesis of the furan analogue 17 of ThDP. Reagents
and conditions: (i) MnO2, CHCl3, 72%; (ii) PhNHCH2CH2CN, NaOMe,
DMSO, MeOH, 76%; (iii) CH3C(NH)NH2·HCl, NaOEt, EtOH, 65%;
(iv) TsCl, pyridine, −5 °C, 72%; (v) TBA3HPP, MeCN, 4 °C, 30%;
(vi) AcCl, AlCl3, DCM, 67%.
A further objective was to show that the furan ring could be will bind ThDP and thus show activity. The plot of the
substituted at C-2. To this end furan 15 was reacted with acetyl percentage residual activity against time is shown in Figure 2.
chloride and AlCl3 and the 2-acetylfuran 18 was obtained in For all concentrations of inhibitor the enzyme lost all activity
67% yield. 1H,13C correlations in the HMBC spectrum of 18 within 30 min, though the inhibition is faster at higher concen-
proved that C-2 and not C-4 had been acetylated, with a key trations.
correlation being from the first CH2 of the CH2CH2OAc side-
chain (C-5a) to the furan carbon that has a hydrogen attached The data-points in Figure 2 fit poorly onto a simple exponential
(C-4). This selective acylation of C-2 will allow the addition of curve, however, a double exponential with 60% of the activity
substituents at this position that match the substituents in the being lost at a fast rate and 40% being lost more slowly fits
enzymic reactions, such as the 2-hydroxyethyl group found in reasonably well; the equation used is
HEThDP (Scheme 1).
Inhibition of pyruvate decarboxylase by furan
17
The faster of these two phases (k1) cannot be measured accu-
The furan analogue of ThDP, 17, was tested as an inhibitor of rately from this data as it is almost all over within the first
pyruvate decarboxylase from Zymomonas mobilis (ZmPDC). minute at all but the lowest concentrations of inhibitor.
ZmPDC is a tetramer, made up of a dimer of dimers, and there However, an estimated value of the second order rate constant
are four active sites per tetramer located at the interface kon for the initial binding of 17 is ca. 1 µM−1 min−1.
2583

Beilstein J. Org. Chem. 2014, 10, 2580–2585.
Figure 3: Recovery of activity for ZmPDC inhibited by furan 17
(1.0 µM) and then incubated with ThDP (1.0 mM). The activity is given
as a percentage of the initial activity of uninhibited enzyme.
Figure 2: Time course inactivation of ZmPDC by various concentra-
tions of furan 17.
20 pM) [12,37]. However, the furan analogue of ThDP has the
advantage over the triazole analogues that it can be function-
alised at the C-2 position.
The apparent two-stage inhibition for PDC was reported by
Mann et al. for the benzene-based ThDP analogue 5 [10] and
the same effect was also observed by Erixon et al. using triazole Conclusion
analogues 3 and 4 [37] and also with a thiazolone analogue by In conclusion, we report here the short synthesis of a furan
Kluger et al. [38]. It may be due to a slow conformational analogue of thiamine, which could be diphosphorylated to give
change after the initial binding or due to communication an analogue of ThDP. Biological evaluation of this diphosphate
between the two active sites in each dimer.
showed that it is a very strong inhibitor of ZmPDC with pico-
molar affinity. Friedel–Crafts acetylation of the furan at C-2
In order to measure a Ki value for furan 17 the enzyme has to be was successful which opens the possibility of synthesising furan
allowed to reach its equilibrium in a competition between analogues of the enzymic reaction intermediates.
binding ThDP and binding the analogue, and then the level of
residual activity will indicate the proportion of active sites that
Supporting Information
have ThDP bound. Therefore, ZmPDC was inactivated by furan
17 (1 µM) and then a large excess of ThDP (1 mM) was added
Supporting Information File 1
and the recovery in activity was followed over 3 days. Some
activity was slowly recovered (Figure 3) but this reached a
plateau within 24 hours at 8.5% of the untreated control (which
remained fully active over this period). The steady activity
regained in the presence of inhibitor is therefore the equilib-
rium position between ThDP binding and inhibitor binding.
Experimental section along with 1H and 13C NMR spectra
for all the compounds synthesised.
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-10-270-S1.pdf]
Acknowledgements
The time-course for recovery of activity fits nicely to single
exponential equation with an apparent first order rate constant
of 0.13 h−1. Using the ratio of recovered activity over unrecov-
ered activity at equilibrium (8.5:91.5), the relative concentra-
tions of ThDP and inhibitor (1000:1) and the previously
reported KD value for ThDP (0.35 µM) [39], one can calculate
that the Ki value of furan analogue 17 is 0.35 × (8.5/91.5) ×
(1/1000) µM = 32.5 pM. This Ki value is more than 10,000-fold
lower than the KD value for ThDP and is in the same range as
previously reported for triazole analogues of ThDP (30 and
This work was supported by a studentship from the Cambridge
Commonwealth Trust (A.I.).
References
1. Schellenberger, A. Biochim. Biophys. Acta 1998, 1385, 177–186.
doi:10.1016/S0167-4838(98)00067-3
2. Kluger, R. Chem. Rev. 1987, 87, 863–876. doi:10.1021/cr00081a001
3. Breslow, R. J. Am. Chem. Soc. 1957, 79, 1762–1763.
doi:10.1021/ja01564a064
4. Breslow, R. J. Am. Chem. Soc. 1958, 80, 3719–3726.
doi:10.1021/ja01547a064
2584

Beilstein J. Org. Chem. 2014, 10, 2580–2585.
5. Jordan, F. Nat. Prod. Rep. 2003, 20, 184–201. doi:10.1039/b111348h
6. Lüdtke, S.; Neumann, P.; Erixon, K. M.; Leeper, F. J.; Kluger, R.;
Ficner, R.; Tittmann, K. Nat. Chem. 2013, 5, 762–767.
doi:10.1038/nchem.1728
33.Lindqvist, Y.; Schneider, G. Curr. Opin. Struct. Biol. 1993, 3, 896–901.
doi:10.1016/0959-440X(93)90153-C
34.Candy, J. M.; Duggleby, R. G. Biochim. Biophys. Acta 1998, 1385,
323–338. doi:10.1016/S0167-4838(98)00077-6
7. Nemeria, N. S.; Chakraborty, S.; Balakrishnan, A.; Jordan, F. FEBS J.
2009, 276, 2432–2446. doi:10.1111/j.1742-4658.2009.06964.x
8. Nemeria, N. S.; Chakraborty, S.; Baykal, A.; Korotchkina, L. G.;
Patel, M. S.; Jordan, F. Proc. Natl. Acad. Sci. U. S. A. 2007, 104,
78–82. doi:10.1073/pnas.0609973104
35.Dobritzsch, D.; König, S.; Schneider, G.; Lu, G. J. Biol. Chem. 1998,
273, 20196–20204. doi:10.1074/jbc.273.32.20196
36.Dyda, F.; Furey, W.; Swaminathan, S.; Sax, M.; Farrenkopf, B.;
Jordan, F. Biochemistry 1993, 32, 6165–6170.
doi:10.1021/bi00075a008
9. Hawksley, D.; Griffin, D. A.; Leeper, F. J.
37.Erixon, K. M.; Dabalos, C. L.; Leeper, F. J. Org. Biomol. Chem. 2008,
6, 3561–3572. doi:10.1039/b806580b
J. Chem. Soc., Perkin Trans. 1 2001, 144–148. doi:10.1039/b006962k
10.Mann, S.; Perez Melero, C.; Hawksley, D.; Leeper, F. J.
Org. Biomol. Chem. 2004, 2, 1732–1741. doi:10.1039/b403619k
11.Jordan, F.; Li, H. J.; Brown, A. Biochemistry 1999, 38, 6369–6373.
doi:10.1021/bi990373g
38.Kluger, R.; Gish, G.; Kauffman, G. J. Biol. Chem. 1984, 259,
8960–8965.
39.Diefenbach, R. J.; Duggleby, R. G. Biochem. J. 1991, 276, 439–445.
12.Erixon, K.; Dabalos, C.; Leeper, F. J. Chem. Commun. 2007, 960–962.
doi:10.1039/b615861g
License and Terms
13.Hashmi, A. S. K.; Schwarz, L.; Choi, J.-H.; Frost, T. M.
Angew. Chem., Int. Ed. 2000, 39, 2285–2288.
This is an Open Access article under the terms of the
Creative Commons Attribution License
doi:10.1002/1521-3773(20000703)39:13<2285::AID-ANIE2285>3.0.CO
;2-F
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
14.Hashmi, A. S. K.; Frost, T. M.; Bats, J. W. J. Am. Chem. Soc. 2000,
122, 11553–11554. doi:10.1021/ja005570d
15.Istrate, F. M.; Gagosz, F. Beilstein J. Org. Chem. 2011, 7, 878–885.
doi:10.3762/bjoc.7.100
16.Li, E.; Yao, W.; Xie, X.; Wang, C.; Shao, Y.; Li, Y. Org. Biomol. Chem.
2012, 10, 2960–2965. doi:10.1039/c2ob07173h
17.Hoffmann, M.; Miaskiewicz, S.; Weibel, J.-M.; Pale, P.; Blanc, A.
Beilstein J. Org. Chem. 2013, 9, 1774–1780. doi:10.3762/bjoc.9.206
18.Wang, T.; Shi, S.; Hansmann, M. M.; Rettenmeier, E.; Rudolph, M.;
Hashmi, A. S. K. Angew. Chem., Int. Ed. 2014, 53, 3715–3719.
doi:10.1002/anie.201310146
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
19.Wang, T.; Shi, S.; Rudolph, M.; Hashmi, A. S. K. Adv. Synth. Catal.
2014, 356, 2337–2342. doi:10.1002/adsc.201400356
20.Wang, T.; Huang, L.; Shi, S.; Rudolph, M.; Hashmi, A. S. K.
Chem. – Eur. J. 2014, in press. doi:10.1002/chem.201404229
21.Muzart, J. Tetrahedron 2008, 64, 5815–5849.
doi:10.1016/j.tet.2008.04.018
doi:10.3762/bjoc.10.270
22.Shen, H. C. Tetrahedron 2008, 64, 3885–3903.
doi:10.1016/j.tet.2008.01.081
23.Hashmi, A. S. K. Chem. Rev. 2007, 107, 3180–3211.
doi:10.1021/cr000436x
24.Gorin, D. J.; Toste, F. D. Nature 2007, 446, 395–403.
doi:10.1038/nature05592
25.Hashmi, A. S. K. Catal. Today 2007, 122, 211–214.
doi:10.1016/j.cattod.2006.10.006
26.Jiménez-Núñez, E.; Echavarren, A. M. Chem. Commun. 2007,
333–346. doi:10.1039/b612008c
27.Hashmi, A. S. K.; Hutchings, G. J. Angew. Chem., Int. Ed. 2006, 45,
7896–7936. doi:10.1002/anie.200602454
28.Hoffmann-Röder, A.; Krause, N. Org. Biomol. Chem. 2005, 3, 387–391.
doi:10.1039/b416516k
29.Rudolph, M.; Hashmi, A. S. K. Chem. Soc. Rev. 2012, 41, 2448–2462.
doi:10.1039/c1cs15279c
30.Hashmi, A. S. K.; Rudolph, M. Chem. Soc. Rev. 2008, 37, 1766–1775.
doi:10.1039/b615629k
31.Aponick, A.; Li, C.-Y.; Malinge, J.; Marques, E. F. Org. Lett. 2009, 11,
4624–4627. doi:10.1021/ol901901m
32.Ravindar, K.; Reddy, M. S.; Deslongchamps, P. Org. Lett. 2011, 13,
3178–3181. doi:10.1021/ol201102x
2585