Aminopyrazine Pathway to the Moco Metabolite Dephospho Form A

Article abstract of DOI:10.1002/chem.201702274

An efficient synthesis of the molybdopterin/molybdenum cofactor (Moco) oxidation product dephospho Form A is described that assembles the pteridinone system starting from an iodinated aminopyrazine. The sodium salt of dephospho Form A could be purified by precipitation from methanol, which paved the way to the title compound in the 100 mg range. By HPLC, the synthetic material was compared with a sample isolated from a recombinant Moco containing protein. Analysis of dephospho Form A is the only method that allows the quantification of the Moco content of crude cell extracts and recombinant protein preparations.

Full text of DOI:10.1002/chem.201702274

DOI: 10.1002/chem.201702274
Communication
&
Heterocycles
Aminopyrazine Pathway to the Moco Metabolite Dephospho
Form A
Anne Klewe,[a] Tobias Kruse,[b] and Thomas Lindel*[a]
Abstract: An efficient synthesis of the molybdopterin/mo-
lybdenum cofactor (Moco) oxidation product dephospho
Form A is described that assembles the pteridinone
system starting from an iodinated aminopyrazine. The
sodium salt of dephospho Form A could be purified by
precipitation from methanol, which paved the way to the
title compound in the 100 mg range. By HPLC, the syn-
thetic material was compared with a sample isolated from
a recombinant Moco containing protein. Analysis of de-
phospho Form A is the only method that allows the quan-
tification of the Moco content of crude cell extracts and
recombinant protein preparations.
Introduction
Oxidative degradation of proteins containing the molybdenum
cofactor (1, Moco) or molybdopterin (2) affords the 6-alkynyl-
pterin “Form A” (3), which can be dephosphorylated enzymati-
cally to dephospho Form A (4, Scheme 1).[1] Next to Moco/MPT
quantification of recombinant Moco dependent enzymes,[2,3]
quantification of dephospho Form A (4) has also become of
key importance for the determination of the Moco/MPT con-
tent also of those proteins that are involved in the biosynthesis
of Moco.[4,5] The detection limit of 4 by HPLC/fluorescence
spectroscopy is in the picomolar range, which makes that
quantification protocol suitable for both recombinant bio-
chemistry and for the characterization of biological samples,
for example mutant strains.[6,7] The procedure was first report-
ed by Rajagopalan and co-workers,[8] and modified later.[9] The
amounts of dephospho Form A (4) that are routinely isolable
from natural sources do not allow analysis by NMR spectrosco-
Scheme 1. Conversion of the molybdenum cofactor (1, Moco, X=SH or cys-
teine side chain) or the Moco precursor molybdopterin (2) to dephospho
Form A (4); intermediates of the synthetic routes by the Taylor[10] and
Joule[12] groups.
py. Mass spectrometry is difficult due to the high buffer load
caused by the isolation protocol.
There are two syntheses of dephospho Form A (4) that in-
stall the alkynyl side chain at an already assembled aminopteri-
dinone-based precursor. Taylor et al.[10] proceeded via 6-chlor-
opterin (5, Scheme 1), which can be obtained from the rather
polar 2,4-diaminopteridine 8-oxide or from the better soluble
2-amino-4-(pentyloxy)pteridine.[11] Joule et al.[12] derived the
pteridinone-based phosphonate 6 from folic acid that was ex-
tended by a C3 chain, followed by conversion of the resulting
alkene to the required alkyne moiety.[12] The synthesis of alky-
nylpterins via 6-sulfonyloxypteridine has also been reported.[13]
However, we still faced a severe supply problem and decided
to develop a novel route that would reliably provide amounts
around 100 mg. Since dephospho Form A (4) is routinely quan-
tified on an achiral HPLC column, we considered it sufficient to
address the synthesis of rac-4.
[a] A. Klewe, Prof. Dr. T. Lindel
Institute of Organic Chemistry
TU Braunschweig
Hagenring 30
38106 Braunschweig (Germany)
E-mail: th.lindel@tu-braunschweig.de
[b] Dr. T. Kruse
Institute of Plant Biology
TU Braunschweig
Results and Discussion
Spielmannstr. 7
38106 Braunschweig (Germany)
Initially, we investigated the synthesis of 6-iodopterin (10).
Methyl 3-aminopyrazine-2-carboxylate (7, Scheme 2) was ob-
tained from the corresponding carboxylic acid by esterification
with methanol in the presence of sulfuric acid.[14] Iodination of
Supporting information and the ORCID identification number(s) for the au-
thor(s) of this article can be found under https://doi.org/10.1002/
chem.201702274.
Chem. Eur. J. 2017, 23, 1 – 5
1
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Communication
Scheme 2. Attempted synthesis of 6-iodopterin (10).
7 (N-iodosuccinimide (NIS), 2 equiv) afforded methyl 3-amino-
6-iodopyrazine-2-carboxylate (8) in 92% yield.[15] In DMSO, the
product precipitated after addition of water to the reaction
mixture. In the HMBC spectrum of 8, 5-H correlated to C-3 and
C-6, but only weakly to C-2, supporting the desired regioselec-
tivity. Attempts to exploit the ester function of 8 by condensa-
tion with guanidine in the presence of NaOMe and subsequent
cyclization to the pterin[16] failed. There was also the chance to
react methyl 3-chloro-6-iodopyrazine-2-carboxylate (9) with
guanidinium bicarbonate, a reaction that had worked for the
non-iodinated analogue.[17] After replacement of the amino
group of 8 by a chloro substituent (45%), treatment of 9 with
guanidinium bicarbonate at 1408C afforded a complex mixture
of fluorescent products, among which we isolated only 5% of
10.
As an alternative, we attempted to obtain guanidino partial
structure of 6-iodopterin by guanylation of the amino group of
8. However, neither cyanamide nor carbamimidic chloride were
effective, even at 1708C under neat conditions. Reaction of an-
thranilic acid with cyanamide had provided the corresponding
aminoquinazolinone.[18,19] Also, carbamimidic chloride reacted
with ethyl anthranilate in good yield.[20]
Scheme 3. Addition of aminoarenes to ethoxycarbonylisothiocyanate, cycli-
zation reactions.
to 82% by employing 9 equivalents of SCNCO2Et. Lacking the
methoxycarbonyl group, 2-aminopyrazine (19) gave a yield of
60% (20) with 1 equivalent of SCNCO2Et. In a test reaction,
methyl 2-(3-(ethoxycarbonyl)thioureido)nicotinate (16) was
converted to the pyrido[2,3-d]pyrimidine derivative 21 in good
yield (71%), indicating the general feasibility of the pathway.
Unfortunately, the improved conditions could not be trans-
ferred to the case of the iodinated pyrazine derivative 8, and it
was not possible to iodinate the pyrazine-based thiourea 17 at
C-6, either. Instead, reaction of 17 with NIS in DMSO afforded
the thiadiazolo[2,3-a]pyrazine 22 (Scheme 3). Both NH signals,
which were present in the starting material at d=11.7 and
11.9 ppm, had disappeared, and the CH carbons had experi-
enced a downfield shift of 11–13 ppm. Mechanistically, iodina-
tion of the sulfur should occur first, followed by nucleophilic
attack of the pyrazine nitrogen N4 at the iodinated sulfur.[25]
The alkynyl moiety at C6 had to be introduced prior to the
formation of the thiourea. Methyl 3-amino-6-iodopyrazine-2-
carboxylate (8) was alkynylated under Sonogashira conditions
with a variety of terminal alkynes, such as but-3-yn-1-ol or 4-
ethynyl-2,2-dimethyl-1,3-dioxolane. Ethynyltrimethylsilane (23)
afforded alkynyl pyrazine derivative 24 (85%, Scheme 4), which
we considered a suitable intermediate towards dephospho
Form A (4). Treatment of 24 with excess SCNCO2Et in MeCN
provided the thioureido pyrazine carboxylate 25. Introduction
of the remaining nitrogen and ring closure to ethynyl pterin
26 was achieved on reaction of 25 with excess of hexamethyl-
disilazane in the presence of EDCI (76%). Desilylation of 26
with tetra-n-butylammonium fluoride trihydrate (TBAF·3H2O) in
MeCN afforded the terminal alkyne 27 (48%), which proved to
be the end point of the sequence since all attempts to
hydroxylalkylate 27 were unsuccessful.
We turned to the condensation of ethoxycarbonylisothiocya-
nate (SCNCO2Et) with amino-substituted heterocycles, since
Stanovnik et al. had reacted benzyl 3-aminopyrazine-2-carbox-
ylate with SCNCO2Et in boiling chloroform and obtained the
corresponding 3-thioureidopyrazine-2-carboxylate, which was
cyclized further to 2-thioxo-2,3-dihydropteridin-4(1H)-one.[21]
Chowdhury et al. had condensed methyl 3-aminopyrazine-2-
carboxylate with ethyl isothiocyanatoacetate.[22] Fabis et al. had
accessed N-3 unsubstituted quinazolones from anthranilic acid
esters and SCNCO2Et, followed by activation of the thiourea
with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI)
and reaction with hexamethyldisilazane.[23] The reaction is as-
sumed to start with the conversion of the thiourea moiety to a
carbodiimide by treatment with EDCI.[24]
In the absence of ring nitrogen atoms, starting from methyl
anthranilate (11), we obtained the thiourea 12 in moderate
yield (52%, Scheme 3). Introduction of one nitrogen into the
ring system did not hamper the condensation with SCNCO2Et
when compared to methyl anthranilate, as shown by products
14 and 16 that were obtained in 69% and 60% yield from 13
and 15, respectively. The presence of two nitrogen atoms in
methyl 3-aminopyrazine-2-carboxylate (7) diminished the yield
(25% of pyrazine-based thiourea 17), which could be improved
&
&
Chem. Eur. J. 2017, 23, 1 – 5
www.chemeurj.org
2
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communication
Scheme 4. Model study providing the 6-ethynylpterin derivative 27.
Thus, we chose to install the complete C4 side-chain prior to
construction of the pteridinone bicycle. Sonogashira coupling
of mono-TBS-protected but-3-yne-1,2-diol[10c] with 8 afforded a
mixture of regioisomeric silylated alkynyl pyrazines, resulting
from partial migration of the TBS group (40% combined yield).
Therefore, we replaced the OTBS (tert-butyldimethylsilyl) with
the OTBDPS (tert-butyldiphenylsilyl) group, which is less prone
to transetherification. For the Sonogashira coupling of 8 with
alkyne 28[12b] (Scheme 5), it also proved to be beneficial to re-
place NEt3 with iPr2NH. The yield of TBDPS-protected alkynyl-
pyrazine 29 was improved from 36 to 90%. We were pleased
to find that TBDPS-protected 6-alkynyl-3-aminopyrazine 29
could be converted to the corresponding alkynylated pterin
derivative 31 in 51% yield over two steps by treatment with
excess SCNCO2Et affording 30 and then hexamethyldisilazane
in the presence of EDCI in MeCN at room temperature.
Scheme 5. Synthesis of rac-dephospho Form A (rac-4).
7.13 ppm (Rajagopalan). It is known that an upfield shift of 7-H
occurs on deprotonation of pterin systems with 1 n NaOD.[26]
In the 13C NMR spectrum, chemical shifts of C-2 (164.4 ppm)
and C-4 (169.6 ppm) were also indicative for deprotonation.
For the neutral pterin, an upfield shift of about 10 ppm would
be expected for both signals.[27] Sodium salt 33 was converted
to the neutral form rac-4 by submitting an aqueous solution
to a weakly acidic cation exchanger column (Merck ion ex-
changer IV). The fluorescent material was collected, concentrat-
Desilylation of 31 (TBAF·3H2O in MeCN, RT, 10 min) afforded
the ethoxycarbonyl-protected dephospho Form A 32. It was
possible to purify 32 by chromatography and to isolate the
product in the high yield of 93% (Scheme 5). Problems arose
when trying to remove the ethoxycarbonyl group with either
NaOMe/MeOH or aqueous NaOH/MeOH, which afforded a
complex mixture of products. By reverting the order of depro-
tection steps, refluxing of 31 with aqueous NaOH (1m)/MeOH
(1:15) and precipitation with EtOH afforded a solid product,
which was treated with MeOH in an ultrasonic bath. The re-
maining solid precipitate revealed to be the sodium salt of rac-
dephospho-Form A (33, 62%), which was obtained in high
purity. From the supernatant solution, we isolated a mixture of
the monosilylated dephospho Form A derivatives 34 and 35
after chromatography on silica (16 and 5%), which were con-
verted to another 14% of rac-dephospho Form A (rac-4) by
treatment with TBAF·3H2O in CHCl3.
1
ed, and resubmitted to NMR spectroscopy. Now, the H NMR
data ([D6]DMSO) agreed well with those reported by Rajagopa-
lan and Taylor. The UV/Vis spectra (see the Supporting Informa-
tion) of the synthesized rac-4 in 0.1n HCl and 0.1n NaOH are
in accordance to those reported earlier by Rajagopalan and
Taylor. In 0.1m HCl, absorption maxima occur at 205 nm (log e
3.9), 271 nm (log e 3.9), 301 nm (log e 3.5), and 343 nm (log e
3.7). In 0.1m NaOH, peaks at 270 nm (log e 4.1), 297 nm (log e
3.8), and 381 nm (log e 3.76) are observed. Characteristically,
the long-wavelength absorption maximum at 343 nm in 0.1 n
HCl is shifted to 381 nm in 0.1n NaOH.
We compared our synthetic rac-dephospho Form A with de-
phospho Form A obtained by oxidation of recombinant Neuro-
spora crassa nitrate reductase Moco-dimer domain. The latter
was recombinantly expressed and purified from the Moco-ac-
cumulating E. coli strain TP1000 (see the Supporting Informa-
tion).[28]
1
The H NMR spectrum ([D6]DMSO) of rac-dephospho Form A
sodium salt (33) deviated from that of neutral dephospho
Form A (4, see the Supporting Information). One NH signal was
absent and the pterin hydrogen 7-H appeared at 8.35 ppm in
our case, compared to 8.72 for Rajagopalan’s sample and 8.71
to Taylor’s. The NH2 group appeared at 5.90 ppm, compared to
Chem. Eur. J. 2017, 23, 1 – 5
www.chemeurj.org
3
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Communication
Figure 1 shows the HPLC elution profiles of synthetic and
isolated dephospho Form A. Since the isolated material com-
monly still contained minor amounts of an unknown compo-
nent eluting earlier than dephospho Form A, we collected the
main peak for comparative analysis by fluorescence spectros-
copy. The fluorescence absorption and emission maxima of
synthetic and isolated 4 were identical. Upon excitation at
302 nm the maximum fluorescence emission of both isolated
and synthetic 4 was reached at 450 nm.
Keywords: cofactors · heterocycles · molybdenum · pteridine ·
synthesis
[1] R. R. Mendel, J. Biol. Chem. 2013, 288, 13165–13172.
[2] P. Ringel, J. Krausze, J. van den Heuvel, U. Curth, A. J. Pierik, S. Herzog,
R. R. Mendel, T. Kruse, J. Biol. Chem. 2013, 288, 14657–14671.
[3] J. C. Chi, J. Roeper, G. Schwarz, K. Fischer-Schrader, J. Biol. Inorg. Chem.
2015, 20, 277–286.
[4] J. Krausze, C. Probst, U. Curth, J. Reichelt, S. Saha, D. Schafflick, D. W.
Heinz, R. R. Mendel, T. Kruse, Biochem. J. 2016, 474, 163–178.
[5] J. Kuper, A. Llamas, H.-J. Hecht, R. R. Mendel, G. Schwarz, Nature 2004,
430, 803–806.
[6] G. Schwarz, J. A. Santamaria-Araujo, S. Wolf, H.-J. Lee, I. M. Adham, H.-J.
Grçne, H. Schwegler, J. O. Sass, T. Otte, P. Hꢁnzelmann, R. R. Mendel, W.
Engeland, J. Reiss, Hum. Mol. Genet. 2004, 13, 1249–1255.
[7] C. Probst, P. Ringel, V. Boysen, L. Wirsing, M. M. Alexander, R. R. Mendel,
T. Kruse, Fungal Genet. Biol. 2014, 66, 69–78.
[8] a) J. L. Johnson, B. E. Hainline, K. V. Rajagopalan, B. H. Arison, J. Biol.
Chem. 1984, 259, 5414–5422; b) K. Clinch, D. K. Watt, R. A. Dixon, S. M.
Baars, G. J. Gainsford, A. Tiwari, G. Schwarz, Y. Saotome, M. Storek, A. A.
Belaidi, J. A. Santamaria-Araujo, J. Med. Chem. 2013, 56, 1730–1738.
[9] G. Schwarz, D. H. Boxer, R. R. Mendel, J. Biol. Chem. 1997, 272, 26811–
26814.
[10] a) E. C. Taylor, P. S. Ray, I. S. Darwish, J. L. Johnson, K. V. Rajagopalan, J.
Am. Chem. Soc. 1989, 111, 7664–7665; b) E. C. Taylor, R. Dçtzer, J. Org.
Chem. 1991, 56, 1816–1822; c) E. C. Taylor, S. Goswami, Tetrahedron
Lett. 1991, 32, 7357–7360; see also d) S. Goswami, A. C. Maity, S. Chak-
raborty, M. K. Das, Tetrahedron Lett. 2013, 54, 2373–2376.
[11] D. Mohr, Z. Kazimirczuk, W. Pfleiderer, Helv. Chim. Acta 1992, 75, 2317–
2326.
[12] a) J. R. Russell, C. D. Garner, J. A. Joule, Tetrahedron Lett. 1992, 33, 3371–
3374; b) B. Bradshaw, D. Collison, C. D. Garner, J. A. Joule, Org. Biomol.
Chem. 2003, 1, 129–133.
Figure 1. HPLC elution profiles of rac-dephospho Form A (black,
c=5 ngmLÀ1, 4 pmol were analyzed) and of a dephospho Form A sample
isolated from the Neurospora crassa nitrate reductase Moco-dimer do-
main (red, c=5.6 ngmLÀ1, 4.56 pmol were analyzed; the Form A pool ana-
lyzed was derived from three independently obtained protein preparations);
detection by fluorescence at 451 nm (excitation wavelength 302 nm), sta-
tionary phase RP-18 Hypersil 120A ODS (4.6ꢃ250; 5 mm particle size), mobile
phase 5 mm aq. NH4OAc/MeOH (85:15), flow rate 1 mLminÀ1
.
[13] W. Nxumalo, A. Dinsmore, Heterocycles 2013, 87, 79–81.
[14] H. Duan, M. Ning, X. Chen, Q. Zou, L. Zhang, Y. Feng, L. Zhang, Y. Leng,
J. Shen, J. Med. Chem. 2012, 55, 10475–10489.
[15] W. D. Arnold, A. Gosberg, Z. Li, R. W. Steensma, M. E. Wilson, WO
2006015124, 2006.
Conclusion
In summary, we have devised a new pathway to rac-dephos-
pho Form A (rac-4). Differing from the routes developed by
the Taylor and Joule groups, our route assembles the pterin bi-
cycle of dephospho Form A after the installation of the C4 side
chain. Starting from methyl 3-aminopyrazine-2-carboxylate (7)
rac-dephospho Form A (rac-4) was obtained in five steps and
8% overall yield. With more than 100 mg of synthesized mate-
rial in hand, sufficient amounts required for quantitative bio-
chemistry are now available.
[16] M. Lang, J.-P. Schoeni, C. Pont, J.-P. Fleury, Helv. Chim. Acta 1986, 69,
793–802.
[17] G. P. G. Dick, H. C. S. Wood, J. Chem. Soc. 1955, 1379–1382.
[18] Y.-K. Kim, H.-K. Na, S.-J. Kwack, S.-R. Ryoo, Y. Lee, S. Hong, S. Hong, Y.
Jeong, D.-H. Min, ACS Nano 2011, 5, 4550–4561.
[19] A. V. Dolzhenko, A. V. Dolzhenko, W.-K. Chui, J. Heterocycl. Chem. 2008,
45, 173–176.
[20] D. J. McNamara, E. M. Berman, D. W. Fry, L. M. Werbel, J. Med. Chem.
1990, 33, 2045–2051.
ˇ
[21] B. Koren, B. Stanovnik, M. Tisler, Heterocycles 1987, 26, 689–697.
[22] A. Z. M. S. Chowdhury, Y. Shibata, M. Morita, K. Kaya, K. Hiratani, J. Heter-
ocycl. Chem. 2001, 38, 1173–1177.
[23] C. Lecoutey, C. Fossey, S. Rault, F. Fabis, Eur. J. Org. Chem. 2011, 2785–
2788.
[24] C. D. Hupp, J. J. Tepe, J. Org. Chem. 2009, 74, 3406–3413.
[25] B. Koren, B. Stanovnik, M. Tisler, Org. Prep. Proced. 1975, 7, 55–59.
[26] A. Dieffenbacher, W. von Philipsborn, Helv. Chim. Acta 1969, 52, 743–
763.
[27] U. Ewers, H. Gꢂnther, L. Jaenicke, Chem. Ber. 1974, 107, 3275–3286.
[28] T. Palmer, C.-L. Santini, C. Iobbi-Nivol, D. J. Eaves, D. H. Boxer, G. Giorda-
no, Mol. Microbiol. 1996, 20, 875–884.
Acknowledgements
The project was funded by the Deutsche Forschungsgemein-
schaft. We thank Dr. Phillip Ringel for providing the construct
allowing recombinant expression of the N. crassa nitrate reduc-
tase Moco-dimer domain.
Conflict of interest
Manuscript received: May 18, 2017
Accepted manuscript online: July 7, 2017
Version of record online: && &&, 0000
The authors declare no conflict of interest.
&
&
Chem. Eur. J. 2017, 23, 1 – 5
www.chemeurj.org
4
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communication
COMMUNICATION
&
Heterocycles
A. Klewe, T. Kruse, T. Lindel*
&& – &&
Aminopyrazine Pathway to the Moco
Metabolite Dephospho Form A
Nine million bicycles of pterin: Quan-
tification of the molybdenum cofactor
(Moco) in crude cell extracts and re-
combinant protein preparations requires
a synthetic standard of dephospho
Form A, which was synthesized on a
new route via an iodinated aminopyra-
zine. The alkynyl moiety had to be intro-
duced prior to the assembly of the
pterin bicycle.
Chem. Eur. J. 2017, 23, 1 – 5
www.chemeurj.org
5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ