Pyridine to aniline: An exceptional biologically driven rearrangement

Article abstract of DOI:10.1016/j.bmcl.2013.01.086

During the course of our study on the innovative ligand for nicotinic acetylcholinergic receptors, LNAChR, and in order to assess activity and toxicity profiles of the drug's metabolites, synthesis of the main metabolites was undertaken. This synthesis work was done in parallel by organic chemistry and by biotransformation of LNAChR. Filamentous fungus Aspergillus alliaceus (NRRL 315) neatly afforded three of the main metabolites, one of which arose from a very unexpected and very uncommon rearrangement.

Full text of DOI:10.1016/j.bmcl.2013.01.086

Bioorganic & Medicinal Chemistry Letters 23 (2013) 2217–2219
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Pyridine to aniline: An exceptional biologically driven
rearrangement
a
a
b
b
Thomas Arnauld a, , Jean-Yves Beaumal , François Lefoulon , Alain Petit , Tristan Renaud
⇑
a Synthesis Department, Technologie Servier, 27 Rue Eugène Vignat, 45000 Orléans, France
b Structural Analysis Department, Technologie Servier, 27 Rue Eugène Vignat, 45000 Orléans, France
a r t i c l e i n f o
a b s t r a c t
Article history:
During the course of our study on the innovative ligand for nicotinic acetylcholinergic receptors, LNAChR,
and in order to assess activity and toxicity profiles of the drug’s metabolites, synthesis of the main metab-
olites was undertaken. This synthesis work was done in parallel by organic chemistry and by biotransfor-
mation of LNAChR. Filamentous fungus Aspergillus alliaceus (NRRL 315) neatly afforded three of the main
metabolites, one of which arose from a very unexpected and very uncommon rearrangement.
Ó 2013 Elsevier Ltd. All rights reserved.
Received 31 October 2012
Revised 17 January 2013
Accepted 20 January 2013
Available online 29 January 2013
Keywords:
Kost–Sagitullin rearrangement
Pyridinium rearrangement
Pyridinium ring opening
Drug metabolism
Filamentous fungi
LNAChR acts as a partial agonist of central nicotinic receptors1
and is capable of inhibiting agonist-mediated activation of a7
4b2 and
3b2 receptors2 in a non-competitive manner. LNAChR
improves mnesic retention in two models of episodic memory in
the adult rat: social recognition and object recognition by acting
on the process of learning and consolidation.
During the course of preclinical development studies of any ac-
tive pharmaceutical ingredient (API), identification of the main
metabolites is routinely obtained from the combination of two
powerful methods, LC/MS and LC/NMR, which, when run under
the same LC conditions, allow a fairly reliable structure attribution
to each and every detectable peak of the chromatogram. Generally
speaking, it is expected from a metabolic degradation process that
most metabolites would be redox products of the active ingredient,
and their conjugates (hydroxylated, glucurono-, sulfo-, etc.).
A great number of metabolites were detected from in vitro and
in vivo experiments with LNAChR. Out of the whole set of metab-
olites that were detected, only the three predominant ones were to
be synthesized on a gram scale in order to get an absolute proof of
their structure but also in order to assess their in vivo pharmaco-
logical and toxicological activities.
(Scheme 1). Structure of the latter product was definitely ascribed
after comparison with the commercially available compound. As
for the two former products, their structure needed also to be val-
idated against authentic samples, based on Mass and NMR data.
For these two metabolites, both methods showed similar pro-
file, though significant differences were seen. By positive Electro-
Spray, both compounds displayed a MH+ = 306, which accounts
for the presence of one additional oxygen atom on the LNAChR
molecule (as suggested by HRMS). By NMR, they both displayed
two sets of aromatic signals; one fitting for a p-bromobenzoyl
group and another showing two neighbouring and one lone pro-
tons.3 Although this second set of protons was quite different for
,
a
a
Br
I–
N+
O
LNAChR
O
Br
OH
N+
LC/MS and LC/NMR analyses of urine and faeces samples of rats
dosed with radiolabelled 14C-LNAChR permitted to hypothesize on
the structure of the three main metabolites: 4- and 5-hydroxylated
species (2 and 3, respectively) and p-bromobenzoic acid (1)
Br
Br
O
HO
N+
1
3
O
OH
2
⇑
Corresponding author. Tel.: +33 238 238 000.
E-mail address: thomas.arnauld@fr.netgrs.com (T. Arnauld).
Scheme 1. Main metabolites.
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.01.086

2218
T. Arnauld et al. / Bioorg. Med. Chem. Lett. 23 (2013) 2217–2219
Br
Br
dine N-oxide (Scheme 3). Nucleophilic displacement of the nitro
group by sodium methoxide was easily performed. Deprotonation
of the exocyclic methyl group by potassium hydride and single at-
tack on methyl p-bromobenzoate gave the expected coupling prod-
uct, in a moderate yield, although in a reproducible manner.
Reduction of the N-oxide by titanium(0) prepared in situ, methyl-
ation by methyl trifluoromethanesulfonate and final deprotection
by concentrated HBr under microwave conditions afforded the ex-
pected 4-hydroxylated derivative (2).
N+
N+
O
OH
Enol
Ketone
Br
H
N+
With the 4- and 5-hydroxylated species in hand (2 and 3, resp.)
and in order to validate the structure hypotheses made for the two
metabolites, HPLC and spectral data had to be compared with LC/
MS and LC/NMR data of the metabolites from in vitro and/or
in vivo assays. There was indeed a perfect correlation in the latter
case (3), allowing ultimate identification of the 5-hydroxylated
metabolite. For the 4-hydroxy-derivative (2) however, results were
quite different in terms of retention time by HPLC, fragmentation
by Mass Spectrometry and chemical shifts by 1H NMR. This clearly
dismissed the 4-hydroxy-derivative (2) as a metabolite of LNAChR.
Out of the three main metabolites, one, called ‘Metabolite X’, re-
mained therefore unidentified, although it displayed a few com-
mon features with compounds 2 and 3. By Mass Spectrometry,
all three compounds have the same MH+ = 306. By 1H NMR, two
sets of aromatic protons are detected, one corresponding to the
p-bromobenzoyl moiety, and another one showing a set of two
neighbouring protons and one lone proton (fitting for a substituted
O
O
Enone
Acidic
Basic
medium
Enone
medium
Br
N
Scheme 2. Mesomeric structures.
the two metabolites, we eventually concluded that they belonged
to the 4- and the 5-hydroxylated species (2 and 3, respectively).
Before going any further, it must be pointed out that LNAChR is
a highly conjugated system from which at least two mesomeric
forms can coexist and can be characterized by 1H NMR (Scheme
2). pH has a key effect on the equilibrium, which can be totally dis-
placed from one form to the other when going from one end to the
other of the pH scale. Therefore, no assumption could be made
hydroxypyridine). In all cases, the presence of protons in
a-posi-
tion to the carbonyl was unclear. As we mentioned earlier (Scheme
2), at least two mesomeric forms can be seen by 1H NMR for the
parent drug LNAChR and some of its derivatives, witnessing of a
high resonance effect. On this account, it was not unusual to detect
from the NMR integrations of the protons in
bonyl group.
a position to the car-
Attempt to obtain the expected metabolites by direct biocon-
version of LNAChR was undertaken. A total of nineteen strains of
filamentous fungi were screened, most of which gave essentially
p-bromobenzoic acid as the main product. Over a one-week period,
out of the nineteen strains, six led to a complete consumption of
the starting material, and one only, Aspergillus alliaceus (NRRL
315), was found to yield an oxidation product of LNAChR, as iden-
tified by LC/MS (MLNAChRH+ + 16 = 306). In this transformation, only
two other products were detected: p-bromobenzoic acid on the
one hand and a heavier metabolite (MLNAChRH+ + 96 = 386) on the
other. The former bioconversion product was isolated and fully
characterized as the 5-hydroxylated metabolite (3). As for the lat-
ter product, it was left aside for the time being.
the two aliphatic protons in
a-position to the carbonyl at varying
chemical shifts and often with a very inaccurate integration. This
is essentially true considering that the NMR are run exclusively
in D2O, allowing a rapid exchange of the labile protons with a deu-
terium atom from the solvent.
By looking more in depth at the Mass Spectrometry data of the
actual ‘Metabolite X’, it appeared that this compound displayed a
very specific fragmentation pattern that had been seen in no other
metabolite, such as loss of a bromine radical and loss of a water
molecule. This tended to prove that a major structural change
had occurred in the molecule, although not in the p-bromobenzoyl
moiety which was clearly identified as a fragment by Electrospray
Mass Spectrometry, but rather on the pyridyl part of the molecule.
Such a fragmentation pattern had previously been seen during
the bioconversion study with Aspergillus alliaceus. It belonged to
the second major product which displayed a MH+ = 386. It then oc-
curred to us that this Mass could correspond to a ‘metabolite X’
precursor, like a sulfated species (LNAChR + OSO3: MH+ = 386), as
suggested by HRMS (Scheme 4).
The 4-hydroxylated derivative (2) was eventually obtained by
conventional chemistry in five steps from 2-methyl-4-nitropyri-
– O
O
+
N
O
O
O
a
b
+
N
+
+
N
N
O–
O–
O–
5
Br
4
6
Br
O
O
N
N+
O
O
d
c
O
+
N
LNAChR
8
Br
7
Br
Aspergillus
alliaceus
OH
O
Br
O
O
S
O
O
e
+
N
O
OH
N+
OH
O
Br
O
Br
9
Metabolite X
2
HO
Br
1
3
Scheme 3. Reagents and conditions: (a) Na/MeOH, 0 °C; (b) KH, 9, rt; (c) TiCl4/Mg,
reflux then 0 °C; (d) CF3SO3Me; (e) HBr/H2O, W.
l
Scheme 4. Bioconversion with Aspergillus alliaceus.

T. Arnauld et al. / Bioorg. Med. Chem. Lett. 23 (2013) 2217–2219
2219
OH
NH
HO
HO
HO
O
O
O
O
O
O
O
Br
O
H2O
+
N
+
N
P450
+
N
HN
Br
Br
Br
Br
10
LNAChR
-H2O
-H2O
OH
HO
Br
O
+
N
Br
NH
O
3
11
Scheme 5. Pyridinium rearrangement.
‘Kost–Sagitullin rearrangement’ were later applied to pyrimidines
and reported in a review by Danagulyan.7
OH
OH
O
a
b
c
As a final proof of structure for metabolite 11, a synthesis work
was undertaken as shown in Scheme 6. 4-Hydroxyaniline was N-
Boc protected and then N- and O-methylated. Boc-oriented
lithiation of the ring allowed a nucleophilic attack on methyl
p-bromobenzoate, yielding benzophenone 15, which could be
O-methyl and N-Boc deprotected in one step using concentrated
HBr under microwave irradiation.
NH
NH2
N
BOC
BOC
OH
13
12
14
O
Br
Br
d
As expected, spectral and LC data were found to fit perfectly
those from the metabolite detected in the urine samples and from
the desulfated metabolisation product of the bioconversion.
In conclusion, two major metabolites of nicotinic ligand of acet-
ylcholinergic receptors LNAChR were synthesized using conven-
tional organic chemistry and biotransformation with filamentous
fungus Aspergillus alliaceus. Of these two metabolites, one was
shown to be the product of a very uncommon rearrangement from
pyridinium to aniline, the first example of a biologically driven
‘Kost–Sagitullin rearrangement’.
NH
O
N
O
11
BOC
15
Scheme 6. Reagents and conditions: (a) (Boc)2O, DMF, rt; (b) NaH, CF3SO3Me, ether,
rt; (c) t-BuLi, p-Br(C6H4)CO2Me, ether, À78 °C; (d) HBr, W.
l
An upscaled reaction with Aspergillus alliaceus was therefore
undertaken and yielded, as expected, a mixture of the 5-hydroxyl-
ated metabolite (3), p-bromobenzoic acid (1) and the unknown sul-
fated species, which after preparative HPLC neatly decomposed by
losing spontaneously sulfur trioxide SO3.
This new species was eventually identified as benzophenone 11
and possessed all the analytical features of ‘Metabolite X’ (Mass,
NMR, Retention Time).
It is noteworthy that opening of a pyridinium ring by hydrolysis
of the cyclic Schiff base like C@N double bond has been postulated
a few number of times in the literature.4 In our case, this is cer-
tainly the key step in the mechanistic hypothesis, in which the
rearrangement would occur after oxidation by cytochrome P450
(Scheme 5).5 The epoxide intermediate will hydrolyze, leading to
the diol, then to the aldehyde with concomitant opening of the
pyridinium ring. The aldehyde functionality is then free to undergo
cyclization by internal Knoevenagel reaction with the methylene in
Acknowledgment
The authors wish to thank Xavier Poisson for helpful discussion
on metabolisation pathways.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2013.01.
086.
References and notes
a
-position to the ketone to yield 11 as ‘Metabolite X’.
It is our hypothesis that sulfatation occurred at the final stage of
1. Meth-Cohn, O.; Yu, C.-Y.; Lestage, P.; Lebrun, M.-C.; Caignard, D.-H.; Renard, P.
EP 1050531.
this mechanism, but we have no indication whatsoever that this
has not occurred anywhere else in the process. The fact that com-
pounds 3 and 11 are the main metabolites in animal species and in
biotransformation by Aspergillus alliaceus stress on the hypothesis
that these two compounds are produced from the same intermedi-
ate which could be the diol 10. Spectral data (1H NMR and Mass
spectrometry) and LC retention time for 11 are in complete agree-
ment with the ones observed for the original metabolite detected
in the urine and faeces samples.
Examples of such rearrangement in the literature from a pyrid-
inium salt to an aniline are scarce and to the best of our knowledge
have been exclusively reported by Sagitullin et al. under conven-
tional chemistry conditions.6 Some examples of the now-called
2. Hogg, R. C.; Raggenbass, M.; Bertrand, D. Rev. Physiol. Biochem. Pharmacol. 2003,
147, 1.
3. Proton sets for pyridinium and aniline moieties (1H NMR (400 MHz, DMSO-d6) d
ppm): (2) 8.65 (d, J = 7.2 Hz, 1H), 7.28 (d, J = 2.8 Hz, 1H), 7.23 (dd, J = 7.2 Hz,
2.8 Hz, 1H) (3) 8.62 (d, J = 2.4 Hz, 1H), 7.97 (dd, J = 8.8 Hz, 2.4 Hz, 1H), 7.85 (d,
J = 8.8 Hz, 1H). (11) 7.01 (dd, J = 9.0 Hz, 2.9 Hz, 1H), 6.75 (d, J = 2.9 Hz, 1H), 6.71
(d, J = 9.0 Hz, 1H)
4. van der Plas, H. C. Tetrahedron 1985, 4, 237.
5. (a) Balogh, G. T.; Keserü, G. M. Arkivoc 2004, 7, 124; (b) Higuchi, T.; Hirobe, M. J.
Mol. Catal. 1996, 113, 403.
6. (a) Shkil, G. P.; Lusis, V.; Muceniece, D.; Sagitullin, R. S. Tetrahedron 1995, 51,
8599; (b) Shkil, G. P.; Sagitullin, R. S. Tetrahedron Lett. 1994, 35, 2075. and
references cited therein.
7. Danagulyan, G. G. Chem. Heterocycl. Compd. 2005, 41, 1205.