Synthesis and oxidation of 2-hydroxynevirapine, a metabolite of the HIV reverse transcriptase inhibitor nevirapine

Article abstract of DOI:10.1039/c1ob06052j

Nevirapine (11-cyclopropyl-5,11-dihydro-4-methyl-6H-dipyrido[3,2-b: 2′,3′-e][1,4]diazepin-6-one, NVP) is a non-nucleoside HIV-1 reverse transcriptase inhibitor used to prevent mother-to-child transmission of the virus. However, severe hepatotoxicity and serious adverse cutaneous effects have raised concerns about the safety of NVP administration. NVP metabolism yields several phenol-type derivatives conceivably capable of undergoing further metabolic oxidation to electrophilic quinoid species that could react with bionucleophiles. The covalent adducts thus formed might be at the genesis of toxic responses. As an initial step to test this hypothesis, we synthesized the phenolic metabolite, 2-hydroxy-NVP, and investigated its oxidation in vitro. Using potassium nitrosodisulfonate and sodium periodate as model oxidants, we obtained evidence for fast generation of an electrophilic quinone-imine, which readily underwent hydrolytic conversion to fully characterized spiro derivatives, 1′-cyclopropyl-4-methyl-1H,1′H-spiro[pyridine-2, 2′-pyrido[2,3-d]pyrimidine]-3,4′,6(3′H)-trione in aqueous media and 1′-cyclopropyl-4-methyl-1′H,2H-spiro[pyridine-3,2′- pyrido[2,3-d]pyrimidine]-2,4′,6(1H,3′H)-trione in non-aqueous media. The spiro compound generated in aqueous solution underwent subsequent hydrolytic degradation of the NVP ring system, whereas the one formed in non-aqueous media was stable to hydrolysis. The product profile observed with the chemical oxidants in aqueous solution was replicated using lactoperoxidase-mediated oxidation of 2-hydroxy-NVP. These observations suggest that metabolic activation of NVP, via Phase I oxidation to 2-hydroxy-NVP and subsequent generation of a quinone-imine, could occur in vivo and play a role in NVP-induced toxicity.

Full text of DOI:10.1039/c1ob06052j

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 7822
www.rsc.org/obc
PAPER
Synthesis and oxidation of 2-hydroxynevirapine, a metabolite of the HIV
reverse transcriptase inhibitor nevirapine†
Alexandra M. M. Antunes,*a David A. Novais,a J. L. Ferreira da Silva,a Pedro P. Santos,a
M. Conceic¸a˜o Oliveira,a Frederick A. Belandb and M. Matilde Marques*a
Received 30th June 2011, Accepted 18th August 2011
DOI: 10.1039/c1ob06052j
Nevirapine (11-cyclopropyl-5,11-dihydro-4-methyl-6H-dipyrido[3,2-b:2¢,3¢-e][1,4]diazepin-6-one, NVP)
is a non-nucleoside HIV-1 reverse transcriptase inhibitor used to prevent mother-to-child transmission
of the virus. However, severe hepatotoxicity and serious adverse cutaneous effects have raised concerns
about the safety of NVP administration. NVP metabolism yields several phenol-type derivatives
conceivably capable of undergoing further metabolic oxidation to electrophilic quinoid species that
could react with bionucleophiles. The covalent adducts thus formed might be at the genesis of toxic
responses. As an initial step to test this hypothesis, we synthesized the phenolic metabolite,
2-hydroxy-NVP, and investigated its oxidation in vitro. Using potassium nitrosodisulfonate and sodium
periodate as model oxidants, we obtained evidence for fast generation of an electrophilic
quinone-imine, which readily underwent hydrolytic conversion to fully characterized spiro derivatives,
1¢-cyclopropyl-4-methyl-1H,1¢H-spiro[pyridine-2,2¢-pyrido[2,3-d]pyrimidine]-3,4¢,6(3¢H)-trione in
aqueous media and 1¢-cyclopropyl-4-methyl-1¢H,2H-spiro[pyridine-3,2¢-pyrido[2,3-d]pyrimidine]-
2,4¢,6(1H,3¢H)-trione in non-aqueous media. The spiro compound generated in aqueous solution
underwent subsequent hydrolytic degradation of the NVP ring system, whereas the one formed in
non-aqueous media was stable to hydrolysis. The product profile observed with the chemical oxidants in
aqueous solution was replicated using lactoperoxidase-mediated oxidation of 2-hydroxy-NVP. These
observations suggest that metabolic activation of NVP, via Phase I oxidation to 2-hydroxy-NVP and
subsequent generation of a quinone-imine, could occur in vivo and play a role in NVP-induced toxicity.
The starting CART regimens recommended by World Health
Introduction
Organization guidelines typically include
a non-nucleoside
reverse transcriptase inhibitor (NNRTI).4 In 1996, nevi-
rapine (11-cyclopropyl-5,11-dihydro-4-methyl-6H-dipyrido[3,2-
b:2¢,3¢-e]diazepin-6-one (1, NVP, Scheme 1) was the first NNRTI
approved by the US FDA for use in combination therapy for
HIV-1 infection.5 The high efficacy levels of the drug, favorable
lipid profile, and suitability for use during pregnancy have
since granted NVP-based regimens a significant role in HIV-
1 treatment strategies.6,7 Currently, NVP is one of the most
prescribed antiretroviral drugs in the developing world, both as a
single drug to prevent mother-to-child HIV transmission and as a
component of CART.8–11 While the immediate-release formulation
currently on the market is approved for twice-daily dosing, a
more convenient once-daily alternative was approved by the US
FDA in March 2011.12 This new option should be particularly
beneficial in resource-limited countries, where adherence is a
recurrent problem, and may lead to an increased worldwide use of
the drug.
The chronic treatment of individuals infected with the human
immunodeficiency virus type 1 (HIV) is expected to remain
a vital need in the near future, since both eradication of the
virus and an ending to new infections are unlikely scenarios.
Despite the indisputable benefits of the currently recommended
combined antiretroviral therapy (CART), and although individual
susceptibilities differ among patients,1 concerns with the adverse
effects of long-term CART are emerging.2,3 Thus, an under-
standing of the molecular mechanisms underlying the toxicity of
antiretroviral drugs is essential for establishing dependable risk-
benefit relationships that can guide decisions on treatment options.
aCentro de Qu´ımica Estrutural, Instituto Superior Te´cnico, Universidade
Te´cnica de Lisboa, 1049-001, Lisboa, Portugal. E-mail: alexandra.antunes@
ist.utl.pt; Tel: 351-21-8419388, matilde.marques@ist.utl.pt; Tel: 351-21-
8419200; Fax: 351-21-8464455
bDivision of Biochemical Toxicology, National Center for Toxicological
Research, Jefferson, AR 72079, USA
† Electronic supplementary information (ESI) available: supplementary
data and crystallographic data for compound 18. CCDC reference number
821404. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/c1ob06052j
Despite its clinical efficacy and suitability for use during
pregnancy13 and breastfeeding,14 NVP is associated with a variety
of toxic responses. The most severe is hepatotoxicity, which can
7822 | Org. Biomol. Chem., 2011, 9, 7822–7835
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
covalent adducts. Evidence for the binding of NVP metabolites
to bionucleophiles in vivo was presented by Srivastava et al.31
who identified two NVP mercapturates, the major one through
NVP-C3 and the minor one through NVP-C12 (10 and 11,
respectively, Scheme 2), in the urine of HIV-positive patients
administered NVP as part of a standard antiretroviral therapeutic
regimen. While the minor mercapturate presumably resulted from
the previously proposed bioactivation pathway through 12-OH-
NVP, it was hypothesized31 that the formation of the NVP-C3
adduct with N-acetylcysteine occurred by nucleophilic attack of
glutathione on either the quinone methide 8 or an arene oxide
intermediate (12, Scheme 2). This report represented the first
indication that metabolic pathways other than 12-hydroxylation
could have a role in NVP-induced toxicity. In particular, it
can not be excluded that the phenolic metabolites (e.g., 2-OH-
NVP), generated from CYP-mediated oxidation to arene oxide
(e.g., 12) and/or diradical (e.g., 13) precursors, may undergo
metabolic activation to quinone/semiquinone electrophiles (e.g.,
14) capable of reacting with bionucleophiles through Michael-type
addition and/or Schiff-base formation, leading to covalent adduct
formation (Scheme 2).32–34
Scheme 1 Structures of NVP (1) and the NVP metabolites and derivatives
(2–9) mentioned in the text.
To gain insight into the potential role of phenolic metabolites in
the toxicity associated with NVP administration, we synthesized
2-OH-NVP and investigated its oxidation in vitro. We report herein
evidence for the formation of an electrophilic quinone-imine
intermediate, using both model chemical oxidants [potassium
nitrosodisulfonate (Fre´my’s salt) and sodium periodate] and
lactoperoxidase catalysis.
be fatal; the most common side effect is skin rash, which may be
life threatening and lead to drug discontinuation.15,16 Moreover,
although NVP is neither mutagenic nor clastogenic in standard in
vitro assays, it induces hepatoneoplasias in rodents.17 In addition,
while no direct correlation between NVP administration and
the occurrence of cancer in humans has been reported, recent
epidemiological data on the incidence of non-AIDS defining
cancers in individuals undergoing CART suggest that the use of
NNRTIs is a risk factor.18 These observations raise concerns about
prolonged administration of the drug.
Results and discussion
Synthesis of 2-OH-NVP
While the reasons for the adverse effects of NVP are still
unclear, increasing evidence suggests that metabolic activation to
highly reactive electrophiles, prone to react with bionucleophiles,
has a role in the initiation of the toxic responses. For instance,
although circumstantial, the report of fast clinical recovery of a
patient suffering from NVP-induced toxic epidermal necrolysis
and toxic hepatitis upon treatment with a continuous infusion
of human immunoglobulins and a high dose (300 mg day-1)
of N-acetylcysteine,19 supports the involvement of reactive NVP
metabolites in the toxic response. Indeed, it is plausible that admin-
istration of N-acetylcysteine, an excellent nucleophile, detoxified
electrophilic NVP metabolites through the formation of readily
excreted mercapturate conjugates.
In all species investigated, cytochrome P450 (CYP)-mediated
Phase I metabolism of NVP yields 2-, 3-, 8-, and 12-OH-
NVP, and 4-carboxy-NVP (2–6, Scheme 1); the majority of
these metabolites undergo a detoxifying glucuronidation and
subsequent excretion.20–24 It has been suggested that 12-OH-
NVP is the species responsible for a skin rash in rats that
resembles the rash in humans,25,26 and sulfotransferase-catalyzed
conversion of 12-OH-NVP into 12-sulfoxy-NVP (7), possibly
followed by hydrogen sulfate elimination to an NVP quinone
methide (8), has been proposed as the bioactivation pathway
in the process.25,27 Using 12-mesyloxy-NVP (9) as a synthetic
surrogate for 12-sulfoxy-NVP, we demonstrated the potential of
this species to react in vitro with DNA and proteins28–30 yielding
Two distinct methods are reported in the literature for the
preparation of 2-OH-NVP (2), starting from either NVP or 2-
chloro-NVP. However, multiple synthetic steps are involved in
both instances, yielding 2 with an overall yield of less than 10%.35
For the ultimate goal of investigating the oxidation of 2-OH-
NVP, it was essential to have a considerable amount of this
NVP metabolite; for this reason, our initial efforts focused on the
development of a new and more efficient synthetic methodology.
The addition of 8.0 molar equivalents of silver acetate
and 4.8 molar equivalents of iodine to an NVP solution in
dichloromethane resulted in an efficient method for selective
acetoxylation at the aromatic NVP C2, affording 2-acetoxy-NVP
(15, Scheme 3) after only 10 min of reflux. Upon subsequent
hydrolysis, 2-OH-NVP was obtained in an overall 85% yield. This
fast and cost-efficient two-step strategy represents a significant
improvement when compared with the already available methods
for C2 hydroxylation of NVP. The iodine/silver acetate system has
found some limited synthetic use, namely in the synthesis of anti-36
and syn-diols37 from alkenes, in the preparation of b-iodoketones,38
in the acetoxylation of porphyrins,39 and in the iodination of
N-heterocycles;40 however, to the best of our knowledge, this is
the first time it is employed for the acetoxylation of a pyridine
ring. We found the efficiency of NVP acetoxylation to be strongly
dependent on the order of reagent addition. When iodine was
the first reagent to be added, acetoxylation was almost negligible,
which rules out activation via direct nucleophilic attack on iodine,
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7822–7835 | 7823
©

View Article Online
Scheme 2 Hypothetic pathways for in vivo generation of covalent adducts from 2-hydroxy-NVP (2). GSH, glutathione; SULT, sulfotransferase.
with formation of an iodopyridinium species (e.g. 16, Scheme 3, or
its di-iodo analogue).41 Our observations suggest an initial reaction
of silver acetate with iodine, affording acetyl hypoiodite (17),
which is a known iodination reagent42 and alcohol oxidant.43 The
transient formation of this species is consistent with the copious
precipitation of a yellow solid (presumed to be silver iodide) after
reagent addition, similar to what has been reported by Barnett
and co-workers.42 Moreover, the fact that a similar compound,
acetyl hypofluorite, is reported to promote the acetoxylation
of pyridine rings44,45 reinforces this hypothesis. Light-induced
homolytic cleavage of the I–O bond in acetyl hypoiodite is reported
to occur during alcohol oxidation;43 however, the formation
of 2-acetoxy-NVP also occurred in the dark, which supports
an ionic, rather than a free radical, mechanism. Presumably,
the iodopyridinium analogue (16) was formed by reaction of
NVP with acetyl hypoiodite, followed by a Chichibabin-type
nucleophilic addition of acetate on the NVP C2 and subsequent
HI elimination.44,45
Chemical oxidation reactions
The metabolic oxidation of 2-OH-NVP to quinoid species is a
plausible event, similar to what is observed with other phenolic
metabolites and/or parent drugs; the electrophilic metabolites
thus generated are widely thought to be involved in the toxicity
of numerous xenobiotics.32–34 To address the potential role of
oxidation products from 2-OH-NVP in the adverse responses
associated with NVP therapy, we investigated the oxidation of
Scheme 3 Proposed mechanism for the synthetic conversion of NVP (1)
into 2-OH-NVP (2).
7824 | Org. Biomol. Chem., 2011, 9, 7822–7835
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Table 1 Experimental conditions used for the preliminary oxidation assays of 2-OH-NVP (10 mg, 35 mmol) with Fre´my’s salt or sodium periodate
Entry
Oxidant
Solvent A
Solvent B
1
3
4
5
6
7
8
9
Fre´my’s salt (12 mg, 45 mmol)
acetonitrile (800 mL)
acetonitrile (800 mL)
acetonitrile (800 mL)
DMF (500 mL)
acetonitrile (800 mL)
acetonitrile (800 mL)
acetonitrile (800 mL)
DMF (500 mL)
100 mM phosphate (pH 10; 400 mL)
100 mM phosphate (pH 7.4; 400 mL)
100 mM phosphate (pH 5; 400 mL)
DMF (500 mL)
100 mM phosphate (pH 10; 400 mL)
100 mM phosphate (pH 7.4; 400 mL)
100 mM phosphate (pH 5; 400 mL)
DMF (500 mL)
NaIO4 (7.5 mg, 35 mmol)
this metabolite in vitro and characterized the major products. As
model oxidants, we chose the radical potassium nitrosodisulfonate
(Fre´my’s salt) and sodium periodate, which are frequently em-
ployed to obtain quinones from phenolic compounds.46–48 The use
of Fre´my’s salt has the advantage of mimicking the one-electron
oxidation involved in enzyme-mediated metabolic oxidations49,50
whereas periodate oxidation is thought to proceed by an ionic
mechanism.51
Preliminary reactions with Fre´my’s salt were conducted at
room temperature using 1.3 molar equivalents of the oxidant
and a biphasic system of ethyl acetate and 100 mM phosphate
buffer (pH 7.4 or 10). Due to the low solubility of 2-OH-NVP
and the difficulty involved in HPLC monitoring of a biphasic
mixture, alternative solvent systems were subsequently screened.
These included DMF and acetonitrile/100 mM phosphate buffer
at different pHs (5, 7.4, and 10). Sodium periodate oxidation
assays were conducted under similar conditions, using equimolar
amounts of oxidant and 2-OH-NVP (Table 1). In each instance,
reactions were monitored by HPLC-DAD at least 1 and 24 h
after addition of the oxidant. The yield of each product was
calculated on the basis of the HPLC peak area, corrected for the
corresponding molar absorptivity at the monitored wavelength
(254 nm). For the purpose of spectroscopic characterization of
the major products, larger scale reactions were performed and the
products were isolated by semi-preparative HPLC or PTLC (cf.
Experimental section).
Although the pattern of oxidation products was considerably
affected by solvent and pH, spiro compounds [either 18 or 19
(a/b/c)] were always observed (Scheme 4; vide infra for the
spectroscopic characterization). Despite the probable mechanistic
differences, the formation of these spiro compounds, regardless of
the oxidant used, suggests a fast generation of a transient elec-
trophilic quinone-imine (20, Scheme 4), which was not detected.
Interestingly, 18 was only observed when the oxidations were
conducted in aqueous systems, whereas its isomer(s) 19 (a/b/c)
appeared in DMF only (Fig. 1). The formation of the two spiro
compounds can be explained on the basis of nucleophilic attack
by water on positions C4a and C11a of the quinone-imine (20),
yielding the ring-opened intermediates 21 and 22, respectively,
which subsequently underwent intramolecular ring closure to 18
and 19 (Scheme 4). In particular, the formation of 19 in DMF in
the absence of added water can only be explained with involvement
of a highly electrophilic intermediate (i.e., the quinone-imine) that
can readily react with the residual water present.
of 2-(cyclopropylamino)nicotinic acid (23), most likely stemming
from initial water attack on C4a of the quinone-imine to yield 21,
the nicotinamide precursor of the spiro compound 18. However,
possibly due to decreased availability of the amide nitrogen
for nucleophilic attack as a result of strong DMF-nicotinamide
association via hydrogen bonding,52 intramolecular ring closure
to give 18 did not occur. Instead, consecutive hydrolysis steps
from 21 afforded the nicotinic acid 23. This process presumably
occurred with concomitant formation of 6-hydroxypyridine-2,3-
dione (Scheme 4) and/or its tautomers, but these species were not
detected, either due to further degradation or low intensity of the
chromophores. By contrast, when the oxidations were performed
in aqueous media we found no evidence for the formation of 19 or
its putative precursor, the ring-opened species 22. Steric hindrance
caused by the cyclopropyl substituent may have been a factor in
delaying initial nucleophilic attack on C11a of the quinone-imine,
compared to attack on C4a. A higher rate of nucleophilic attack
at C4a could plausibly shift the equilibria toward the formation of
18 in aqueous media, thus preventing the formation of 19.
When the reactions were conducted in aqueous systems, the
product profiles were considerably affected by the pH. For
instance, in the presence of Fre´my’s salt, the major product formed
after 1 h in basic medium (pH 10) was the spiro compound 18; this
material had nearly disappeared after 24 h, when the major prod-
uct corresponded to the pyridopyrimidine derivative 24, which
had increased approximately 4-fold (Fig. 1 and Scheme 5). By
contrast, the amount of the spiro derivative 18 formed after 1 h at
pH 7.4 was ca. 4-fold lower than at pH 10 but remained essentially
unchanged with time; under these conditions compound 24 was
not detected (Fig. 1). Moreover, the amount of spiro compound
18 formed after 1 h at pH 5 was approximately 10- and 3-fold
lower than at pH 10 and 7.4, respectively, but increased ca. 4-
fold after 24 h; compound 24 was detected only in trace amounts
after 24 h at pH 5 (Fig. 1). The overall oxidation efficiency with
Fre´my’s salt was found to decrease in the order pH10 > pH7.4 >
pH5. Taken together, these observations are consistent with the
proposed mechanism for the formation of the spiro compound
18 (Scheme 4), involving a transient quinone-imine intermediate
(20), which is expected to undergo base catalysis. The susceptibility
of 18 to degradation in alkaline media also suggests base
catalysis.
Contrasting with the results obtained with Fre´my’s salt, the
formation of the spiro compound 18 was maximal at pH 5 in the
presence of sodium periodate, indicating that periodic acid was the
oxidant.53 As such, there was a slight increase in the amount of the
spiro compound 18 between 1 and 24 h at pH 5. While the extent
of oxidation was low at both pH 7.4 and pH 10, the changes
The reasons for the solvent-determined selectivity toward 18 or
19 are not entirely clear. When the oxidations were performed in
DMF, besides the spiro tautomers 19, we observed the formation
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7822–7835 | 7825
©

View Article Online
Scheme 4 Proposed mechanism for the formation of the spiro derivatives 18 and 19 and the nicotinic acid 23, involving the transient generation of a
quinone-imine (20). The spiro compound 18 was formed in aqueous solution only, and the spiro compound 19 was formed in DMF.
To investigate further the degradation of 18 we isolated the
compound and monitored its behaviour in solution, both in the
absence and in the presence of Fre´my’s salt. Under the same
solvent conditions, the results were identical with or without
oxidant, which confirmed that oxidative processes were not
involved. In DMF/100 mM phosphate buffer, pH 10 (Fig. 2)
the pyridopyrimidine derivative 24 was already observed after
5 min of incubation; after 1 h, 18 and 24 were present in a ratio
close to 1 : 1 and after 24 h, the conversion of 18 into 24 was
almost complete (Fig. 2). Although this conversion was faster in
basic medium, 24 was also detected after 1 h in DMF/phosphate
buffer, pH 7.4; with longer incubation times (24 h) it was present
in both DMF/phosphate buffer, pH 5, and DMF alone (Fig. 2).
Interestingly, a residual amount of the spiro compound 18 was
detected after 24 h at pH 10, but not pH 7.4, despite the lower
rate of conversion at neutral pH (Fig. 2). This suggests some
extent of competitive base-catalyzed reversible conversion of 18
Scheme 5 Proposed mechanism for the hydrolytic degradation of the
spiro compound 18, leading to the irreversible formation of 24 and 26.
to the ring-opened species 21 and/or the quinone-imine 20, in
a sequence of events mirroring the formation of 18 from 20
displayed in Scheme 4. Moreover, given the significant stability of
18 under mildly acidic conditions and in DMF, our observations
are consistent with a base-catalyzed hydrolytic conversion of 18
to 24 (Scheme 5), presumably involving initial abstraction of the
in product profiles with time confirmed the stability of 18 in
neutral media and its susceptibility to degradation at alkaline pH
(Fig. 1).
7826 | Org. Biomol. Chem., 2011, 9, 7822–7835
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Fig. 1 Relative percentages of the compounds detected by HPLC-DAD in oxidation reactions of 2-OH-NVP, with Fre´my’s salt (panels a–c in
acetonitrile/100 mM phosphate buffer at pH 5, 7.4, and 10, respectively; panel d in DMF) and sodium periodate (panels e–g in acetonitrile/100 mM
phosphate buffer at pH 5, 7.4, and 10, respectively; panel h in DMF). The reactions were monitored at 254 nm and the quantifications were based on
peak intensities, corrected for the molar absorptivities of the different compounds at 254 nm.
N3¢-amide proton, with concomitant ring contraction to 25 and
subsequent collapse of this intermediate to yield two stable species,
24 and the maleimide 26. It should be noted that we were unable
to detect 26 in any of the assays, probably due to low intensity of
the chromophore at the wavelength (254 nm) used to monitor the
reaction mixtures by HPLC-DAD.
In contrast to its isomer, the spiro derivative 19 (a/b/c) was
very stable under the hydrolytic conditions used to study the
degradation of 18, both in the presence and in the absence of
Fre´my’s salt (not shown). Solvent exposure of the N3¢-amide
proton may be lower in 19, due to the vicinity of the methyl
substituent, thus preventing the initiation of the process. However,
it is more plausible that a base-catalyzed ring-contraction in 19 is
not favoured because it would require cleavage of a C–C bond,
rather than the C–N bond scission observed for 18.
Taking into consideration the phenol-type structure of
2-OH-NVP and the indirect evidence that an electrophilic
quinone-imine (20) is an intermediate in the oxidation of
this compound by both Fre´my’s salt and sodium periodate,
we investigated if a similar transformation could occur under
biologically plausible conditions. Toward this end, we incu-
bated 2-OH-NVP with hydrogen peroxide in the presence of
lactoperoxidase. The resulting mixture was analyzed by HPLC-
DAD. By comparison of the retention times and distinctive UV
profiles with those of the previously characterized standards, we
could easily identify the pyridopyrimidine derivative 24 (Fig. 3,
panels A and B). A signal with the retention time of the spiro
compound 18, not detected in the absence of lactoperoxidase,
was also present; however, there were slight differences in the
UV spectrum compared to the standard (not shown), possibly
due to co-elution with another component of the mixture. To
resolve this uncertainty, the incubation mixture was analyzed
by LC-ESI-MS. Based upon retention time, detection of the
protonated molecule (m/z 299), and MS/MS fragmentation
pattern (Fig. 3, panels C and D), we obtained clear evidence for the
formation of the spiro compound 18. These results demonstrate
that lactoperoxidase has the ability to catalyze the oxidation of
2-OH-NVP, yielding the same products obtained with Fre´my’s
salt and sodium periodate. Moreover, the data support our initial
hypothesis that the metabolism of NVP to 2-OH-NVP, and
subsequent metabolic activation to 20, is plausible in vivo, and
could be a factor in some of the toxic events associated with NVP
administration.
Enzyme-mediated oxidations
Lactoperoxidase, a member of the heme peroxidase family of
enzymes, is secreted from mammary, salivary, and other mucosal
glands and is known to catalyze the oxidation of a number of or-
ganic and inorganic substrates mediated by endogenous hydrogen
peroxide.54 It has been proposed that oxidation of the phenolic
ring of estradiol by lactoperoxidase, with production of free
radicals, may be involved in breast carcinogenesis.55,56 In addition,
it has been speculated that lactoperoxidase-catalyzed activation of
xenobiotic carcinogens (e.g., aromatic and heteroaromatic amines)
could also be a factor in the initiation of breast cancer.57
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7822–7835 | 7827
©

View Article Online
Fig. 2 Relative percentages of the compounds detected by HPLC-DAD during the stability assessment of the spiro compound 18 in different conditions.
The reactions were monitored at 254 nm and the quantifications were based on peak intensities, corrected for the molar absorptivities of the different
compounds at 254 nm.
Fig. 3 HPLC-DAD chromatograms of: A) the reaction mixture obtained after incubation of 2-OH-NVP with lactoperoxidase; B) standard 24; and
LC-MS/MS spectra of the protonated spiro molecule 18 (m/z 299) in: C) the incubation mixture containing 2-OH-NVP and lactoperoxidase; D) the
synthetic standard of the spiro compound.
Spectroscopic characterization of the oxidation products from
2-OH-NVP
molecules at m/z 299. Compound 19 had a greater propensity
to be sodiated under the ionization conditions used, showing an
intense [M + Na]+ peak at m/z 321. The MS/MS spectra of
the protonated molecules of both compounds yielded common
product ions, as a result of preferential protonation of the basic
pyridine nitrogen in both instances (cf. Scheme S1, Electronic
Supplementary Material†).
The most distinctive feature in the NMR spectra of the spiro
compounds was the presence of high field quaternary carbons (at
82.6 and 74.7 ppm, for 18 and 19, respectively; cf. Experimental
section and Electronic Supplementary Material†). This observa-
tion was the first evidence for a loss of aromaticity in ring A;
moreover, the resonances obtained were compatible with spiro
An initial comparison of the 1H and 13C NMR spectra of all the
products from 2-OH-NVP oxidation (cf. Experimental section)
promptly allowed the conclusion that ring C (Scheme 1) remained
unchanged in these derivatives, whereas a substantial degradation
of rings A and B had occurred during the oxidation process.
The structural similarity between the two isomeric spiro com-
pounds, 18 and 19, was initially inferred during the HPLC-DAD
separation process, whereupon the two compounds displayed very
close retention times and similar UV profiles (Fig. 4). The ESI-
MS spectra were consistent with isomers, exhibiting protonated
7828 | Org. Biomol. Chem., 2011, 9, 7822–7835
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Fig. 4 UV spectra of 2-OH-NVP (2), the spiro compounds 18 and 19, the nicotinic acid 23, and compound 24, isolated in oxidation reactions of
2-OH-NVP. The spectra were obtained online, by HPLC-DAD, in acetonitrile/0.1% aqueous formic acid.
quaternary carbons bound to at least one electron withdrawing
substituent.58 Although definite proof for the structure of 18 was
obtained by X-ray diffraction (vide infra), it should be noted
that this structure could be solved by analysis of the 2- and 3-
and shows aromaticity, with similar values for all bond lengths.
Bond lengths and angles (cf. Table S2, Electronic Supplementary
Material†) are well within the expected range, as judged from
extensive analysis of the values included in the Cambridge
Structural Database (CSD).59 The crystal structure of compound
18 revealed a solvent crystallization molecule (dimethylketone)
for two molecules of compound. Fig. 7 displays the interactions
between two molecules of 18 (one of isomer R and the other one
1
bond HMBC H-13C correlations (Fig. 5). In particular, 2-bond
correlations were observed between N3¢H (9.04 ppm) and both
the spiro quaternary carbon (C2; 82.6 ppm) and the amide C4¢
(161.5 ppm); likewise, a 3-bond correlation between the same
exchangeable proton and C4¢a (111.3 ppm) was detected. Further
crucial evidence was provided by the observation of 3-bond
correlations between the exchangeable N1H and both the olefinic
C5 (136.5 ppm) and the carbonyl C3 (191.2 ppm). All other
proton and carbon resonances were consistent with the proposed
structure.
of isomer S) and one solvent molecule. Thus, an R2 (8) synthon
2
is formed by N–H . . . O hydrogen bonds (N1R/S ◊ ◊ ◊ O9S/R = 2.895
A; H1R/S ◊ ◊ ◊ O9S/R = 2.004 A; N1R/S –H1R/S ◊ ◊ ◊ O9S/R = 177.1◦)
between the two enantiomers. In addition, a short C–H . . . O
˚
˚
˚
contact with the solvent (Cmethyl ◊ ◊ ◊ O9 = 3.378 A; Hmethyl ◊ ◊ ◊ O9 =
2.620 A; Cmethyl–Hmethyl ◊ ◊ ◊ O9 = 134.2◦) is also observed. Interaction
˚
The X-ray crystallographic data indicated that 18 crystallizes
between two molecules of the same enantiomer was seen as well,
establishing another R2 (8) synthon by N–H ◊ ◊ ◊ O hydrogen bonds
˚
in the monoclinic C2/c space group: a = 21.528(10) A; b =
2
◦
◦
◦
˚
˚
˚
˚
8.433(4) A; c = 17.211(8) A; a = 90 ; b = 94.650(16) ; g = 90 ;
(N3¢ ◊ ◊ ◊ O12¢ = 2.862 A; H3¢ ◊ ◊ ◊ O12¢ = 1.940 A; N3¢–H3¢ ◊ ◊ ◊ O12¢ =
V = 3114(3) A ; Z = 4. After refinement the residual factor, R1,
164.6◦) (cf. Electronic Supplementary Material†).
3
˚
had a value of 0.0591 (cf. Table S1, Electronic Supplementary
Material†). The molecule is composed of two ring systems, a
6-membered ring and a bicyclic unit consisting of two fused 6-
membered rings, with the two systems connected at C2 (Fig. 6).
The figure shows the molecular structure of the R- enantiomer;
it should be noted that, since C2/c is a centrosymmetric space
group, the S- enantiomer is generated by symmetry operations.
The average planes of the two rings display an angle of 78.63(7)◦,
a parameter that, together with the values of the angles between the
4 bonds originating at C2 [ranging from 106.5(2)◦ to 111.3(2)◦)],
confirms that the geometry around this atom is nearly tetrahedral.
It is noteworthy that both the single 6-membered ring and the
inner ring of the fused system show significant atomic deviations
We were unable to obtain crystals of the spiro compound
19 suitable for X-ray diffraction analysis. Therefore, the definite
structural assignment was based essentially on the 1H-13C HMBC
correlations (Fig. 8). The proton signal at 6.23–6.25 ppm (H5)
showed three-bond correlations with the methyl carbon (18.5
ppm) and the spiro quaternary carbon (C3; 74.7 ppm). Three-
bond correlations were also observed between the methyl protons
(2.14 ppm) and both the spiro carbon (C3; 74.7 ppm) and C5
(123.9 ppm). Based upon these correlations, and ignoring poten-
tial tautomerism in the pyrimidinone-type ring, three different
tautomeric structures consistent with the data can be proposed
for 19 (Scheme 4). Evidence for the co-existence of more than
one isomer in equilibrium can be inferred from the presence of
1
from planarity (-0.09225 to 0.1180 A and -0.1522 to 0.2721 A,
respectively) and large bond length differences, due to the presence
of sp3 nitrogen atoms; on the contrary, the outer ring is planar
multiple resonances for a few specific signals in the H and 13C
˚
˚
NMR spectra [e.g., H5 (an apparent multiplet at 6.23–6.25 ppm),
C4¢ (164.8 and 164.9 ppm), and C2/C6 (169.5 and 169.6 ppm).
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7822–7835 | 7829
©

View Article Online
Fig. 5 Expanded region of the 1H-13C HMBC spectrum of the spiro compound 18, displaying the 3-bond connectivities between the N1H proton and
carbons C3 and C5, and the 2- and 3-bond connectivities between the N3¢H proton and the spiro (C2) and C4¢a carbons, respectively.
Fig. 7 Diagram displaying the interactions between two molecules of the
spiro compound 18 (one of isomer R, left, and another of isomer S, right)
and one molecule of solvent (dimethylketone).
of 13C NMR signals in 24, when compared to 2-OH-NVP, was
the first evidence for the loss of ring A. Additionally, the presence
Fig. 6 ORTEP diagram, drawn with 50% probability ellipsoids, showing
the atomic labelling scheme for the spiro compound 18.
of a one-proton singlet at 8.64 ppm, correlating with a carbon
1
signal at 156.3 ppm in the H-13C HSQC spectrum, suggested
The two C4¢ resonances were identified on the basis of three-
bond HMBC correlations with H5¢ (8.07 ppm) but the C2 and C6
resonances could not be assigned unambiguously, due to the lack
of HMBC correlations involving these quaternary carbons.
By comparison with the spiro compounds, the characterization
of the pyridopyrimidine derivative 24 and of the nicotinic acid
23 was straightforward. A considerable decrease in the number
the presence of an imine group in the molecule. Conclusive
evidence for the structural assignment was obtained from the
1H-13C HMBC 3-bond correlations observed between H2 and
the carbonyl C4 (170.2 ppm), the quaternary C8a (152.2 ppm),
and the cyclopropyl C-9 (31.8 ppm) (cf. Experimental section
and Electronic Supplementary Material†). Moreover, the ESI
mass spectral data were entirely consistent with structure 24;
7830 | Org. Biomol. Chem., 2011, 9, 7822–7835
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Fig. 8 Expanded region of the 1H-13C HMBC spectrum of the spiro compound 19, displaying the 3-bond connectivities between the spiro carbon C3
and both the methyl and olefinic (H5) protons.
both the protonated (m/z 188) and sodiated (m/z 210) molecules
were observed and MS/MS of the protonated molecule resulted
in characteristic fragmentations, involving sequential loss of the
cyclopropyl moiety, HCN and CO (cf. Scheme S2, Electronic
Supplementary Material†).
particular 2-OH-NVP, may play a role in NVP-induced toxicity.
The toxicological significance of quinone-imine intermediates
has long been recognized.61 Representative examples of ther-
apeutic drugs activated via quinone-imines are the analgesic
acetaminophen (paracetamol),62–64 the nonsteroidal antiinflam-
matory agent diclofenac,65 and the cyclooxygenase-2 inhibitor
lumiracoxib.66,67 Indeed, an intermediate quinone-imine has been
postulated to be at the onset of fatal liver damage prompted by
overmedication with paracetamol, presumably due to covalent
protein binding when the sacrificial detoxifying agent, glutathione,
has been depleted.
Our data indicate that the intermediate quinone-imine 20
underwent prompt nucleophilic attack by water, which supports
the likelihood of competitive reaction with bionucleophiles in
vivo, affording covalent adducts. Quinone-imines are known to be
especially prone to nucleophilic attack by sulfhydryl groups, and
thus glutathione conjugation represents the typical detoxification
process for these electrophiles.68 However, given that HIV-infected
patients are reported to have depleted levels of glutathione,69
this will hamper an efficient detoxification of quinone-imine
metabolites, which will conceivably become available to react with
sulfhydryl groups (or other nucleophilic side chains) in proteins,
eliciting toxic responses.
The most interesting end products of oxidative degradation
of 2-OH-NVP were the spiro compounds 18 and 19. Naturally
occurring bioactive spiro compounds are well known,70 and the
generation of spiro derivatives as oxidation products from other
drugs and toxicants has also been reported.71,72 Considering that
19 only formed in non-aqueous media and proved to be stable
under hydrolytic and oxidative conditions, it is not anticipated
to have biological significance. On the contrary, 18 was formed
in aqueous media, including under lactoperoxidase catalysis, and
was very susceptible to hydrolysis, which suggests that it may be
The nicotinic acid 23 has previously been reported as a
zwitterion60 but the structural characterization data were not
provided. We isolated this compound in very small amounts,
which precluded obtaining a 13C NMR spectrum with good
signal/noise ratio. Nonetheless, carbon resonance assignments
were performed on the basis of the correlations observed in
the more sensitive inverse bidimensional experiments, HSQC
and HMBC. The identification of 23 relied on the recognition
of 1H NMR signal patterns from a 2,3-disubstituted pyridine
and a cyclopropylamino group (cf. Experimental section), which
confirmed that the original ring C of 2-OH-NVP was essentially
intact. A relevant three-bond correlation, observed in the 1H-13C
HMBC spectrum between H4 (8.29 ppm) and a quaternary carbon
at 162.4 ppm, indicated that a carbonyl group was still bound to
the pyridine ring (not shown). Labile protons were not detected,
presumably due to fast exchange on the NMR timescale, but the
ESI mass spectrum yielded a signal at m/z 379, entirely consistent
with a sodiated dimer of 23. Given the presence of both an amino
and a carboxylic acid group in the molecule, this propensity for
dimerization was not unexpected.
Conclusions
The product profile obtained upon chemical and enzymatic
oxidation of 2-OH-NVP is consistent with the formation of a
transient quinone-imine intermediate (20). The presence of this
electrophilic species in the reaction media supports the hypothesis
that an activation pathway from phenolic NVP metabolites, in
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7822–7835 | 7831
©

View Article Online
prone to reacting with bionucleophiles. Although the significance
of this pathway to toxicity remains to be established, the formation
of electrophilic 2-OH-NVP derivatives under lactoperoxidase
catalysis suggests that it could play a role in the perinatal setting,
due to NVP administration concurrently with breastfeeding and
the fact that nevirapine readily enters breast milk.13
1H NMR spectra were recorded on Bruker Avance III 400 or 500
spectrometers (Bruker BioSpin GmbH, Rheinstetten, Germany)
operating at 400 and 500 MHz, respectively. 13C NMR spectra
were recorded on the same instruments, operating at 100.62
and 125.77 MHz, respectively. Chemical shifts are reported in
ppm downfield from tetramethylsilane, and coupling constants
(J) are reported in Hz; the subscripts ortho and meta, refer to
ortho and meta couplings, respectively, and asterisks are used
to indicate a second isomer. The presence of labile protons
was confirmed by chemical exchange with D2O. Resonance and
structural assignments were based on the analysis of coupling
patterns, including the 13C-1H coupling profiles obtained in bidi-
mensional heteronuclear multiple bond correlation (HMBC) and
heteronuclear single quantum coherence (HSQC) experiments,
performed with standard pulse programs.
Experimental
General remarks
Chemicals. NVP was purchased from Cipla (Mumbai, India).
All other commercially available reagents and enzymes were
acquired from Sigma-Aldrich Qu´ımica, S.A. (Madrid, Spain),
unless specified otherwise, and were used as received. Whenever
necessary, solvents were purified by standard methods.73
X-ray crystallographic data for compound 18 were collected
from crystals using an area detector diffractometer (Bruker AXS-
KAPPA APEX II) equipped with an Oxford Cryosystem open-
flow nitrogen cryostat at 150 K and graphite-monochromated Mo-
Instrumentation. Melting temperatures were measured in a
Leica Galen III hot stage apparatus and are uncorrected. Infrared
(IR) spectra were recorded on a Perkin-Elmer 683IR spectrometer
(Waltham, MA); group frequencies are reported in cm-1. The
UV measurements were recorded on a Perkin Elmer Lambda
35 UV/VIS spectrophotometer. HPLC was conducted on an
Ultimate 3000 Dionex system consisting of an LPG-3400A
quaternary gradient pump and a diode array spectrophotometric
detector (Dionex Co., Sunnyvale, CA) and equipped with a
Rheodyne model 8125 injector (Rheodyne, Rohnert Park, CA).
HPLC analyses were performed with a Luna C18 (2) column
(250 mm ¥ 4.6 mm; 5 mm; Phenomenex, Torrance, CA), at a
flow rate of 1 mL min-1. Semipreparative HPLC separations were
conducted with a Luna C18 (2) column (250 mm ¥ 10 mm; 5 mm;
Phenomenex) at a flow rate of 3 mL min-1. A 30-min linear gradient
from 5 to 70% acetonitrile in 0.1% aqueous formic acid, followed
by a 2-min linear gradient to 100% acetonitrile and an 8-min
isocratic elution with acetonitrile, was used in all instances. The
UV absorbance was monitored at 254 nm.
LC-ESI-MS/MS analyses were performed with a Varian system
consisting of a ProStar 410 autosampler, two 210-LC chromato-
graphy pumps, a ProStar 335 diode array detector, and a 500-MS
ion trap mass spectrometer, with an ESI ion source (Varian, Inc.,
Palo Alto, CA). Data acquisition and processing were performed
using Varian MS Control 6.9 software. The samples were injected
onto the column via a Rheodyne injector with a 20 mL loop.
Separations were conducted at 30 ◦C, using a Luna C18 (2) column
(150 mm ¥ 2 mm, 3 mm; Phenomenex). The mobile phase was
delivered at a flow rate of 200 mL min-1, using a 5-min isocratic
elution with 5% acetonitrile in 0.1% aqueous formic acid, followed
by a 30-min linear gradient from 5 to 70% acetonitrile, a 2-
min linear gradient to 100% acetonitrile, and an 8-min isocratic
elution with acetonitrile. The mass spectrometer was operated
in the positive ESI mode; the optimized operating parameters
were: ion spray voltage, +5.2 kV; capillary voltage, 20 V; and RF
loading, 90%. Nitrogen was used as the nebulizing and drying
gas, at pressures of 50 and 30 psi, respectively; the drying gas
temperature was 350 ◦C. MS/MS spectra were obtained with an
isolation window of 1.5 Da, excitation energy values between 0.9
and 1.2 V, and an excitation time of 10 ms. High resolution ESI
mass spectra were obtained on a Bruker Apex Ultra FTICR mass
spectrometer (Bruker Daltonics, Billerica, MA) at the FCUL node
of the Portuguese MS network.
˚
Ka (l = 0.71073 A) radiation. Cell parameters were retrieved using
Bruker SMART software and refined with Bruker SAINT74 on all
observed reflections. Absorption corrections were applied using
SADABS.75 The structures were solved by direct methods using
SIR 9776 and refined with full-matrix least-squares refinement
against F2 using SHELXL-97.77 All the programs are included in
the WINGX package (version 1.70.01).78 All non-hydrogen atoms
were refined anisotropically, and the hydrogen atoms were inserted
in idealized positions, riding on the parent C atom, except for the
methyl hydrogens, whose orientation was refined from electron
density, allowing the refinement of both C–C torsion angles and
C–H distances, and the hydrogen atoms bonded to nitrogens,
which were found directly in the density map. Drawings were made
with ORTEP3 for Windows.79 Intermolecular interactions were
analysed using Mercury 1.4.2 (Build 2).80 Plane calculations were
made using Parst.81,82 The crystals had reasonable quality and
diffracting power, presenting low R2 values (0.123) that allowed
low R values (R1 all = 0.139 and R1 obs = 0.059). The structure was,
therefore, unequivocally determined and is in good agreement
with the remaining spectral characterization data for 18. Relevant
details of the X-ray data analysis are displayed in the Electronic
Supplementary Material† section.
Synthesis of 11-cyclopropyl-2-hydroxy-4-methyl-5,11-dihydro-6H-
dipyrido[3,2-b:2¢,3¢-e][1,4]diazepin-6-one (2-OH-NVP, 2)
11-Cyclopropyl-4-methyl-6-oxo-6,11-dihydro-5H-dipyrido[3,2-
b:2¢,3¢-e][1,4]diazepin-2-yl acetate (2-acetoxy-NVP, 15). To a so-
lution of NVP (250 mg, 0.94 mmol) in dichloromethane (100 mL)
at 104 ◦C (temperature of the oil bath), were added silver acetate
(1.25 g, 7.5 mmol) and then iodine (1.15 g, 4.5 mmol), and the
resulting solution was stirred at 104 ◦C for 5 min. The yellow solid
formed upon cooling to room temperature was applied onto a sil-
ica bed. Iodine was removed with dichloromethane and the prod-
uct was eluted with ethyl acetate/dichloromethane (1 : 1; 150 mL)
and recovered in quantitative yield upon evaporation. Mp = 242–
245 ◦C (AcOEt); IR (KBr): 1774 (C O, ester), 1654 (C O,
amide); 1H NMR (CDCl3): d 8.73 (1H, s, NH), 8.52 (1H, dd, Jortho
=
4.8 and Jmeta = 1.8, H9), 8.10 (1H, dd, Jortho = 7.8 and Jmeta = 1.8,
H7), 7.08 (1H, dd, Jortho = 7.8 and J¢ortho = 4.8, H8), 6.73 (1H, s, H3),
7832 | Org. Biomol. Chem., 2011, 9, 7822–7835
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
3.69–3.67 (1H, m, H13), 2.41 (3H, s, C12H3), 2.34 (3H, s, COCH3),
0.98–0.96 (2H, m, H14+H15), 0.50–0.49 (2H, m, H14+H15); 13
(3 mg); Mp = 172–174 ◦C; Retention time (Rt) = 13.2 min; UV
(acetonitrile): lmax 210, 219, 325 nm; H NMR (DMSO-d6): d
1
C
NMR (CDCl3): d 169.1 (OC O), 168.6 (C6), 160.3 (C10a), 152.8
(C2/C11a), 152.6 (C2/C11a), 152.0 (C9), 142.7 (C4), 140.3 (C7),
123.0 (C4a), 120.1 (C6a), 119.0 (C8), 113.5 (C3), 29.5 (C13), 21.2
(COCH3), 18.1 (C12), 8.9 (C14 + C15); MS (ESI): m/z 347 [M
+ Na]+, 325 [MH]+, 283 [MH2 - COCH3]+. HRMS calcd for
[C17H16N4O3 + H]+: 325.129517. Found: 325.12941.
9.41 (1H, s, N1H), 9.04 (1H, s, N3¢H), 8.36 (1H, d, Jortho = 3.8,
H7¢), 8.00 (1H, d, Jortho = 7.1, H5¢), 6.97–6.94 (1H, m, H6¢), 6.85
(1H, s, H5), 2.12–2.09 (1H, m, H9), 1.93 (3H, s, CH3), 0.95–0.87
(1H, m, H10/H11), 0.87–0.81 (2H, m, H10/H11), 0.36–0,35
(1H, m, H10/H11); 13C NMR (DMSO-d6): d 191.2 (C3), 161.6
(C6), 161.5 (C4¢), 156.0 (C8¢a), 151.9 (C7¢), 142.1 (C4), 136.5
(C5), 135.0 (C5¢), 115.7 (C6¢), 111.3 (C4¢a), 82.6 (C2), 25.8 (C9),
14.3 (CH3), 9.1 (C10/C11), 8.4 (C10/C11). MS (ESI): m/z 299
[MH]+. HRMS calcd for [C15H14N4O3 + H]+: 299.113867. Found:
299.11375. For the X-ray diffraction data see the Results and
Discussion and Electronic Supplementary Material† section.
2-OH-NVP (2). The entire amount of 15 (ca. 0.94 mmol),
obtained as described above, was dissolved in methanol (10 mL),
10% KOH was added to the solution, and the mixture was stirred at
room temperature for 15 min. The methanol was then evaporated
and the pH was adjusted to 7, allowing the precipitation of a
white solid, which was isolated by centrifugation and washed with
ethanol to afford 2 (225 mg, 0.80 mmol, 85% overall yield). Mp =
330–332 ◦C (ethanol) (lit.36 >270 ◦C); IR (KBr): 3195 (O–H and
N–H), 1654 (C O); UV (acetonitrile): lmax 234, 314 nm; 1H NMR
(DMSO-d6): d 10.56 (1H, bs, OH), 9.60 (1H, s, NH), 8.47–8.46
(1H, m, H9), 7.97–7.95 (1H, m, H7), 7.16–7.13 (1H, m, H8), 6.29
(1H, s, H3), 3.55 (1H, bs, H13), 2.22 (3H, s, CH3), 0.85 (2H, bs,
H14 + H15), 0.35 (2H, bs, H14 + H15); 13C NMR (DMSO-d6):
d 167.4 (C6), 160.8 (C10a), 159.8 (C2), 152.4 (C11a), 151.3 (C9),
145.0 (C4), 140.3 (C7), 121.6 (C6a), 119.6 (C8), 117.6 (C4a), 106.9
(C3), 29.2 (C13), 18.2 (CH3), 9.0 (C14/C15), 8.9 (C14/C15); MS
(EI): m/z 282 [M]+ (76.7%), 281 [M - H]+ (100.0%), 267 [M -
CH3]+ (60.4%), 253 [M - HCO]+ (30.3%).
At pH 10:
Compound 18, obtained in 8.3% yield (3.5 mg); and
1-Cyclopropylpyrido[2,3-d]pyrimidin-4(1H)-one (24). Obtain-
ed in 10.5% yield (2.8 mg); Mp = 210 ◦C; Rt, 9.3 min; UV
(acetonitrile): lmax 221, 309, 319 nm; 1H NMR (DMSO-d6): d 8.91
(1H, dd, Jortho = 4.6 and Jmeta = 1.9, H7), 8.64 (1H, s, H2), 8.43 (1H,
dd, Jortho = 7.8 and Jmeta = 1.9, H5), 7.61 (1H, dd, Jortho = 7.8 and
J¢ortho = 4.6, H6), 3.54–3.48 (1H, m, H9), 1.23–1.04 (4H, m, H10 +
H11); 13C NMR (DMSO-d6): d 170.2 (C4), 156.3 (C2), 154.9 (C7),
152.2 (C8a), 137.8 (C5), 123.6 (C6), 115.8 (C4a), 31.8 (C9), 7.4
(C10 + C11); MS (ESI): m/z 210 [M + Na]+, 188 [MH]+. HRMS
calcd for [C10H9N3O + H]+: 188.081838. Found: 188.08150.
Oxidations of 2-OH-NVP
WithFre´my’s salt inDMF. Toasolutionof2-OH-NVP(60mg,
213 mmol) in DMF (3 mL) was added a solution of Fre´my’s salt
(74 mg, 276 mmol) in DMF (3 mL). Following incubation at room
temperature for 24 h, the mixture was purified by semi-preparative
HPLC, yielding:
General method for preliminary reactions with Fre´my’s salt
and sodium periodate. To a solution of 2-OH-NVP (10 mg,
35.5 mmol) in solvent A was added a solution of the oxidant
in solvent B (Table 1). The reaction mixture was monitored by
HPLC-DAD at 254 nm after 1 and 24 h.
1¢-Cyclopropyl-4-methyl-1¢H,2H-spiro[pyridine-3,2¢-pyrido[2,3-
d]pyrimidine]-2,4¢,6(1H,3¢H)-trione [19 (a/b/c)]. Obtained in
3% yield (2 mg); Mp > 300 ◦C; Rt, 13.5 min; UV (acetonitrile),
Reaction with lactoperoxidase. To a solution of 2-OH-NVP
(10 mg, 35.5 mmol) in 100 mM phosphate, pH 7.4 (200 mL) were
added a solution of hydrogen peroxide (0.5%, 50 mL) and a solution
of lactoperoxidase (EC 1.11.1.7; 0.4–0.7 units mL-1 in 100 mM
phosphate, pH 7.4; 14 mL). The resulting mixture was incubated
overnight at 37 ◦C and then analyzed by LC-DAD and LC-MS. A
control assay was performed using the same conditions in the
absence of lactoperoxidase.
lmax 222, 331 nm; 1H NMR (methanol-d4): d 8.33 (1H, dd, Jortho
=
5.0 and Jmeta = 2.0, H7¢), 8.07 (1H, dd, Jortho = 7.5 and Jmeta = 2.0,
H5¢), 6.91 (1H, dd, Jortho = 7.5 and J¢ortho = 5.0, H6¢), 6.25–6.23
(1H, m, H5), 2.36–2.30 (1H, m, H9¢), 2.14 (3H, s, CH3), 0.95–0.81
(2H, m, H10¢/H11¢), 0.75–0.73 (2H, m, H10¢/H11¢); 13C NMR
(methanol-d4): d 173.3 (C2/C6), 169.6 (C2/C6), 169.5 (C2*/C6*),
164.9 (C4¢), 164.8 (C4¢*), 159.8 (C8¢a), 155.7 (C4), 153.5 (C7¢),
137.0 (C5¢), 123.9 (C5), 116.6 (C6¢), 112.0 (C4¢a), 74.7 (C3), 28.8
(C9¢), 18.5 (CH3), 9.0 (C10¢/C11¢), 8.9 (C10¢/C11¢). MS (ESI)
m/z 321 [M + Na]+, 299 [MH]+. HRMS calcd for [C15H14N4O3 +
H]+: 299.113867. Found: 299.11388.
Oxidation reactions on a preparative scale
With Fre´my’s salt in phosphate buffer. To a solution of 2-OH-
NVP (40 mg, 142 mmol) in ethyl acetate (8 mL) was added a
solution of Fre´my’s salt (48 mg, 179 mmol) in 100 mM phosphate
buffer (4 mL; pH 5, 7.4, or 10) and the mixture was stirred
overnight at room temperature. Following phase separation and
additional extraction with ethyl acetate, the organic layers were
dried over anhydrous sodium sulfate and the products were
isolated by PTLC on silica (1/1 dichloromethane/ethyl acetate).
The novel products characterized from these reactions were:
2-(Cyclopropylamino)pyridine-3-carboxylic acid (23). Obtain-
ed in 4% yield (1.5 mg); Rt = 12.0 min. UV (acetonitrile): lmax 222,
1
245 nm; H RMN (DMSO-d6): d 8.73 (1H, dd, Jortho = 4.7 and
Jmeta = 1.6, H6), 8.29 (1H, Jortho = 7.7 and Jmeta = 1.6, H4), 7.31 (1H,
Jortho = 7.7 and J¢ortho = 4.7, H5), 2.78–2.76 (1H, m, H8), 1.12–1.08
(2H, m, H9 + H10) 0.81–0.75 (2H, m, H9 + H10); MS (ESI): m/z:
379 [2 M + Na]+.
At pH 7.4
With sodium periodate. To a solution of 2-OH-NVP (60 mg,
213 mmol) in DMF (3 mL) was added a solution of NaIO4
(45 mg, 210 mmol) in DMF (3 mL). Following incubation at room
1¢-Cyclopropyl-4-methyl-1H,1¢H-spiro[pyridine-2,2¢-pyrido[2,3-
d]pyrimidine]-3,4¢,6(3¢H)-trione (18). Obtained in 7% yield
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7822–7835 | 7833
©

View Article Online
temperature for 24 h, the mixture was purified by semi-preparative
HPLC affording the spiro compound 19 in 1.5% yield (1 mg).
10 M. Lallemant, G. Jourdain, S. Le Coeur, J. Y. Mary, N. Ngo-Giang-
Huong, S. Koetsawang, S. Kanshana, K. McInstosh and V. Thaineua,
for the Perinatal HIV Prevention Trial (Thailand) Investigators,
N. Engl. J. Med., 2004, 351, 217.
11 S. Lockman, R. L. Shapiro, L. M. Smeaton, C. Wester, I. Thior, L.
Stevens, F. Chand, J. Makhema, C. Moffat, A. Asmelash, P. Ndase, P.
Arimi, E. van Widenfelt, L. Mazhani, V. Novitsky, S. Lagakos and M.
Essex, N. Engl. J. Med., 2007, 356, 135.
12 FDA, Approval of Viramune XR (nevirapine) 400 mg extended release
tablet, 2011 (http://www.fda.gov/ForConsumers/ByAudience/For-
PatientAdvocates/HIVandAIDSActivities/ucm248800.htm; accessed
May 16, 2011).
Assessment of the stability of the spiro compounds (18 and 19).
To a solution of the spiro compound (18 or 19; 1.5 mg; 5 mmol) in
DMF (100 ml) were added 100 ml of DMF or 100 mM phosphate
buffer (pH 5, 7.4, or 10). A similar assay was performed in
DMF in the presence of Fre´my’s salt. The resulting mixtures were
monitored by HPLC after 5 min, 1 h, and 24h.
Abbreviations. CART, combined antiretroviral therapy; CYP,
cytochrome P450; DMF, N,N-dimethylformamide; ESI, elec-
trospray ionization; Fre´my’s salt, potassium nitrosodisulfonate;
HIV, human immunodeficiency virus type 1; HPLC-DAD, high
performance liquid chromatography with diode array detection;
MS/MS, tandem mass spectrometry; NNRTI, non-nucleoside re-
verse transcriptase inhibitor; NVP, nevirapine; PTLC, preparative
thin layer chromatography.
13 M. Mirochnick, D. F. Clarke and A. Dorenbaum, Clin. Pharmacokinet.,
2000, 39, 281.
14 Six Week Extended-Dose Nevirapine (SWEN) Study Team, Lancet,
2008, 372, 300.
15 M. S. Baylor and R. Johann-Liang, J. Acquir. Immune Defic. Syndr.,
2004, 35, 538.
16 M. Popovic, J. M. Shenton, J. Chen, A. Baban, T. Tharmanathan, B.
Mannargudi, D. Abdulla and J. P. Uetrecht, Handb. Exp. Pharmacol.,
2010, 196, Part 3, 437.
R
17 VIRAMUNEꢀ (nevirapine), Physicians’ Desk Reference, 2009, 63rd
ed, pp 873–881, Physicians’ Desk Reference Inc., Montvale, NJ.
18 T. Powles, D. Robinson, J. Stebbing, J. Shamash, M. Nelson, B.
Gazzard, S. Mandelia, H. Møller and M. Bower, J. Clin. Oncol., 2009,
27, 884.
Acknowledgements
We thank the Portuguese NMR Network (IST-UTL Center)
and the Portuguese MS Network (IST-UTL Center) for pro-
viding access to the facilities. This work was supported in
part by Fundac¸a˜o para a Cieˆncia e a Tecnologia (FCT),
Portugal, through pluriannual funds to Centro de Qu´ımica
Estrutural (PEst-OE/QUI/UI0100/2011) and research grants
PPCDT/QUI/56582/2004 and PTDC/QUI-QUI/113910/2009,
and by Interagency Agreement Y1ES1027 between the National
Center for Toxicological Research/Food and Drug Adminis-
tration and the National Institute of Environmental Health
Sciences/National Toxicology Program. The opinions expressed
in this paper do not necessarily represent those of the U.S. Food
and Drug Administration.
19 P. Claes, M. Wintzen, S. Allard, P. Simons, A. De Coninck and P. Lacor,
Eur. J. Intern. Med., 2004, 15, 255.
20 P. Riska, M. Lamson, T. MacGregor, J. Sabo, S. Hattox, J. Pav and J.
Keirns, Drug. Metab. Dispos., 1999, 27, 895.
21 P. S. Riska, D. P. Joseph, R. M. Dinallo, W. C. Davidson, J. J. Keirns
and S. E. Hattox, Drug. Metab. Dispos., 1999, 27, 1434.
22 D. A. Erickson, G. Mather, W. F. Trager, R. H. Levy and J. J. Keirns,
Drug. Metab. Dispos., 1999, 27, 1488.
23 Z. Liu, P. Fan-Havard, Z. Xie, C. Ren and K. K. Chan, Rapid Commun.
Mass Spectrom., 2007, 21, 2734.
24 C. Ren, P. Fan-Havard, N. Schlabritz-Loutsevitch, Y. Ling, K. K. Chan
and Z. Liu, Biomed. Chromatogr., 2010, 24, 717.
25 J. Chen, B. M. Mannargudi, L. Xu and J. Uetrecht, Chem. Res. Toxicol.,
2008, 21, 1862.
26 J. M. Shenton, M. Teranishi, M. S. Abu-Asab, J. A. Yager and J. P.
Uetrecht, Chem. Res. Toxicol., 2003, 16, 1078.
27 B. Wen, Y. Chen and W. L. Fitch, Drug Metab. Dispos., 2009, 37, 1557.
28 A. M. M. Antunes, M. P. Duarte, P. P. Santos, G. Gamboa da Costa,
T. M. Heinze, F. A. Beland and M. M. Marques, Chem. Res. Toxicol.,
2008, 21, 1443.
Notes and references
29 A. M. M. Antunes, A. L. A. Godinho, I. L. Martins, G. C. Justino,
F. A. Beland and M. M. Marques, Chem. Res. Toxicol., 2010, 23, 888.
30 A. M. M. Antunes, A. L. A. Godinho, I. L. Martins, M. C. Oliveira,
R. A. Gomes, A. V. Coelho, F. A. Beland and M. M. Marques, Chem.
Res. Toxicol., 2010, 23, 1714.
1 F. Vidal, F. Gutie´rrez, M. Gutie´rrez, M. Olona, V. Sa´nchez, G. Mateo, J.
Peraire, C. Vilade´s, S. Veloso, M. Lo´pez-Dupla and P. Domingo, AIDS
Rev., 2010, 12, 15.
2 A. Calmy, B. Hirschel, D. A. Cooper and A. Carr, Lancet, 2007, 370,
12.
31 A. Srivastava, L.-Y. Lian, J. L. Maggs, M. Chaponda, M. Pirmohamed,
D. P. Williams and B. K. Park, Drug Metab. Dispos., 2010, 38, 122.
32 J. L. Bolton, M. A. Trush, T. M. Penning, G. Dryhurst and T. J. Monks,
Chem. Res. Toxicol., 2000, 13, 135.
33 T. J. Monks and D. C. Jones, Curr. Drug Metab., 2002, 3, 425.
34 S. Bittner, Amino Acids, 2006, 30, 205.
35 K. G. Grozinger, D. P. Byrne, L. J. Nummy, M. D. Ridges and A.
Salvagno, J. Heterocyclic Chem., 2000, 37, 229.
36 R. C. Cambie, G. J. Potter, P. S. Rutledge and P. D. Woodgate, J. Chem.
Soc., Perkin Trans. 1, 1977, 530.
37 A. D’Alfonso, M. Pasi, A. Porta, G. Zanoni and G. Vidari, Org. Lett.,
2010, 12, 596.
38 E. Ciganek and J. C. Calabrese, J. Org. Chem., 1995, 60, 4439.
39 W. Zhang, M. N. Wicks and P. L. Burn, Org. Biomol. Chem., 2008, 6,
879.
3 C. J. Fichtenbaum, Current HIV/AIDS Reports, 2010, 7, 92.
4 World Health Organization, WHO guidelines on HIV/AIDS
(http://www.who.int/rpc/guidelines/hiv_aids/en/index.html;
accessed May 16, 2011).
5 Food and Drug Administration, FDA approves nevirapine to treat HIV.
1996, News release T96-44, June 24, 1996.
6 M. A. Thompson, J. A. Aberg, P. Cahn, J. S. G. Montaner, G.
Rizzardini, A. Telenti, J. M. Gatell, H. F. Gu¨nthard, S. M. Hammer,
M. S. Hirsch, D. M. Jacobsen, P. Reiss, D. D. Richman, P. A.
Volberding, P. Yeni and R. T. Schooley, JAMA, 2010, 304, 321.
7 Panel on Treatment of HIV-Infected Pregnant Women and Pre-
vention of Perinatal Transmission. Recommendations for Use of
Antiretroviral Drugs in Pregnant HIV-1-Infected Women for Ma-
ternal Health and Interventions to Reduce Perinatal HIV Trans-
mission in the United States. May 24, 2010; pp 1–117. Available
at http://aidsinfo.nih.gov/ContentFiles/PerinatalGL.pdf. Accessed
May 16, 2011.
8 E. Marseille, J. G. Kahn, F. Mmiro, L. Guay, P. Musoke, M. G. Fowler
and J. B. Jackson, Lancet, 1999, 354, 803.
9 J. B. Jackson, P. Musoke, T. Fleming, L. A. Guay, D. Bagenda, M.
Allen, C. Nakabiito, J. Sherman, P. Bakaki, M. Owor, C. Ducar, M.
Deseyve, A. Mwatha, L. Emel, C. Duefield, M. Mirochnick, M. G.
Fowler, L. Mofenson, P. Miotti, M. Gigliotti, D. Bray and F. Mmiro,
Lancet, 2003, 362, 859.
40 M. Iglesias, O. Schuster and M. Albrecht, Tetrahedron Lett., 2010, 51,
5423.
41 S. Aronson, P. Epstein, D. B. Aronson and G. Wieder, J. Phys. Chem.,
1982, 86, 1035.
42 J. R. Barnett, L. J. Andrews and R. M. Keefer, J. Am. Chem. Soc., 1972,
94, 6129.
43 T. R. Beebe, B. A. Barnes, K. A. Bender, A. D. Halbert, R. D. Miller,
M. L. Ramsay and M. W. Ridenour, J. Org. Chem., 1975, 40, 1992.
44 S. Rozen, D. Hebel and D. Zamir, J. Am. Chem. Soc., 1987, 109, 3789.
7834 | Org. Biomol. Chem., 2011, 9, 7822–7835
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
45 D. Hebel and S. Rozen, J. Org. Chem., 1991, 56, 6298.
46 J. G. Marrero, L. San Andre´s and J. G. Luis, Chem. Pharm. Bull., 2005,
53, 1524.
47 J. M. Saa´, J. Morey and C. Rubido, J. Org. Chem., 1986, 51, 4471.
48 M. Sugumaran, Arch. Biochem. Biophys., 2000, 378, 404.
49 H. Zimmer, D. C. Lankin and S. W. Horgan, Chem. Rev., 1971, 71, 229.
50 J. Zielonka, H. Zhao, Y. Xu and B. Kalyanaraman, Free Radical Biol.
Med., 2005, 39, 853.
66 Y. Li, J. G. Slatter, Z. Zhang, Y. Li, G. A. Doss, M. P. Braun, R. A.
Stearns, D. C. Dean, T. A. Baillie and W. Tang, Drug Metab. Dispos.,
2008, 36, 469.
67 P. Kang, D. Dalvie, E. Smith and M. Renner, Chem. Res. Toxicol.,
2009, 22, 106.
68 D. J. Waldon, Y. Teffera, A. E. Colletti, J. Liu, D. Zurcher, K. W.
Copeland and Z. Zhao, Chem. Res. Toxicol., 2010, 23, 1947.
69 D. M. Townsend, K. D. Tew and H. Tapiero, Biomed. Pharmacother.,
2003, 57, 145.
70 R. Pradhan, M. Patra, A. K. Behera, B. K. Mishra and R. K. Behera,
Tetrahedron, 2006, 62, 779.
51 E. Adler and R. Magnusson, Acta Chem. Scand., 1959, 13, 505.
52 A. Johnston, A. J. Florence and A. R. Kennedy, Acta Crystallogr., Sect.
E: Struct. Rep. Online, 2005, 61, o1509.
53 S. W. Weidman and E. T. Kaiser, J. Am. Chem. Soc., 1966, 88, 5820.
54 H. Kohler and H. Jenzer, Free Radical Biol. Med., 1989, 6, 323.
55 H. J. Sipe Jr., S. J. Jordan, P. M. Hanna and R. P. Mason, Carcinogenesis,
1994, 15, 2637.
56 E. M. Ghibaudi, E. Laurenti, P. Beltramo and R. P. Ferrari, Redox
Rep., 2000, 5, 229.
57 K. M. Gorlewska-Roberts, C. H. Teitel, J. O. Lay Jr., D. W. Roberts
and F. F. Kadlubar, Chem. Res. Toxicol., 2004, 17, 1659.
58 A. S. Kyei, K. Tchabanenko, J. E. Baldwin and R. M. Adlington,
Tetrahedron Lett., 2004, 45, 8931.
59 F. H. Allen, O. Kennard, D. G. Watson, L. Brammer, A. G. Orpen and
R. Taylor, J. Chem. Soc., Perkin Trans. 2, 1987, S1.
60 R. F. Boswell, B. F. Gupton and Y. S. Lo, United States Patent 6,680,383,
2004.
71 E. J. Land, A. Perona, C. A. Ramsden and P. A. Riley, Org. Biomol.
Chem., 2005, 3, 2387.
72 M. Dracˇ´ınsky´, J. Cvacˇka, M. Semanska´, V. Mart´ınek, E. Frei and M.
Stiborova´, Chem. Res. Toxicol., 2009, 22, 1765.
73 D. D. Perrin and W. L. F. Armarego, 1988, Purification of Laboratory
Chemicals, 3rd ed., pp 1–391, Pergamon Press, Oxford, U. K.
74 SMART and SAINT: Area Detector Control and Integration Software,
Bruker AXS, Madison, WI, USA, 2004.
75 G. M. Sheldrick, SADABS, Program for Empirical Absorption Cor-
rection of Area Detectors (Version 2.10), University of Go¨ttingen,
Germany, 2004.
76 A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giacov-
azzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori and R. Spagna,
J. Appl. Crystallogr., 1999, 32, 115.
61 D. Porubek, M. Rundgren, R. Larsson, E. Albano, D. Ross, S. D.
Nelson and P. Molde´us, Adv. Exp. Med. Biol., 1986, 197, 631.
62 I. A. Blair, A. R. Boobis and D. S. Davies, Tetrahedron Lett., 1980, 21,
4947.
63 W. Chen, L. L. Koenigs, S. J. Thompson, R. M. Peter, A. E. Rettie,
W. F. Trager and S. D. Nelson, Chem. Res. Toxicol., 1998, 11, 295.
64 L. P. James, P. R. Mayeux and J. A. Hinson, Drug Metab. Dispos., 2003,
31, 1499.
77 G. M. Sheldrick, SHELXL-97, a Computer Program for the Re-
finement of Crystal Structures, University of Go¨ttingen, Go¨ttingen,
Germany, 1997.
78 L. J. Farrugia, J. Appl. Crystallogr., 1999, 32, 837.
79 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
80 I. J. Bruno, J. C. Cole, P. R. Edgington, M. Kessler, C. F. Macrae, P.
McCabe, J. Pearson and R. Taylor, Acta Crystallogr., Sect. B: Struct.
Sci., 2002, 58, 389.
65 J. S. van Leeuwen, G. Vredenburg, S. Dragovic, T. F. J. Tjong, J. C. Vos
and N. P. E. Vermeulen, Toxicol. Lett., 2011, 200, 162.
81 M. Nardelli, Comput. Chem., 1983, 7, 95.
82 M. Nardelli, J. Appl. Crystallogr., 1995, 28, 659.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7822–7835 | 7835
©