Semisynthesis and antiplasmodial activity of the quinoline alkaloid aurachin E

Article abstract of DOI:10.1021/np8004612

A one-step synthesis of the rare aurachin E (1) from the easily accessible aurachin C (2) and cyanogen bromide is described. 3-Bromocarbamoylquinoline (5) is formed in a side reaction with concomitant loss of the 3-farnesyl residue. In an alternative approach, aurachin D (3) was reacted with phosgene and sodium azide to form the imidazolone ring of 1 via n-acylation. Unexpectedly, the initial reaction occurred at the carbonyl group of 3 to give 1H-pyrrolo[3,2-c]quinoline 4. The reaction sequence represents a novel route to this type of compound. Aurachin E, contrary to other aurachins, combines a high in vitro antiplasmodial activity with low cytotoxicity and absence of mitochondrial respiratory inhibition.

Full text of DOI:10.1021/np8004612

J. Nat. Prod. 2008, 71, 1967–1969
1967
Semisynthesis and Antiplasmodial Activity of the Quinoline Alkaloid Aurachin E
Gerhard Ho¨fle,* Bettina Bo¨hlendorf,† Thomas Fecker, Florenz Sasse, and Brigitte Kunze
Helmholtz Centre for Infection Research (preViously GBF, Gesellschaft fu¨r Biotechnologische Forschung),
Inhoffenstrasse 7, 38124 Braunschweig, Germany
ReceiVed July 23, 2008
A one-step synthesis of the rare aurachin E (1) from the easily accessible aurachin C (2) and cyanogen bromide is
described. 3-Bromocarbamoylquinoline (5) is formed in a side reaction with concomitant loss of the 3-farnesyl residue.
In an alternative approach, aurachin D (3) was reacted with phosgene and sodium azide to form the imidazolone ring
of 1 via N-acylation. Unexpectedly, the initial reaction occurred at the carbonyl group of 3 to give 1H-pyrrolo[3,2-
c]quinoline 4. The reaction sequence represents a novel route to this type of compound. Aurachin E, contrary to other
aurachins, combines a high in Vitro antiplasmodial activity with low cytotoxicity and absence of mitochondrial respiratory
inhibition.
The aurachins form a large family of isoprenoid quinoline
alkaloids isolated from the myxobacteria Stigmatella aurantiaca1
and Stigmatella erecta.2 Their biosynthesis has been investigated
recently by a feeding study1 and by cloning part of the biosynthesis
gene cluster.3 Whereas anthranilic acid and acetate were directly
identified as precursors of the quinoline nucleus and the side chain
of the predominant aurachins A-D, the formation of the minor
aurachins followed directly from biosynthetic considerations.1 Only
aurachin E (1), with its extra imidazolone ring, did not fit in this
scheme.
Scheme 1. Proposed Reactive Intermediates I and II in the
Formation of Aurachin E (1) from Aurachin C (2) and Aurachin
D (3), Respectively (Farn ) farnesyl)
different, and HRMS indicated the unexpected elemental composi-
tion C25H32N2. Thus, in a formal sense, the oxygen of 3 (C25H33NO)
had been replaced by nitrogen, and another double-bond equivalent
introduced. As the NMR spectra showed only modifications at C-1′
to C-3′ of the farnesyl side chain, the azidocarbonyl group must
have been introduced in the C-4 position, followed by insertion of
the intermediate nitrene in the adjacent double bond to give an
intermediate N-acylaziridine.5 Rearrangement and extrusion of CO2
eventually leads to pyrroloquinoline 4 (Scheme 2). Characteristic
signals in the 1H NMR spectrum of 4 are a sharp singlet at δ 6.47
for H-1′, a sextet at δ 3.04 for H-3′, and a doublet at δ 1.40 for
H3-13′. Moreover, all 1H and 13C NMR signals of the heterocyclic
moiety are nearly identical with those reported for 2-butyl-4-methyl-
1H-pyrrolo[3,2-c]quinoline.6 Preliminary work indicated that this
novel synthesis of pyrroloquinolines has broad applicability.7 At
this point, we became aware of a recent report on another case of
exclusive O-acylation of 2-alkylquinolones with ethyl chlorofor-
mate.8 Thus, depending on structural restraints and reaction
conditions either N-4 or O-acylation8 as in our case is observed
with 4-quinolones. The presence of pyridine or DMAP had no
influence on the outcome of the reaction, and no trace of aurachin
E (1) was detected.
Following an alternative approach via intermediate I, aurachin
C (2) was treated with cyanogen bromide in CH2Cl2. A rapid
reaction occurred, producing a complex mixture that on TLC
showed the conspicuous fluorescent spot for 1. After removal of
some insoluble material by filtration, chromatographic separation
on silica gel gave 1 in 14% yield, and in a later fraction 5% of 3.
The insoluble material (5) showed an IR band at 1812 cm-1 and
liberated HBr when dissolved in methanol. After partition between
ethyl acetate and buffer a neutral compound was obtained, which
from its IR band at 1767 cm-1, elemental composition C11H8N2O2,
and NMR spectra was identified as the cyclic carbamate 7 (δCdO
153.1). From these results the original material must have been
carbamoyl bromide, which on dissolution in DMSO or protic
solvents rapidly cyclized to the hydrobromide of carbamate 7. A
On the basis of the fact that feeding of anthranilic acid increased
the production of aurachin C (2) up to 20-fold,1b 2,3-diaminobenzoic
acid, the presumed precursor of 1, was fed to S. aurantiaca. As
this had no influence on the production of 1, it was hypothesized
that it might be derived from aurachin C (2) or D (3) by an unknown
biosynthetic mechanism. Alternatively, it could not be excluded
that 1 was an artifact formed during fermentation or isolation of
the aurachins, e.g., via reactive intermediates such as I or II
(Scheme 1). Particularly nitrene II was suspected as an intermediate,
being formed from aurachin D (3), phosgene, and sodium azide.
Both compounds may have been present during downstream
processing of the fermenter: phosgene in the chloroform used for
extraction and sodium azide used for equipment sterilization. The
fact that 4-quinolones are attacked at the nitrogen by methyl and
benzyl chloroformate4 suggested this was a possibility. Indeed, 3
reacted smoothly with phosgene followed by sodium azide to give
an 85% yield of a new compound, 4, showing the bright blue
fluorescence of 1. However, its chromatographic behavior was
* To whom correspondence should be addressed. Tel: +49-531-
61811040. Fax: +49-531-61812697. E-mail: gho@helmholtz-hzi.de.
† Current address: DSM Nutritional Products Ltd., P.O. Box 3255, Bldg.
203, CH-4002 Basel, Switzerland.
10.1021/np8004612 CCC: $40.75
2008 American Chemical Society and American Society of Pharmacognosy
Published on Web 10/16/2008

1968 Journal of Natural Products, 2008, Vol. 71, No. 11
Notes
Scheme 2. Reaction of Aurachin D (3) with Phosgene and Sodium Azide (Farn ) farnesyl)
fourth compound observed in significant amounts in the proton
NMR spectrum of the reaction mixture was identified as farnesyl
bromide (6). The characteristic doublet at δ 3.95 and triplet at δ
5.50 for 1-H2 and 2-H were in good agreement with the expected
values,15 and in the HPLC/MS a peak appearing at the end of the
gradient showed negative molecular ions of 6 at m/z 281/283.
Table 1. Inhibition of Mitochondrial Respiration, Cytotoxicity,
and in Vitro Antiplasmodial Activity of Aurachins A-E,
Chloroquine, and Artemisinine
antiplasmodial activityc
inhibition of
IC50 (ng/mL)
respirationa
cytotoxicityb
aurachin
IC50 (µg/mL) IC50 (µg/mL) W-2 clone D-6 clone
A
B
C (2)
D (3)
E (1)
28d
1.4
3.2
1.8
1.3
25
14
>40
>12 000
36
26
370
13
35
>12 000
78
0.90
250
0.4
Scheme 3. Reaction of Aurachin C (2) with Cyanogen Bromide
14d
16d
18d
1100e
>8100e
>8100e
chloroquine
artemisinine
1.2
1.1
0.43
a NADH oxidation was tested in submitochondrial particles of bovine
heart as described previously.1,15 b Serial dilutions of the compounds
were incubated with L929 mouse fibroblasts in 96-well plates. Metabolic
activities were measured after 5 days using a MTT assay.16 c The
activity was tested against Plasmodium falciparum Indochina W-2 and
Sierra Leone D-6 clones according to
a
procedure described
previously.14 d Values were taken from ref 1. e Protein concentration was
46 µg/mL, and the rate of NADH oxidation in the control was 1.6
µmol·mg-1 ·min-1
.
investigations into the mechanism of action and SAR studies are
needed to ascertain whether 1 is a useful lead for the development
of an antimalarial drug.
Experimental Section
Mechanistically these four compounds are derived from an
intermediate radical pair III, which recombines through either C-8
or C-3 (Scheme 3). Whereas the 8-isocyanato intermediate cyclizes
to give aurachin E (1), the corresponding C-3 isomer loses a farnesyl
cation followed by addition of hydrogen bromide to give carbamoyl
bromide 5.8 The latter on aqueous workup affords the cyclic
carbamate 7. The farnesyl cation reacts with bromide to give 6,
and in a side reaction, the quinolone radical in III abstracts a
hydrogen atom from the solvent to give aurachin D (3).
General Experimental Procedures. UV spectra were recorded on
a Shimadzu UV-2102 PC scanning spectrometer. IR spectra were
measured on a Nicolet 20DXB FT-IR spectrometer. NMR spectra were
recorded on Bruker ARX-400 and DMX-600 NMR spectrometers. EI
and DCI mass spectra (reactant gas NH3) were obtained on a Finnigan
MAT 95 spectrometer, with high-resolution data acquired using peak
matching (M/DM ) 10000). HPLC/DAD/ESIMS: PE Sciex Api-2000.
LC/MS; Nucleodur C18 column, 5 µm, 2 × 125 mm (Machery-Nagel)
(0.3 mL/min CH3CN/5 mM NH4OAc buffer pH 5.5, isocratic 5:95 for
5 min, then gradient to 95:5 in 30 min). Analytical TLC: aluminum
sheets silica gel Si 60 F254; Merck, detection UV absorption at 254
nm. Cyanogen bromide was supplied by EGA Chemie, Steinheim,
Germany.
4-Methyl-2-[(E)-1,5,9-trimethyldeca-4,8-dienyl]-1H-pyrrolo[3,2-
c]quinoline (4). Aurachin D (3) (110 mg, 0.3 mmol) was dissolved in
a solution of phosgene in toluene (1 mL, 20%, 2 mmol) and kept at
room temperature for 24 h. Excess phosgene and solvent were
evaporated in a stream of N2, the residue was dissolved in acetone (2
mL), NaN3 (130 mg, 2 mmol) was added, and the suspension was stirred
at 80 °C for 24 h. The acetone was evaporated and the residue extracted
with CH2Cl2. TLC of the extract indicated the presence of some educt
3 and a single more polar product. The reaction mixture was separated
by flash chromatography on silica gel (3.3 g) with a CH2Cl2/MeOH
gradient of 98:2 to 50:50. In addition to 3 (20 mg, 18%), 4 was obtained
as a tan-colored oil (77 mg, 71%, based on recovered starting material
85%; TLC Rf ) 0.33, CH2Cl2/MeOH, 9:1).
Compound 4: colorless solid; TLC, red-violet spot on staining with
vanillin/H2SO4 followed by heating to 120 °C; UV (MeOH) λmax (log
ε) 210 (4.50), 250 (4.60), 256 (4.64), 284 (3.83), 328 nm (3.41); IR
(KBr) νmax 3427, 2963, 2924, 2853, 1594, 1569, 1515 cm-1; 1H NMR
(CDCl3, 300 MHz) δ 9.90 (1H, br s, NH), 8.10 (1H, J ) 7.5 Hz, d, J
) 7.5 Hz, H-8), 8.02 (1H, d, J ) 6.5 Hz, H-5), 7.49 (1H, t, J ) 7.5
Hz, H-7), 7.40 (1H, t, J ) 7.5 Hz, H-6), 6.47 (H, s, H-1′), 5.09 (H, m,
H-6′, H-10′), 3.04 (1H, tq, J ) 7.5 Hz, H-3′), 2.88 (3H, s, H3-9),
1.90-2.10 (6H, m, H2-5′, H2-8′, H2-9′), 1.60-1.90 (2H, m, H2-4′), 1.67
Even though, the yield of 1 is low, gram amounts can now be
produced in a single step from the easily accessible aurachin C
(2)1a and introduced in biological tests. From early on antimalaria
activity has been suspected for the aurachins.11 A first in Vitro
screening against Plasmodium falciparum provided by the WHO
(Geneva) showed good activity for aurachins C and E comparable
to that of chloroquine and artemisinine (Table 1). Remarkably, the
cytotoxicity of aurachin E (1) was also as low, and, in contrast to
aurachins A-D, no inhibition of mitochondrial respiration nor
antibacterial activity was observed. While the antiplasmodial activity
of aurachins B-D may be ascribed to the inhibition of respiration
of the malaria parasites, another mechanism must be operative with
1. For unknown reasons, no in ViVo activity was observed in a
murine malaria model with Plasmodium berghei at 100 mg/kg,
whereas chloroquine showed an ED90 of 2.8 mg/kg.14 Nevertheless,

Notes
Journal of Natural Products, 2008, Vol. 71, No. 11 1969
(3H, br s, H3-12′), 1.60 (3H, br s, H3-15′), 1.52 (3H, br s, H3-14′), 1.40
(3H, d, J ) 7.5 Hz, H3-13′); 13C NMR/HMBC correlation (CDCl3, 75
MHz) δ 153.8 (C-2/H-9, H-1′), 144.2 (C-2′/H-1′, H-3′, H-4′, H3-13′),
143.0 (C-8a/H-7, H-8), 135.8 (C-7′/H2-5′, H2-8′, H3-14′), 134.4 (C-4/
H-5, H-1′), 131.5 (C-11′/H2-9′, H3-12′, H3-15′), 128.5 (C-8/H-6), 126.3
(C-7/H-5), 124.9 (C-6/H-8), 124.3 (C-10′/H2-8′, H2-9′, H3-12′, H3-15′),
123.9 (C-6′/H2-4′, H2-5′, H2-8′, H3-14′), 120.9 (C-3/H-1′, H3-9), 119.7
(C-5/H-7), 116.9 (C-4a/H-5, H-6, H-8), 99.0 (C-1′/H-3′), 39.7 (C-8′/
H-6′, H-9′, H-10′, H3-14′), 37.4 (C-4′/H-3′, H2-5′, H3-13′), 32.8 (C-3′/
H-1′, H-4′, H-5′, H3-13′), 26.6 (C-5′/H-3′, H-4′, H-6′), 25.7 (C-9′, C-12′/
H2-8′, H-10′, H3-12′, H3-15′), 22.3 (C-9), 20.9 (C-13′/H-3′, H2-4′), 17.8
(C-15′/H3-12′), 16.1 (C-14′/H-6′, H2-8′); HRESI/MS [M + H]+ m/z
361.2604 (calcd for C25H33N2, 361.2644).
Hydrobromide of 7: 1H NMR (DMSO-d6 solution of 5, 400 MHz)
δ 8.25; 8.22 (2H, d, J ) 7 Hz, H-5, H-8), 7.99; 7.87 (2H, t, J ) 7 Hz,
H-6, H-7), 2.48 (3H, s, H3-9); 13C NMR (DMSO-d6 solution of 5, 100
MHz) δ 153.1 (C-10), 148.4 (C-4), 142.4 (C-8a), 137.4 (C-2), 132.1
(C-8), 129.0 (C-7), 124.6 (C-3), 121.5 (C-6), 120.7 (C-5), 112.8 (C-
4a), 17.4 (C-9).
Acknowledgment. We thank S. Reinecke, K. Schober, and A. Ritter
for technical assistance, and C. Kakoschke, B. Jaschok-Kentner, R.
Christ, and U. Felgentra¨ger for recording NMR and mass spectra. We
further thank Dr. V. Wray and Dr. M. Nimtz for helpful discussions,
and Dr. P. I. Trigg, WHO (Geneva) for providing the antimalaria tests.
References and Notes
Reaction of 2 with Cyanogen Bromide. A solution of BrCN (0.6
g, 5.7 mmol) was added quickly to a solution of 2 (2.0 g, 5.3 mmol)
in 30 mL of CH2Cl2 with rapid stirring. The color turned red-brown
and faded within a few minutes, and a yellow precipitate formed slowly.
After 30 min the suspension was filtered and the crystals were washed
with CH2Cl2 to give 390 mg (37%) of carbamoylbromide 5. The
combined solutions were separated by flash chromatography on silica
gel. Elution with a gradient of CH2Cl2/MeOH from 100:0 to 95:5 gave
1 (310 mg, 14%; TLC Rf 0.53, CH2Cl2/MeOH, 95:5) and in a later
fraction yielded 3 (93 mg, 5%). Both compounds were identified by
comparison with authentic samples.1a
(1) (a) Kunze, B.; Ho¨fle, G.; Reichenbach, H. J. Antibiot. 1987, 40, 258–
265; note a missprint in the caption of Table 5: it should read 1.6
µmol·mg-1 ·min-1 instead of 1.6 nmol·mg-1 ·min-1. (b) Ho¨fle, G.;
Kunze, B. J. Nat. Prod. 2008, in press.
(2) Ho¨fle, G.; Irschik, H. J. Nat. Prod. 2008, in press.
(3) Sandmann, A.; Dickschat, J.; Jenke-Kodama, H.; Kunze, B.; Dittmann,
E.; Mu¨ller, R. Angew. Chem., Int. Ed. 2007, 46, 2712–2716.
(4) (a) Alvarez, M.; Ajana, W.; Lo´pez-Calahorra, F.; Joule, J. A. J. Chem.
Soc., Perkin Trans. 1 1994, 917, 919. (b) Shintani, R.; Yamagami,
T.; Kimura, T.; Hayashi, T. Org. Lett. 2005, 5317–5319.
(5) Rhouati, S.; Bernou, A. J. Chem. Soc., Chem. Commun. 1989, 730,
732.
Aurachin E (1): TLC, bright blue fluorescence on irradiation at 366
(6) Konradin, C.; Dohle, W.; Rodrigues, A. L.; Schmid, B.; Kochel, P.
Tetrahedron 2003, 59, 1571–1587.
1
nm; H NMR (CDCl3, 400 MHz) δ 8.90 (1H, br s, NH), 7.86 (1H, s,
(7) Alternatively, depending on the location of double and triple bonds
in the C-3 substituent, other heterocycles are formed in good yields:
Bo¨hlendorf, B.; Ho¨fle, G., unpublished results.
(8) Woschek, A.; Mahout, M.; Mereiter, K.; Hammerschmidt, F. Synthesis
2007, 1517, 1522.
(9) From a mechanistic point of view this reaction is not new. According
to Hamana et al.,10 4-substituted quinoline-N-oxides react with
cyanogen bromide in ethanol to give among other products C-3 and
C-8 ethyl carbamates.
J ) 7.5 Hz, H-5), 7.36 (1H, t, J ) 7.5 Hz, H-6), 7.23 (1H, d, J ) 7.5
Hz, H-7), 5.00-5.10 (3H, m, H-2′, H-6′, H-10′), 3.41 (2H, d, J ) 6.5
Hz, H2-1′), 1.85-2.11 (8H, m, H2-4′, H2-5′, H2-8′, H2-9′, 2.82 (3H, s,
H3-10), 1.82 (3H, br s, H3-13′), 1.65 (3H, br s, H3-15′), 1.57 (3H, br s,
H3-14′), 1.56 (3H, br s, H3-12′).
3-Bromocarbonylamino-2-methyl-1H-quinoline-4-one (5): pale
yellow powder, mp 245 °C; IR (KBr) νmax 3460 (m), 3400 (m), 3100
r 2700 (s), 1812 (vs), 1667 (m), 1621 (m), 579 cm-1 (s).
(10) Hamana, M.; Kumadaki, S. Chem. Pharm Bull. 1974, 1506, 1518.
(11) This was communicated in a seminar given at the University of Bonn
in 1994; it was cited in a review12 and stimulated synthetic work.13
(12) Koch, M. Dtsch. Apotheker Zeitung 1995, 135, 4023–4032.
(13) Pochanakom, K. Doctoral Thesis, University of Bonn, Germany, 1999.
(14) Milhous, W. K.; Weatherly, N. F.; Bowdre, J. H.; Desjardins, R. E.
Antimicrob. Agents Chemother. 1985, 27, 525–530. For the 4-day/sc
in ViVo study, Plasmodium berghei N-strain was used.
(15) Surup, F.; Shojaei, H.; v. Zezschwitz, P.; Kunze, B.; Grond, S Bioorg.
Med. Chem. 2008, 1738, 1746.
4-Methyl-3H-oxazolo[4,5-c]quinoline-2-one (7). A fresh solution
of 5 (20 mg) in 1 mL of methanol was evaporated at 30 °C to dryness.
The residue was partitioned between pH 7 buffer and ethyl acetate.
The organic phase was dried with MgSO4 to give 11 mg of 7 as a
colorless solid (TLC Rf 0.32, CH2Cl2/MeOH, 95:5; decomposition on
attempted crystallization from organic solvents); UV (MeOH) λmax (log
ε) 228 (4.38), 316 (3.36), 338 (3.40); IR (KBr) νmax 3200-2850, 1767,
1
1650, 1610 cm-1; H NMR (CD3OD, 300 MHz) δ 8.06, 8.02 (2Η, d,
J ) 7.5 Hz, H-5, H-8), 7.72, 7.65 (2Η, t, J ) 7.5 Hz, H-6, H-7), 2.77
(3Η, s, H3-9); HPLC/ESIMS, [M + H]+ m/z 201; EIMS m/z 200 M+
(100%), 145 (18), 117 (16); HREIMS m/z 200.0584 (calcd for
C11H8N2O2, 200.0586).
(16) Vintonyak, V. V.; Cala`, M.; Lay, F.; Kunze, B.; Sasse, F.; Maier,
M. E. Chem.-Eur. J. 2008, 14, 3709–3720.
NP8004612