Synthetic remodeling of the chartreusin pathway to tune antiproliferative and antibacterial activities

Article abstract of DOI:10.1021/ja4080024

Natural products of the benzonaphthopyranone class, such as chartreusin, elsamicin A, gilvocarcin, and polycarcin, represent potent leads for urgently needed anticancer therapeutics and antibiotics. Since synthetic protocols for altering their architectur

Full text of DOI:10.1021/ja4080024

Article
pubs.acs.org/JACS
Synthetic Remodeling of the Chartreusin Pathway to Tune
Antiproliferative and Antibacterial Activities
Nico Ueberschaar,† Zhongli Xu,† Kirstin Scherlach,† Mikko Metsa-Ketela,‡ Tom Bretschneider,†
̈
̈
Hans-Martin Dahse,§ Helmar Gorls,∥ and Christian Hertweck*,†,⊥
̈
†Department of Biomolecular Chemistry and Department of Infection Biology, Leibniz Institute for Natural Product Research and
§
Infection Biology, HKI, 07745 Jena, Germany
‡Department of Biochemistry and Food Chemistry, University of Turku, 20014 Turku, Finland
∥Friedrich Schiller University, Institute for Inorganic and Analytical Chemistry, 07743 Jena, Germany
⊥Friedrich Schiller University, Chair for Natural Product Chemistry 07743 Jena, Germany
S
* Supporting Information
ABSTRACT: Natural products of the benzonaphthopyranone
class, such as chartreusin, elsamicin A, gilvocarcin, and
polycarcin, represent potent leads for urgently needed
anticancer therapeutics and antibiotics. Since synthetic
protocols for altering their architectures are limited, we
harnessed enzymatic promiscuity to generate a focused library
of chartreusin derivatives. Pathway engineering of the
chartreusin polyketide synthase, mutational synthesis, and
molecular modeling were employed to successfully tailor the
structure of chartreusin. For the synthesis of the aglycones,
improved synthetic avenues to substituted coumarin building blocks were established. Using an engineered mutant, in total 11
new chartreusin analogs (desmethyl, methyl, ethyl, vinyl, ethynyl, bromo, hydroxy, methoxy, and corresponding (1→2) abeo-
chartreusins) were generated and fully characterized. Their biological evaluation revealed an unexpected impact of the ring
substituents on antiproliferative and antibacterial activities. Irradiation of vinyl- and ethynyl-substituted derivatives with blue light
resulted in an improved antiproliferative potency against a colorectal cancer cell line. In contrast, the replacement of a methyl
group by hydrogen caused a drastically decreased cytotoxicity but markedly enhanced antimycobacterial activity. Furthermore,
mutasynthesis of bromochartreusin led to the first crystal structure of a chartreusin derivative that is not modified in the glycoside
residue. Beyond showcasing the possibility of converting diverse, fully synthetic polyphenolic aglycones into the corresponding
glycosides in a whole-cell approach, this work identified new chartreusins with fine-tuned properties as promising candidates for
further development as therapeutics.
analog, elsamicin A or elsamitrucin (2, Figure 1) is produced by
an unidentified actinomycete.11 Both elsamicin A and a
semisynthetic prodrug of chartreusin, IST-622 (3), have
reached phase II clinical trials.12,13 Pharmacological studies
revealed that chartreusin and its derivatives exert their
antitumor activities through binding to DNA, radical-mediated
single strand breaks, and inhibition of topoisomerase II.14
Chartreusin and elsamicin A recognize almost the same G+C-
rich DNA sequences.14 Furthermore, analogs modified in the
disaccharide portion of chartreusin retained the antileukemic
effect in vivo.15 Consequently, there is strong evidence that the
unusual polycyclic aromatic aglycone chartarin represents a key
element for bioactivity.9,15 Interestingly, the benzonaphthopyr-
anone chromophore is also found in potent antitumoral agents
that belong to the gilvocarcin family of polyketide glycosides
(4−9, Figure 1).16−18 Yet, there is a complete lack of
INTRODUCTION
■
Natural products and derivatives thereof represent the prime
source of antitumoral agents.1 Most potent chemotherapeutics
either interfere with the cell cycle or interact with DNA, thus
triggering apoptotic processes.2 A large number of antiprolifer-
ative bacterial metabolites share an architecture consisting of a
DNA-intercalating unit and sugar residues that serve to increase
solubility and binding to DNA.3−5 The aromatic polyketide
glycoside chartreusin (1, Figure 1) from the soil-dwelling
bacterium Streptomyces chartreusis6 is a prominent example of
this natural design. Chartreusin consists of a disaccharide
(fucose and digitalose) and an unusual benzonaphthopyranone
aglycone, named chartarin. While chartreusin was first
investigated because of its antibacterial activity,7,8 further
studies revealed its potent antiproliferative activity against
various tumor cell lines, such as murine P388 and L1210
leukemia, and B16 melanoma cells.9 Given the remarkably high
antitumor activity in vitro, there has been considerable interest
in natural and semisynthetic chartreusin derivatives.10 A natural
Received: August 2, 2013
Published: October 21, 2013
© 2013 American Chemical Society
17408
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
alkyl substituent, and to date the total synthesis of chartreusin
or elsamicin A has not been achieved. Thus, with the intention
of harnessing the biosynthetic potential of the chartreusin
producer, we have previously elucidated the genetic basis for
chartreusin biosynthesis.19 Knowledge of the biosynthetic genes
has set the basis for a targeted manipulation of this promising
family of antitumoral agents. This potential has recently been
highlighted by the design of a vinyl-substituted chartreusin
analog, which is photoactivatable with visible light.20
Here we report the successful combination of chemical
synthesis and biosynthesis to generate a focused library of
chartreusin derivatives that are not readily accessible by total
synthesis. Evaluation of these new compounds reveals the
impact of ring substitution on antiproliferative and antibacterial
activities of chartreusin analogs and the prospective for
photoactivation.
RESULTS AND DISCUSSION
■
Combinatorial Biosynthesis of Homochartreusin
Employing Anthracycline Synthase Genes. Variables in
the biosynthetic pathway of chartreusin were considered as a
way to alter the ring substitution of chartreusin (Figure 2A).
Earlier stable isotope labeling studies21 and our pathway
dissection19 revealed that the methyl substituent at C-1
represents the methyl group of the acetyl starter unit utilized
by the cha polyketide synthase (PKS).19,21 The product of the
cha PKS is a decaketide that is enzymatically shaped by cyclases
Figure 1. Chemical structures of chartreusin-type and gilvocarcin-type
glycosides. Chartreusin (1), elsamicin A (2) and the semisynthetic
derivative IST-622 (3), gilvocarcin M(4), E (5), V (6), H (7), HE (8),
and polycarcin (9).
information on the impact of the chartarin ring substitution on
antitumoral and antibacterial activities. Notably, there are no
reported efficient synthetic methods available for altering the
Figure 2. Combinatorial biosynthesis of ethyl-substituted chartreusin. (A) Organization of the chartreusin (cha) and aclacinomycin (akn)
biosynthetic gene cluster. Gene cassettes used for complementation are highlighted. (B) Model of chartreusin biosynthesis and engineered hybrid
pathway leading to homochartreusin (14). (C) HPLC-MS profiles (λ = 422 nm) of extracts from cultures of (i) heterologous host expressing the
complete cha biosynthesis gene cluster and regulator gene chaR1 (S. albus::pSC5P21/pXU343), (ii) ΔchaABC mutant (S. albus::pXU41/pXU343),
(iii) ΔchaABC mutant supplemented with anthracycline biosynthesis gene cassette aknBCDE2F (S. albus::pXU41/pMMK1).
17409
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
Figure 3. Convergent assembly of chartreusin and norchartreusin. (A) Synthesis of the coumarins 15 and 19 via a formylation-Wittig cascade or a
halogenation-Heck cascade: Reagents and conditions: (a) polyphosphoric acid (PPA), hexamethylenetetramine, 45 min, 140 °C; (e) DEtA, Ph3P
CH−COOEt, 20 min, reflux, 11% for R1 = Me, 7% for R1 = H over 2 steps; (c) 2,4,4,6-tetrabromocyclohexa-2,5-dienone, AcOH, 5 h, 50 °C, 31%;
(d) methyl acrylate, Pd(OAc)2, P(o-Tol)3, DEtA, 30 min, reflux, 78%; (e) phthalide (22), coumarin (20 or 21), (KOt-Bu, 18 h, −60 °C to rt, 66%
for R = Me, 92% for R = H; (g) BBr3, CH2Cl2, 5 min, rt, 93−99%. (B) HPLC profiles (λ = 400 nm) of extracts from cultures of S. albus::pXU41/
pXU343 (ΔchaABC mutant and regulator gene) supplemented with the following substrates: (i) DMSO (negative control), (ii) synthetic chartarin
(12) in DMSO.
and aromatases to form an anthracyclic ring system (5)
reminiscent of the doxorubicin and aclacinomycin aglycones
(Figure 2B).22 However, in contrast to the chartreusin
intermediate, these canonical anthracyclines are, in part,
assembled from propionate starter units.23 It was reasoned
that it would be possible to exchange the chartreusin methyl
substituent for an ethyl group by swapping the PKS priming
units.24 First, in order to permit stable chartreusin production
in the heterologous host, chaR1, a specific cha pathway activator
gene was cloned into an expression vector (yielding pXU343),
and the activator was coexpressed using a constitutive
promoter. Next the genes coding for the aclacinomycin (11)
(akn) PKS (aknBCD) and the genes required for propionate
starter unit selection and incorporation (aknE2 and aknF)25
were cloned to give expression plasmid pMMK1. To construct
a cha PKS null mutant, the genes coding for the minimal PKS
(chaABC) were excised from the cha biosynthetic gene cluster
(Figure 2A) using PCR targeting.26 The resulting cha PKS null
mutant (S. albus::pXU41) lost the ability to produce
chartreusin or any related metabolites (Figure 2C, trace ii).
However, the chartreusin chemotype could be restored by
coexpressing the chaABC gene cassette on a replicative plasmid
(pKJ01).27 In stark contrast, the strain harboring the
incomplete cha gene cluster and the aknBCDE2F cassette
produced a new metabolite with the expected molecular mass
of 654 amu (640 amu for chartreusin) (Figure 2C, trace iii).
From MS/MS fragmentation analyses, which revealed the
known glycoside cleavage pattern for chartreusin, it was
inferred that the aglycone structure was altered. To elucidate
its structure, the compound was isolated from a large-scale
fermentation (20 L) by a combination of different chromato-
graphic techniques. The molecular composition of compound
14, C33H34O14, was deduced from HRESI-MS measurements
1
and H and 13C NMR data. The NMR spectra showed high
similarity to those of 1, suggesting a similar core structure.
However, the DEPT135 spectrum displayed a signal for a
methylene carbon, which is not present in chartreusin, and
1H,1H-COSY measurements showed the vicinity of the
methylene protons to a methyl function. The connectivity of
this ethyl substituent as well as the full assignment of the
17410
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
structure of 14 was confirmed through the analyses of the
HSQC and HMBC data (Supporting Information).
Scheme 1. Convergent Assembly of 1-Substituted
Chartarins
a
Mutasynthesis of a Desmethyl Analog of Chartreusin
(Norchartreusin). Whereas manipulation of the starter unit
selection allowed for the successful exchange of the C-1 methyl
for an ethyl group, biosynthetic engineering could not be used
to produce a chartreusin analog lacking the C-1 methyl group,
as the PKS cannot utilize a one-carbon starter unit.28 To
overcome this limitation a mutasynthesis approach was
employed by complementing the ΔchaABC mutant with
synthetic chartarin analogs. First, mutasynthesis was explored
using the natural substrate.
The Hauser annulation of coumarin 21 and phthalide 22
building blocks has proven a viable method to construct the
pentacyclic ring system of the chartreusin aglycone.20,29 Fusion
of coumarin 21 and phthalide 22 under Hauser conditions gave
10-methoxymethyl-protected chartarin (24) (yield 66%), which
was rapidly (<5 min) and nearly quantitatively deprotected by
boron tribromide. Pure synthetic chartarin (12) was dissolved
in dimethylsulfoxide (DMSO) and added to cultures of the
ΔchaABC mutant (S. albus::pXU41/pXU343). HPLC-MS
monitoring of the fermentation broth extract showed that the
exogenously supplied aglycone was incorporated by the mutant
and processed by glycosylation to restore chartreusin biosyn-
thesis (Figure 3B, trace ii).
For the synthesis of the desmethyl analog, norchartarin (25),
we sought to apply the same protocol. We realized, however,
that the typically employed route to the coumarin 20 involved a
reaction step with only unsatisfactory yields (8.3%). Because of
the limitations of reported formylation reactions for non-
activated aromatic systems30 an alternative route was under-
taken using a halogenation-Heck reaction sequence.31 In order
to direct the halogen into position C-2, several reagents and
conditions were tested. Best results were achieved using 2,4,4,6-
tetrabromocyclohexa-2,5-dienone in acetic acid at 50 °C, which
yielded the desired product in good quantities and with
pronounced regioselectivity. The subsequent Heck reaction
provided the coumarin in high yield (78%). Thus, this new
protocol is over three times more effective than the previously
reported one.
Coumarin 20 was then subjected to the Hauser reaction.
After deprotection of 23 the expected structure 25 was verified
by NMR and MS analyses, and the surrogate was added to the
mutant broth for glycosylation. HPLC-MS analysis indicated
the formation of a new compound (27) with a molecular mass
of 626 amu, which pointed to the successful production of the
desmethyl variant (27). To verify the identity of 27, the new
compound was isolated from a 7 L fermentation culture
supplemented with 70 mg of 25 (100% turnover, 13% isolated
yield). The HRESI-MS, ESI-MS,2 1D and 2D NMR data as well
as the characteristic UV−Vis spectrum fully supported the
structure of norchartreusin (27).
Mutasynthesis of 1-Ethyne, 1-Bromo, 1-Hydroxy, and
1-Methoxy Derivatives. To introduce an ethynyl residue at
position C-1 the corresponding 6-bromo coumarin (32) was
prepared (Scheme 1). Bromination of methyl 3-hydroxyben-
zoate yielded 28, which was formylated under Duff
conditions.29,32 As the reaction in polyphosphoric acid proved
to be unsatisfactory trifluoroacetic acid and microwave
irradiation were used, which proved to be more efficient.
Unexpectedly arduous, however, was the synthesis of the
bromo coumarin 32. The Wittig reaction was hampered by
a
Reagents and conditions: (a)CCl4, Br2, 5 h, rt, 57%; (b)
trifluoroacetic acid, hexamethylenetetramine, 5−15 min, 120 °C,
29−48%; (c) DMA, Ph3PCH−COOMe, 20 min, reflux, 72−83%;
(d) MeI, K2CO3, DMF, 2 h, rt, 84%; (e) MOMCl, N,N-
diisopropylethylamine (DIPEA), CH2Cl2, 2 h, rt, 96%; (f) Tf2O,
DIPEA, CH2Cl2, 2 h, rt, 78%; (g) trimethylsilylacetylene, Pd(PPh3)4,
CuI, N,N-diisopropylamine (DIPA), THF, 18 h, reflux, 63%; (h)
phthalide (22), coumarin (32, 34, 35 or 37), (KOt-Bu, 18 h, −60 °C
to rt, 59−71%; (i) Cs2CO3, MeOH, 2 h, rt; (j) BBr3, CH2Cl2, 5 min,
98%.
unwanted transesterification, formation of the coumaric acid
ester, and dehalogenation.
By utilizing a series of experiments to optimize the reaction
conditions, we found that the regularly used solvent, diethyl
aniline, induces the unwanted trans-esterification. Its sub-
stitution with dimethyl aniline resulted in only in a minor
increase in yield (42% dimethylaniline (DMA) vs 36%
diethylaniline (DEtA)). The main challenge associated with
this step appeared to be the lactonization of the o-coumaric
ester and that the electron-withdrawing bromine may be
exchanged for a hydrogen (with formation of compound 20)
under basic and thermal conditions.33 It became obvious that
the Wittig reagent needed to be exchanged for an active ester or
a related olefination reagent. While there are a large number of
active esters, only a few are available in combination with a
Wittig reagent.34,35 Use of N-hydroxysuccinimidyl ester and the
recently described polymer-bound reagent34 raised the yield to
61% (Table 1). Particularly advantageous is the workup of this
setup, which only requires filtration though a silica gel pad. We
found that the use of Bestmann’s ylide is even more efficient,
providing the bromo coumarin 32 in 71% yield.
With the bromocoumarin at hand the next step was to
introduce the ethynyl residue by means of a Sonogashira
coupling with trimethylsilylacetylene.36 Despite numerous
attempts using diverse conditions the target compound could
not be accessed in this way. Since the synthesis of the more
reactive iodinated analog would have been intricate, the
corresponding triflate was prepared (see Scheme 1). Starting
17411
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
Table 1. Survey of Coumarin Synthesis Starting from the
Salicylaldehyde 30a; PS = Polystyrene
Figure 4. Crystal structure of bromochartreusin (46). Each unit cell
includes two identical molecules, one molecule acetone and one
molecule 3-pentanone.
from methyl 2,5-dihydroxybenzoate (29) the established
reaction sequence involving formylation and Wittig reaction
was employed to produce hydroxycoumarin 33, which was then
converted into triflate 36. Transformation of 36 into the
trimethylsilyl (TMS)-protected ethyne (37), annulation, and
deprotection provided the desired product. The ethyne was
finally processed into the chartreusin derivative by mutasyn-
thesis, and the new compound was isolated and fully
characterized.
The hydroxycoumarin 33 proved to be useful for the
mutasynthesis of two more chartreusin derivatives. The
corresponding methoxymethyl (MOM)- and methoxy-substi-
tuted coumarins (34 and 35) were prepared and yielded
hydroxychartarin (43) and methoxychartarin (44) after
annulation and deprotection. Finally, the bromo-, hydroxy-,
and methoxy-substituted chartarins (42, 43, and 44) were
successfully glycosylated using the ΔchaABC mutant. All new
compounds were characterized using NMR, MS, and UV−Vis
techniques.
The structure of 1-bromochartreusin (46) was confirmed by
X-ray analysis. Initially, we noted that tiny yellow crystals
formed in fractions from open column chromatography. By
regulating the diffusion process the crystallization conditions
were optimized, and it was possible to isolate crystals suitable
for X-ray analysis. Given the presence of the bromine
substituent it was possible to infer the absolute configuration
of 46 (Figure 4) and thus authenticate the glycoside residues of
the mutasynthesis products. Notably, this is the first crystal
structure of a chartreusin derivative that is not modified in the
glycoside residue. As a consequence, this structure proved to be
most useful for modeling experiments.
Potential Structural Role of Chartarin Substituents.
To unravel a potential structural role of substituents on the
pharmacophore the bromo-substituted chartreusin and other
analogs were docked into a canonical B-DNA between the T
and G nucleotides of the sequence GCGTATGATGCG.
Biding of the glycosyl residue supports intercalation of chartarin
between the DNA bases. We noted a cavity at the distal end of
the aglycone, which would, in theory, allow binding of 2-
substituted chartreusins with the DNA (highlighted in Figure
5).
Figure 5. DNA modeling of (1→2) abeo-homochartreusin (71) into
double-stranded DNA. View of the major groove, highlighting free
space around the substituent at position 2 (marked in yellow).
Design of 2-Substituted Chartreusin Analogs. Encour-
aged by the substrate tolerance of the chartreusin glycosylating
machinery and the results of the modeling experiments, we next
decided to introduce a series of residues at position C-2. This
substitution pattern is found naturally in gilvocarcin-type
antitumor agents (4−9, Figure 1). In analogy to these natural
archetypes we aimed to prepare chartreusin derivatives with
methyl-, ethyl-, or vinyl groups at position C-2 of the
aglycone.18 It was envisaged that using 7-bromo-substituted
coumarin as a start point would allow for the introduction of
various substituents. Since it is not viable to directly brominate
methyl 3-hydroxybenzoate (15) at this position, methyl 3-
nitrobenzoate (50) was halogenated using dibromoisocyanuric
́
acid (Scheme 2). The aniline 52 was generated by Bechamp
reduction of the nitro group,37,38 followed by formation and
hydrolysis of the diazonium salt. The well-established Duff
formylation was used to prepare the substituted salicylic
17412
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
substituted derivative 73 was hydrogenated using Lindlar’s
catalyst. The identity of all new compounds was verified by
their physicochemical data (HRESI-MS, ESI-MS2, 1D-, 2D-
NMR, UV−Vis), which were in full agreement with the
proposed structures of (1→2) abeo-chartreusins (70-74).
Evaluation of the Antiproliferative and Cytotoxic
Activities of the Novel Chartreusin Variants. The
biological activities of all new chartreusin analogs were analyzed
in a panel of assays employing different cancer cell lines,
including human umbilical vein endothelial cells (HUVEC),
human chronic myeloid leukemia in blast crisis (K-562), human
cervix carcinoma (HeLa), human colon adenocarcinoma (HT-
29), and human melanoma (Mel-HO) cell lines. We found that
most of the chartreusin derivatives exhibit strong activity
against leukemia and melanoma cell lines, whereas HUVEC
and HeLa cells were substantially less sensitive toward this
compound class.
Scheme 2. Convergent Assembly of 2-Substituted
Chartarins
a
Comparative activity profiling revealed that the various
derivatives differ in their cytostatic potency and thus allowed
for first insights into structure−activity relationships. We found
that the substituent at position C-1 of the aglycone plays a
crucial role for pharmacological activity. Whereas homochar-
treusin (14) and chartreusin (1) were found to be similarly
potent, norchartreusin (27) proved to be almost inactive. 1-
Vinylchartreusin (75) and hydroxychartreusin (43), in contrast,
exhibit strong antiproliferative effects against K-562 cells (GI50:
3.1 and 4.8 μM, respectively). Interestingly, the 1-ethynyl
analogue and 1-methoxychartreusin do not display any
cytostatic properties. Furthermore, the brominated derivatives
are less potent than chartreusin and homochartreusin. Analogs
with substituents at position C-2 of the aglycone were also
found to possess cytostatic properties. (1→2) Abeo-chartreusin
(70) is as potent as the natural congener. A surprising result
was that (1→2) abeo-ethynylchartreusin (73) shows signifi-
cantly increased antiproliferative and cytotoxic activities
compared to its 1-ethynyl counterpart (45).
In addition to the antitumoral activity, the chartreusin
derivatives also possess strong antibiotic properties including
against methicillin-resistant Staphylococcus aureus (MRSA) and
vancomycin-resistant Enterococcus (VRE). 1-Vinyl- and 2-
ethynyl-chartreusin are as active as the natural product and
are even more potent than the reference antibiotic cipro-
floxacin. A similar scenario becomes apparent in case of the
VRE assay. Compounds 1, 75 (1-vinyl), and the 1-hydroxy
derivative 43 are highly active against this strain (MIC 9.59−
9.76 μM). An unexpected result is the antibacterial activity of
norchartreusin (27) against Mycobacterium vaccae (2.49 μM).
As this compound possesses low cytotoxicity, it may be a
promising lead compound for developing an antibiotic.
Impact of Aglycone Substitution and Light on
Antitumoral Activity. The vinyl-substituted chartreusin
derivative 75 exhibits a substantially increased antiproliferative
activity upon photo activation.20 To evaluate the effect of the
substitution site, we analyzed the potential photoreactive
properties of the 2-vinyl derivative (74) as well as the
corresponding alkyne 73. In the assay a blue gallium(III)
nitride (GaN) LED was used since the emission wavelength
(420 nm) of the LED matches with the absorption maximum of
the chartreusins. Whereas the antiproliferative effect of 1, 73−
75 against colon adenocarcinoma cell line HT-29 is rather low
in the absence of light (7.0−9.4 μM), incubation in the
presence of the blue LED substantially increased the activity of
(1→2) abeo-vinylchartreusin (74) (GI50 3.1 vs 9.4 μM) (Figure
a
Reagents and conditions: (a) H2SO4, dibromoisocyanuric acid, 2 h, rt,
86%; (b) Fe, MeOH, HCl, 1 h, reflux, 92%; (c) H2SO4, NaNO2, 0 °C,
then reflux, 44%; (d) PPA, hexamethylenetetramine, 45 min, 140 °C,
36%; (e) DEtA, Ph3PCH−COOEt, 20 min, reflux, 89%; (f) Me2Zn,
Me2N(CH2)2OH, Pd(dppf)Cl2·CH2Cl2, THF, 1h, 60 °C, 94%, dppf =
1,1′-bis(diphenylphosphino)ferrocene; (g) Et2Zn, (CH3)2N-
(CH2)2OH, Pd(dppf)Cl2·CH2Cl2, THF, 1 h, 60 °C, 61%; (h)
vinylboronic acid pinacol ester, Pd(PPh3)4, NaHCO3, H2O, dioxane,
2.5 h, 95 °C, 72%;(i) trimethylsilylacetylene, Pd(PPh3)4, CuI, DIPA,
THF, 18 h, reflux, 83%; (j) phthalide (22), coumarin(56−60), (KOt-
Bu, 18 h, −60 °C to rt, 60−92%; (k) Cs2CO3, MeOH, 2 h, rt; (l) BBr3,
CH2Cl2, 5 min 86−96%; (m) Lindlar catalyst, MeOH, H2, ultrasonic,
2 h, rt, 57%.
aldehyde 54, which was then converted into the coumarin 56
by Wittig olefination. Using a modified Negishi coupling
procedure (via dimethylzinc or diethylzinc) bromocoumarin 56
was converted into methyl- and ethyl-substituted derivatives. A
Suzuki reaction using vinylboronic acid pinacol ester was used
to introduce the vinyl group. All four coumarins 56−60 were
subjected to the Hauser annulation protocol. Whereas the alkyl-
substituted and brominated coumarins were readily converted,
the vinyl derivative 64 could not be produced in this way due to
the instability of the reactant and/or the product under these
conditions.
Next, we sought to convert 2-bromochartarin 66 into its
corresponding vinyl derivative. However, despite testing over
80 different conditions, such as Stille- (n-Bu3Sn-vinyl) or
Suzuki reactions (O’Sheas reagent, vinylboronic acid pinacol
ester)20,39−41 in different solvents, with additives and different
catalysts, the vinyl group could not be installed. To bypass this
hurdle, a TMS-protected ethynyl group was introduced into the
coumarin via a Sonogashira reaction, which granted access to
the vinyl residue by alkyne reduction.
DMSO solutions of (1→2) abeo-chartarins 66−69 were
individually added to the mutant broths, and the biotransfor-
mations yielded all four expected derivatives (70-73). To
prepare the missing vinyl-substituted candidate the ethynyl-
17413
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
Table 2. Survey of on Antimicrobial (MIC value [μM]), Cytotoxic (CC50 [μM] for HeLa), and Antiproliferative (GI50 [μM] for
a
HUVEC, K-562, HT-29, and human melanoma cell line (MEL-HO) Activities of Chartreusin and Its Derivatives
MRSA
w/o
VRE
w/o
M. vaccae
HUVEC
w/o
K-562
w/o
HeLa
w/o
HT-29
MEl-HO
light
w/o
w/o
7.96
w
w/o
1.56
w
1
19.5
38.2
>100
70.9
38.9
9.76
19.1
79.8
35.4
9.73
1.22
9.56
2.49
8.87
2.43
6.24
7.18
>80
65.8
35.0
>78
>77
5.78
26.6
25.1
7.84
16.9
32.8
1.56
2.14
31.6
8.79
4.82
>78
64.1
2.03
11.6
8.65
3.23
4.90
3.06
12.6
23.1
70.7
48.2
25.5
>78
>77
14.2
58.1
26.1
13.4
33.7
24.7
7.65
2.34
14
27
42
43
44
45
70
71
72
73
74
75
cip
van
tet
dox
epi
im
>100
78.1
>100
70.9
19.2
76.6
19.2
37.7
1.08
52.0
76.9
39.0
>100
35.4
4.80
38.3
9.59
2.35
>100
>100
19.2
4.87
19.1
17.7
19.2
19.2
4.78
0.60
33.2
26.9
8.61
9.35
7.05
1.69
3.06
0.61
2.00
1.03
0.11
1.38
0.48
0.08
0.12
0.29
22.1
0.12
0.09
0.2
0.45
0.64
78.6
a
Test strains: MRSA 134/93; VRE 1528; M. vaccae 10670. Cip: ciprofloxacin; van: vancomycin; tet: tetracyclin; dox: doxorubicin; epi: epirubicin;
im: imatinib.
6). Similarly, the activity of (1→2) ethynylchartreusin (73) was
5-fold higher in the presence of light (GI50 1.7 vs 8.6 μM).
chartreusins, thus corroborating their interaction with DNA
(see Supporting Information).
CONCLUSION
■
In this study we have designed 11 new chartreusin analogs by
merging the strengths of chemical synthesis with biosynthesis
and biological target modeling. Since the ring substitution,
namely the methyl residue, cannot be exchanged by current
synthetic protocols, we explored, as an alternative, biosynthetic
engineering and mutasynthesis. Initially, by deletion of the PKS
genes in the cha gene locus and cross-complementation with
anthracycline biosynthesis genes a polyketide primed with
propionate in lieu of acetate was created, and thus an ethyl
substituent was installed in the chartarin structure. The strategy
of exchanging PKS starter unit cassettes is a powerful approach
that has been employed in several impressive biosynthetic
engineering studies.23,44−47 Beyond granting access to a
chartreusin analog the successful cross-complementation of
the cha PKS mutant with an anthracycline PKS has an added
value in the context of biosynthesis. Our findings reinforce the
biosynthetic model involving the rearrangement of an
anthracyclic polyketide scaffold into the rare benzonaphthopyr-
anone chromophore.19 While it would be, in principle, feasible
to expand the size of the priming acyl chain, it is not viable to
shorten the chain in order to yield the corresponding
chartreusin analog lacking an alkyl residue (desmethyl or
norchartreusin (27)). Thus, a chemo-biosynthetic route was
explored, where the cha PKS null mutant, which is not capable
of producing the natural chartarin aglycone, was complemented
with a synthetic surrogate. For preparation of the chartarin
cores, synthetic protocols toward substituted coumarins were
optimized, which led to more efficient routes, higher yields, and
improved regioselectivity. In all cases the mutant accepted the
exogenously supplied aglycone and completed the pathway
yielding chartreusin derivatives. Conceptionally related muta-
synthesis approaches have been successfully applied for the
Figure 6. Evaluation of antiproliferative effect and growth inhibition
values on colorectal cancer cell line HT-29 with and without light.
Interestingly, an electrophoretic mobility assay (EMSA) using a
defined DNA fragment and a HPLC-HRMS-guided photo-
product assay did not give any evidence for a [2 + 2] photo
adduct with thymine or other nucleobases.20,42 Apparently
photo activation of the ethynyl derivative opens an alternative
reaction channel. It should be highlighted that the ethynyl
functionality has been unprecedented in this compound class.
Finally, the interaction between the chartreusin variants and
DNA, which is a key mechanism for native 1, was investigated.
Buffered solutions of chartreusin and its derivatives were
titrated with a DNA solution, while the UV−Vis spectrum was
monitored.43 This experiment revealed a clear band shift for all
17414
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
Notes
exchange of starter and extender units in polyketide and
nonribosomal peptide pathways.24,28,48−59 In this work we
report an unparalleled systematic conversion of fully synthetic
polyphenolic aglycones into the corresponding glycosides on a
preparative scale. This efficient chemo-biosynthetic route not
only led to the first crystal structure of a chartreusin derivative
with a native glycosyl residue but also generated a library of
aglycone-modified chartreusins. Their biological evaluation
grants fresh insights into the role of ring substitution of the
DNA intercalating unit. For the first time it was possible to
deduce structural prerequisites for antitumoral activities of the
benzonaphthopyranones. Three new chartreusin derivatives
(14, 70, and 73) show potent antitumoral activity. This finding
that norchartreusin lacks cytostatic properties highlights the
crucial role of the aglycone substitution pattern for the
pharmacological activity. Moreover, we have demonstrated
that altering the substituent at position C-1 of the aglycone
causes a substantial enhancement in antiproliferative cell
selectivity compared to the analogs modified at position C-2.
Notably, all chartreusin derivatives interfere with DNA, and two
new chartreusin derivatives show an increased antiproliferative
activity when irradiated with blue light. One of these
compounds harbors an ethynyl group in lieu of an aliphatic
group, which has not yet been reported for any photoreactive
agents employed in in vitro cell assays. The most surprising
finding was that the simple lack of the methyl group causes a
dramatic decrease in cytotoxicity, but strongly enhances
antibacterial activities. Specifically, the desmethyl variant 27 is
active against mycobacteria, which are in the focus of antibiotic
research due to the rise of tuberculosis and other life-
threatening infectious diseases. Thus, the blend of biological
and synthetic methods not only granted access to a library of
chartreusins that are otherwise not accessible but also yielded
promising candidates for further development as antitumoral or
antibacterial agents (Figure 7).
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank U. Knupfer for performing the fermentation, C.
̈
Karkowski for downstream processing, A. Perner for MS
analysis, H. Heinecke for NMR measurements, and E.-M.
Neumann for assistance in biological assays. Financial support
by the BMBF (GenBioCom in the GenoMik program) to C.H.,
the EMBO foundation and the Academy of Finland to M.M.-K.
is gratefully acknowledged.
REFERENCES
■
(1) Newman, D. J.; Cragg, G. M. J. Nat. Prod. 2012, 75, 311−335.
(2) Souza, A. C.; de Fatima, A.; da Silveira, R. B.; Justo, G. Z. Curr.
Drug Targets 2012, 13, 1072−1082.
(3) Cragg, G. M.; Kingston, D. G. I.; Newman, D. J. Anti Cancer
Agents From Natural Products; Taylor & Francis: Abingdon, U.K.,
2005.
(4) Wong, A.; Toth, I. Curr. Med. Chem. 2001, 8, 1123−1136.
(5) Kren, V.; Martinkova, L. Curr. Med. Chem. 2001, 8, 1303−1328.
(6) Weissman, K. Chem. Biol. 2005, 12, 512−514.
(7) Uchida, H.; Nakakita, Y.; Enoki, N.; Abe, N.; Nakamura, T.;
Munekata, M. J. Antibiot. 1994, 47, 655−667.
(8) Uchida, H.; Nakakita, Y.; Enoki, N.; Abe, N.; Nakamura, T.;
Munekata, M. J. Antibiot. 1994, 47, 648−654.
(9) Kon, K.; Sugi, H.; Tamai, K.; Ueda, Y.; Yamada, N. J. Antibiot.
1990, 43, 372−382.
(10) Portugal, J. Curr. Med. Chem. 2003, 3, 411−420.
(11) Lorico, A.; Long, B. H. Eur. J. Cancer 1993, 29A, 1985−1991.
(12) Asai, G.; Yamamoto, N.; Toi, M.; Shin, E.; Nishiyama, K.;
Sekine, T.; Nomura, Y.; Takashima, S.; Kimura, M.; Tominaga, T.
Cancer Chemother. Pharmacol. 2002, 49, 468−472.
(13) Verweij, J.; Wanders, J.; Nielsen, A. L.; Pavlidis, N.; Calabresi,
F.; Huinink, W. T.; Bruntsch, U.; Piccart, M.; Franklin, H.; Kaye, S. B.
Ann. Oncol. 1994, 5, 375−376.
(14) Salas, X.; Portugal, J. FEBS Lett. 1991, 292, 223−228.
(15) Takai, M.; Uehara, Y.; Beisler, J. A. J. Med. Chem. 1980, 23,
549−553.
(16) Hou, J.; Liu, P.; Qu, H.; Fu, P.; Wang, Y.; Wang, Z.; Li, Y.;
Teng, X.; Zhu, W. J. Antibiot. 2012, 65, 523−526.
(17) Shepherd, M. D.; Liu, T.; Mendez, C.; Salas, J. A.; Rohr, J. Appl.
Environ. Microbiol. 2011, 77, 435−441.
(18) Li, Y. Q.; Huang, X. S.; Ishida, K.; Maier, A.; Kelter, G.; Jiang, Y.;
Peschel, G.; Menzel, K. D.; Li, M. G.; Wen, M. L.; Xu, L. H.; Grabley,
S.; Fiebig, H. H.; Jiang, C. L.; Hertweck, C.; Sattler, I. Org. Biomol.
Chem. 2008, 6, 3601−3605.
(19) Xu, Z.; Jakobi, K.; Welzel, K.; Hertweck, C. Chem. Biol. 2005,
12, 579−588.
(20) Ueberschaar, N.; Dahse, H.-M.; Bretschneider, T.; Hertweck, C.
Angew. Chem., Int. Ed. 2013, 52, 6185−6189.
Figure 7. Overview on the impact of ring substituents on biological
activity. Orange arrows highlight groups that support antiproliferative
activity against K-562 leukemia cells, dark-blue arrows highlight groups
that support antibacterial activity toward M. vaccae.
(21) Canham, P. L.; Vining, L. C.; McInnes, A. G.; Walter, J. A.;
Wright, J. L. C. Can. J. Chem. 1977, 55, 2450−2457.
(22) Hertweck, C.; Luzhetskyy, A.; Rebets, Y.; Bechthold, A. Nat.
Prod. Rep. 2007, 24, 162−190.
(23) Moore, B. S.; Hertweck, C. Nat. Prod. Rep. 2002, 19, 70−99.
(24) Xu, Z.; Schenk, A.; Hertweck, C. J. Am. Chem. Soc. 2007, 129,
6022−6030.
ASSOCIATED CONTENT
* Supporting Information
Experimental details, physicochemical data, biological assays,
MS- and NMR spectra. This material is available free of charge
via the Internet at http://pubs.acs.org.
■
(25) Raty, K.; Kantola, J.; Hautala, A.; Hakala, J.; Ylihonko, K.;
Mantsala, P. Gene 2002, 293, 115−122.
S
(26) Gust, B.; Challis, G. L.; Fowler, K.; Kieser, T.; Chater, K. F. Proc.
Natl. Acad. Sci. U.S.A. 2003, 100, 1541−1546.
(27) Jakobi, K.; Hertweck, C. J. Am. Chem. Soc. 2004, 126, 2298−
2299.
AUTHOR INFORMATION
Corresponding Author
christian.hertweck@hki-jena.de.
■
(28) Kennedy, J. Nat. Prod. Rep. 2008, 25, 25−34.
(29) Mal, D.; Patra, A.; Roy, H. Tetrahedron Lett. 2004, 45, 7895−
7898.
17415
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416

Journal of the American Chemical Society
Article
(30) Bagno, A.; Kantlehner, W.; Scherr, O.; Vetter, J.; Ziegler, G. Eur.
J. Org. Chem. 2001, 2001, 2947−2954.
(31) Valizadeh, H.; Vaghefi, S. Heterocycl. Comm. 2010, 16, 113−120.
(32) Harayama, T.; Nakatsuka, K.; Nishioka, H.; Murakami, K.;
Hayashida, N.; Ishii, H. Chem. Pharm. Bull. 1994, 42, 2170−2173.
(33) Hekmatshoar, R.; Sajadi, S.; Heravi, M. M. J. Chin. Chem. Soc.
2008, 55, 616−618.
(34) Henkel, B. Synlett 2008, 355−358.
(35) Schobert, R. Naturwissenschaften. 2007, 94, 1−11.
(36) Kloss, F.; Kohn, U.; Jahn, B. O.; Hager, M. D.; Gorls, H.;
̈
Schubert, U. S. Chem. Asian J. 2011, 6, 2816−2824.
(37) Khan, A.; Hecht, S. Chem.Eur. J. 2006, 12, 4764−4774.
(38) Schwetlick, K. Organikum; WILEY-VCH: Weinheim, 2001; Vol.
21.
(39) Blanco, B.; Sedes, A.; Peon, A.; Lamb, H.; Hawkins, A. R.;
Castedo, L.; Gonzalez-Bello, C. Org. Biomol. Chem. 2012, 10, 3662−
3676.
(40) Coleman, C. M.; O’Shea, D. F. J. Am. Chem. Soc. 2003, 125,
4054−4055.
(41) Li, G.; Tao, Z.-F.; Tong, Y.; Przytulinska, M. K.; Kovar, P.;
Merta, P.; Chen, Z.; Zhang, H.; Sowin, T.; Rosenberg, S. H.; Lin, N.-
H. Bioorg. Med. Chem. Lett. 2007, 17, 6499−6504.
(42) McGee, L. R.; Misra, R. J. Am. Chem. Soc. 1990, 112, 2386−
2389.
(43) Krueger, W. C.; Pschigoda, L. M.; Moscowitz, A. J. Antibiot.
1986, 39, 1298−1303.
(44) Hertweck, C. Angew. Chem., Int. Ed. 2009, 48, 4688−4716.
(45) Lee, T. S.; Khosla, C.; Tang, Y. J. Am. Chem. Soc. 2005, 127,
12254−12262.
(46) Xu, Z.; Schenk, A.; Hertweck, C. J. Am. Chem. Soc. 2007, 129,
6022−6030.
(47) Cropp, T. A.; Wilson, D. J.; Reynolds, K. A. Nat. Biotechnol.
2000, 18, 980−983.
(48) Eichner, S.; Knobloch, T.; Floss, H. G.; Fohrer, J.; Harmrolfs,
K.; Hermane, J.; Schulz, A.; Sasse, F.; Spiteller, P.; Taft, F.; Kirschning,
A. Angew. Chem., Int. Ed. 2012, 51, 752−757.
(49) Kusebauch, B.; Brendel, N.; Kirchner, H.; Dahse, H. M.;
Hertweck, C. ChemBioChem 2011, 12, 2284−2288.
(50) Weist, S.; Sussmuth, R. D. Appl. Microbiol. Biotechnol. 2005, 68,
̈
141−150.
(51) Werneburg, M.; Busch, B.; He, J.; Richter, M. E. A.; Xiang, L. K.;
Moore, B. S.; Roth, M.; Dahse, H. M.; Hertweck, C. J. Am. Chem. Soc.
2010, 132, 10407−10413.
(52) Weissman, K. J. Trends. Biotechnol. 2007, 25, 139−142.
(53) Eustaq
3936−3938.
́
uio, A. S.; Moore, B. S. Angew. Chem., Int. Ed. 2008, 47,
(54) Mo, S.; Kim, D. H.; Lee, J. H.; Park, J. W.; Basnet, D. B.; Ban, Y.
H.; Yoo, Y. J.; Chen, S. W.; Park, S. R.; Choi, E. A.; Kim, E.; Jin, Y. Y.;
Lee, S. K.; Park, J. Y.; Liu, Y.; Lee, M. O.; Lee, K. S.; Kim, S. J.; Kim,
D.; Park, B. C.; Lee, S. G.; Kwon, H. J.; Suh, J. W.; Moore, B. S.; Lim,
S. K.; Yoon, Y. J. J. Am. Chem. Soc. 2011, 133, 976−985.
(55) Murphy, A. C.; Fukuda, D.; Song, Z.; Hothersall, J.; Cox, R. J.;
Willis, C. L.; Thomas, C. M.; Simpson, T. J. Angew. Chem., Int. Ed.
2011, 50, 3271−3274.
(56) Kirschning, A.; Hahn, F. Angew. Chem., Int. Ed. 2012, 51, 4012−
4022.
(57) Taft, F.; Harmrolfs, K.; Nickeleit, I.; Heutling, A.; Kiene, M.;
Malek, N.; Sasse, F.; Kirschning, A. Chem.Eur. J. 2012, 18, 880−886.
(58) Almabruk, K. H.; Lu, W.; Li, Y.; Abugreen, M.; Kelly, J. X.;
Mahmud, T. Org. Lett. 2013, 15, 1678−1681.
(59) Moss, S. J.; Stanley-Smith, A. E.; Schell, U.; Coates, N. J.; Foster,
T. A.; Gaisser, S.; Gregory, M. A.; Martin, C. J.; Nur-E-Alam, M.;
Piraee, M.; Radzom, M.; Suthar, D.; Thexton, D. G.; Warneck, T. D.;
Zhang, M. Q.; Wilkinson, B. Med. Chem. Commun. 2013, 4, 324−331.
17416
dx.doi.org/10.1021/ja4080024 | J. Am. Chem. Soc. 2013, 135, 17408−17416