Visible-Light Catalytic Photooxygenation of Monoterpene Indole Alkaloids: Access to Spirooxindole-1,3-oxazines

Article abstract of DOI:10.1002/chem.201801882

Few natural oxindole alkaloids possess an exceptional spiro-[(1,3)oxazinan-3,6′-oxindole] core structure, which results from an unusual oxidative indole rearrangement. The Rauvolfia alkaloid reserpine can be converted into the spirooxindole-1,3-oxazines dioxyreserpine and trioxyreserpine through efficient visible-light catalytic photooxygenation with anthraquinone photocatalysts. A mechanistic investigation sheds new light on the photooxidative rearrangement of reserpine and related monoterpene indole alkaloids, and the spirooxindole-1,3-oxazine products can be valorized by reductive ring opening, to obtain cis-decahydroisoquinolines as new enantiopure synthetic building blocks, as demonstrated for dioxyreserpine.

Full text of DOI:10.1002/chem.201801882

A Journal of
Accepted Article
Title: Visible Light Catalytic Photooxygenation of Monoterpene Indole
Alkaloids: Access to Spirooxindole-1,3-Oxazines
Authors: Thorsten von Drathen, Frank Hoffmann, and Malte Brasholz
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201801882
Link to VoR: http://dx.doi.org/10.1002/chem.201801882
Supported by

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
Visible Light Catalytic Photooxygenation of Monoterpene Indole
Alkaloids: Access to Spirooxindole-1,3-Oxazines
Thorsten von Drathen,[a] Frank Hoffmann[b] and Malte Brasholz*[c]
Abstract: Few natural oxindole alkaloids possess an exceptional
spiro-[(1,3)oxazinan-3,6'-oxindole] core structure, which results from
an unusual oxidative indole rearrangement. The Rauvolfia alkaloid
reserpine can be converted into the spirooxindole-1,3-oxazines
dioxyreserpine and trioxyreserpine through efficient visible-light
catalytic photooxygenation with anthraquinone photocatalysts. A
mechanistic investigation sheds new light on the photooxidative
rearrangement of reserpine and related monoterpene indole alkaloids,
and the spirooxindole-1,3-oxazine products can be valorized by
reductive ring-opening, to obtain cis-decahydroisoquinolines as new
enantiopure synthetic building blocks as demonstrated for dioxy-
reserpine.
As the first natural product containing the spirooxindole-1,3-
oxazine substructure III, melodinoxanine (2) was isolated from the
stem and leaves of Melodinus henryi from Yunnan province in
China,4 and it was speculated that the biosynthetic precursor to 2
is isoreserpiline (1) which is also present in M. henryi (Scheme 2).
Second, mitradiversifoline (4) was isolated from Mitragyna
diversifolia and its putative progenitor is isopaynantheine (3),
which is a major component in this plant.5 A third known but
synthetic monoterpene spirooxindole-1,3-oxazine is dioxy-
reserpine (6) which has been isolated after the daylight-triggered
autoxidation of the Rauvolfia alkaloid reserpine (5). Awang and
co-workers reported in 1990 that when a chloroform solution of 5
was left standing for a prolonged period of time (15 days) under
air and in ambient light, product 6 could be isolated in 28% yield
at about 50% conversion of 5, and the structural assignment of 6
was based on NMR spectroscopy.6
Introduction
Spirocyclic 2-oxindoles constitute an important class of
oxygenated indole alkaloids, and many of these natural products
possess valuable biological properties such as antihypertensive,
analgesic, antitumor, and antiviral activity.1 The typical core
structure encountered in monoterpene spiro-2-oxindoles is the
spiro[pyrrolidine-3,3’-oxindole] motif I which can biosynthetically2
and chemically3 be accessed from a parent tetrahydro-β-
carboline II by oxidation followed by 1,2-rearrangement. However,
a few natural spiro-2-oxindoles possess a peculiar spiro-
[(1,3)oxazinan-3,6'-oxindole] ring system III whose biosynthesis is
yet to be elucidated, but putatively proceeds via a divergent and
unusual oxidative indole rearrangement, with concomitant
incorporation of two oxygen atoms into the product structure
(Scheme 1).
Scheme 1. Divergent oxidative rearrangements of indole alkaloids.
Scheme 2. Monoterpene spirooxindole-1,3-oxazines 2, 4, 6 and their indole
alkaloid precursors.
[a]
[b]
[c]
M.Sc. Thorsten von Drathen
Department of Chemistry, Institute of Organic Chemistry
University of Hamburg
Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
Dr. Frank Hoffmann
Department of Chemistry, Institute of Inorganic Chemistry
University of Hamburg
Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
Prof. Dr. Malte Brasholz
Institute of Chemistry – Organic Chemistry
University of Rostock
Albert-Einstein-Str. 3A, 18059 Rostock, Germany
E-mail: malte.brasholz@chemie.uni-hamburg.de
ORCID: 0000-0001-8184-646X
We investigated the oxidation of reserpine (5) to
dioxyreserpine (6) and report here a rapid and selective visible
light-driven catalytic photooxygenation of 5, leading to 6 with
attractive yield, and we could for the first time unambiguously
confirm the proposed structure of 6 by X-ray crystallography.
Further, we report that dioxyreserpine (6) can be readily
converted into a new secondary photooxygenation product which
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
we term trioxyreserpine. In addition, we demonstrate that 6 can
be chemically dissected to harness its unique cis-deca-
hydroisoquinoline substructure as a new and valuable chiral
building block for organic synthesis. Finally, we present results of
During an initial screening, irradiation of solutions of 5 in
MeOH, MeCN, THF and HOAC under air and using blue
fluorescent lamps resulted in no conversion after 90 min reaction
time (which is at least partly due to the poor solubility of 5 in these
solvents), but switching to CHCl3, CH2Cl2 or PhCH3 produced
mixtures of unreacted 5 along with dioxyreserpine (6), the known
C7 hydroperoxyindolenine 7 as well as the hydroxyindolenine 8
(a sample of which was prepared independently for comparison
with 7, see SI) in varying ratios. As outlined in Table 1, reserpine
(5) was then irradiated under an atmosphere of O2 in CHCl3 using
compact fluorescent lamps (λmax 450 and 375 nm / 2×18 W each)
as well as 460 nm/10 W blue LEDs (entries 1-5; see also Table
S1 for additional information). Full conversion was generally
observed using CFL lamps and reaction times of 150 min, but
prolonged reaction times led to a slow decomposition of product
6 while at the same time, hydroperoxide 7 appeared to have
slowly been converted into hydroxyl compound 8 by O-O bond
cleavage (entries 1-3). The reaction under blue LED irradiation
proceeded more slowly with 83% conversion after 150 min, and
with a significant shift of selectivity towards hydroperoxide 7 (entry
5). As we employed the longwave-absorbing 9,10-
anthraquinones A-D as photocatalysts in the reaction, we were
pleased to observe a general increase in selectivity in favor of the
desired product 6 (entries 6-12). We identified 1,5-diamino-
anthraquinone C (1,5-AAQ) as the optimal catalyst, and using
blue LED irradiation for 150 min with catalyst quantities of 1-
3 mol-% in MeCN produced dioxyreserpine (6) in isolated yields
of 60-63% after chromatography, as a single stereoisomer
(entries 11 and 12). Notably, when using 3 mol-% of C, the
catalyst showed >99.8% absorption of the incident 460 nm light
(reserpine 5 shows hardly any absorption in this spectral area,
see Figure S3) and product 6 was formed in 90-95% purity as
evidenced by 1H-NMR of the crude reaction mixture, which
otherwise showed no trace of possible byproducts 7 and 8. All
crude NMR spectra were further carefully examined for the
possible presence of the conceivable C2-C7 cleavage product 9
and the previously described anhydronium bases 3,4-
dehydroreserspine (10) and lumireserpine (11),8 however,
compounds 9-11 were not detected in any of our experiments
including those run under catalytic conditions.
a
mechanistic study which encompassed the catalytic
photooxygenation of the C3-epimer of 6, isoreserpine, as well as
the alkaloid mitragynine from Mitragyna speciosa, which led to a
refined mechanistic picture.
Results and Discussion
We studied the visible light-induced photooxygenation of
reserpine (5) under catalyst-free conditions as well as in the
presence of various anthraquinone photocatalysts, potent organic
photooxidants which we utilized previously in the photocatalytic
oxidation of indoles and other heterocycles.7
Table 1. Variation of reaction conditions for the visible light photooxygenation of
(–)-reserpine (5).
light
solvent
yield of
6 [%]b
entry
catalyst
source
(nm)
t [min]
6/7/8 [%]a
(additive)
1
2
CHCl3
CHCl3
CHCl3
CHCl3
CHCl3
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
-
CFL (450)
CFL (450)
CFL (450)
CFL (375)
LED (460)
CFL (375)
CFL (450)
CFL (450)
CFL (450)
LED (460)
LED (460)
LED (460)
150c
300
900
150
150d
150
150
150
150
150
150
150
52/48/0
50/12/38
57/0/43
42/0/58
17/46/37
67/0/33
100/0/0
69/0/31
53/0/47
100/0/0
100/0/0
91/0/9
37
29
-
3
-
17
4
-
23
5
-
15
6
A (5%)
B (5%)
C (5%)
D (5%)
C (5%)
C (3%)
C (1%)
26
7
26
8
58
9
43
On the occasion, the X-ray crystal structure of 6 could be
obtained showing that the initial NMR based structural
assignment6 was in fact correct, including the relative
configuration at the C7 and C3 centers of the newly formed 1,3-
oxazine ring (CCDC No. 1552387).9
10
11
12
63
63 (60)e
67 (63)e
While we observed during the catalyst-free photooxygenation
of 5 that prolonged reaction times led to decomposition of
dioxyreserpine (6), traces of a new secondary photooxygenation
product were detected in crude NMR spectra when using catalyst
C and extended reaction times. When pure dioxyreserpine (6)
was subjected to the typical reaction conditions (3 mol-% catalyst
All reactions were performed with c = 18 µM for substrate 5 and under an
atmosphere of O2. Light sources: CFL = compact fluorescent lamp, 2×18 W,
45050 nm or 2×18 W, 37525 nm; LED = blue LED assembly, 10.1 W,
46015 nm. [a] Ratio determined by 1H NMR analysis. [b] Determined by 1H
NMR against CH2Br2 standard. [c] Conversion of 5 93%. [d] Conversion of 5
83% [e] Isolated yield after chromatography.
This article is protected by copyright. All rights reserved.

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
C, O2, MeCN, blue LED), the new photoproduct could be obtained
in 42% yield after 17 h, and it was identified as the C14 ketone
which we term trioxyreserpine (12, Scheme 3). The mechanism
of this remarkably regioselective Csp3-H oxygenation is currently
under investigation; a C3-C14 alkene is a conceivable primary
intermediate. Additional new synthetic transformations of
dioxyreserpine (6) encompass electrophilic bromination with NBS
to furnish the C10 bromide 13 in 59% yield, and the reduction of
6 with lithium aluminium hydride which cleaved the aryl and
methyl esters to give hydroxymethyl compound 14 in 77% yield.
On the other hand, using sodium borohydride in methanol, a
reductive opening of the 1,3-oxazine ring could be achieved to
give the N-alkylated cis-decahydroisoquinoline 15 in 85% yield,
and we were pleased to find that its N-dealkylation could readily
be accomplished using phenyl chloroformate and proton
sponge,10 to furnish the N-carbamoyl-cis-decahydroisoquinoline
16 in 43% yield. Hence, this highly valuable enantiopure synthetic
building block could be made accessible by chemical degradation
of reserpine (5) in three simple and straightforward steps.
18O2, both the autoxidation reaction in CHCl3 (condition a), and
the reaction with catalyst 1,5-AAQ (C, 3mol-%) in MeCN
(condition b) produced dioxyreserpine (6) with ≥90%
incorporation of two 18O labels (m/z 645 for doubly labelled
[6+H]+), whereas hydroperoxide 7, isolable only from the catalyst-
free photooxidation, was obtained as a ~1:1 mixture of the singly
and doubly 18O-labelled compounds (m/z 645 and 643 for [7+H]+;
Scheme 4a and Figures S4-8). We attribute the latter observation
to O-O bond photolysis of hydroperoxide 7 under the reaction
conditions followed by oxygen exchange with trace water in the
reaction mixture.
(m/z) 645:643:641
(m/z) 645:643:641
cond. a 41:57:2
cond. a
cond. b
90:10:0
90:9:1
cond. a decomposition
cond. b 40% conv. After 150 min
Scheme 4. Isotope labelling and control experiments.
Reserpine (5) possesses a relatively long-lived triplet state
with an energy of ca. 265 kJ/mol12 and therefore it may self-
sensitize singlet oxygen during its autoxidation. In fact,
deliberately inducing the singlet oxygenation of 5 with rose bengal
or methylene blue gives comparable results (see Table S1). When
the catalyst-free oxidation of 5 was performed in the presence of
an excess of 2-methyl-2-butene, the allylic hydroperoxides 17 and
Scheme 3. Photooxygenation of dioxyreserpine (6) to trioxyreserpine (12) and
synthesis of enantiopure cis-decahydroisoquinoline 16.
The mechanistic investigation of the photooxygenation of
reserpine (5) began with isotope labelling experiments in the
presence of 18O2, 18OH2, D2O as well as in d3-MeCN. In the
anthraquinone catalyzed reaction, no deuterium incorporation
was observed with d3-MeCN as the reaction solvent or with added
D2O, and also no 18O labelling when 18OH2 (5-10 equiv.) was
added to the mixture (under 16O2 atmosphere), which clearly ruled
out any long-lived cationic reaction intermediates.11 When the
photooxygenation of 5 was carried out under an atmosphere of
1
18 were produced along with 6 evidencing the presence of O2
(Scheme 4b). Substituted anthraquinones (AQs) A-D are also
1
13
potential sensitizers of O2 but at the same time their excited
states readily engage in hydrogen abstraction reactions.14,7b We
propose that the predominant pathway for initiation of the catalytic
photooxygenation of 5 is the abstraction of the indole N-H
hydrogen atom by the excited state catalyst AQ* to generate a C7
This article is protected by copyright. All rights reserved.

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
indolyl radical 19,7b which is trapped stereoselectively by O2 to
give the peroxyradical 20 as key intermediate (attack of oxygen
occurs from the side of the β-positioned C3 hydrogen), and this
pathway is supported by several observations. When the
photooxygenation of 5 with catalyst C was carried out in DMF
instead of MeCN, a complete switch of selectivity occurred and
hydroperoxide 7 emerged as the only product, which can be
attributed to the much stronger hydrogen atom donation ability of
DMF (homolytic BDE of 310 kJ/mol compared to 406 kJ/mol for
MeCN15), causing an immediate trapping of intermediate
peroxyradical 20 to give 7 (Scheme 4c). On the other hand, when
pure hydroperoxide 7 was subjected to the typical reaction
conditions for the photooxygenation of 5 catalyzed by C (condition
b), 7 was slowly but cleanly converted into dioxyreserpine (6),
which may be explained by a HAT reaction between 7 and the
excited catalyst AQ* to liberate peroxyradical 20 as the key
species, to enter the subsequent rearrangement pathway leading
to 6 (Scheme 4d). By comparison, irradiation of hydroperoxide 7
in CHCl3 alone (condition a) only led to decomposition of 7.
Stereochemically, dioxyreserpine (6) has the same C7-
configuration as hydroperoxide 7 and hydroxyindolenine 8, while
during the rearrangement, a formal inversion occurs at the former
β-configured C3 of reserpine (5). Isopaynantheine (3) and
mitradiversifoline (4) show the identical stereochemical
relationship, but isoreserpiline (1) is α-configured at C3 while
melodinoxanine (2) has again the same core configuration as 4
and 6. To clarify the stereochemical course of the
photooxygenation, we employed the C3-α-epimer of 5,
isoreserpine (21), as well as the alkaloid mitragynine (22) in the
catalytic reaction.
Compared to spirooxindole oxazines 2, 4 and 6, dioxyiso-
reserpine (23) and dioxymitragynine (24) show a double inverted
configuration at the C3 and C7 stereocenters as evident from
NMR analysis in solution and the X-ray crystal structure of 23
(Figure 1, CCDC No. 1820245),9 which proved that 23 and 24
were formed via the same mechanism as 6, beginning with a
stereoselective attack of oxygen from the side of the C3-α-
hydrogen. However, the formation of the new oxazine ring
occurred in such fashion that the C and D rings in 23 and 24 are
cis-fused, causing the compounds to suffer from sterically highly
unfavorable 1,3-diaxial interactions, and thus explaining the poor
yield and selectivity observed in the photooxygenation of C3-α-
configured substrates 21 and 22. At the same time, this result
strongly suggests that if isoreserpiline (1, Scheme 1) is the true
biosynthetic precursor to melodinoxanine (2), it must undergo
isomerization to its β-configured epimer reserpiline prior to
oxidation, possibly under acid catalysis.16
Any plausible mechanism for the rearrangement occurring
during the photooxygenation of tetrahydro-β-carboline alkaloids
leading to the spiro-[(1,3)oxazinan-3,6'-oxindole] ring system III
but not to the ‘regular’ spiro[pyrrolidine-3,3’-oxindole] motif I must
account for the key observations made during this study. N-Methyl
reserpine did not undergo the analogous rearrangement under
catalytic or catalyst-free conditions, confirming the necessity of an
unprotected indole nitrogen. Apart from byproducts 7 and 8, no
other reaction intermediates could be detected in the
photooxygenation of reserpine (5), and the attempted nucleophilic
trapping of possible radical cationic or cationic intermediates
failed. This points towards a rapid reaction sequence involving
radical or possibly even carbene intermediates, and further it
renders a previous proposal of the biosynthetic intermediacy of
the ‘regular’ spiro-2-oxindole structure
I on the way to
spirooxazines III highly unlikely.5 Further, a previously assumed
dioxetane intermediate4,6 seems not to be involved at least in the
case of reserpine (5), since no trace of keto amide 9 could be
detected and the clean stereochemical inversion at C3 can be
explained only with difficulty in this scenario, which at the same
time involves an unlikely simultaneous homolytic cleavage of the
O-O and C2-C3 bonds. On the other hand, the observation of the
C2-C7 cleavage product in the photooxygenation of isoreserpine
(21) clearly shows that mechanistic branching must occur after
the initial radical oxygenation step.
A proposed overall reaction mechanism is depicted in
Scheme 5, for the case of C3-β-configuration as found in
reserpine (5). The catalytic photooxygenation is initiated by
hydrogen abstraction from indole precursor II by excited catalyst
AQ* followed by stereoselective radical oxygenation with O2 (vide
supra).7b The resulting peroxyradical 20’ is less stable than its C7
epimer derived from α-configured II due to cis-fused C and D
rings,16 however, its peroxy group is properly oriented to engage
in an intramolecular 1,5-hydrogen atom transfer (HAT), to give a
stabilized and strain-relieved allylic radical 25. Subsequent HAT
to 25 generates a reactive dienamine 26 which is epoxidized,
followed by ring-opening of epoxide 27 to the zwitterion 28. At this
stage, the key scission of the C2-C3 bond occurs and we propose
that intermediate 28 rearranges to give the cyclic (alkyl)(amino)
carbene 29,17 which undergoes final ring-closure by proton
Whereas isoreserpine (21) is unreactive under catalyst-free
conditions,6 full conversion was observed in the presence of
catalyst C. However, in sharp contrast to reserpine (5), its epimer
21 reacted very unselectively to give dioxyisoreserpine (23) in
only ca. 10% yield along with the C7 hydroperoxyindolenine and
the C2-C7 cleavage product derived from 21 (see SI section for
details). The photooxygenation of mitragynine (22) followed the
same trend, giving rise to dioxymitragynine (24) in 24% yield.
Figure 1. Structures of dioxyisoreserpine (23) and dioxymitragynine (24).
This article is protected by copyright. All rights reserved.

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
transfer18 and nucleophilic trapping of the iminium ion through the
chair-like transition structure 30, to yield the final product with the
observed stereochemistry. Alternatively, carbene 29 may react
via direct OH-insertion19 with the C7-hydroxy group.
1.97-2.02 (m, 2 H, 19-Ha, 20-H), 2.26-2.42 (m, 3 H, 19-Hb, 6-Hb, 15-H),
2.54 (dd, J = 2.7, 11.7 Hz, 1 H, 21-Ha), 2.64-2.76 (m, 3 H, 5-Ha, 16-H, 21-
Hb), 3.03 (mc, 1 H, 5-Hb), 3.50 (s, 3 H, 33-H), 3.67 (s, 3 H, 34-H), 3.79 (s,
3 H, 35-H), 3.82-3.89 (m, 2 H, 17-H, 18-H), 3.92 (s, 3 H, 31-H), 3.93 (s, 6
H, 30-H, 32-H), 4.59 (dd, J = 3.2, 9.7 Hz, 1 H, 3-H), 4.99-5.08 (m, 1 H, 18-
H), 6.37 (d, J = 2.2 Hz, 1 H, 12-H), 6.52 (dd, J = 2.2, 8.2 Hz, 1 H, 10-H,
7.26 (d, J = 8.2 Hz, 1 H, 9-H), 7.34 (s, 2 H, 25-H, 29-H) ppm. 13C NMR
(CDCl3, 150 MHz): δ = 29.3 (t, C14), 30.3 (t, C19), 31.8 (t, C6), 34.0 (d,
C20), 36.1 (d, C15), 47.9 (t, C5), 51.7 (d, C16), 51.8 (q, C34), 55.6 (q, C35),
56.3 (q, C30, C32), 57.3 (t, C21), 60.86 (q, C33), 60.93 (q, C31), 74.4 (s,
C7), 77.8 (d, C17), 78.1 (d, C18), 87.0 (d, C3), 97.2 (d, C12), 106.8 (d,
C25, C29), 107.4 (d, C10), 122.7 (s, C8), 125.2 (d, C9), 125.4 (s, C24),
141.1 (s, C13), 142.3 (s, C27), 153.0 (s, C26, C28), 161.3 (s, C11), 165.4
(s, C23), 171.8 (s, C22), 178.1 (s, C2) ppm. IR: ꢀ = 3355 (N-H), 2955, 2920,
2850 (C-H), 1715 (C=O), 1660, 1630, 1460, 760 cm-1. HRMS (ESI): m/z
[M+H]+ calc. for [C33H41N2O11]+ 641.2705, found 641.2772.
Trioxyreserpine (12): In a 10 mL crimp cap vial, dioxyreserpine (6; 35.4
mg; 55.3 μmol) was dissolved in 3.0 mL of a freshly prepared stock solution
of 1,5-AAQ (C) in dry MeCN (0.13 mg/mL equivalent of 0.54 μmol/mL) to
give a 18 mM solution of 6 containing 3 mol-% of catalyst. The vial was
sealed and O2 was briefly bubbled through the reaction mixture via cannula,
then an O2-balloon was fitted to the cannula. The reaction mixture was
stirred rapidly with irradiation inside a blue LED assembly (10.1 W, 294 lm,
460±15 nm) at ambient temperature for 17 h. The solvent was removed
under reduced pressure, and the crude product was purified by column
chromatography (silica gel, EtOAc/CHCl3 1:1) to give 15.3 mg (42%) of
trioxyreserpine (12) as a colorless resin.
Scheme 5. Proposed mechanism for the catalytic photooxygenation of C3-β-
configured tetrahydro-β-carbolines.
Conclusions
Rf = 0.30 (EtOAc/CHCl3 1:1); [α]D = −69.4°, c 0.16, CHCl3. 1H NMR (CDCl3,
600 MHz): δ = 1.79 (td, J = 2.2, 13.9 Hz, 1 H, 6-Ha), 1.91 (dd, J = 12.4,
13.0 Hz, 1 H, 19-Ha), 2.07 (mc, 1 H, 19-Hb), 2.33 (dt, J = 4.9, 13.9 Hz, 1 H,
6-Hb), 2.44 (mc, 1 H, 20-H), 2.53 (dd, J = 5.2, 10.3 Hz, 1 H, 16-H), 2.82-
2.88 (m, 2 H, 5-Heq, 21-Heq), 3.10 (dd, J = 3.0, 12.3 Hz, 1 H, 21-Hax), 3.36
(mc, 1 H, 5-Hax), 3.48-3.52 (m, 1 H, 15-H), 3.63 (s, 3 H, 33-H), 3.68 (s, 3 H,
34-H), 3.65 (s, 3 H, 34-H), 3.76 (s, 3 H, 35-H), 3.92 (s, 3 H, 31-H), 3.94 (s,
6 H, 30-H, 32-H), 4.06 (dd, J = 9.3, 10.3 Hz, 1 H, 17-H), 4.99 (ddd, J = 5.0,
9.3, 12.2 Hz, 1 H, 18-H), 5.31 (s, 1 H, 3-H), 6.28 (d, J = 2.2 Hz, 1 H, 12-H),
6.52 (dd, J = 2.2, 8.2 Hz, 1 H, 10-H), 7.25 (d, J = 8.2 Hz, 1 H, 9-H), 7.33
(s, 3 H, 25-H, 29-H, N-H) ppm. 13C NMR (CDCl3, 150 MHz): δ = 31.3 (t,
C6), 31.6 (t, C19), 35.9 (d, C20), 46.5 (d, C-16), 48.2 (t, C5), 51.0 (d, C15),
51.9 (q, C34), 55.5 (q, C-35), 55.8 (t, C21), 56.3 (q, C30, C32), 60.8 (q,
C33), 60.9 (q, C31), 74.7 (s, C7), 77.1 (d, C17), 78.1 (d, C18), 87.8 (d, C3),
97.1 (d, C12), 106.8 (d, C25, C29), 107.4 (d, C10), 121.7 (s, C8), 125.2 (s,
C24), 125.6 (d, C9), 141.3 (s, C13), 142.3 (s, C27), 153.0 (s, C26, C28),
161.5 (s, C11), 165.4 (s, C23), 171.2 (s, C22), 177.4 (s, C2), 200.9 (s,
C14) ppm. IR: ꢀ = 3305 (N-H), 2940, 2840 (C-H), 1715 (C=O), 1635, 1590,
1505, 1335, 1250, 1155, 755 cm-1. HRMS (ESI): m/z [M+H]+ calc. for
[C33H39N2O12]+ 655.2498, found 655.2508.
In summary, we developed efficient catalytic photooxygenation
procedures to access the valuable spirooxindole-1,3-oxazines
dioxyreserpine (6) and trioxyreserpine (12) from their parent
indole alkaloid reserpine (5). Dioxyreserpine (6) could be
valorized by reductive ring-opening and N-dealkylation to harness
its cis-decahydroisoquinoline core structure 16 as a new
enantiopure building block. The mechanistic investigation of the
photooxygenation of 5 compared to its C3-epimer 21 as well as
mitragynine (22) allowed to develop
a new mechanistic
hypothesis for their unusual photooxidative rearrangements, with
implications that may enhance the understanding of the
biosynthesis of related natural spirooxindole-1,3-oxazines.
Experimental Section
Dioxyreserpine (6): In a 10 mL crimp cap vial, (−)-reserpine (5; 22.0 mg;
36.1 μmol) was suspended in 2.0 mL of a freshly prepared stock solution
of 1,5-AAQ (C) in dry MeCN (0.13 mg/mL equivalent of 0.54 μmol/mL) to
give a 18 mM solution of 5 containing 3 mol-% of catalyst. The vial was
sealed and O2 was briefly bubbled through the reaction mixture via cannula,
then an O2-balloon was fitted to the cannula. The reaction mixture was
stirred rapidly with irradiation inside a blue LED assembly (10.1 W, 294 lm,
460±15 nm) at ambient temperature for 150 min, whereupon the mixture
turned homogeneous. The solvent was removed under reduced pressure,
and the crude product was purified by column chromatography (silica gel,
petroleum ether/ethyl acetate 1:0→1:3) to give 14.6 mg (63%) of
dioxyreserpine (6) as a pale yellow solid.
N-Alkyl-cis-decahydroisoquinoline 15: Dioxyreserpine (6, 108 mg,
169 μmol) was dissolved in MeOH (27 mL) under N2 and the solution was
cooled to 0 °C. Sodium borohydride (63.9 mg, 169 μmol, 10.0 eq.) was
added, and the mixture was allowed to warm to r.t. and stirred overnight.
The mixture was diluted with H2O, extracted with EtOAc (3×), dried over
MgSO4, filtered and concentrated. The crude mixture was then purified by
column chromatography (silica gel, PET/ethyl acetate 1:1→0:1) to give
92.3 mg (85%) 15 as a colorless film.
Rf = 0.38 (EtOAc); [α]D = −41.0°, c 0.58, CHCl3. 1H NMR (CDCl3, 500
MHz): δ = 1.26-1.33 (m, 1 H, 8-Ha), 1.34-1.40 (m, 1 H, 4-Ha), 1.63 (br s, 1
Rf = 0.60 (EtOAc); m.p. 150 °C; [α]D = −60.0°, c 0.50, CHCl3 (lit. −78.8°, c
0.10, CHCl3).[6] 1H NMR (CDCl3, 400 MHz): δ = 1.44 (td, J = 3.5, 12.7 Hz,
1 H, 14-Ha), 1.72 (mc, 1 H, 14-Hb), 1.80 (td, J = 1.7, 13.8 Hz, 1 H, 6-Ha),
This article is protected by copyright. All rights reserved.

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
H, OH), 1.81 (ddd, J = 2.9, 6.0, 14.6 Hz, 1H, 2’-Ha), 1.86-2.00 (m, 3 H, 4-
Hb, 2’-Hb, 3-Ha), 2.02-2.15 (m, 3 H, 4a-H, 8a-H, 8-Hb), 2.22 (d, J = 11.5 Hz,
1 H, 1-Ha), 2.54 (mc, 1 H, 1’-Ha), 2.73 (dd, J = 4.5, 11.0 Hz, 1 H, 5-H), 3.10
(d, J = 11.5 Hz, 1 H, 1-Hb), 3.16 (mc, 1 H, 1’-Hb), 3.32-3.38 (m, 1 H, 3-Hb),
3.49 (s, 3 H, C6-OMe), 3.74 (s, 3 H, CO2Me), 3.76 (s, 3 H, C6’’-OMe), 3.79
(dd, J = 9.5, 11.0 Hz, 1 H, 6-H), 3.88 (s, 3 H, Ar-OMe), 3.89 (s, 6 H, Ar-
OMe), 4.98-5.06 (m, 1 H, 7-H), 6.38 (d, J = 2.2 Hz, 1 H, 7’’-H), 6.51 (dd, J
= 2.2, 8.3 Hz, 1 H, 5’’-H), 7.16 (d, J = 8.3 Hz, 1 H, 4’’-H), 7.28 (s, 2 H, Ar-
H), 7.70 (s, 1 H, NH) ppm. 13C NMR (CDCl3, 100 MHz): δ = 23.3 (t, C4),
29.7 (t, C8), 32.1 (t, C-2‘), , 34.7 (d, C8a), 37.1 (d, C4a), 51.7 (q, CO2Me),
52.1 (d, C5), 53.7 (t, C1‘), 54.5 (t, C3), 55.5 (q, C6’’-OMe), 56.2 (q, Ar-
OMe), 57.9 (t, C1), 60.89 (q, Ar-OMe), 60.92 (q, C6-OMe), 77.2 (s, C3’’),
77.9 (2 d, C6, C7), 97.5 (C7’’), 106.97 (d, Ar), 107.05 (d, C5’’), 124.3 (s,
C3a’’), 124.8 (d, C4’’), 125.2 (s, Ar), 140.9 (s, C7a’’), 142.3 (s, Ar), 152.9
(s, Ar), 160.8 (s, C6’’), 165.5 (s, OCOAr), 172.1 (s, CO2Me), 180.5 (s,
C2’’) ppm. IR: ꢀ = 3310 (OH), 2940, 2840 (C-H), 1715 (C=O), 1630, 1590,
1505, 1250, 1125, 725 cm-1. HRMS (ESI): m/z [M+H]+ calc. for
[C33H43N2O11]+ 643.2861, found 643.2955.
M. U. Rahman, N. K. Nyola, Synth. Commun. 2016, 46, 1643-1664; (c)
B. Yu, D.-Q. Yu, H.-M. Liu, Eur. J. Med. Chem. 2015, 97, 673-698 (d) X.-
H. Zhong, L. Xiao, Q. Wang, B.-J. Zhang, M.-F. Bao, X.-H. Cai, L. Peng,
Phytochem. Lett. 2014, 10, 55-59; (e) Y.-J. Wu, Top. Heterocyclic Chem.
2010, 26, 1-29; (f) S. Peddibhotla, Curr. Bioactive Comp. 2009, 5, 20-38.
(a) G. A. Cordell, Introduction to Alkaloids. A Biogenetic Approach. Wiley-
Interscience, New York 1981, pp. 656-690; (b) R. Ahmad, F. Salim in
Studies in Natural Product Chemistry, Vol. 45, A.-u.-Rahman, Ed.;
Elsevier, Amsterdam 2015, pp. 485-525.
[2]
[3]
(a) M. M. M. Santos, Tetrahedron 2014, 70, 9735-9757; (b) B. M. Trost,
M. K. Brennan, Synthesis 2009, 3003-3025; (c) L. A. Paquette, J. E.
Hofferberth, in Organic Reactions Vol. 62, pp. 477-567, Wiley 2004.; (d)
C. Marti, E. M. Carreira, Eur. J. Org. Chem. 2003, 2209-2219; e) E.
Schendera, S. Lerch, T. von Drathen, L.-N. Unkel, M. Brasholz, Eur. J.
Org. Chem. 2017, 3134-3138.
[4]
[5]
[6]
[7]
M. Kitajima, S. Ohara, Noriyuki Kogure, Y. Wu, R. Zhang, H. Takayama,
Heterocycles 2012, 85, 1949-1959.
X.-F. Cao, J.-S. Wang, X.-B. Wang, J. Luo, H.-Y. Wang, L.-Y. Kong,
Phythochemistry 2013, 96, 389-396.
D. V. C. Awang, B. A. Dawson, M. Girard, A. Vincent, I. Ekiel, J. Org.
Chem. 1990, 55, 4443-4448.
cis-Decahydroisoquinoline 16: N-Alkyl-cis-decahydroisoquinoline 15
(29.5 mg, 45.9 μmol) was dissolved in DCE (2 mL) under N2 and the
solution was cooled to 0 °C. A solution of proton sponge (30.0 mg,
140 μmol, 3.05 eq.) and phenyl chloroformate (11.8 µL, 94.2 μmol,
2.05 eq.) in DCE (3 mL) was added dropwise. The mixture was stirred at
r.t. for 1 h and 2.5 h at 100 °C. The solvent was removed under reduced
pressure, the residue was diluted with water and extracted with EtOAc (3×).
The combined organic layers were washed with cold 1 M HCl (3×), H2O
(3×), brine (3×) and dried over MgSO4. The solvent was removed under
reduced pressure, and the crude mixture was purified by column
chromatography (silica gel, PET/ethyl acetate 1:0→1:1) to give 11.0 mg
(43%) 16 as a colorless film.
Rf = 0.35 (PET/EtOAc 1:1); [α]D = −30.3°, c 0.46, CHCl3. 1H NMR (CDCl3,
600 MHz): δ = 1.36-1.46 (m, 4-Ha), 1.79-1.98 (m, 2 H, 4-Hb, 8-Ha), 2.04-
2.14 (m, 2 H, 8-Hb, 8a-H), 2.23-2.32 (m, 1 H, 4a-H), 2.77 (dd, J = 4.7, 10.8
Hz, 1 H, 5-H), 2.75,* 2.89* (2 t, J = 12.0 Hz, 1 H, 3-Ha), 3.03,* 3.19* (2 d,
J = 12.5 Hz, 1 H, 1-Ha), 3.52 (s, 3 H, CH-OMe), 3.76 (s, 3 H, CO2Me), 3.80-
3.89 (m, 1 H, 6-H), 3.92, 3.93 (2 s, 9 H, Ar-OMe), 4.13-4.25 (m, 1 H, 1-Hb),
4.33-4.46 (m, 1 H, 3-Hb), 5.02-5.13 (m, 1 H, 7-H), 6.99-7.12 (m, 2 H, Ph)
7.14-7.21 (m, 1 H, Ph) 7.27-7.37 (m, 4 H, Ar, Ph) ppm. *Signals marked
show rotamer splitting. 13C NMR (CDCl3, 150 MHz): δ = 22.5 (t, C4), 28.8
(t, C8), 34.4 (d, C8a), 37.2 (d, C4a), 44.3 (t, C3), 48.9 (t, C1), 51.9 (q,
CO2Me), 52.1 (d, C5), 56.3 (q, Ar-OMe), 61.0 (q, Ar-OMe, CH-OMe), 77.5
(d, C7), 77.7 (d, C6), 106.9 (d, Ar), 121.6 (d, Ph), 125.1 (s, Ar), 125.3 (d,
Ph), 129.2 (d, Ph), 142.4 (s, Ar), 151.2 (s, Ph), 153.0 (s, Ar), 154.0 (s,
NCO), 165.4 (s, OCOAr), 171.8 (s, CO2Me) ppm. IR: ꢀ = 2940, 2840 (C-
H), 1710 (C=O), 1590, 1505, 1250, 1210, 725 cm-1. HRMS (ESI): m/z
[M+H]+ calc. for [C29H36NO10]+ 558.2334, found 558.2347.
a) S. Lerch, L.-N. Unkel, P. Wienefeld, M. Brasholz, Synlett 2014, 2673-
2680; b) S. Lerch, L.-N. Unkel, M. Brasholz, Angew. Chem. Int. Ed. 2014,
53, 6558-6562; c) F. Rusch, L.-N. Unkel, D. Alpers, F. Hoffmann, M.
Brasholz, Chem. Eur. J. 2015, 21, 8336-8340.
[8]
a) J. Bayer, Pharmazie 1958, 13, 468-469; b) S. Ljungberg, J. Pharm.
Belg. 1959, 14, 115-125; c) G. E. Wright, T. Y. Tang, J. Pharmaceut. Sci.
1972, 61, 299-300; d) N. Jamil, H. Afrozrizvi, I. Ahmed, A. E. Beg,
Pharmazie 1983, 38, 467-469; e) M. A. Muñoz, D. González-Arjona, M.
Balón, J. Chem. Soc., Perkin Trans. 2 1991, 453-456.
[9]
The crystal structure data can be retrieved from the Cambridge
Crystallographic Data Centre, www.ccdc.cam.ac.uk.
[10] a) R. A. Olofson, R. C. Schnur, L. Bunes, J. P. Pepe, Tetrahedron Lett.
1977, 18, 1567-1570; b) D. M. Zimmerman, B. E. Cantrell, J. K. Reel, S.
K. Hemrick-Luecke, R. W. Fuller, J. Med. Chem. 1986, 29, 1517-1520.
[11] Reserpine (5) shows two pH-dependent irreversible oxidations in cyclic
(1)
voltammetry, Eox at +0.6 to 0.8 V vs. SCE (in organic solvent/H2O
mixtures) corresponding to one-electron oxidation of the indole nucleus
(2)
and Eox at around +1.40 V being the oxidation of its tertiary amine
nitrogen atom, see: a) J. A. Fournier, A. B. Wolk, M. A. Johnson, Anal.
Chem. 2013, 85, 7339-7344; b) D. M. Stanković, E. Mehmeti, L. Svorc,
K. Kalcher, Int. J. Electrochem. Sci. 2015, 10 1469-1477; values
converted to SCE scale. For excited catalyst 3C*, E*red can be estimated
around +0.8 V, see: c) J. Ritter, H.-U. Borst, T. Lindner, M. Hauser, S.
Brosig, K. Bredereck, U. E. Steiner, D. Kühn, J. Kelemen, H. E. A. Kramer,
J. Photochem. Photobiol. A: Chem. 1988, 41, 227-244. Hence,
photoelectron transfer between 3C* and the indole nucleus of 5 is
principally possible, but Eox(2) of the tertiary amine nitrogen is too high for
PET to occur with 3C*.
[12] B. Savory, J. H. Turnbull, J. Photochem. 1983, 23, 171-181.
[13] a) I. M. Byteva, G. P. Gurinovich, O. L. Golomb, V. V. Karpov, Zh. Prikl.
Spektrosk. 1986, 44, 589-593; b) K. Gollnick, S. Held, D. O. Mártire, S.
E. Braslavsky, J. Photochem. Photobiol. A: Chem. 1992, 69, 155-165.
[14] a) K. Hamanoue, T. Nakayama, A. Tanaka, Y. Kajiwara, H. Teranishi, J.
Photochem. 1986, 34, 73-81; b) H. Pal, D. Palit, T. Mukherjee, J. P. Mittal,
J. Photochem. Photobiol. A: Chem. 1991, 62, 183-193; c) S. Kamijo, G.
Takao, K. Kamijo, T. Tsuno, K. Ishiguro, T. Murafuji, Org. Lett. 2016, 18,
4912-15; d) T. Yamaguchi, Y. Kudo, S.-I. Hirashima, E. Yamaguchi, N.
Tada, T. Miura, A. Itoh, Tetrahedron Lett. 2015, 56, 1973-1975; e) T.
Yamaguchi, E. Yamaguchi, A. Itoh, Org. Lett. 2017, 19, 1282-1285.
[15] Y. R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC
Press, Boca Raton, 2007.
Acknowledgements
The authors thank the Deutsche Forschungsgemeinschaft (DFG
grant no. BR3748/2-1) for financial support of this research.
Keywords: alkaloids • indoles • photooxygenation • rearrange-
ments
[16] L.-H. Zhang, A. K. Gupta, J. M. Cook, J. Org. Chem. 1989, 54, 4708-
4712.
[1]
Reviews: (a) M. Kaur, M. Singh, N. Chadha, O. Silakari, Eur. J. Med.
Chem. 2016, 123, 858-894; (b) P. Saraswat, G. Jeybalan, M. Z. Hassan,
This article is protected by copyright. All rights reserved.

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
[17] The transformation 28→29 is reminiscent of the classical route to cyclic
(alkyl)(amino) carbenes (CAACs) by deprotonation of iminium ions, see:
V. Lavallo, Y. Canac, C. Präsang, B. Donnadieu, G. Bertrand, Angew.
Chem. Int. Ed. 2005, 44, 5705-5709.
[19] X-H insertion reactions of (alkyl)(amino) carbenes have been described,
for a review see: M. Melaimi, R. Jazzar, M. Soleilhavoup, G. Bertrand,
Angew. Chem. Int. Ed. 2017, 56, 10046-10068.
[18] S. T. Belt, C. Bohne, G. Charette, S. E. Sugamori, J. C. Scaiano, J. Am.
Chem. Soc. 1993, 115, 2200-2205.
This article is protected by copyright. All rights reserved.

10.1002/chem.201801882
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Thorsten von Drathen, Frank Hoffmann,
Malte Brasholz*
Page No. – Page No.
Visible Light Catalytic
Photooxygenation of Monoterpene
Indole Alkaloids: Access to
Spirooxindole-1,3-Oxazines
Photooxidative rearrangement: Few natural oxindole alkaloids possess an
exceptional spirooxindole-1,3-oxazine core structure, which results from an unusual
oxidative indole rearrangement. The Rauvolfia alkaloid reserpine can be converted
into the photoproducts dioxyreserpine and trioxyreserpine through efficient visible-
light catalytic photooxygenation with anthraquinone photocatalysts.
This article is protected by copyright. All rights reserved.