Bifunctional Doscadenamides Activate Quorum Sensing in Gram-Negative Bacteria and Synergize with TRAIL to Induce Apoptosis in Cancer Cells

Article abstract of DOI:10.1021/acs.jnatprod.0c01003

New cyanobacteria-derived bifunctional analogues of doscadenamide A, a LasR-dependent quorum sensing (QS) activator in Pseudomonas aeruginosa, characterized by dual acylation of the pyrrolinone core structure and the pendant side chain primary amine to form an imide/amide hybrid are reported. The identities of doscadenamides B-J were confirmed through total synthesis and a strategic focused library with different acylation and unsaturation patterns was created. Key molecular interactions for binding with LasR and a functional response through mutation studies coupled with molecular docking were identified. The structure-activity relationships (SARs) were probed in various Gram-negative bacteria, including P. aeruginosa and Vibrio harveyi, indicating that the pyrrolinone-N acyl chain is critical for full agonist activity, while the other acyl chain is dispensable or can result in antagonist activity, depending on the bacterial system. Since homoserine lactone (HSL) quorum sensing activators have been shown to act in synergy with TRAIL to induce apoptosis in cancer cells, selected doscadenamides were tested in orthogonal eukaryotic screening systems. The most potent QS agonists, doscadenamides S10-S12, along with doscadenamides F and S4 with partial or complete saturation of the acyl side chains, exhibited the most pronounced synergistic effects with TRAIL in triple negative MDA-MB-231 breast cancer cells. The overall correlation of the SAR with respect to prokaryotic and eukaryotic targets may hint at coevolutionary processes and intriguing host-bacteria relationships. The doscadenamide scaffold represents a non-HSL template for combination therapy with TRAIL pathway stimulators.

Full text of DOI:10.1021/acs.jnatprod.0c01003

pubs.acs.org/jnp
Article
Bifunctional Doscadenamides Activate Quorum Sensing in Gram-
Negative Bacteria and Synergize with TRAIL to Induce Apoptosis in
Cancer Cells
Xiao Liang, Qi-Yin Chen, Gustavo M. Seabra, Susan Matthew, Jason C. Kwan, Chenglong Li,
Valerie J. Paul, and Hendrik Luesch*
Cite This: J. Nat. Prod. 2021, 84, 779−789
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: New cyanobacteria-derived bifunctional analogues
of doscadenamide A, a LasR-dependent quorum sensing (QS)
activator in Pseudomonas aeruginosa, characterized by dual
acylation of the pyrrolinone core structure and the pendant side
chain primary amine to form an imide/amide hybrid are reported.
The identities of doscadenamides B−J were confirmed through
total synthesis and a strategic focused library with different
acylation and unsaturation patterns was created. Key molecular
interactions for binding with LasR and a functional response
through mutation studies coupled with molecular docking were
identified. The structure−activity relationships (SARs) were
probed in various Gram-negative bacteria, including P. aeruginosa
and Vibrio harveyi, indicating that the pyrrolinone-N acyl chain is
critical for full agonist activity, while the other acyl chain is dispensable or can result in antagonist activity, depending on the bacterial
system. Since homoserine lactone (HSL) quorum sensing activators have been shown to act in synergy with TRAIL to induce
apoptosis in cancer cells, selected doscadenamides were tested in orthogonal eukaryotic screening systems. The most potent QS
agonists, doscadenamides S10−S12, along with doscadenamides F and S4 with partial or complete saturation of the acyl side chains,
exhibited the most pronounced synergistic effects with TRAIL in triple negative MDA-MB-231 breast cancer cells. The overall
correlation of the SAR with respect to prokaryotic and eukaryotic targets may hint at coevolutionary processes and intriguing host−
bacteria relationships. The doscadenamide scaffold represents a non-HSL template for combination therapy with TRAIL pathway
stimulators.
he tumor necrosis factor-related apoptosis-inducing
Tligand (TRAIL), also known as Apo-2L and TNFSF10,
is a TNF family cytokine that can induce cell apoptosis and
cause programmed cell death by binding to the death receptors
DR4 (TRAIL-RI) and DR5 (TRAIL-RII).1,2 Moreover, there
are studies showing that TRAIL can selectively cause cancer
cell death without detrimental effects on normal cells.2,3 This
selectivity has made TRAIL a promising candidate for cancer
therapy4 and stimulated intensive studies toward the develop-
ment of therapeutic agents targeting the TRAIL signaling
pathway,1 including recombinant TRAIL proteins and
monoclonal antibody agonists specific for DR4 (TRAIL-RI)
and DR5 (TRAIL-RII).5 However, these agents did not
achieve satisfying anticancer activities in clinical investiga-
tions,6 because many cancer cells have developed resistance
toward TRAIL, thus compromising the efficacy of TRAIL-
therapy.4,7
sensitizing the cells to TRAIL therapy.7 Recently, we have
isolated and reported a new QS activator, doscadenamide A8
(1a, Figure 1), from the cyanobacterium Moorena bouillonii
collected at Finger Reef, Guam.9 Doscadenamide A (1a)
possesses an unique structural skeleton featuring a pyrrolinone
core with a pendant side chain linker (pyLys-OMe), and this
core structure is doubly acylated with the same carboxylic acid,
(R)-2-methyl-oct-7-ynoic acid (Moya), (1) on the side chain
amine (Moya1) and (2) the five-membered amide ring
structure (Moya2), to form amides and imides, respectively
(Figure 1).9 Our initial investigation indicated that doscade-
namide A (1a) is an unprecedented nonhomoserine lactone
Special Issue: Special Issue in Honor of A. Douglas
Kinghorn
Received: September 14, 2020
Published: January 22, 2021
Previously, the bacterial quorum sensing (QS) activator N-
(3-oxododecanoyl)-L-homoserine lactone (C12) was reported
to enhance the TRAIL-induced apoptosis in cancer cells, thus
© 2021 American Chemical Society and
American Society of Pharmacognosy
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789
779

Journal of Natural Products
pubs.acs.org/jnp
Article
Figure 1. Structures of doscadenamides, including naturally occurring doscadenamides and synthetic structural analogues of doscadenamide A. The
core structure/general skeleton is diversified through different acylation patterns. Additional structures are characterized by inversion of the
pyrrolinone ring configuration, transposed pyrrolinone functional groups or cyclization. Compounds included in the SAR study described in this
manuscript are highlighted in yellow. *Natural products that were synthesized.
(HSL) QS agonist that is able to activate the QS signaling
pathway in Pseudomonas aeruginosa in a LasR dependent
manner.9 LasR is a transcriptional activator of Pseudomonas
aeruginosa virulence genes and belongs to the LuxR protein
family that was named after the regulator of bioluminescence
(lux) in Vibrio fischeri.10,11
Here we describe the identification of additional natural
doscadenamides (B−I) from the same cyanobacterium and the
synthesis of strategically diversified analogues to probe the
structure−activity relationship (SAR) with respect to QS
modulation in several different Gram-negative bacterial
systems and interrogated the tentatively causative molecular
interactions with bacterial receptors. We then evaluated a
subset of the focused doscadenamide library in cancer cells and
successfully established a synergistic activity with TRAIL in
invasive triple negative breast cancer cells.
RESULTS
■
Isolation and Characterization of Natural Doscade-
namides B−J. The isolation of doscadenamide A (1a, Figure
1) from M. bouillonii was reported previously.9 A chemical
investigation of the original extracts enabled the identification
of several related compounds that differed in the degree of
unsaturation and methylation pattern (Figure 1). 1D and 2D
NMR analysis (Tables S1−S7) coupled with high resolution
electrospray ionization mass spectrometry (HRESIMS) and
780
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
Scheme 1. Total Synthesis of Doscadenamide H (1h)
synthesis of selected family members unambiguously estab-
lished the structures.
unsaturation (Figure S29, Table S6). Due to the difficulty of
purifying doscadenamide H (1h, Figure 1), we synthesized the
compound with the proposed structure (Scheme 1) using our
established synthetic route9 and compared its 1H NMR
spectrum with the isolated fraction, confirming the dominant
presence of doscadenamide H (1h) in the natural product
fraction (Figure S30). The proposed doscadenamides I (1i)
and J (1j) coeluted under various HPLC conditions and were
present in roughly equal abundance (1:1 mixture). To confirm
our hypothesis, we synthesized the two compounds,
established the structures and compared their 1H NMR
spectra with the coeluting natural products, confirming their
presence (Figure S31, Table S7).
Synthesis of a Strategic Focused Library to Probe the
SAR. In addition to doscadenamides that are biosynthesized
through natural diversification, we aimed to prepare a
complementary compound set that possessed different degrees
of methylation and/or unsaturation in the carboxylic acid
chains (1k−1n), lacked one of the carboxylic acid chains (2a,b
vs 3a−3c), exhibited different pyrrolinone configuration (1o)
or transposed pyrrolinone functionalization (1p), as well as a
cyclized version (4a) (Figure 1). We termed these new
synthetic analogues doscadenamides S4−S15 (Figure 1). The
previously reported diastereomers of 1a were named
doscadenamides S1−S3 (Supporting Information, Figure S32).
To explore the contribution of each structural characteristic
in doscadenamide A (1a, Figure 1) to the interaction with its
target and its biological activity, we synthesized most of these
structural analogues (doscadenamides F−J, 1f−1j) following a
similar approach as previously described9 for 1a and
exemplified by the synthesis of doscadenamide H (1h, Scheme
1). Each target compound can be obtained in three main steps:
pyrrolinone construction, carboxylic acid activation, and
amide/imide coupling. For certain analogues (1f, 1k, 1l, 1m,
2b, 3b, and 3c), we synthesized additional carboxylic acids as
specific building blocks to prepare the desired products. The
detailed synthesis steps are described in the Supporting
Information.
Briefly, we isolated five pure structural analogues,
doscadenamides B−F (1b−1f, Figures 1 and S1−28, Tables
S1−5) as well as impure doscadenamides G−J (1g−j, Figures
1 and S29−31, Tables S6−7). Doscadenamide B (1b, Figure
1) features an alkene moiety in the R2 side chain. The
HRESIMS spectrum of 1b in positive mode exhibited a [M +
H]+ peak at m/z 455.2897 with ten degrees of unsaturation;
while the 1H NMR spectrum of 1b in CDCl3 further confirmed
the presence of characteristic signals corresponding to the
alkene proton (δH 5.94 ppm) and the attached methyl group
(δH 1.90 ppm) (Figures S1−6, Table S1). Doscadenamide C
(1c, Figure 1) and doscadenamide D (1d, Figure 1) feature a
vinyl terminus at R1 and R2 side chain, respectively. The
HRESIMS spectrum of 1c and 1d in positive mode exhibited a
[M + H]+ peak at m/z 459.3218 and m/z 459.3226 with eight
1
degrees of unsaturation, while the H NMR spectrum of 1c
and 1d in CDCl3 suggested the presence of a vinyl group at
4.96 and 4.94 ppm, respectively (Figures S7−17, Tables S2−
3). Doscadenamide E (1e, Figure 1) not only features a vinyl
terminus at the R1 side chain but also exhibits the loss of the α-
methyl group compared to doscadenamide A (1a, Figure 1).
The HRESIMS spectrum of 1e in positive mode exhibited a
[M + H]+ peak at m/z 445.3069 with eight degrees of
1
unsaturation, and the H NMR spectrum of 1e in CDCl3
indicated the characteristic signals of a vinyl group at 4.96 ppm
and methylene at 2.94 ppm (Figures S18−22, Table S4).
Doscadenamide F (1f, Figure 1) possesses two vinyl termini,
which was supported by its HRESIMS spectrum in the positive
mode showing a [M + H]+ peak at m/z 461.3362 with seven
degrees of unsaturation and its 1H NMR spectrum in CDCl3 at
4.95 and 4.97 ppm for characteristic signals of vinyl groups
(Figure 23−28, Table S5). Doscadenamide G (1g, Figure 1)
includes a R1 side chain with a vinyl terminus without α-
methylation, while doscadenamide H (1h, Figure 1) includes a
R2 side chain with a vinyl terminus without α-methylation. The
structural similarity of 1g and 1h complicated the purification
of each compound. We were only able to obtain pure
Our preliminary study indicated that the diastereomeric
doscadenamides S1−S3 (Figure S32) behave similarly in
modulating QS in Pseudomonas aeruginosa.9 To perform the
SAR study of doscadenamides systematically, we selected
several analogues to compose a focused library for primary
1
doscadenamide G (1g, Figure 1), and the H NMR spectrum
of 1g in CDCl3 indicated the presence of a vinyl group at 4.97
ppm. The HRESIMS spectrum of 1g in positive mode shows a
[M + H]+ peak at m/z 445.3063 with eight degrees of
781
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
Figure 2. Docked structure of doscadenamide A into LasR pocket. (A) Image showing the protein surface, clipped to show the buried binding
pocket (blue). (B) Cartoon diagram of the protein in the same orientation as in A. The mutated residues Tyr56, Trp60, and Asp73 are shown in
orange. (C) Detailed view of the binding mode. The system is rotated to better show the interactions. (D) Ligand interaction diagram of the
binding mode. Regions in green are hydrophobic, light blue are polar, red is negatively charged, and dark blue are positively charged. Purple arrows
indicate hydrogen bonds. Numbering was adapted from the work of Zou et al.13
Figure 3. QS activity in different systems. (A) Activity of doscadenamide A (1a) in E. coli reporter gene assay using pSB1075 (wild-type) and point
mutants (Y56F, W60F, and D73N). (B) Normalized pyocyanin production in P. aeruginosa mutant PAO-JP1 after treatment with doscadenamide
A (1a) and its structural analogues for 6 h at 100 μM. C12 was used as the positive control. (C) Measurement of QS activating activity of
doscadenamide A (1a) and its structural analogues in V. harveyi after 8 h treatment, presented by ΔLUM/LUM (DMSO) [ΔLUM =
LUM(doscadenamide analogue) − LUM(DMSO)] normalized by cell viability measured with OD600. AI-1 is a reported quorum sensing
autoinducer in this system (Figure S33) and was used as the positive control.
studies (Figure 1, highlighted in yellow). To explore the
contribution of the unsaturated alkyne moiety to the
bioactivity of doscadenamide A (1a, Figure 1), we included
structural analogues with two unsaturated alkene termini (1f,
Figure 1) and two saturated carboxylic acid moieties (1k,
Figure 1). To further investigate the significance of each side
chain in doscadenamides, we included structural analogues 2a
and 3a−3c, possessing only one carboxylic acid chain and with
different degrees of unsaturation. To confirm the previously
hypothesized9 importance of the linear structure of doscade-
namide A (1a, Figure 1), the cyclic compound 4a was also
included in the library.
QS Modulation by Doscadenamides in Pseudomonas
aeruginosa and LasR Molecular Interaction Studies. We
782
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
Figure 4. Three-dimensional representation of 3a−c (A) in the binding pocket (3a, cyano; 3b, pink; 3c, orange), and diagram showing the
interactions with the surrounding residues (B) for 3a. Regions in green are hydrophobic, light blue are polar, red is negatively charged, and dark
blue are positively charged. Purple arrows indicate hydrogen bonds. The binding modes of the other agonists (3b,c) are similar and are shown in
Figure S36.
previously determined that doscadenamide A (1a) activates
QS, which was established using the 3-oxo-C12-HSL-
responsive reporter plasmid pSB1075,12 a plasmid encoding
LasR (the C12 HSL receptor in Pseudomonas aeruginosa) and
containing a light-producing luxCDABE cassette expressed in
Escherichia coli. The activity was abolished in a strain
(pTIM5319) that lacked the AHL domain.
We first interrogated potential binding modes of doscade-
namide A (1a). Figure 2A,B show the LasR protein
highlighting the deep binding pocket with doscadenamide A
in place. Figure 2C,D show a representative pose of the
binding mode. Compared to the structure of LasR with the
autoinducer N-3-oxo-dodecanoyl acyl homoserine lactone
(C12, Figure S33), the interactions of the C12-polar
headgroup are replaced by the amide bond in the
doscadenamide A pyrrolinone-N side chain, which forms H-
bonds to Tyr56 and Asp73 (Figure S34) (numbering for the
protein following Zou and Nair13).
We screened three mutant strains (Y56F, W60F, and D73N)
in our LasR reporter gene assay in E. coli to study the
molecular interaction between doscadenamide A (1a, Figure
1) and LasR in concert with docking. The assay result (Figure
3A) indicated the strong binding contribution of the Tyr56
and Asp73 residues, consistent with molecular docking results
(Figure 2C,D).
No bonds to Trp60 are formed. However, the acyl chain
may form hydrophobic interactions with Trp88, Tyr93, and
Phe101; the latter on the α5 helix. The pyrrolinone is buried
deeper in the binding pocket, surrounded by Leu36, Gly38,
Tyr47, Ala50, Ile52 Val76, and Ala127. Tyr47, in the L3 loop
that closes the binding site, is in position to establish a side-to-
face π interaction with the heterocycle. The pyrrolinone-N side
chain extends further in the tunnel toward the right side of the
figure, surrounded by Leu40, Asp46, Tyr47 (backbone),
Arg71, and Thr80. Most of those interactions are the same
as detected for the long hydrophobic tail in the autoinducer.
To follow up our previous study regarding the QS
modulatory activity of doscadenamide A (1a) on wild-type
P. aeruginosa and extend our reporter gene assay studies, we
examined the QS activation of different doscadenamides using
a QS systems deficient strain, PAO-JP1.14 PAO-JP1 is a P.
aeruginosa mutant with a lasl deletion that cannot produce
C12.15 To investigate the activating effect of doscadenamide A
on P. aeruginosa without interference from the C12 produced
by P. aeruginosa itself, we examined the pyocyanin production
of P. aeruginosa mutant PAO-JP1 after treatment with
doscadenamide A and selected structural analogues (Figure
1) for 6 h at 100 μM.
Compounds selected for this bioassay possessed a different
number of carboxylic acid chains (1a vs 2a vs 3a) or different
degrees of unsaturation (1a/f/k vs 3a−c). As shown in Figure
3B, when normalized for cell number, these doscadenamides
(Figure 1) can all activate the pyocyanin production of PAO-
JP1 but to a different extent at the concentration of 100 μM.
Doscadenamide A (1a) induced 3-fold QS activation, similarly
to the positive control C12. Compound 2a, which possesses
only one side chain R1, exhibits minimal QS-activating effect,
while 3a−c with acylation of the amide core showed the most
potent activation, 3-fold, 4-fold, and 5-fold, indicating that this
chain is predominantly responsible for the QS agonist activity
in P. aeruginosa. The degree of unsaturation appears to play a
minor role in a chain-dependent fashion, as the trends for 1a/
f/k vs 3a−c were opposite, possibly indicating the dependence
of both chains to the overall net effect. The overall trend was
the same even without normalization for bacterial cell count
(Figure S35), except that 2a did not show any level of
activation, further underscoring the critical importance of
Moya2 (regardless of degree of saturation) for QS activation.
Moreover, these results are consistent with our docking results,
which suggest that unsaturation of doscadenamides would not
obviously impact binding.
Thus, we focused on compounds 3a−3c in our molecular
analysis. Figure 4A shows representative docked poses for 3a−
c, and Figure 4B displays the binding mode of the agonist 3a to
LasR. The binding modes of the other agonists (3b,c) are
similar and are shown in Figure S36. The amide bond in all
those ligands make the same interactions as in 1a, with the
compound anchoring in pocket by the two H-bonds between
Tyr56 and Asp73 and the Moya1 side chain, and the
pyrrolinone surrounded by the same residues as in 1a.
Furthermore, in the agonist compounds, the Moya2 side
chain extends toward the right side of the figure, similar to the
783
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
Figure 5. Doscadenamide A (1a) and its structural analogues sensitize breast cancer MDA-MB-231 cells in combination with TRAIL. (A) Dose−
response analysis of TRAIL on MDA-MB-231 cells. (B, C) MDA-MB-231 cell viability after 3 h pretreatment with DMSO (solvent control), C12
or doscadenamides, at (B) 50 and (C) 25 μM, followed by combined TRAIL (20 ng/mL) treatment for 24 h. (D) Δ Bliss independence
calculations for MDA-MB-231 cells cotreated with TRAIL (20 ng/mL) and doscadenamides. MDA-MB-231 cells were treated with doscadenamide
A (1a) and its structural analogues for 3 h, followed by treatment with TRAIL (20 ng/mL) for 24 h, C12 was included as positive control. “Δ Bliss
independence” is the difference between observed growth inhibition and Bliss expectation. Values greater than zero represent a synergistic response,
represented by red highlights. Bliss expectation is C = (A + B) − (A × B), where A and B are the growth inhibition fractions of two compounds at a
given dose. Cell viability was quantified using MTT assay. (E) A representative result of Western blot analysis of protein extracts from breast cancer
MDA-MB-231 cells after treatment with TRAIL (20 ng/mL), doscadenamide A (1a), and selected analogues (50 μM) and their combinations as
shown.
doscadenamide A case, establishing similar interactions with
the L3 loop and stabilizing the protein. However, missing the
N side chain, the antagonist 2a is incapable of interacting with
the L3 loop (Figure S37).
LuxR binds to the autoinducer and activates the transcription
of luciferase enzymes for bioluminescence production.21 In P.
aeruginosa, two pairs of LuxI/LuxR homologues, namely LasI/
LasR and RhlI/RhlR, exist to regulate the QS signaling
pathway.17 For these two signaling systems, the RhlI/RhlR
system is subordinate to the LasI/LasR system.21 Two
acylhomoserine lactones (AHLs), C12 and N-butyryl-homo-
serine lactone (C4) (Figure S33) are required to trigger the
expression of RhlR and its downstream target genes including
virulence factor production, while the expression of LasR is
independent of the AHL-mediated QS signaling pathway, thus
making LasR a representive target for QS inhibition.22
Compared to V. fischeri and P. aeruginosa, V. harveyi contains
a more complex QS cascade involving three parallel regulating
QS-signaling pathways and responds to three different
autoinducers.23,24 Even though the homologous LuxI/LuxR
system of V. fischeri has never been identified in V. harveyi, V.
harveyi can produce and detect autoinducer 1 (AI-1), N-(3-
hydroxybutanoyl) homoserine lactone,23 which shares signifi-
cant similarity to C12 and C4 (Figure S33). More importantly,
Tyr56 and Asp73 are conserved in both LasR and LuxR
systems. Therefore, we adopted V. harveyi as a practical model
QS Modulation by Doscadenamides in other Gram-
Negative Bacteria: Vibrio harveyi. We then aimed to
determine if doscadenamides modulate QS in other Gram-
negative bacteria. To extend our investigation of doscadena-
mides on QS modulation, we adopted a bioluminescent marine
bacterium, Vibrio harveyi, as our model system to examine the
effect of doscadenamides on modulating the bioluminescence
production. V. harveyi has been reported to be responsive to
bacterial QS activator N-(3-oxododecanoyl)-L-homoserine
lactone (C12) and thus serves as a robust model for quorum
sensing related research.16,17
Most Gram-negative bacteria possess QS circuits similar to
the marine symbiotic bioluminescent bacterium Vibrio fischeri,
the QS system which has been studied the most.18 As the V.
fischeri population grows, it produces and releases autoinducer
molecules to accumulate and thus eliciting bioluminescence.19
There are two proteins, LuxI and LuxR, in V. fischeri to regulate
the QS signaling pathway. LuxI regulates the production of the
autoinducer, N-(3-oxohexanoyl)-homoserine lactone (C6).20
784
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
to understand and validate the QS modulating activity of
doscadenamide A (1a) and its analogues (Figure 1).
the parent compound, doscadenamide A (1a). As expected, the
cyclized version 4a was completely inactive in these assays.
Compared to doscadenamide A (1a), the analogues with lower
unsaturation number (1f, 1k, 3b, and 3c) not only performed
better as synergistic agents with TRAIL but also exhibited
inherent weak cytotoxic effects on MDA-MB-231 cells. This
finding provides an approach to optimize doscadenamides with
improved cytotoxic activities, although the active concentration
is still relatively high.
Almost all compounds activated QS in this biological
context to a certain extent (Figure 3C), with 1a and 3a
activating across the concentration range tested (10−200 μM).
However, 3c only had a minor activating trend and for 1f and
1j the activities at ≤100 μM were also reduced, suggesting that
increased saturation of the terminal units decreases activity for
both chains. Most notably, 2a was not able to activate the
bioluminescent response in V. harveyi but, instead, consistently
inhibited QS signaling in V. harveyi, whereas the same
compound was essentially inactive in P. aeruginosa. In both
cases, 2a was not able to activate QS, providing common
ground in the two Gram-negative bacteria. The same trend in
V. harveyi was observed without normalization for cell number
(Figure S38), indicating that the differential effect is not an
artifact of normalization.
These results indicate that both acylations in doscadenamide
A are involved in the QS modulating activity, with partially
overlapping but also bacteria-specific trends. Both chains
contribute differently to the final biological activity output,
even in opposing or interacting fashion, suggesting a more
complex interaction between doscadenamide A and its
molecular targets in various bacterial systems, potentially
regulating their respective activity.
Synergy with TRAIL in MDA-MB-231 Breast Cancer
Cells. TRAIL has been regarded as an attractive therapeutic
anticancer agent; however, many cancer cells have developed
resistance to TRAIL, impacting the treatment efficacy.4 Given
that C12 was reported to exhibit a synergistic effect on TRAIL-
induced apoptosis in cancer cells,7 we examined the potential
synergistic effect of doscadenamide A (1a) and its analogues
on sensitizing cancer cells. First, we generated the dose−
response curve of TRAIL in MDA-MB-231 breast cancer cells,
which were partially responsive to TRAIL. Cell viability
remained 50% even at the highest concentration (500 ng/mL)
after 24 h treatment (Figure 5A). A concentration of 20 ng/
mL was selected, under which TRAIL would reduce cell
viability by approximately 20% (Figure 5A). MDA-MB-231
cells were pretreated with C12, doscadenamides (25 or 50
μM) or solvent control for 3 h and then stimulated with
TRAIL for 24 h (Figure 5B,C). Cell viability was measured and
ΔBliss independence of each compound with TRAIL in a
doscadenamide concentration-dependent manner calculated.
Most analogues exhibited synergistic effect with TRAIL at 50
μM, except 2a (Figure 1), and among all the analogues, the
“saturated” analogues 1k and 3c (Figure 1) behaved as the best
agents to sensitize MDA-MB-231 cells (Figure 5D). To further
investigate the pro-apoptotic effect of doscadenamide A (1a,
Figure 1), the poly(ADP-ribose) polymerase (PARP) cleavage
was also analyzed using Western blot (Figure 5E) to indicate
programmed cell death. The decrease in cell viability correlated
with the presence of cleaved PARP in cell extracts after
cotreatment, suggesting that doscadenamide A (1a, Figure 1)
and its analogues can effectively improve TRAIL-induced
apoptosis in MDA-MB-231 cells. Collectively, our SAR study
revealed that low degrees of unsaturation lead to high potency
of the analogues to ameliorate TRAIL resistance. Importantly,
the R2 side chain is more critical for the bioactivities of
doscadenamide analogues than the R1 side chain (Figure 1);
2a showed no effect on cell viability (Figure 5B,C) or effect on
PARP cleavage (Figure 5E). Compound 3c showed enhanced
activity compared with the doubly acylated versions, including
DISCUSSION
■
Providing perspective of our findings, doscadenamides were
isolated from a natural product “superproducer” with
tremendous biosynthetic capability, Moorena bouillonii from
Apra Harbor, Guam,25 which previously yielded several
apratoxins,26−29 lyngbyabellins,30−33 lyngbyalosides,33,34 lain-
golide B,33 apramides,35,36 lyngbyastatin 2,37 lyngbyapep-
tins,32,33 and apratyramide.38 The lipopeptides apramides
possess partial structural similarity to doscadenamides.
Doscadenamide A (1a) possesses two C8-alkynoic acid
moieties (Moya) that are present in apramides35 and several
natural products from marine cyanobacteria, representing a
biosynthetic signature motif.35,39−43 Both units in doscadena-
mides appear to play complementary or opposing functions in
different biological contexts, overall modulating the biological
outcome. Similar to apramides A−G,35 natural diversification
by the cyanobacterial biosynthetic machinery is exemplified by
the production of numerous additional doscadenamides (B−J,
1b−1j) with minor structural changes that fine-tune
bioactivities. To further diversify the doscadenamide family,
structurally and functionally, we have developed a robust
scheme to synthesize different structural and configurational
doscadenamide analogues, including different degrees of
unsaturation (1k−n, 2a−b, 3a−c). Notably, a related
pyrrolinone core with two different pendant side chains is
also the scaffold of the sponge-derived sintokamides, and
natural and synthetic diversification has been reported44,45 for
this class of natural products.
The degree of terminal unsaturation in C8 carboxylic acid
chains has previously been demonstrated to modulate activity
toward prokaryotic and eukaryotic cells. For instance,
pitipeptolides A, D−F39 (alkynoic side chain) and pitipepto-
lides B (alkenoic acid acid) and C (alkanoic acid chain) were
reported to display differential cancer cytotoxic activity and
antimycobacterial activity, depending on the degree of
unsaturation. Both doscadenamides and pitipeptolides can be
regarded as natural product examples that show that different
degrees of unsaturation lead to fine-tuned biological activities.
We have shown that doscadenamides can modulate the
quorum sensing signaling pathway in P. aeruginosa, acting as
non-HSL inducers. An HSL-type QS inducer bearing a C8
(saturated) carboxylic acid (C8-HSL) has been previously
identified from a cyanobacterium,22,46,47 while other (partic-
ularly branched/substituted) cyanobacterial carboxylic acids or
derivatives have usually been found to act as inhibitors of
quorum sensing.22 To further understand the molecular
interaction of doscadenamide A and its bacterial receptor
LasR in P. aeruginosa, we performed interactive mutation and
molecular docking studies. Molecular docking is a powerful
method to model the interactions between small molecules and
proteins. There are, however, two main limitations that must
be considered in such a study. The first limitation relates to the
standard docking procedure where different conformations of
785
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
the ligand are tested against the same structure of a rigid target
whereas, in reality, the target may adjust to better
accommodate each ligand. To circumvent this limitation, we
included the effect of target flexibility by first using molecular
dynamics to sample the conformational space of LasR,
followed by clustering to select 15 different structures of the
protein, to which the ligands were docked. Another important
limitation is in the approximate nature of the scoring functions
used, which are generally unable to reliably discriminate
between ligand poses with similar binding energies. Here we
collected all the structures within 3 kcal/mol of the lowest
binding energy as reasonable structures, and selected the ones
showing the interactions that were experimentally determined
to be important.
We identified two amino acid residues in LasR from P.
aeruginosa that play a critical role in binding with
doscadenamide A and performed molecular modeling for
further understanding of the binding mode between
doscadenamides and LasR. The experimental mutagenesis
results show that Tyr56 and Asp73 are required for the binding
of 1a to LasR. The selected docked structures confirm the
importance of residues Tyr56 and Asp73 for ligand
recognition, by establishing H-bonds to the atoms in the
amide bond in the Moya1 side chain and anchoring the
ligands. The docking results show very similar interaction
patterns of these with all the ligands, indicating that these
interactions must be important for ligand recognition, but not
necessarily for activation. This is in agreement with Zou and
Nair,13 who concluded (based on the crystal structures) that
activation is a result of the interactions of the long
hydrophobic tail of the autoinducer with the residues of the
L3 loop, which closes the active site and stabilizes LasR,
allowing dimerization.13 In agreement with this mechanism,
the results show that the Moya2 side chains in the agonists 1a
and 3a−c are able to establish the required interactions with
the L3 loop to stabilize the structure and promote
dimerization, activating LasR. On the other hand, the 2a is
still able to bind but, missing this side chain, is incapable of
forming the necessary interactions to promote dimerization
and activation (Figure S37).
The mutagenesis analysis here did not confirm the
requirement of interactions to Trp60 for ligand binding, and
the dockings show that none of the ligands considered here
establish H-bonds to Trp60. The analysis did show, however, a
dose-dependent effect on the W60F mutant. The Trp60
residue is located on the back of the binding pocket, and most
of its contacts with the ligand are hydrophobic in nature. The
MD simulations also show interactions of Trp60, which is
located in the α3 helix, with residues in the opposed α5 helix
(95−108), which stabilize the α5 helix. In the W60F mutant,
both sides become more mobile. It was suggested that the local
structure around Trp60 might be in contact with transcription
factors and, thus, involved in the process of transcription-
activation.13 This increased flexibility of the binding site may
be responsible for the slight increased response of the W60F
mutant with respect to the WT by interfering with this
mechanism. Finally, although these results agree with the
experimental data on LasR, it is important to note that other
factors that cannot be assessed by the techniques used here
must play an important role, such as access to the binding site
and neighbor interactions in the biological assembly.
Furthermore, some of the experimental results here were
obtained for a different system, Vibrio harveyi. Unfortunately,
there is still not enough information about the receptor in this
organism to build a reliable model. Nevertheless, we have
confirmed that the Moya2 side chain plays an important role in
the interactions of doscadenamides with LasR, largely
contributing to the bioactivity in Gram-negative bacteria.
These compounds also have effects against mammalian cells,
although the eukaryotic target remains to be determined.
Importantly, the improved efficiency of TRAIL-induced
apoptosis after pretreatment with doscadenamides would
provide a new template to the development of therapeutic
agents toward anticancer therapy to alleviate and overcome
drug resistance. Sensitizing cancer cells to chemotherapy is an
important concept to improve the therapeutic window of
treatments.48 We have shown that doscadenamides synergize
with TRAIL, an inducer of the extrinsic apoptotic pathway.49
The dual activity on QS and cancer cells reveal an interesting
interplay of molecular pathways in bacteria and eukaryotes.
Through total synthesis, we were able to discover new
doscadenamide analogues to display improved anticancer
activity, especially (or only) when combined with TRAIL.
From a medicinal chemistry perspective, these analogues
revealed a critical element to optimize the in vitro activity,
although the potency is low and floppiness may hinder
significant improvement. Nevertheless, cotreatment of dosca-
denamide-type agents and TRAIL may provide a template for a
new treatment modality. On the other hand, doscadenamides
with more potent cytotoxic activity were discovered through
efficient total synthesis, which demonstrates our developed
synthetic route as a robust approach to produce diverse
analogues with various biological activities.50
EXPERIMENTAL SECTION
■
General Experimental Procedures. All commercial reagents
were used without further purification unless otherwise noted. Flash
column chromatography was performed with Fisher 170−400 mesh
silica gel. NMR spectra were recorded on a Bruker Avance III 600
MHz spectrometer or a Bruker Avance Neo-600 spectrometer with a
1
broadband Prodigy cryogenic probe. Chemical shifts for H NMR
spectra are reported in ppm relative to the signal residual CDCl3 at
7.26 ppm. Chemicals shifts for 13C NMR spectra are reported in ppm
relative to the center line of the CDCl3 triplet at 77.16 ppm. Optical
rotation was measured on a PerkinElmer 341 polarimeter (Na D line)
using a microcell of 1 dm path length. HRMS was conducted using a
Thermo Fisher Q Exactive Focus mass spectrometer.
Biological Material. The cyanobacterium Moorena bouillonii25
was collected in May 2005 by SCUBA in shallow waters at Finger
Reef, Apra Harbor, Guam. This is a recollection of VP417 (16S rRNA
gene sequence deposited under GenBank accession nos. AY049750
and AY049751)27 and previously yielded doscadenamide A.9
voucher specimen is located at the Smithsonian Marine Station.
A
Extraction and Isolation. The sample was extracted with CH2Cl2
and MeOH (2:1) and the extract (10 g) was fractionated using
column chromatography on silica gel, eluting with CH2Cl2 containing
increasing concentrations of iPrOH to afford 16 fractions. Fractions 3,
4, and 5 (2% iPrOH in CH2Cl2; 400 mg, 1.54 g, and 100 mg) and
fractions 6 (5% iPrOH in CH2Cl2; 71 mg) were individually subjected
to semipreparative HPLC (Phenomenex Phenyl-hexyl, 250 × 10 mm,
5 μ, 2.0 mL/min; PDA detection) using a MeOH−H2O linear
gradient (90−100% MeOH in 30 min and 100% MeOH for 10 min).
1
Fractions were pooled based on retention times, H NMR analysis,
and low-resolution MS measurements to afford impure doscadena-
mide A (1a) containing fractions. These fractions were further
purified with HPLC (Ultracarb, 250 × 10 mm, 5 μ, 2.0 mL/min; PDA
detection) using a MeOH−H2O linear gradient (90−100% MeOH in
30 min and 100% MeOH for 10 min) to afford a series of compounds
(1a, 4.5 mg, tR = 11.0 min; 1b, 0.3 mg, tR = 10.0 min; 1c, 0.3 mg, tR =
786
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
13.7 min; 1d, 0.5 mg, tR = 14.7 min; 1e, 0.2 mg, tR = 11.9 min; 1f, 0.2
mg, tR = 13.9 min; 1g, 0.3 mg, tR = 12.4 min; 1h, present in a mixture,
0.2 mg, tR = 12.6 min; 1i/1j, binary mixture, 0.6 mg, tR = 9.4 min).
reader. AI-1 and DMSO were used as positive control and negative
control, respectively.
Synergistic Effect Evaluation of Doscadenamide Analogues
with TRAIL. MDA-MB-231 cells were seeded in 12-well plates at a
density of 100 000 cells per well. The next day, cells were pretreated
with doscadenamide analogues (50 μM and 25 μM), positive controls
(C12, 50 μM) or solvent control (DMSO) for 3 h. Then the cells
were treated with TRAIL (20 ng/mL) for 24 h. The whole cell lysates
were collected using PhosphoSafe Buffer (EMD Chemicals). Protein
concentrations were measured with the BCA Protein Assay kit
(Thermo Fisher Scientific). Lysates containing equal amounts of
protein were separated by NuPAGE 4−12% Bis-Tris protein gels
(Thermo Fisher Scientific), transferred to polyvinylidene difluoride
membranes, probed with primary and secondary antibodies. The
membranes were visualized using Li-Cor imaging system. PARP and
actin antibodies were from Cell Signaling. Secondary mouse and
rabbit antibodies were from Invitrogen.
Ensemble Docking to LasR. The initial structure for LasR was
obtained from its crystal structure in complex with the autoinducer
mimic TP1 (PDBID: 3IX4), obtained at 1.8 Å resolution.13 The
protonation states of histidines were determined by the H++ server at
pH 7.3 (http://biophysics.cs.vt.edu/H++).51 Mutants were obtained
by mutating the corresponding residue in the PDB file before
preparing the system.
All MD calculations used a similar protocol with the programs from
the AMBER18 suite.52 The protein was neutralized by adding 7 Na+
ions and placed in an octahedral box with 7405 TIP3P water
molecules, ensuring a minimal distance of 10 Å between the protein
atoms and the box limits. The system was then submitted to two
stages of energy minimization. First, only water molecules were
allowed to move for 100 steps. Then all atoms were allowed to move
for 10 000 minimization steps. After that the system was heated from
0 to 310 K during 600 ps at constant volume, then kept at the final
temperature for an extra 400 ps. This was followed by 2 ns
equilibration at 310 K and constant pressure with the Monte Carlo
barostat and 2 ps relaxation time. Finally, data was acquired during
two 100 ns runs with the same conditions (total 200 ns production).
All MD runs used a Langevin thermostat with collision frequency of
2.0 ps−1, a 2 fs time step, and a 10 Å cutoff for nonbonded
interactions. The trajectory obtained from the last 200 ns of
simulation was split into 15 clusters using a hierarchical agglomerative
algorithm, based on the root-mean-square distance between the
frames, considering only the residues around the binding site, with the
program CPPTRAJ, part of the AMBER18 suite. For each cluster, a
representative structure was selected as the structure with the lowest
cumulative distance to all other frames in the same cluster. All the
input files used are in the Supporting Information.
Doscadenamide A (1a). White solid; [α]20 + 40 (c 0.07,
D
MeOH)].9
Doscadenamide B (1b). White solid; [α]20 + 127 (c 0.05,
D
MeOH); 1H and 13C NMR data, see Table S1; HRESIMS m/z [M +
H]+ 455.2897 (calcd for C27H39N2O4, 455.2904).
Doscadenamide C (1c). White solid; [α]20 + 116 (c 0.02,
D
MeOH); 1H and 13C NMR data, see Table S2; HRESIMS m/z [M +
H]+ 459.3218 (calcd for C27H43N2O4, 459.3217).
Doscadenamide D (1d). White solid; [α]20 + 118 (c 0.03,
D
MeOH); 1H and 13C NMR data, see Table S3; HRESIMS m/z [M +
H]+ 459.3226 (calcd for C27H43N2O4, 459.3217).
Doscadenamide E (1e). White solid; [α]20 + 115 (c 0.02,
D
MeOH); 1H and 13C NMR data, see Table S4; HRESIMS m/z [M +
H]+ 445.3069 (calcd for C26H41N2O4, 445.3061).
Doscadenamide F (1f). White solid; [α]20 + 114 (c 0.03,
D
MeOH); 1H and 13C NMR data, see Table S5; HRESIMS m/z [M +
H]+ 461.3362 (calcd for C27H45N2O4, 461.3374).
Doscadenamide G (1g). White solid; [α]20 + 110 (c 0.06,
D
MeOH); 1H and 13C NMR data, see Table S6; HRESIMS m/z [M +
H]+ 445.3063, (calcd for C26H41N2O4, 445.3061).
Doscadenamide H (1h). White solid; HRESIMS m/z [M + H]+
445.3063, (calcd for C26H41N2O4, 445.3061).
Doscadenamide I/J (1i/1j). White solid; H and 13C NMR data,
1
see Table S7; HRESIMS m/z [M + H]+ 443.2916, (calcd for
C26H39N2O4, 443.2910).
Quorum Sensing Reporter Assays with Doscadenamide A
(1a). 1a, in a dose−response manner, and EtOH control were added
to the corresponding wells in a 96-well plate and the solvent was
allowed to evaporate. Then 100 μL of an overnight culture of E. coli
expressing plasmid wild-type pSB1075 (a luxCDABE reporter
construct encoding LasR) or an overnight E. coli culture expressing
reporter pSB1075 mutant (Y56F, W60F or D73N; obtained from Dr.
Max Teplitski, UF), was added to each well. The plate was incubated
at 37 °C for 6 h before the measurement of luminescence.
Pyocyanin Quantification in P. aeruginosa Strain PAO-JP1.
An overnight culture of P. aeruginosa strain PAO-JP1 (obtained from
Dr. Everett C. Pesci, Department of Microbiology and Immunology,
Brody School of Medicine, East Carolina University, and Dr. Abdul
Hamood, Department of Surgery and Immunology and Molecular
Microbiology, School of Medicine, Texas Tech University Health
Science Center) was diluted 10-fold before being transferred to
another culture tube containing 890 μL LB broth and 10 μL testing
compound (10 μM final concentration) or EtOH control and
incubated at 37 °C with shaking for 6 h. The culture was then spun
down for 10 min at the maximum speed, and the supernatant was
collected and filtered using 0.2 μM Eppendorf filters. A 500 μL
portion of the sterile supernatant was added to 500 μL CHCl3 in an
Eppendorf tube. Tube shaking allowed for the extraction of pyocyanin
in the CHCl3 layer. This layer was then added to 150 μL of 0.2 N HCl
in another Eppendorf tube. After shaking, the aqueous layer and the
organic layer were separated. A 100 μL portion of the aqueous layer
was transferred to a clear bottomed 384-well plate, and the
absorbance was measured at 385 nm to quantify the pyocyanin
production. Data was analyzed using GraphPad Prism 5 software.
Bioluminescence Modulation in V. harveyi. Bioluminescent V.
harveyi BB120 strain (wild-type, obtained from Dr. Julie L. Meyer,
Soil and Water Sciences Department, Institute of Food and
Agricultural Sciences, University of Florida) was cultivated in AB
media overnight (15−16 h) at 30 °C (OD600 2.0−2.5). The
overnight cultivated strain (2 μL) was diluted to 10 mL by AB media
(OD600 < 0.1). A 100 μL portion of the diluted BB120 strain was
distributed to each well of 96-well white plate. The solution of each
tested compound in DMSO (0.5 μL) at different concentrations was
added to BB120 cultures on plate. Triplicate analysis was conducted
for each compound and each concentration. After addition, bacterial
cultures were incubated at 30 °C for 7 h. Then, the emitted
luminescence and OD600 was measured on a SpectraMax M5 plate
All dockings were done in triplicate with AutoDock Vina 4.2,53
with exhaustiveness of 128 and a grid box of dimensions 18.75,
24.375, and 18.75 Å (x, y, z), centered on the binding site. Results
were analyzed using a combination of UCSF Chimera,54 Jupyter
notebooks (Python), and in-house scripts, available at GitHub.
(https://github.com/gmseabra/lasr). Only poses within 3 kcal/mol
from the pose with lowest Vina score in each case were considered.
The final poses shown were selected to display the experimentally
validated interactions.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.jnatprod.0c01003.
Experimental details of the chemical synthesis, Tables
S1−7, and Figures S1−107, including NMR spectra and
assay data (PDF)
787
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
reporter assays. We also thank Dr. Everett C. Pesci
(Department of Microbiology and Immunology, Brody School
of Medicine, East Carolina University) and Dr. Abdul Hamood
(Department of Surgery and Immunology and Molecular
Microbiology, School of Medicine, Texas Tech University
Health Science Center) for providing the PAO-JP1 strains.
AUTHOR INFORMATION
Corresponding Author
Hendrik Luesch − Department of Medicinal Chemistry and
Center for Natural Products, Drug Discovery and
Development (CNPD3), University of Florida, Gainesville,
Florida 32610, United States; orcid.org/0000-0002-
4091-7492; Email: luesch@cop.ufl.edu
■
DEDICATION
■
Authors
Dedicated to Dr. A. Douglas Kinghorn, The Ohio State
University, for his pioneering work on bioactive natural
products.
Xiao Liang − Department of Medicinal Chemistry and Center
for Natural Products, Drug Discovery and Development
(CNPD3), University of Florida, Gainesville, Florida 32610,
United States
Qi-Yin Chen − Department of Medicinal Chemistry and
Center for Natural Products, Drug Discovery and
Development (CNPD3), University of Florida, Gainesville,
Florida 32610, United States
Gustavo M. Seabra − Department of Medicinal Chemistry and
Center for Natural Products, Drug Discovery and
Development (CNPD3), University of Florida, Gainesville,
Florida 32610, United States; orcid.org/0000-0002-
1303-4483
Susan Matthew − Department of Medicinal Chemistry,
University of Florida, Gainesville, Florida 32610, United
States
Jason C. Kwan − Department of Medicinal Chemistry,
University of Florida, Gainesville, Florida 32610, United
States
Chenglong Li − Department of Medicinal Chemistry and
Center for Natural Products, Drug Discovery and
Development (CNPD3), University of Florida, Gainesville,
Florida 32610, United States
Valerie J. Paul − Smithsonian Marine Station, Fort Pierce,
Florida 34949, United States; orcid.org/0000-0002-
4691-1569
REFERENCES
■
(1) von Karstedt, S.; Montinaro, A.; Walczak, H. Nat. Rev. Cancer
2017, 17, 352−366.
(2) Srivastava, R. K. Neoplasia 2001, 3, 535−546.
(3) French, L.; Tschopp, J. Nat. Med. 1999, 5, 146−147.
(4) Trivedi, R.; Mishra, D. P. Front. Oncol. 2015, 5, 69.
(5) Amarante-Mendes, G.; Griffith, T. Pharmacol. Ther. 2015, 155,
117−131.
(6) Legler, K.; Hauser, C.; Egberts, J. H.; Willms, A.; Heneweer, C.;
Boretius, S.; Röcken, C.; Gluer, C. C.; Becker, T.; Kluge, M.; Hill, O.;
̈
Gieffers, C.; Fricke, H.; Kalthoff, H.; Lemke, J.; Trauzold, A. Cell
Death Dis. 2018, 9, 445.
(7) Kravchenko, V.; Garner, A. L.; Mathison, J.; Seit-Nebi, A.; Yu, J.;
Gileva, I. P.; Ulevitch, R.; Janda, K. D. ACS Chem. Biol. 2013, 8,
1117−1120.
(8) The Spanish historic influence on the collection site Guam
coupled with the structural features, the acylation with two (“dos”)
carboxylic acid chains (“cadenas”) to form the final amide, led us to
coin the name of this compound class.
(9) Liang, X.; Matthew, S.; Chen, Q. Y.; Kwan, J. C.; Paul, V. J.;
Luesch, H. Org. Lett. 2019, 21, 7274−7278.
(10) Kiratisin, P.; Tucker, K. D.; Passador, L. J. Bacteriol. 2002, 184,
4912−4919.
(11) Swartzman, E.; Silverman, M.; Meighen, E. A. J. Bacteriol. 1992,
174, 7490−7493.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.jnatprod.0c01003
(12) Winson, M. K.; Swift, S.; Fish, L.; Throup, J. P.; Jørgensen, F.;
Chhabra, S. R.; Bycroft, B. W.; Williams, P.; Stewart, G. S. FEMS
Microbiol. Lett. 1998, 163, 185−192.
Author Contributions
(13) Zou, Y.; Nair, S. K. Chem. Biol. 2009, 16, 961−970.
(14) Zhao, J.; Jiang, H.; Cheng, W.; Wu, J.; Zhao, J.; Wang, J.; Dong,
L. J. Basic Microbiol. 2015, 55, 671−678.
X.L.: total synthesis, biological experiments with MDA-MB-
231, experimental design, and manuscript writing. Q.-Y.C.:
supervision of total synthesis and biological experiments with
PAO-JP1 and V. harveyi. G.M.S.: molecular docking. S.M.:
isolation and structure elucidation. J.C.K.: biological evaluation
using reporter assays. C.L.: supervision of molecular docking.
V.J.P.: sample collection and identification and manuscript
editing. H.L.: project supervision, experimental design, and
manuscript editing. All authors have given approval to the final
version of the manuscript.
(15) Smith, R. S.; Harris, S. G.; Phipps, R.; Iglewski, B. J. Bacteriol.
2002, 184, 1132−1139.
(16) Aharoni, R.; Bronstheyn, M.; Jabbour, A.; Zaks, B.; Srebnik, M.;
Steinberg, D. Bioorg. Med. Chem. 2008, 16, 1596−1604.
(17) Naik, S. P.; Scholin, J.; Ching, S.; Chi, F.; Herpfer, M. J. Agric.
Food Chem. 2018, 66, 40−44.
(18) Miller, M.; Bassler, B. Annu. Rev. Microbiol. 2001, 55, 165−199.
(19) Engebrecht, J.; Nealson, K.; Silverman, M. Cell 1983, 32, 773−
781.
(20) Eberhard, A.; Burlingame, A. L.; Eberhard, C.; Kenyon, G. L.;
Nealson, K. H.; Oppenheimer, N. J. Biochemistry 1981, 20, 2444−
2449.
Notes
The authors declare the following competing financial
interest(s): The University of Florida has filed a patent
application related to the subject of the manuscript.
(21) Engebrecht, J.; Silverman, M. Proc. Natl. Acad. Sci. U. S. A.
1984, 81, 4154−4158.
ACKNOWLEDGMENTS
(22) Kwan, J. C.; Meickle, T.; Ladwa, D.; Teplitski, M.; Paul, V.;
Luesch, H. Mol. BioSyst. 2011, 7, 1205−1216.
■
We acknowledge financial support from the National Institutes
of Health (National Cancer Institute grant R01CA172310)
and the Debbie and Sylvia DeSantis Chair professorship
(H.L.). We thank Dr. Max Teplitski and Dr. Julie Meyer (Soil
and Water Sciences Department, Institute of Food and
Agricultural Sciences, University of Florida) for providing E.
coli reporter and V. harveyi strains and support with the
(23) Henke, J. M.; Bassler, B. L. J. Bacteriol. 2004, 186, 6902−6914.
(24) Waters, C. M.; Bassler, B. L. Genes Dev. 2006, 20, 2754−2767.
(25) Tronholm, A.; Engene, N. Notulae Algarum 2019, 122, 1−2.
(26) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.; Corbett,
T. H. J. Am. Chem. Soc. 2001, 123, 5418−5423.
(27) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J. Bioorg.
Med. Chem. 2002, 10, 1973−1978.
788
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789

Journal of Natural Products
pubs.acs.org/jnp
Article
(28) Matthew, S.; Schupp, P. J.; Luesch, H. J. Nat. Prod. 2008, 71,
1113−1116.
Omelyan, I.; Onufriev, A.; Pan, F.; Qi, R.; Roe, D. R.; Roitberg, A.;
Sagui, C.; Schott-Verdugo, S.; Shen, J.; Simmerling, C. L.; Smith, J.;
Salomon-Ferrer, R.; Swails, J.; Walker, R. C.; Wang, J.; Wei, H.; Wolf,
R. M.; W, X.; Xiao, L.; York, D. M.; Kollman, P. A. Amber18 /
AmberTools18; University of California: San Francisco, 2018.
(53) Trott, O.; Olson, A. J. J. Comput. Chem. 2009, 31, 455−461.
(54) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.;
Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. J. Comput. Chem. 2004,
25, 1605−1612.
(29) Wu, P.; Cai, W.; Chen, Q. Y.; Xu, S.; Yin, R.; Li, Y.; Zhang, W.;
Luesch, H. Org. Lett. 2016, 18, 5400−5403.
(30) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.; Mooberry,
S. L. J. Nat. Prod. 2000, 63, 611−615.
(31) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J. J. Nat.
Prod. 2000, 63, 1437−1439.
(32) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.
Tetrahedron 2002, 58, 7959−7966.
NOTE ADDED AFTER ASAP PUBLICATION
(33) Matthew, S.; Salvador, L. A.; Schupp, P. J.; Paul, V. J.; Luesch,
H. J. Nat. Prod. 2010, 73, 1544−1552.
■
This paper was published ASAP on January 22, 2021, with a
compound naming error in the Abstract text and in a Results
heading. The corrected version was reposted on March 3,
2021.
(34) Chang, C. F.; Stefan, E.; Taylor, R. E. Chem. - Eur. J. 2015, 21,
10681−10686.
(35) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J. J. Nat.
Prod. 2000, 63, 1106−1112.
(36) Williams, P. G.; Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul,
V. J. J. Nat. Prod. 2003, 66, 595−598.
(37) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J. J. Nat.
Prod. 1999, 62, 1702−1706.
(38) Cai, W.; Salvador-Reyes, L. A.; Zhang, W.; Chen, Q. Y.;
Matthew, S.; Ratnayake, R.; Seo, S. J.; Dolles, S.; Gibson, D. J.; Paul,
V. J.; Luesch, H. ACS Chem. Biol. 2018, 13, 91−99.
(39) Montaser, R.; Paul, V. J.; Luesch, H. Phytochemistry 2011, 72,
2068−2074.
(40) Salvador, L. A.; Biggs, J. S.; Paul, V. J.; Luesch, H. J. Nat. Prod.
2011, 74, 917−927.
(41) Mevers, E.; Liu, W. T.; Engene, N.; Mohimani, H.; Byrum, T.;
Pevzner, P. A.; Dorrestein, P. C.; Spadafora, C.; Gerwick, W. H. J.
Nat. Prod. 2011, 74, 928−936.
(42) Gunasekera, S. P.; Ross, C.; Paul, V. J.; Matthew, S.; Luesch, H.
J. Nat. Prod. 2008, 71, 887−890.
(43) Gunasekera, S. P.; Owle, C. S.; Montaser, R.; Luesch, H.; Paul,
V. J. J. Nat. Prod. 2011, 74, 871−876.
(44) Sadar, M. D.; Williams, D. E.; Mawji, N. R.; Patrick, B. O.;
Wikanta, T.; Chasanah, E.; Irianto, H. E.; Soest, R. V.; Andersen, R. J.
Org. Lett. 2008, 10, 4947−4950.
(45) Yan, L.; Banuelos, C. A.; Mawji, N. R.; Patrick, B. O.; Sadar, M.
D.; Andersen, R. J. J. Nat. Prod. 2020, DOI: 10.1021/acs.jnat-
prod.0c00921.
(46) Sharif, D. I.; Gallon, J.; Smith, C. J.; Dudley, E. ISME J. 2008, 2,
1171−1182.
(47) Kwan, J. C.; Teplitski, M.; Gunasekera, S. P.; Paul, V. J.;
Luesch, H. J. Nat. Prod. 2010, 73, 463−466.
(48) Leary, M.; Heerboth, S.; Lapinska, K.; Sarkar, S. Cancers 2018,
10, 483.
(49) Sp, N.; Kang, D. Y.; Kim, D. H.; Yoo, J. S.; Jo, E. S.; Rugamba,
A.; Jang, K. J.; Yang, Y. M. Anticancer Res. 2020, 40, 3209−3220.
(50) (a) While the manuscript was under review, doscadenamide A
was also described from other M. bouillonii collections by Leber et
al.50b In that paper, structures of doscadenamide analogues
(nomenclature “B−J”) were proposed merely based on low-resolution
MS2 fragmentation data, without isolation and rigorous structural
analysis. These cannot be seriously considered as characterized
natural products and remain hypothetical, so we did not take them
into consideration. Our nomenclature (Figure 1) is also consistent
with the patent literature: PCT/US2020/043495 (priority to July 26,
2019). (b) Leber, C. A.; Naman, C. B.; Keller, L.; Almaliti, J.; Caro-
Diaz, E. J. E.; Glukhov, E.; Joseph, V.; Sajeevan, T. P.; Reyes, A. J.;
Biggs, J. S.; Li, T.; Yuan, Y.; He, S.; Yan, X.; Gerwick, W. H. Mar.
Drugs 2020, 18, 515.
(51) Anandakrishnan, R.; Aguilar, B.; Onufriev, A. V. Nucleic Acids
Res. 2012, 40 (Web Serverissue), W537−W541.
(52) Case, D. A.; Ben-Shalom, I. Y.; Brozell, S. R.; Cerutti, D. S.;
Cheatham, I. T. E.; Cruzeiro, V. W. D.; Darden, T. A.; Duke, R. E.;
Ghoreishi, D.; Gilson, M. K.; Gohlke, H.; Goetz, A. W.; Harris, R.;
Homeyer, N.; Izadi, S.; Kovalenko, A.; Kurtzma, T.; Lee, T. S.;
LeGrand, S.; Li, P.; Lin, C.; Liu, J.; Luchko, T.; Luo, R.; Mermelstein,
D. J.; Merz, K. M.; Miao, Y.; Monard, G.; Nguyen, C.; Nguyen, H.;
789
https://dx.doi.org/10.1021/acs.jnatprod.0c01003
J. Nat. Prod. 2021, 84, 779−789