Total synthesis and antileukemic evaluations of the phenazine 5,10-dioxide natural products iodinin, myxin and their derivatives

Article abstract of DOI:10.1016/j.bmc.2017.02.058

A new efficient total synthesis of the phenazine 5,10-dioxide natural products iodinin and myxin and new compounds derived from them was achieved in few steps, a key-step being 1,6-dihydroxyphenazine di-N-oxidation. Analogues prepared from iodinin, including myxin and 2-ethoxy-2-oxoethoxy derivatives, had fully retained cytotoxic effect against human cancer cells (MOLM-13 leukemia) at atmospheric and low oxygen level. Moreover, iodinin was for the first time shown to be hypoxia selective. The structure-activity relationship for leukemia cell death induction revealed that the level of N-oxide functionality was essential for cytotoxicity. It also revealed that only one of the two phenolic functions is required for activity, allowing the other one to be modified without loss of potency.

Full text of DOI:10.1016/j.bmc.2017.02.058

Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Total synthesis and antileukemic evaluations of the phenazine
5,10-dioxide natural products iodinin, myxin and their derivatives
Elvar Örn Viktorsson a, Bendik Melling Grøthe b, Reidun Aesoy c, Misbah Sabir c, Simen Snellingen b,
Anthony Prandina a, Ove Alexander Høgmoen Åstrand a, Tore Bonge-Hansen b, Stein Ove Døskeland d,
Lars Herfindal c, Pål Rongved a,
⇑
a Department of Pharmaceutical Chemistry, School of Pharmacy, University of Oslo, PO Box 1068 Blindern, N0316 Oslo, Norway
b Department of Chemistry, University of Oslo, PO Box 1033, Blindern, 0315 Oslo, Norway
c Centre for Pharmacy, Department of Clinical Science, University of Bergen, Jonas Lies vei 87, N-5009 Bergen, Norway
d Department of Biomedicine, University of Bergen, Jonas Lies vei 91, N-5009 Bergen, Norway
a r t i c l e i n f o
a b s t r a c t
Article history:
A new efficient total synthesis of the phenazine 5,10-dioxide natural products iodinin and myxin and new
compounds derived from them was achieved in few steps, a key-step being 1,6-dihydroxyphenazine di-
N-oxidation. Analogues prepared from iodinin, including myxin and 2-ethoxy-2-oxoethoxy derivatives,
had fully retained cytotoxic effect against human cancer cells (MOLM-13 leukemia) at atmospheric
and low oxygen level. Moreover, iodinin was for the first time shown to be hypoxia selective. The struc-
ture-activity relationship for leukemia cell death induction revealed that the level of N-oxide functional-
ity was essential for cytotoxicity. It also revealed that only one of the two phenolic functions is required
for activity, allowing the other one to be modified without loss of potency.
Received 4 January 2017
Revised 22 February 2017
Accepted 25 February 2017
Available online xxxx
Keywords:
Phenazines
Phenazine 5,10-dioxides
Acute myeloid leukemia
Hypoxia selectivity
Ó 2017 Published by Elsevier Ltd.
1. Introduction
Iodinin (2) and myxin (3) (Fig. 1) are well-known natural prod-
ucts.4 Iodinin was first identified by Davis in 1939 within bacterial
Phenazines (PHz’s, Fig. 1) are N-heterocyclic aromatic com-
pounds found widely in terrestrial and aquatic environments.
Genera of Pseudomonas and Streptomycetes are among the richest
natural sources, although most of the >6000 PHz’s described are
of synthetic origin.1–3 Hundreds of PHz structures display biologi-
cal activities, such as antitumor-,4–6 anti-infective-7,8, anti-inflam-
matory,9 and immune-suppressive effects.10 Pseudomonas
aeruginosa’s pyocyanin induces neutrophil death via mitochondrial
reactive oxygen species and mitochondrial acid sphingomyeli-
nase.11 Various modes of action have been assigned for their bio-
cultures of Chromobacterium iodinum16 (later re-classified as
Brevibacterium iodinum), and its antimicrobial features described
shortly after by McIlwain.4,17,18 Since then, various genera of bacte-
ria producing the deep purple pigment have been identified.1,4
Recently, iodinin (2) was isolated from marine actinomycetes bac-
terium collected from sediment samples in the Trondheim fjord
of Norway.19,20 Screens showed iodinin (2) to possess cytotoxic
features with high selectivity towards acute myeloid leukemia
(AML) and acute promyelocytic leukemia (APL) cells (EC50 values
for apoptotic cell death up to 40 times lower than for various nor-
mal cells). Remarkably, iodinin (2) promoted apoptotic cell death
in cell-lines characterized by genetic features of poor patient prog-
nosis.5,20 Myxin (3) displays an unusually attractive antimicrobial
profile for a PHz; potent broad spectrum activity against both
Gram positive and -negative bacteria and fungi.7
Earlier findings indicate iodinin (2) and myxin (3) to have abil-
ity to interact with DNA through intercalation, especially with the
C-G base pair of DNA, as documented for myxin (3),6 which can
inhibit DNA template controlled RNA synthesis.21 Moreover, myxin
(3) has recently been shown to cause radical mediated DNA strand
cleavage and to promote cytotoxic events in the human colorectal
cancer cell line HCT-116 under aerobic and anaerobic conditions.13
logical
effects,
like
DNA-intercalation,6
inhibition
of
topoisomerases,1 metal chelation12 and production of reactive oxy-
gen species (ROS).13 The latter is most prominent for the subclass
of PHz 5,10-dioxides. Some of these oxides can be bio-reduced to
form free hydroxyl radicals, and more so under hypoxic condi-
tions.13–15 These features render the PHz 5,10-dioxides particularly
valuable as starting points for drug discovery, aimed to identify
potential antitumor agents which selectively promote apoptotic
cell death at low levels of oxygen saturation.
⇑
Corresponding author.
E-mail address: pal.rongved@farmasi.uio.no (P. Rongved).
http://dx.doi.org/10.1016/j.bmc.2017.02.058
0968-0896/Ó 2017 Published by Elsevier Ltd.
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

2
E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
growth and incomplete vascularization.24 In addition, cancer cells
H
O-
N+
O
within hypoxic regions are often resistant to chemotherapy due
to upregulation of hypoxia-induced survival gene products.25 How-
ever, this challenge can also provide a window of opportunity for
pro-drugs activated selectively in oxygen deficit tumor tissues.26
AML is the most common malignant myeloid cancer in adults,
and its yearly prevalence increases to near 18 cases per 100,000
at high age.27 Advances in treatments like allogeneic stem cell
transplantation have mostly benefited younger patients, and the
preferred chemotherapy today was developed more than 40 years
ago.28,29 Current standards in clinic, such as daunorubicin (5)
(Fig. 2), are unselective and have toxic-, sometimes lethal, side
effects.4,28 Thus, new therapeutic agents which target leukemia
more selectively are called for. On forehand, iodinin (2) has an
attractive profile as lead candidate in the quest of finding new
and more selective therapies against AML.5 However, the com-
pound is not readily available and is known to be notoriously insol-
uble in aqueous and organic media.7
Our primary goal in the present work was to develop an iodi-
nin-based new potential drug candidate for AML therapy. To
achieve this, it was mandatory to develop a new effective synthetic
route towards iodinin (2) from readily available chemical sources.
In the synthetic process, the crucial step would be the oxidation
forming both N-oxides simultaneously on the 1,6-di-O-substituted
PHz scaffold. Judged by synthetic procedures in prior litera-
ture,13,30 this step was anticipated to be the most challenging to
achieve efficiently. Due to structural resemblance, myxin (3) would
also be prepared from iodinin (2) as myxin’s (3) antileukemic
effects were unknown by forehand. A second goal was to explore
the N-oxide role of importance in terms of cytotoxicity and further,
if the phenols of iodinin (2) could be derivatized by alkylation
without compromising the biological activity on leukemic cell
death.
10
9
6
1
N
8
7
2
3
N
N+
4
5
OR O-
1
Phenazine ( )
2
R = -H; iodinin ( )
R = -Me; myxin (3)
Fig. 1. The phenzine core (1) with numbered positions and the phenazine
5,10-dioxide natural produdcts iodinin (2) and myxin (3).
O-
O
OH
O
N+
N
OH
O
N+ NH2
O-
tirapazamine (4)
OMe O
daunorubicin (5)
OH
O
OH
NH2
Fig. 2. The structure of tirapazamine, a hypoxia selective cytotoxic prodrug and
daunorubicin, a DNA-intercalator used for the treatment of acute myeloid leukemia.
The mode of action is reported to be the same or similar as for tira-
pazamine (4) (Fig. 2),13 one of the most documented hypoxia-
selective cytotoxic agents known. Tirapazamine (4) has been sub-
jected to phase I, II and III clinical trials to treat various forms of
cancer, and clinical efficacy reported in some cases, especially in
combination with cis-platin.22 Chowdhury et al. showed that Phe-
nazine 5,10-dioxides may undergo reduction of a single electron
intracellularly.13 These events afford oxygen sensitive radicals of
the heterocyclic aromatic N-oxide. Thus under normoxic condi-
tions, the radicalized derivative is immediately converted back to
the mother compound. In hypoxic environment however, life span
of such drug-radical is prolonged and can lead to the formation of
highly cytotoxic species through loss of one N-oxide from the
scaffold.13,23
2. Results and discussion
2.1. Synthesis
The first phase of development of a synthetic route was aimed
to provide 1,6-dimethoxyphenazine (9) from readily available
commercial sources. To achieve this, two different approaches
were explored, both of which avoid formation of PHz regioisomers
(Scheme 1). On one hand, a Pd-catalyzed double Buchwald-
Hartwig C-N cross-coupling of 2-bromo-3-methoxyaniline (6)
Exploiting the advantage of tumor hypoxia in cancer therapy is
an active research field, in part due to the fact that areas of tumor
tissue can be very oxygen deficient as the result of rapid cell
OMe
O-
OH
NH2
O2N
N+
Br
N+
NH2
OMe
6
8
OMe
7
OMe O-
3
i
ii
v
O-
OH
OMe
OH
N
N
N
N
N+
iii
iv
N+
OMe
OH
OH O-
9
10
2
Scheme 1. Synthesis outline and reaction conditions: i) Pd(II)-Brettphos, KHMDS (cat.) Cs2CO3, PhMe, reflux 24 h, 79%. ii) KOH, PhMe, 80 °C, 6%. iii) BBr3, reflux, 5 h. >99% iv)
mCPBA, PhMe, 80 °C, 6 h, 76%. v) MeI, K2CO3, 18-Crown-6, DMF, rt, 24 h, 61–72%.
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
3
O-
N+
OH
N+
OH O-
2
Scheme 2. A plausible mechanism of action for the mCPBA oxidation where the
phenol hydrogen atoms of 1,6-dihydroxyphenazine (10) have a coordinating effect
as the per-acid approaches which results in the formation of an N-O bond.
MeI (3 eq)
DMF
K2CO3 (1.5 eq)
rt, 24 h
18-Crown-6 (1.5 eq)
was performed, based on the reported work of Winkler et al.31
,
O-
N+
OH
O-
OMe
yielding the self-condensed product in 79% yield on a one gram
scale (scheme 1). After screening of various Pd-catalysts (Table S-
1 in Supplementary information), the best results were obtained
using the commercially available Pd(II)-Brettphos precatalyst. Acti-
vating bases (Et3N, KOtBu, LiHMDS and KHMDS) in combination
with the precatalyst were screened, and KHMDS was found to keep
the catalyst active for the longest time, also providing instant acti-
vation upon complete base addition. On the other hand, a Wohl-
Aue reaction32,33 was undertaken condensing o-anisidine (7) and
2-nitroanisole (8) under strong alkaline conditions in toluene. Iso-
lated yields from the crude black tar disappointingly afforded 6%
yield of 9. Although low yields were as expected from litera-
ture,13,33 one can argue that gram quantities were acquired from
inexpensive starting materials using this method.
N+
N+
N+
OMe O-
3
11
OMe O-
Entry a : 72%
Entry b : 61%
Entry a : 12%
Entry b : 5%
O-
OH
OMe
N+
N
N
N+
OMe O-
13
OMe
12
Entry a : n.d.
Entry b : 18%
Entry a : 10%
Entry b : 12%
In the 2nd step of the synthesis, 9 was subjected to demethylat-
ing conditions refluxing in neat BBr3,34 which afforded 1,6-dihy-
droxyphenazine (10) quantitatively and reproducibly.
Scheme 3. Isolated compounds from two identical entries marked entry a and
entry b. Compounds 11, 12 and 13 were isolated as byproducts resulting from
dialkylation, deoxygenation or both as in 13. Table S-3 (Supplementary informa-
tion) lists reaction conditions for more entries achieving this transformation.
As the route to 10 was established, the synthesis of iodinin (2)
was solely depending on the di-N-oxidation of the PHz scaffold.
Early studies conducted by Yosioka and Kidani, 30 afforded iodinin
(2) in 21% yield where 30% H2O2 in the presence of acetic anhy-
dride was employed. This procedure did not prove to be repro-
ducible in our hands. Several oxidizing reagents were tested
where only mCPBA proved to be efficient (Table S-2 Supplementary
information provides an overview over 30 entries). mCPBA in
dichloroethane at 80 °C afforded 21% yield of iodinin (2) while
yields were dramatically increased with toluene as solvent, which
resulted in 76% yield of isolated material. In order to achieve this,
mCPBA was added portion-wise every hour during the 6 h reaction
time span. A proposed mechanism for the reaction is depicted in
Scheme 2. Purification of iodinin (2) was easily achieved using fil-
tration and washing with a series of solvents, which do not dissolve
iodinin (2) like aqueous NaHCO3, MeOH and Et2O. The robustness
of the reaction is demonstrated by the fact that it is efficient on
more than 1 g scale. As many PHz 5,10-dioxides are biologically
active and of growing interest, especially in terms of cancer and
antibacterial research, we believe this oxidation approach might
pave the way for higher synthetic versatility of compounds within
literature containing the PHz 5,10-dioxide scaffold.
Methylating conditions were explored in order to synthesize
myxin (3) from iodinin (2). The transformation was executed using
a combination of MeI, K2CO3, and 18-Crown-6-ether in DMF which
previously has been used to mono-O-methylate structurally
related phenazine 5.10-dioxides.15 This method avoids using
Me2SO4 and KOtBu in hexamethylphosphoric triamide (HMPT)
which was used by Weigele’s et al. to convert iodinin (2) cultured
in bacteria to myxin (3).7,35
The highest yields of myxin (3) were obtained where iodinin (2)
was exposed to 3 eq. of MeI in presence of 1.5 eq of K2CO3/18-
crown-6 in DMF (room temperature, 24 h). This method gave 62–
72% yield of isolated myxin (3) (see entries a and b, Scheme 3).
However, yields proved difficult to reproduce as the solubility of
myxin (3) is greater than of iodinin (2) and formation of di-O-
methylated- (11) and deoxygenated by-products 12 and 13 are
unavoidable. These compounds were isolated as well and tested
for biological activity.
OH
O- OH
N+
N
mCPBA
PhMe, 80 °C
20h
N
N
OMe
OMe
(20%)
14
12
Scheme 4. An alternative route towards myxin (2) was explored. Upon mCPBA
oxidation of 14, only formation of the mono-oxide 12 was detected by TLC analysis.
This route was therefore abandoned.
An alternative route to myxin (3) was explored in our attempt
to oxidize the myxin-scaffold 1-hydroxy-6-methoxyphenazine
(14) (Scheme 4). However, mCPBA oxidation of 14 under the same
conditions as used for the oxidation of 1,6-dihydroxyphenazine
(10) to give iodinin (2), afforded only the mono-oxidized product
1-hydroxy-6-methoxyphenazine 10-oxide (12). We propose two
main reasons for this phenomenon. First, the methoxy group in
the 6th position does not support the peracid mechanism as
depicted in Scheme 2 in the case of 10. Secondly, myxin (3) turned
out to be unstable at temperatures approaching 80 °C since the
molecule is not as stable as iodinin (2) due to absence of stabilizing
intramolecular forces between the 6th positioned phenolic proton
and the N-oxide placed at the 5th position (Fig. 1). Similar prob-
lems were also reported by Zhao and co-workers, attempting to
oxidize 14 by the aid of isolated oxidation enzymes from
bacteria.36
As a step further, iodinin (2) was alkylated using ethyl bromoac-
etate affording 1-hydroxy-6-(2-ethoxy-2-oxoethoxy)-phenazine
5,10-dioxide (15) (Scheme 5). In addition, the dialkylated bypro-
duct 16 was isolated. By similar means, analogues 17 and 18 were
afforded from reaction of iodinin with tert-butyl bromoacetate to
afford the mono- and dialkylated analogues 17 and 18. tert-Butyl
ester derivatives 17 and 18 were individually treated with aqueous
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

4
E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
Table 1
O-
OR
O-
OH
EC50 values ( SEM) of phenazine-analogues on Molm13 cells.
N+
N+
EC50 (mM)
a
N+
O-
Compound
19% O2
2% O2
N+
O
OH O-
2 (iodinin)
3 (myxin)
4 (TPZ)
9
10
11
12
13
14
15
16
17
18
19
20
2.0 0.07
1.4 0.30
95 8⁄
0.79 0.10
0.77 0.13
22 2⁄
O
EtO
15,
2
iodinin ( )
R = -H , 17%
16, R = -CH2COOEt, 14%
72 14⁄
100–120⁄
4.0 0.48
>200
84 6⁄
110–140⁄
3.1 0.83
b
O-
N+
OR
O-
OR
22
4.1 0.38
39
5
5.4 0.68
>200
0.57 0.06
2.0 0.47
2.9 0.34
4.2 0.20
N+
9
c
0.49 0.12
0.79 0.06
2.4 0.16
3.3 0.22
N+
O-
O
N+
O-
O
O
O
tBuO
17, R = -H, 43%
18, R = -CH2COOtBu, 47%
HO
54
18
7
3
61
39
4
9
19, R = -H , 92%
, R = -CH2COOH, 88%
20
The cells were left untreated or treated with various doses of iodinin (2) or its
analogues for 24 h before cell viability was assessed. The given EC50 values are
based on WST-proliferation assay except for those marked with ⁄, which are based
on microscopic evaluation of cell death. The data are based on regression analyses
of 3 to 4 experiments and are adjusted relative to untreated control cells in each
experiment.
Scheme 5. Schematic outline for the synthesis of mono- and dialkylated iodinin
esters and carboxylic acids. Reaction conditions: a) ethyl bromoacetate, K2CO3, 18-
Crown-6, DMF, rt, b) tert-Butyl bromoacetate, K2CO3, 18-Crown-6, DMF, rt, c) H3PO4
(85% aq w/w), DCM, rt.
phosphoric acid which gave the free carboxylic acids 19 and 20 in
decent yields. By comparing analogues 15–20 in biological in vitro
studies we could map the importance of this side chain with an
ethyl ester, tert-butyl ester and the free carboxylic acid function.
2.2.2. Structure-activity relationship of phenazine-analogues and their
ability to induce AML cell death
We ensured first that chemically synthesized iodinin (2)
induced AML cell apoptotic death (Fig. 3A) with similar features
as reported previously for iodinin (2) isolated from bacteria.5,20
2.2. Evaluation of iodinin analogue effects on intact leukemia cells
The EC50 value of iodinin (2) was close to 2 lM (normoxia)
whether based on counting of apoptotic cells (Fig. 3A, inset) or
spectroscopic evaluation of decline of enzyme activity by the
WST-1 method (Fig. 3A). Subsequently, all the synthesized phena-
zine-analogues were tested for ability to induce death of the
human AML cell line MOLM-13, starting with hydroxyl sub-
stituents of iodinin (2) in the 1st and 6th position and the N-oxides
in positions 5 and 10 (see Fig. 1 for molecular structure).
2.2.1. Iodinin and its analogues are more potent in 2% than 19% O2
environment
It is important to know if drug candidates are active in a
hypoxic environment, since such an environment is associated
with aggressive cancers and poor prognosis.37 Tirapazamine (TPZ,
4, Fig. 2), which is in phase III clinical trials for chemotherapy
against several solid cancers38,39 resembles iodinin (2) by having
a planar aromatic geometry and two N-oxides. TPZ (4) shows
higher activity at low O2 saturation, thus favouring action in the
tumor environment. The bone marrow tends to be hypoxic40,41
and although the role of hypoxia in leukemia is not as well under-
stood as for solid cancers, it is associated with poor outcome.42 To
know if iodinin (2) or any of its analogues also were more active at
low oxygen saturation, they were tested on MOLM-13 leukemia
cells at 2% and near atmospheric (19%) O2, and the amount of cell
death scored (see Table 1 and Fig. 3). Both iodinin (2) and several of
its analogues, like myxin (3), 11 and 14–17 had increased potency
(lowered EC50 value) at low O2 (Table 1). Remarkably, 1-hydroxy-
6-methoxyphenazine (14) and the corresponding monoxide (12)
show increased efficacy in 2% O2 compared to 19% O2 (Table 1),
despite not containing two N-oxides like myxin. However, the
potency is low compared to myxin (2) and similar to what we
observe for TPZ (4) in hypoxic environment. These compounds
might thus be attractive to investigate further as they show very
low toxicity at normoxic levels (EC50 > 200 mM for both com-
pounds). The effect of low oxygen saturation became even more
apparent for cells pre-incubated at 2% O2 for two days or longer
before being exposed to the compounds (Fig. S-1, suppl. info.).
We conclude that not only iodinin (2) itself,5,20 but also several
of its analogues have enhanced activity under hypoxia. They have
therefore promise to act more strongly on malignant cells in the
hypoxic bone marrow of AML patients than on non-malignant cells
in normoxic tissues. To our best knowledge, this is the first time
iodinin (2) is shown to be hypoxia selective.
When a methoxy group replaced the 6th positioned hydroxyl
functionality resulting in the compound myxin (3), the cytotoxic
potential was found to be unaltered compared to iodinin (2) at
2% O2 (Table 1). However, if both OH-moieties were substituted
for methoxy groups (compound 11), the potency decreased consid-
erably (Table 1). These results are of interest, as myxin (3) is
already described as an unusually potent antimicrobial agent, pos-
sessing broad-spectrum activity against Gram-positive and -nega-
tive bacteria while iodinin (2) only has weak to moderate
antimicrobial effects.7 Yet, we observe them to possess approxi-
mately equal potencies in inhibition of growth of human AML cells
(MOLM-13) at 2% O2. The far lower activity of tirapazamine (TPZ,
4), which lacks the oxygen atoms (hydroxy- or methoxy-) adjacent
to the N-oxide functionalities present in iodinin (2) and myxin (3),
suggests that an electron donating oxygen atom in adjacent posi-
tion to the N-oxide may play an important role for the cytotoxic
properties. Moreover, as long as one of the hydroxyl groups is
unsubstituted, phenol and methoxy group O-substitution seem to
result in equal activity. This indicates that one phenolic oxygen
of iodinin (2) may be functionalized without reducing the cytotoxic
activity.
Next, we wanted to investigate if the two N-oxides were essen-
tial for the cytotoxicity of compound 3 (myxin). We demonstrated
that the removal of one N-oxide (compound 12) decreased the
cytotoxicity markedly, and that removal of both N-oxides produced
a compound (14) with even less cytotoxicity (Fig 3B and Table 1).
These data underpin the importance of di-N-oxide functionality for
the activity of the PHz compound class.
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
5
Fig. 3. The cytotoxic effect of iodinin (2) and its analogues on the AML cell line MOLM-13. The cells were seeded in 96-well microplates with 20,000 cells/well and
immediately exposed to increasing concentrations of iodinin (2) or analogues under 2% or 19% oxygen (O2) for 24 h. Cell proliferation was assessed by the WST-1 assay and by
microscopy after Hoechst staining of the cell nuclei. A: Cytotoxicity of iodinin at 19% and 2% O2, measured by the proliferation assay WST-1. The inset shows cytotoxicity at
19% measured by microscopic evaluation of nuclear morphology. B: The role of the N-oxides in phenazine-induced AML cell death. The MOLM-13 cells were incubated with
increasing concentration of 1-hydroxy-6-methoxyphenazine without N-oxide (diamonds), with one (circles) or two (triangles) N-oxides for 24 h before assessment of
viability by the WST-1 assay. C: AML cell death inducing activity of iodinin analogues with different substituents on the oxygen in position 6. The data are average of 3–4
experiments and standard error. The lines are from regression analyses as described in the Section 4.
Finally, the phenolic moieties of iodinin (2) were mono- and di-
functionalized by three different substituents resulting in struc-
tures 15–20 (Scheme 5). Substitution of 2-ethoxy-2-oxoethoxy-
moiety to the 6th position (compound 15) enhanced bioactivity
substantially while the presence of this substituent in both posi-
tions 1 and 6 produced cytotoxicity similar to that of iodinin or
myxin (Table 1). A clear trend in these data is therefore that fully
oxidized di-alkylated iodinin results in reduced activity compared
to mono-alkylated derivatives (see 15 vs 16 and 3 vs. 11 in
Table 1).
as solubility and cell membrane permeability while maintaining or
even enhancing the cytotoxic effect by destabilizing and thereby
enhancing the reactivity of the N-oxide. The enhanced reactivity
predicted for compounds 15 and 3 (myxin) may be less important
in cells exposed to hypoxia, whose internal ROS level already is
high.43,44 For targeting cells in hypoxic niches with high internal
ROS production, iodinin (2) may therefore be more selective than
derivatives like myxin (3), whose inherent high reactivity may
activate them at lower ROS levels as found in normoxic niches.
A puzzling finding was that the addition of a 2-(tert-butoxy)2-
oxoethoxy (compound 17) moiety to the 6th positioned oxygen
produced a compound with less cytotoxicity than 15, myxin (3)
and iodinin (2) (see Fig 3C and Table 1). Moreover, if a carboxy-
methylenoxy moiety was added to the oxygen in the 6th position
(compound 19), the cytotoxicity was about 100-fold less than for
the 2-ethoxy-2-oxo-ethoxy substituted analogue 15 (see Fig. 3C
and Table 1). A likely explanation is that the carboxyl-group of
19, predicted to be negatively charged at physiological pH, pre-
vents it from crossing the surface membrane to enter cells. In sup-
port of this suggestion, compounds 15 and 17 that carry more
lipophilic ester side chains were found to be much more potent.
Like myxin (3) both of these analogues have greatly enhanced sol-
ubility compared to iodinin (2) in organic and aqueous media since
they lack the intramolecular ion-dipole/hydrogen forces between
the 5th positioned N-oxide and the phenol hydrogen in the 6th
position of iodinin (2).
Based on the demonstrated importance of both of the N-oxide
functionalities (Fig. 3B) and the fact that the 6th positioned
hydroxy group can be functionalized without drop in cytotoxic
activity, we propose that the ester functionalities of 15, 16, and
17 are cleaved off when the compounds have entered the cells
and become exposed to the ubiquitous cellular hydrolytic
enzymes. If the removal of the moieties happened outside the cells,
we would expect also compound 19 to be equally active. The dif-
ference in bioactivity between the carboxy-methylenoxy- and the
2-ethoxy-2-oxoethoxy-substituted analogues 19 and 15 is
explained if only the latter can cross the membrane to enter the
cell where its ester functionality is cleaved off. The superior bioac-
tivity of analogue 15 compared to iodinin (2) itself is thus most
likely a result of its higher ability to cross biological membranes,
although a destabilized N-oxide in position 5 due to the lack of
intermolecular hydrogen bond from the 6th position might also
contribute. Therefore, the 2-ethoxy-2-oxo-ethoxy side chain (com-
pounds 15, 17 and 19) chain may be used to influence factors such
3. Conclusions
In summary, the chemical part describes for the first time con-
venient synthetic methods of the literature compounds iodinin (2)
and myxin (3), along with a series of compounds 11–20 derived
from them. The key step, revealed to be problematic in previous lit-
erature, was the oxidation of 1,6-dihydroxyphenazine (10), for
which highly efficient oxidation conditions were developed. Since
numerous phenazine 5,10-dioxides potentially have exploitable
cytotoxic and/or antimicrobial actions, these methods might pave
the way for higher synthetic versatility of preparation of novel
compounds of this class. Priority was given to produce derivatives
that can reveal structure-activity relationships relevant to guide
drug development of new anti-leukemic compounds. Synthesized
iodinin (2) and analogues were evaluated in vitro for their ability
to induce death of human AML leukemia cells (MOLM-13). We
found that one, but not both of the two phenolic functions of iodi-
nin (2) can be functionalized synthetically without compromising
the cytotoxic activity. In fact, some such derivatives were more
potent than iodinin (2) itself. In addition, we show that the two
N-oxides are required for full anti-leukemic effect. The cytotoxic
effects provided by iodinin (2) and several of its derivatives
improved when cells were incubated at hypoxia rather than nor-
moxia, indicating preference for cells in hypoxic niches like the
leukemic bone marrow. To our best knowledge, iodinin (2) was
for the first time shown to be hypoxia selective. Of great impor-
tance for the purpose of drug development of this class is the find-
ing that the synthetic ester analogue 1-hydroxy-6-(2-ethoxy-2-
oxoethoxy)phenazine (15) outperformed both iodinin (2) and
myxin (3) in cytotoxicity assays, most likely as the result of greater
cell membrane penetration and more unstable N-oxide in its 5th
position. Work is in progress to establish a library of screening can-
didates for broader SAR studies, and to identify potential drug lead
candidates for more user-friendly therapeutic regimes against
acute myeloid leukemia.
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

6
E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
urated sol. of NaHCO3 (200 mL), H2O (50 mL), MeOH (200 mL) and
4. Experimental
Et2O (200 mL) or until a homogenous dark-purple color was
obtained and clear transparent solvent runs through the filter
paper. The remaining product was collected from the filter paper
and the product was dried in vacuo affording 938 mg (76%) of iodi-
nin (2) as a deep-purple solid with a coppery luster. No further
purification was necessary. Rf: 0.61 (100% DCM). 1H NMR
(600 MHz, DMSO-d6) d 14.25 (s, 2H), 7.92 (dd, J = 9.0, 1.1 Hz, 2H),
7.82 (dd, J = 9.0, 7.8 Hz, 2H), 7.19 (dd, J = 7.8, 1.1 Hz, 2H). 13C
NMR (151 MHz, DMSO-d6) d 152.5, 135.0, 133.4, 126.7, 114.0,
General information regarding the synthetic work is given
within the Supplementary information.
4.1. Synthetic methods
4.1.1. 1,6-Dimethoxyphenazine (9)
A dry round-bottom flask with a reflux condenser was charged
with 2-bromo-3-methoxyanilin (1.00 g, 4.95 mmol), KHMDS
(32 mg, 0.16 mmol, 0.03 eq), BrettPhos Pd G1 Methyl-t-Butyl Ether
adduct (118 mg, 0.15 mmol, 0.03 eq) and Cs2CO3 (3.22 g,
9.88 mmol, 2.0 eq). The system was sealed by a rubber septum
before air was removed under reduced pressure and replaced by
argon (repeated 3 times). Anhydrous toluene (20 mL) was trans-
ferred to the flask and the resulting orange colored suspension
gradually warmed up to reflux and left stirring for 24 h. The heat-
ing was removed and the crude mixture was cooled to reach rt. The
reaction mixture was filtered through a funnel of Celite and undis-
solved material washed with DCM until no yellow solution came
through the funnel. The resulting crude mixture was dry-loaded
on silica and purified by flash column chromatography on silica
gel (1:7 EtOAc/DCM as eluent) affording 468 mg (79%) of the yel-
low solid. Rf: 0.51 (100% EtOAc). 1H NMR (400 MHz, CDCl3) d
7.99 (dd, J = 8.9, 1.1 Hz, 2H), 7.74 (dd, J = 8.9, 7.6 Hz, 2H), 7.09
(dd, J = 7.6, 1.1 Hz, 2H), 4.17 (s, 6H). 13C NMR (101 MHz, CDCl3) d
155.1, 143.2, 137.0, 130.3, 122.2, 107.0, 56.6. HRMS (EI): Exact
107.4. HRMS (EI): Exact mass calculated for
C12H8N2O4:
244.0484, found 244.0487 (ꢁ1.2 ppm). 1H- and 13C NMR spectral
data are in accordance with prior literature based on isolation of
the natural product.20
4.1.4. 1-Hydroxy-6-methoxyphenazine 5,10-dioxide (myxin; 3)
A dry round-bottom flask was charged with iodinin (2) (0.23 g,
0.94 mmol, 1.0 eq), K2CO3 (195 mg, 1.41 mmol, 1.5 eq) and 18-
Crown-6 (373 mg, 1.41 mmol, 1.5 eq). The solids were dispersed
in anhydrous DMF (20 mL) under argon atm and shielded from
light. After 30 min of stirring the resulting mixture had switched
color from dark violet towards emerald green, a dropwise addition
of MeI (0.21 mL, 3.3 mmol, 3.5 eq) was carried out before the mix-
ture was left stirring for an additional period of 24 h at rt. The reac-
tion mixture was concentrated in vacuo, diluted by 200 mL of
NH4Cl (10% aqueous sol.) and the aqueous phase extracted by
EtOAc (4 ꢂ 30 mL). The pooled organic phases were dried with
MgSO4 and filtered. The resulting crude mixture was dry-loaded
on silica and further purified by flash column chromatography on
silica gel (20–80% EtOAc/Heptan) to afford 175 mg (72%) of myxin
(3) as a bright cherry-red solid. Rf: 0.59 (100% EtOAc). 1H NMR
(600 MHz, CDCl3) d 14.59 (broad s, 1H), 8.24 (dd, J = 9.1, 1.1 Hz,
1H), 8.03 (dd, J = 9.0, 1.1 Hz, 1H), 7.68 (dd, J = 9.1, 8.0 Hz, 1H),
7.63 (dd, J = 9.1, 7.9 Hz, 1H), 7.12 (dd, J = 7.9, 1.1 Hz, 1H), 7.08
(dd, J = 8.0, 1.0 Hz, 1H), 4.09 (s, 3H). 13C NMR (151 MHz, CDCl3) d
153.9, 153.8, 138.8, 136.0, 132.6, 131.8, 130.0, 126.0, 115.0,
110.7, 109.9, 109.0, 57.4. HRMS (EI): Exact mass calculated for
mass calculated for
C14H12N2O2: 240.0899, found 240.0894
(1.8 ppm). 1H- and 13C NMR data are in accordance with published
literature.13,45
4.1.2. 1,6-Dihydroxyphenazine (10)
The procedure was performed according to previously
described method by Alonso et al.34 Boron tribromide (5.0 g,
20 mmol) was transferred to a dry round-bottom flask containing
1,6-dimethoxyphenazine (9) (460 mg, 1.91 mmol) and a magnetic
bar under argon atm. The mixture was gradually warmed up
91 °C and refluxed for 5 h. The mixture was cooled on ice bath
and carefully quenched by a dropwise addition of ice water and
allowed to reach rt. The pH of the solution was adjusted to 7 by
1.0 M NaOH. The goldenrod precipitate was filtered off and washed
with cold water. The afforded crude material was dried in vacuo
yielding 425 mg of 1,6-dihydroxyphenazine (>99%). No further
purification was undertaken. Rf: 0.42 (100% EtOAc). 1H NMR
(600 MHz, DMSO-d6) d 10.49 (s, 2H), 7.79–7.70 (m, 4H), 7.19 (dd,
J = 7.2, 1.4 Hz, 2H). 13C NMR (151 MHz, DMSO-d6) d 153.4, 142.1,
135.7, 131.3, 119.2, 110.6. 1H- og 13C NMR data are in accordance
with prior literature.34
C
13H10N2O4: 258.0641, found 258.0643 (ꢁ0.8 ppm). 1H- and 13C
NMR data are in accordance with prior literature.13,46
4.1.5. 1,6-Dimethoxyphenazine 5,10-dioxide (11)
Isolated as a by-product from the same crude material afforded
after the work-up described for the synthesis of myxin (3). Flash
column chromatography on silica gel (0–3% MeOH/DCM) yielded
30 mg (12%) of the orange solid. Rf: 0.06 (100% EtOAc). 1H NMR
(600 MHz, CDCl3) d 8.31 (dd, J = 9.1, 1.1 Hz, 2H), 7.60 (dd, J = 9.1,
7.9 Hz, 2H), 7.08 (dd, J = 7.9, 0.7 Hz, 2H), 4.07 (s, 6H). 13C NMR
(151 MHz, CDCl3) d 153.8, 139.7, 130.8, 129.5, 112.3, 110.5, 57.4.
HRMS (EI): Exact mass calculated for C14H12N2O4: 272.0797, found
272.0795 (0.9 ppm). 1H- and 13C NMR data are in accordance with
prior literature.45
4.1.3. 1,6-Dihydroxyphenazine 5,10-dioxide (iodinin; 2)
A dry round bottomed flask (with a reflux condenser) was
loaded with 1,6-dihydroxyphenazine (10) (1.23 g, 3.79 mmol) at
rt under argon atm. The goldenrod solid was suspended in
100 mL of anhydrous toluene and stirred for 15 min at rt. mCPBA
(2.0 g, ꢀ77% purity; Sigma-Aldrich) was added before the mixture
was shielded from light and gradually warmed to 80 °C and added
0.80 g mCPBA in pulses every hour (i.e. repeated 4 times). After 5 h
at 80 °C, 1.0 g of mCPBA was added and the reaction mixture stir-
red for additional 60 min. The reaction mixture was cooled down
on ice bath before it was concentrated in vacuo to afford dark
and slurry crude material. The resulting crude product was dis-
persed in MeOH/Et2O (1:1) and filtered on a filter paper in rela-
tively small portions. Each portion was roughly washed by 20 mL
NaHCO3 (saturated aqueous sol.), 20 mL MeOH and 20 mL Et2O.
These portions were collected, combined and washed again by sat-
4.1.6. 1-Hydroxy-6-methoxyphenazine 10-oxide (12)
This product was isolated as a by-product from a corresponding
crude material afforded after the work-up described for the syn-
thesis of myxin (3). Flash column chromatography on silica gel
(20–60% EtOAc/Heptane) afforded an orange solid in 18% yield.
Rf: 0.65 (100% EtOAc). 1H NMR (400 MHz, CDCl3) d 13.63 (s, 1H),
8.14 (dd, J = 9.2, 1.1 Hz, 1H), 7.82 (dd, J = 8.8, 1.2 Hz, 1H), 7.70
(dd, J = 8.8, 7.7 Hz, 1H), 7.67 (dd, J = 9.7, 7.7 Hz, 1H), 7.08 (dd,
J = 7.8, 1.2 Hz, 2H), 4.16 (s, 3H). 13C NMR (101 MHz, CDCl3) d
155.9, 152.2, 145.3, 139.1, 133.6, 132.5, 131.3, 125.2, 120.2,
113.0, 109.8, 108.3, 56.9. HRMS (EI): Exact mass calculated for
C
13H10N2O3: 242.0691, found 242.0692 (ꢁ0.1 ppm). 1H- and 13C
NMR data are in accordance with prior literature.13,45
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
7
4.1.7. 1,6-Dimethoxyphenazine 5-oxide (13)
DCM).1H NMR (600 MHz, CDCl3) d 8.41 (dd, J = 9.0, 1.2 Hz, 2H),
7.60 (dd, J = 9.1, 7.8 Hz, 2H), 7.16 (dd, J = 7.9, 1.2 Hz, 2H), 4.88 (s,
4H), 4.30 (q, J = 7.1 Hz, 4H), 1.30 (t, J = 7.2 Hz, 6H).13C NMR
This product was isolated as a by-product from the same crude
material afforded after the work-up described for myxin (2). Flash
column chromatography on silica gel (0–100% EtOAc/Heptane)
afforded 25 mg (10%) of the yellow solid. Rf: 0.18 (100% EtOAc).
1H NMR (600 MHz, CDCl3) d 8.24 (d, J = 9.1 Hz, 1H), 7.92 (d,
J = 8.7 Hz, 1H), 7.67–7.56 (m, 2H), 7.07 (d, J = 7.6 Hz, 1H), 7.00 (d,
J = 7.8 Hz, 1H), 4.15 (s, 3H), 4.07 (s, 3H). 13C NMR (151 MHz, CDCl3)
d 155.9, 153.1, 146.7, 138.7, 137.9, 130.5, 129.8, 129.1, 123.3,
111.1, 108.4, 108.3, 57.1, 56.9. HRMS (EI): Exact mass calculated
for C14H12N2O3: 256.0848, found 256.0851 (ꢁ1.2 ppm).
(151 MHz, CDCl3)
d 168.19, 151.66, 139.86, 130.57, 115.66,
114.69, 77.16, 68.68, 61.75, 14.32. HRMS (TOF ES+): Exact mass
calculated for C20H20N2O8Na [M+Na]+: 439.1117, found 439.1121
(0.82 ppm).
4.1.11. 1-Hydroxy-6-(2-(tert-butoxy)-2-oxoethoxy)phenazine 5,10-
dioxide (17)
A 50 mL dry round bottom flask was charged with iodinin
(158 mg, 0.65 mmol, 1 eq), K2CO3 (0.10 g, 0.72 mmol, 1.1 eq) and
18-Crown-6 (0.19 g, 0.72 mmol, 1.1 eq). The flask was sealed by a
rubber septum, shielded from light and flushed thoroughly with
argon. Anhydrous DMF (5 mL) was added affording an emerald
green dispersion which was stirred for 10 min at rt before tert-
butyl bromoacetate (0.15 mL, 0.98 mmol, 1.5 eq) was added drop-
wise. The resulting mixture was allowed to stir for 2 h at rt before
quenched with NH4Cl (30 mL of 10% aqueous sol.). The crude mix-
ture was transferred to a separatory funnel and extracted by EtOAc
(3 ꢂ 30 mL). The organic phase was washed by brine (2 ꢂ 200 mL),
collected and consequently dry-loaded on silica. Flash column
chromatography on silica (10–25% EtOAc/Heptane) afforded
99 mg (43%) of the deep-red solid. Rf: 0.78 (100% EtOAc). 1H
NMR (300 MHz, CDCl3) d 14.54 (s, 1H), 8.30 (dd, J = 9.1, 1.1 Hz,
1H), 8.02 (dd, J = 9.0, 1.1 Hz, 1H), 7.69–7.57 (m, 2H), 7.11 (dd,
J = 7.9, 1.1 Hz, 1H), 7.06 (dd, J = 7.9, 0.9 Hz, 1H), 4.77 (s, 2H), 1.50
(s, 9H). 13C NMR (151 MHz, CDCl3) d 166.8, 153.8, 151.9, 138.9,
136.0, 132.6, 131.4, 130.4, 126.0, 115.0, 113.4, 112.4, 109.1, 83.1,
68.2, 28.2. HRMS (TOF ES+): Exact mass calculated for
4.1.8. 1-Hydroxy-6-methoxyphenazine (14)
MeI (0.11 mL, 1.82 mmol) was added dropwise to a stirring
solution of 1,6-dihydroxyphenazine (10) (350 mg, 1.65 mmol),
K2CO3 (228 mg, 1.65 mmol) and 18-Crown-6 (436 mg, 1.65 mmol)
in anhydrous DMF (22 mL) at 0 °C. Resulting mixture was left stir-
ring overnight gradually reaching rt concentrated in vacuo. The
resulting crude redissolved in 40 mL of EtOAc and the organic layer
was extracted with 10% NaOH aqueous solution (4 ꢂ 30 mL to sep-
arate 1-hydroxy-6-methoxy phenazine from the di-methylated
byproduct 9). Basic aqueous fractions were pooled and adjusted
to pH 7 by 2 M HCl. The neutral aqueous layer was thereafter
extracted by EtOAc (4x30 mL). Organic phases pooled, washed
with brine (3 ꢂ 200 mL), dried over MgSO4 and concentrated in
vacuo. The resulting crude was purified by flash column chro-
matography (10–50% EtOAc/Heptane) affording 89 mg (24%) of
yellow solid. Rf: 0.33 (1:1 EtOAc/Heptane). 1H NMR (400 MHz,
CDCl3) d 8.19 (s, 1H), 7.94 (dd, J = 8.9, 1.2 Hz, 1H), 7.82 (dd,
J = 8.9, 1.2 Hz, 1H), 7.75 (dd, J = 8.9, 7.5 Hz, 2H), 7.29–7.22 (m,
1H), 7.10 (dd, J = 7.5, 1.2 Hz, 1H), 4.19 (s, 3H). 13C NMR
(101 MHz, CDCl3) d 155.3, 151.5, 142.7, 142.0, 137.7, 134.7,
131.6, 130.7, 121.0, 120.7, 109.5, 107.0, 56.7. 1H og 13C NMR data
are in accordance with litterature.13
C18H18N2O6 Na [M+Na]+: 381.1062, found 381.1058 (ꢁ1.20 ppm).
4.1.12. 1,6-Bis(2-(tert-butoxy)-2-oxoethoxy)phenazine 5,10-dioxide
(18)
A dry roundbottom flask was loaded with iodinin (71 mg,
0.29 mmol, 1 eq), K2CO3 (60 mg, 0.44 mmol, 1.5 eq) and 18-
crown-6-ether (114 mg, 0.44 mmol, 1.5 eq). The sealed flask was
shielded from light and flushed thoroughly with argon before
anhydrous THF (4 mL) was added to disperse the ingredients. After
15 min of rotation at rt, tert-butyl bromoacetate (0.15 mL,
1.02 mmol) was added dropwise. The resulting mixture was
allowed to rotate until no starting material observed by TLC
(16 h). The reaction mixture was diluted by 15 mL of DCM and
dry loaded directly on silica. Flash column chromatography (10–
50% EtOAc in Heptane) afforded 64 mg (47%) of bright-orange-
red solid. Rf: 0.43 (100% EtOAc). 1H NMR (600 MHz, CDCl3) d 8.37
(dd, J = 9.0, 1.1 Hz, 2H), 7.56 (dd, J = 9.1, 7.8 Hz, 2H), 7.07 (dd,
J = 8.0, 1.2 Hz, 2H), 4.76 (s, 4H), 1.47 (s, 18H). 13C NMR (151 MHz,
CDCl3) d 167.0, 151.8, 139.8, 130.4, 130.0, 114.4, 114.1, 82.8,
68.4, 28.2. HRMS (TOF ES+): Exact mass calculated for C24H28N2O8-
Na [M+Na]+: 495.1743, found 495.1729 (2.90 ppm).
4.1.9. 1-Hydroxy-6-(2-ethoxy-2-oxoethoxy)phenazine 5,10-dioxide
(15)
A dry round bottom flask was charged with iodinin (180 mg,
0.73 mmol), K2CO3 (111 mg, 0.80 mmol) and 18-Crown-6
(212 mg, 0.80 mmol) and a magnetic stir bar. The system was
closed by a rubber septum, shielded from light and flushed thor-
oughly with argon. THF (10 mL) was added to disperse ingredients
and the resulting mixture was allowed to rotate for additional
30 min before ethyl bromoacetate (0.07 mL, 0.63 mmol) was added
in a drop-wise manner. The reaction was left rotating at rt for 2 h.
The reaction mixture was quenched by 50 mL of NH4Cl (10% aque-
ous sol.) before the aqueous phase was extracted with EtOAc
(4 ꢂ 30 mL). Combined organic phases were dried over MgSO4, fil-
tered and the resulting solution absorbed onto silica gel. Flash col-
umn chromatography on silica (30–100% EtOAc/Heptane) gave
45 mg (17%) of the cherry-red solid. 1H NMR (600 MHz, CDCl3) d
14.52 (s, 1H), 8.34 (dd, J = 9.1, 1.1 Hz, 1H), 8.03 (dd, J = 9.1,
1.1 Hz, 1H), 7.64 (dt, J = 9.0, 7.8 Hz, 2H), 7.13 (dd, J = 7.8, 1.1 Hz,
2H), 4.88 (s, 2H), 4.31 (q, J = 7.1 Hz, 2H), 1.32 (t, J = 7.1 Hz, 3H)
13C NMR (151 MHz, CDCl3) d 167.94, 153.89, 151.67, 138.95,
135.97, 132.70, 131.35, 130.64, 126.03, 115.09, 114.58,
113.05,109.02, 68.38, 61.87, 14.34. HRMS (TOF ES+): Exact mass
calculated for C16H14N2O6Na [M+Na]+: 353.0749, found 353.0745
(ꢁ1.29 ppm).
4.1.13. 1-(Carboxymethoxy)-6-hydroxyphenazine 5,10-dioxide (19)
1-(2-(tert-Butoxy)-2-oxoethoxy)-6-hydroxyphenazine
5,10-
dioxide (18) (87 mg, 0.08 mmol) was dissolved in 5 mL DCM. The
red solution was cooled on ice bath and 2 mL of H3PO4 (85% v/w
aqueous sol.) added dropwise. The resulting dispersion was left
stirring overnight at reaching rt gradually. The crude mixture
was neutralized and pH adjusted to 8 by sat. NaHCO3 (aqueous
sol.). The red aqueous phase was washed with DCM (3 ꢂ 20 mL,
or until no color was observed within the organic phase). The
resulting aqueous phase was collected and pH adjusted to 1 by
HCl (37% w/w aqueous sol.) The red precipitate was filtered,
washed with H2O (50 mL), MeOH (20 mL) and Et2O (40 mL). The
filtered product was collected affording 67 mg (92%) of the deep-
red solid. No further purification was necessary. 1H NMR
4.1.10. 1,6-Bis(2-ethoxy-2-oxoethoxy)phenazine 5,10-dioxide (16)
This product was isolated as a by-product from the same crude
material afforded after the work-up described for compound 15.
Flash column chromatography, first 30–100% EtOAc/heptane to
elute out compound 15, then switching to 1:7 EtOAc/DCM)
afforded 43 mg (14%) of an orange solid. Rf: 0.23 (3% MeOH/
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

8
E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
(400 MHz, DMSO-d6) d 14.93 (s, 1H), 13.17 (s, 1H), 8.16 (d,
J = 8.9 Hz, 1H), 7.89–7.75 (m, 2H), 7.71 (dd, J = 9.0, 7.8 Hz, 1H),
7.27 (dd, J = 8.0, 1.1 Hz, 1H), 7.12 (dd, J = 7.9, 1.1 Hz, 1H), 4.92 (s,
2H). 13C NMR (101 MHz, DMSO-d6) d 169.3, 153.1, 151.3, 138.5,
135.5, 132.1, 131.9, 130.0, 125.5, 113.9, 113.3, 111.0, 108.2, 66.9.
HRMS (TOF ES+): Exact mass calculated for C14H9N2O6 [MꢁH]ꢁ:
301.0460, found 301.0465 (1.45 ppm).
Acknowledgements
The work with the project and this paper was supported by
NOVO Exploratory Pre Seed Grant no. 4462, the University of Oslo
Innovation Fund 2015, the Norwegian Cancer Society, and the
Western Norway Regional Health Authorities. The School of Phar-
macy, University of Oslo is greatly acknowledged for Ph.D.-scholar-
ships to EÖV and OAHÅ. We thank Dr. Geir Kildahl-Andersen for
valuable assistance on NMR-analysis, and Nina Lied Larsen and
Kirsten Brønstad for assistance on cell culture. Our NMR-labs at
department of Chemistry, University of Oslo are acknowledged as
well for excellent service.
4.1.14. 1,6-Bis(carboxymethoxy)phenazine 5,10-dioxide (20)
1,6-Bis(2-(tert-butoxy)-2-oxoethoxy)phenazine
5,10-dioxide
(113 mg, 0.24 mmol) was dissolved in DCM (4 mL) and pipetted
dropwise to stirring aqueous solution of H3PO4 (3 mL; ꢃ85 wt.%
in H2O). The resulting mixture was stirred for 5 h at room temper-
ature before the mixture was diluted by 50 mL of sat. NaHCO3
aqueous solution and placed in a separatory funnel. The aqueous
phase (orange colored) was washed with 2 ꢂ 20 mL of DCM in
order to wash out unreacted starting material. The pH of the aque-
ous solution was adjusted by dropwise addition of HCl (37% w/v)
and the precipitated product filtered and dried affording 76 mg
(88%) of the wanted orange solid. No further purification was nec-
essary. 1H NMR (600 MHz, DMSO- d6) d 13.17 (s, 2H), 8.17 (dd,
J = 9.1, 1.1 Hz, 2H), 7.72 (dd, J = 9.0, 7.9 Hz, 2H), 7.30 (dd, J = 8.0,
1.2 Hz, 2H), 4.91 (s, 4H).13C NMR (151 MHz, DMSO- d6) d 169.6,
151.4, 139.1, 130.7, 129.3, 114.7, 112.7, 67.6. HRMS (TOF ES+):
Exact mass calculated for C16H10N2O8Na [M-2H+Na]-: 381.0334,
found 381.0339 (1.08 ppm).
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2017.02.058.
References
1. Laursen JB, Nielsen J. Phenazine natural products: biosynthesis, synthetic
analogues, and biological activity. Chem Rev. 2004;104:1663–1686.
2. Mavrodi DV, Blankenfeldt W, Thomashow LS. Phenazine compounds in
fluorescent Pseudomonas Spp. biosynthesis and regulation. Annu Rev
Phytopathol. 2006;44:417–445.
3. Guttenberger N, Blankenfeldt W, Breinbauer R. Recent developments in the
isolation, biological function, biosynthesis, and synthesis of phenazine natural
products. Bioorg Med Chem. 2017.
4. Turner JM, Messenger AJ. Occurrence, biochemistry and physiology of
phenazine pigment production. Adv Microb Physiol. 1986;27:211–275.
5. Myhren LE, Nygaard G, Gausdal G, et al. Iodinin (1,6-dihydroxyphenazine 5,10-
dioxide) from Streptosporangium sp. induces apoptosis selectively in myeloid
leukemia cell lines and patient cells. Marine Drugs. 2013;11:332–349.
6. Hollstein U, Van Gemert RJ. Interaction of phenazines with
polydeoxyribonucleotides. Biochemistry. 1971;10:497–504.
7. Weigele M, Maestrone G, Mitrovic M, Leimgruber W. Antimicrobial agents
structurally related to myxin. Antimicrob Agents Chemother. 1970;10:46–49.
8. Behki RM, Lesley SM. Deoxyribonucleic acid degradation and the lethal effect
by myxin in Escherichia coli. J Bacteriol. 1972;109:250–261.
9. Kondratyuk TP, Park E-J, Yu R, et al. Novel marine phenazines as potential
cancer chemopreventive and anti-inflammatory agents. Marine Drugs.
2012;10:451.
10. Zhao J, Wu Y, Alfred AT, Wei P, Yang S. Anticancer effects of pyocyanin on
HepG2 human hepatoma cells. Lett Appl Microbiol. 2014;58:541–548.
11. Managò A, Becker KA, Carpinteiro A, et al. Pseudomonas aeruginosa pyocyanin
induces neutrophil death via mitochondrial reactive oxygen species and
4.2. Cell Maintenance and experimental conditions
The MOLM-13 cells47 were cultured in RPMI medium (R5886)
enriched with 10% fetal bovine serum (F7524), 0.2 mM L-glutamine
(G7513) and additionally supplemented with 50 IU/mL penicillin
and 0.1 mg/mL streptomycin (P0781) (all from Sigma-Aldrich,
St. Louis, MO, USA). The cells were grown at a density between
75 and 800 ꢂ 103 cl/mL at 37 °C in a humidified atmosphere.
All analogues were dissolved in DMSO at concentrations
between 2.5 and 20 mM before testing. For cytotoxic testing, the
cells were seeded in 96-well Microplates (#167008, Thermo Scien-
tificTM NuncTM MicroWellTM) with 20,000 cells/well in 100
lL med-
mitochondrial
acid
sphingomyelinase.
Antioxid
Redox
Signal.
ium. The cells were exposed to various concentrations of iodinin
or its analogues for 24 h before assessment of viability by the cell
proliferation reagent WST-1, according to the manufacturer’s
instructions (Roche Diagnostics). The cells were next fixed in 2%
buffered formaldehyde (pH 7.4) with the DNA-specific dye Hoechst
33342 (Polysciences Inc.) and scored for percent of dead cells by
UV-microscopy based on nuclear morphology.48 The highest con-
centration of analogue corresponded to 1% of DMSO, which alone
gave less than 7% cell death as judged by nuclear morphology.
EC50 values were determined by analyses of WST-1 signal
results as well as microscopic evaluation of cell death by four-
parameter regression analyses using the SigmaPlot software (Sys-
tat Software Inc. San Jose, CA):
2015;22:1097–1110.
12. Fraser RS, Creanor J. The mechanism of inhibition of ribonucleic acid synthesis
by 8-hydroxyquinoline and the antibiotic lomofungin. Biochem J.
1975;147:401–410.
13. Chowdhury G, Sarkar U, Pullen S, et al. DNA strand cleavage by the phenazine
di-N-oxide natural product myxin under both aerobic and anaerobic
conditions. Chem Res Toxicol. 2012;25:197–206.
14. Lavaggi ML, Cabrera M, Pintos C, et al. Novel phenazine 5,10-dioxides release
(ꢄ)OH in simulated hypoxia and induce reduction of tumour volume in vivo.
ISRN Pharmacol. 2011;2011:314209.
15. Lavaggi ML, Nieves M, Cabrera M, et al. Structural modifications on the
phenazine N, N0-dioxide-scaffold looking for new selective hypoxic cytotoxins.
Eur J Med Chem. 2010;45:5362–5369.
16. Davis JG. Chromobacterium iodinum (n. sp.), Zentralblatt fuer Bakteriologie,
Parasitenkunde,
Naturwissenschaftliche: Allgemeine, Lan. (1939); 100: 273–276.
17. Collins MD, Jones D, Keddie RM, Sneath PHA. Reclassification of
Chromobacterium iodinum (Davis) in Redefined Genus Brevibacterium
Infektionskrankheiten
und
Hygiene,
Abteilung
2,
a
(Breed) as Brevibacterium iodinum nom.rev.; comb.nov. Microbiology.
1980;120:1–10.
ðmax ꢁ minÞ
Y ¼ min þ
À
Á
h
X
EC50
18. McIlwain H. The anti-streptococcal action of iodinin. Naphthaquinones and
anthraquinones as its main natural antagonists. Biochem J. 1943;37:265–271.
19. Oftedal L, Selheim F, Wahlsten M, Sivonen K, Doskeland SO, Herfindal L. Marine
benthic cyanobacteria contain apoptosis-inducing activity synergizing with
daunorubicin to kill leukemia cells, but not cardiomyocytes. Marine Drugs.
2010;8:2659–2672.
20. Sletta H, Degnes KF, Herfindal L, et al. Anti-microbial and cytotoxic 1,6-
dihydroxyphenazine-5,10-dioxide (iodinin) produced by Streptosporangium
sp. DSM 45942 isolated from the fjord sediment. Appl Microbiol Biotechnol.
2014;98:603–610.
1 þ
where Y is the response (WST-1 signal or percent apoptosis), min
and max are minimum and maximum response, X is concentration
of analogue, EC50 equals the point of inflection, i.e. the point that
gives half of maximum response, and h is the Hill’s slope of the
curve. The two methods gave consistent response curves for each
analogue. In cases where the analogues’ color interfered with the
WST-1 signal, cell counting was used.
21. Hollstein U, Butler PL. Inhibition of ribonucleic acid synthesis by myxin.
Biochemistry. 1972;11:1345–1350.
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058

E.Ö. Viktorsson et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
9
22. von Pawel J, von Roemeling R, Gatzemeier U, et al. Tirapazamine plus cisplatin
versus cisplatin in advanced non–small-cell lung cancer: report of the
37. Menon C, Fraker DL. Tumor oxygenation status as a prognostic marker. Cancer
Lett. 2005;221:225–235.
a
international CATAPULT I study group. J Clin Oncol. 2000;18:1351–1359.
23. Cimmino A, Evidente A, Mathieu V, et al. Phenazines and cancer. Nat Product
Rep. 2012;29:487–501.
24. Brown JM, Wilson WR. Exploiting tumour hypoxia in cancer treatment. Nat Rev
Cancer. 2004;4:437–447.
25. Mimeault M, Batra SK. Frequent gene products and molecular pathways altered
in prostate cancer- and metastasis-initiating cells and their progenies and
novel promising multitargeted therapies. Mol Med (Manhasset, NY, United
States). 2011;17:949–964.
26. Wilson WR, Hay MP. Targeting hypoxia in cancer therapy. Nat Rev Cancer.
2011;11:393–410.
38. DiSilvestro PA, Ali S, Craighead PS, et al. Phase III randomized trial of weekly
cisplatin and irradiation versus cisplatin and tirapazamine and irradiation in
stages IB2, IIA, IIB, IIIB, and IVA cervical carcinoma limited to the pelvis: a
Gynecologic Oncology Group study. J Clin Oncol. 2014;32:458–464.
39. Cohen EEW, Rosine D, Haraf DJ, et al. Phase I trial of tirapazamine, cisplatin, and
concurrent accelerated boost reirradiation in patients with recurrent head and
neck cancer. Int J Radiat Oncol Biol Phys. 2007;67:678–684.
40. Chiarini F, Lonetti A, Evangelisti C, et al. Advances in understanding the acute
lymphoblastic leukemia bone marrow microenvironment: from biology to
therapeutic targeting. Biochim Biophys Acta. 2016;1863:449–463.
41. Spencer JA, Ferraro F, Roussakis E, et al. Direct measurement of local oxygen
concentration in the bone marrow of live animals. Nature (London, UK).
2014;508:269–273.
27. Deschler B, Lübbert M. Acute myeloid leukemia: epidemiology and etiology.
Cancer. 2006;107:2099–2107.
28. Lichtman MA. A historical perspective on the development of the cytarabine (7
days) and daunorubicin (3 days) treatment regimen for acute myelogenous
42. Petit C, Gouel F, Dubus I, Heuclin C, Roget K, Vannier JP. Hypoxia promotes
chemoresistance in acute lymphoblastic leukemia cell lines by modulating
death signaling pathways. BMC Cancer. 2016;16:746.
43. Martinez-Sanchez G, Giuliani A. Cellular redox status regulates hypoxia
inducible factor-1 activity. Role in tumour development. J Exp Clin Cancer Res:
CR. 2007;26:39–50.
44. Wu W, Platoshyn O, Firth Amy L, Yuan Jason XJ. Hypoxia divergently regulates
production of reactive oxygen species in human pulmonary and coronary
artery smooth muscle cells. Am J Physiol Lung Cell Mol Physiol. 2007;293:
L952–L959.
45. Weigele M, Leimgruber W. The structure of myxin. Tetrahedron Lett.
1967;8:715–718.
leukemia. The 40th anniversary of
(2013):119–130.
7 + 3. Blood Cells Mol Dis. 2013;50
29. Higby DJ, Wallace Jr HJ, Holland JF. Cis-diamminedichloroplatinum (NSC-
119875): a phase I study. Cancer Chemother Rep. 1973;57:459–463.
30. Yosioka I, Kidani Y. Studies on phenazines. II Synthesis of iodinin. Yakugaku
Zasshi. 1952;72:1301–1303.
31. Winkler JD, Twenter BM, Gendrineau T. Synthesis of substituted phenazines via
palladium-catalyzed aryl ligation. Heterocycles. 2012;84:1345–1353.
32. Wohl A, Aue W. Ueber die Einwirkung von Nitrobenzol auf Anilin bei
Gegenwart von Alkali. Ber Dtsch Chem Ges. 1901;34:2442–2450.
33. Pachter IJ, Kloetzel MC. The Wohl-Aue reaction. I. Structure of benzo [a]
phenazine oxides and syntheses of 1,6-dimethoxyphenazine and 1,6-
dichlorophenazine1. J Am Chem Soc. 1951;73:4958–4961.
34. Mateo Alonso A, Horcajada R, Groombridge HJ, et al. Synthesis of phenazine
derivatives for use as precursors to electrochemically generated bases. Org
Biomol Chem. 2005;3:2832–2841.
35. Leimgruber WM. Alkylation of iodinin. (Nutley, NJ), United States: Hoffmann-La
Roche Inc.; 1975.
46. Edwards OE, Gillespie DC. N-dioxides, a new class of amine oxides. Tetrahedron
Lett. 1966;7:4867–4870.
47. Matsuo Y, Macleod RAF, Uphoff CC, et al. Two acute monocytic leukemia (AML-
M5a) cell lines (MOLM-13 and MOLM-14) with interclonal phenotypic
heterogeneity showing MLL-AF9 fusion resulting from an occult chromosome
insertion, ins(11;9)(q23;p22p23. Leukemia. 1997;11:1469–1477.
48. Myhren L, Nilssen IM, Nicolas V, Doeskeland SO, Barratt G, Herfindal L. Efficacy
of multi-functional liposomes containing daunorubicin and emetine for
treatment of acute myeloid leukemia. Eur J Pharm Biopharm. 2014;88:186–193.
36. Zhao Y, Qian G, Ye Y, et al. Heterocyclic aromatic N-oxidation in the
biosynthesis of phenazine antibiotics from Lysobacter antibioticus. Org Lett.
2016;18:2495–2498.
Please cite this article in press as: Viktorsson E.Ö., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.058