Actinoramide A Identified as a Potent Antimalarial from Titration-Based Screening of Marine Natural Product Extracts

Article abstract of DOI:10.1021/acs.jnatprod.5b00489

Methods to identify the bioactive diversity within natural product extracts (NPEs) continue to evolve. NPEs constitute complex mixtures of chemical substances varying in structure, composition, and abundance. NPEs can therefore be challenging to evaluate efficiently with high-throughput screening approaches designed to test pure substances. Here we facilitate the rapid identification and prioritization of antimalarial NPEs using a pharmacologically driven, quantitative high-throughput-screening (qHTS) paradigm. In qHTS each NPE is tested across a concentration range from which sigmoidal response, efficacy, and apparent EC₅₀s can be used to rank order NPEs for subsequent organism reculture, extraction, and fractionation. Using an NPE library derived from diverse marine microorganisms we observed potent antimalarial activity from two Streptomyces sp. extracts identified from thousands tested using qHTS. Seven compounds were isolated from two phylogenetically related Streptomyces species: Streptomyces ballenaensis collected from Costa Rica and Streptomyces bangulaensis collected from Papua New Guinea. Among them we identified actinoramides A and B, belonging to the unusually elaborated nonproteinogenic amino-acid-containing tetrapeptide series of natural products. In addition, we characterized a series of new compounds, including an artifact, 25-epi-actinoramide A, and actinoramides D, E, and F, which are closely related biosynthetic congeners of the previously reported metabolites.

Full text of DOI:10.1021/acs.jnatprod.5b00489

Article
pubs.acs.org/jnp
Actinoramide A Identified as a Potent Antimalarial from Titration-
Based Screening of Marine Natural Product Extracts
Ken Chih-Chien Cheng,†,□ Shugeng Cao,‡,§,□ Avi Raveh,⊥,□ Ryan MacArthur,† Patricia Dranchak,†
George Chlipala,⊥ Matthew T. Okoneski,⊥,∥ Rajarshi Guha,† Richard T. Eastman,△ Jing Yuan,△
Pamela J. Schultz,⊥ Xin-zhuan Su,△ Giselle Tamayo-Castillo,○ Teatulohi Matainaho,# Jon Clardy,§
David H. Sherman,*,⊥,∥ and James Inglese*,†
†National Center of Advancing Translational Sciences, National Institutes of Health, Rockville, Maryland 20850, United States
‡Department of Pharmaceutical Sciences, Daniel K. Inouye College of Pharmacy, University of Hawai’i at Hilo, Hilo, Hawaii 96720,
United States
§Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, 240 Longwood Avenue, C-643, Boston,
Massachusetts 021151, United States
⊥Life Sciences Institute, University of Michigan, Ann Arbor, Michigan 48109-2216, United States
∥Departments of Medicinal Chemistry, Chemistry, and Microbiology & Immunology, University of Michigan, Ann Arbor, Michigan
48109-2216, United States
△Laboratory of Malaria and Vector Research, National Institute of Allergy and Infectious Diseases, National Institutes of Health,
Bethesda, Maryland 20852, United States
○Unidad Estrategica de Bioprospeccion
́
, Instituto Nacional de Biodiversidad (INBio), Santo Domingo de Heredia, Costa Rica &
CIPRONA-Escuela de Química, Universidad de Costa Rica, 2060 San Pedro, Costa Rica
#School of Medicine and Health Sciences, University of Papua New Guinea, Boroko, Papua New Guinea
S
* Supporting Information
ABSTRACT: Methods to identify the bioactive diversity
within natural product extracts (NPEs) continue to evolve.
NPEs constitute complex mixtures of chemical substances
varying in structure, composition, and abundance. NPEs can
therefore be challenging to evaluate efficiently with high-
throughput screening approaches designed to test pure
substances. Here we facilitate the rapid identification and
prioritization of antimalarial NPEs using a pharmacologically
driven, quantitative high-throughput-screening (qHTS) para-
digm. In qHTS each NPE is tested across a concentration
range from which sigmoidal response, efficacy, and apparent
EC50s can be used to rank order NPEs for subsequent organism reculture, extraction, and fractionation. Using an NPE library
derived from diverse marine microorganisms we observed potent antimalarial activity from two Streptomyces sp. extracts identified
from thousands tested using qHTS. Seven compounds were isolated from two phylogenetically related Streptomyces species:
Streptomyces ballenaensis collected from Costa Rica and Streptomyces bangulaensis collected from Papua New Guinea. Among
them we identified actinoramides A and B, belonging to the unusually elaborated nonproteinogenic amino-acid-containing
tetrapeptide series of natural products. In addition, we characterized a series of new compounds, including an artifact, 25-epi-
actinoramide A, and actinoramides D, E, and F, which are closely related biosynthetic congeners of the previously reported
metabolites.
example, the β-lactam antibacterials produced by fungi such as
Penicillium sp., the antimalarial alkaloid quinine originally
isolated from the bark of the cinchona tree, and artemisinin
isolated from the leaf of Artemisia annua stand as prototypical
examples of highly impactful therapeutic agents.
volved over eons to encode biological activity, natural
Eproducts (NPs) are one of medicine’s oldest and most
enduring sources of drugs. Once the only means to treat disease,
elixirs and traditional medicines ceded considerable ground to
the perceived more efficient synthetic chemical approach to drug
discovery. However, the chemotypes represented by natural
products are among the most versatile for addressing the range of
pathogenic microbes that continue to inflict significant health
and economic burden on human populations worldwide. For
Received: June 1, 2015
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
While NPs were once the mainstay of our pharmaceutical
armamentarium, the relative ready access to synthetic agents has
dampened interest in maintaining a discovery paradigm involving
isolation, structure elucidation, and fermentation of novel
medicinally active NPs.1 However, semisynthetic NPs, synthetic
agents based on or inspired by NPs, and natural products
themselves make up a significant portion of drugs today2,3 and
remain an important resource for the expansion of pharmacology
into emerging disease target space.4 This strongly suggests that
improving the efficiency toward their discovery is an important
means to identifying chemotypes with novel mechanisms of
action.
reinforcing or contrasting outputs can be used to pharmacolog-
ically and mechanistically prioritize active samples, the efficiency
of identifying biologically relevant activities from NP extracts
(NPEs) or other complex mixtures increases significantly. For
example, in this study we aimed to identify active compounds
that targeted Plasmodium falciparum viability from various
geographic localities while selecting for minimal toxicity to
human cells.
RESULTS AND DISCUSSION
■
Quantitative High-Throughput Screening of NPEs. A
library of 16 503 natural product extracts from 5425 culturable
microorganisms prepared by the Amberlite XAD16 extraction
method5,8 was screened across 4 orders of magnitude in
concentration beginning at 15 mg/mL (assay concentration =
43 μg/mL) and diluting by 5-fold to <1 μg/mL as described
previously.5 In the present study we have conducted the qHTS of
this NPE library across six geographically distinct lines (cp250,
Dd2, HB3, 7G8, GB4, and 3D7) of Plasmodium falciparum (Pf)
using a SYBR Green fluorescent DNA staining assay9−11 that
enables the intraerythrocytic measurement of Pf viability (see
Tables S1 and S2). Using this same approach previously applied
to screen several drug/probe libraries12,13 and novel Chemical
Methodologies and Library Development (CMLD) libraries,14
we herein report efforts to explore the more complex case of NPE
libraries.
To this end, we have embarked on a program to develop more
effective approaches to analyze the extracts of marine micro-
organisms, a proven source of bioactive chemical diversity
(Figure 1).5 To maintain diversity, manage source materials, and
The pharmacological resolving power of qHTS7 yielded the
concentration−response curve (CRC) relationships from the
NPE library, as exemplified in Figure 2A for the Dd2 line.
Analysis of the primary qHTS data for the five lines indicated 364
NPEs displayed activity consistent with quality CRCs that we
classify as 1a, which is where the sample displays complete
sigmoidal response over the concentration range tested (Figure
2b, white bar).7 The impact of the qHTS approach to
significantly narrow the consideration of NPEs for further
studies is illustrated by our modeling a follow-up compound
selection process solely based on percent activity at a single
concentration of 43 μg/mL. This retrospective analysis (ex-
plored in depth later in the paper) indicates nearly an order of
magnitude more NPEs would cross the activity threshold (Figure
2b, black and gray bars) for a traditional HTS approach vs qHTS.
To identify broadly acting antimalarial NPEs, we selected
those inhibiting Pf viability across multiple laboratory isolates.
These initially included both chloroquine-sensitive (HB3, 3D7)
and chloroquine-resistant (cp250, Dd2, GB4, 7G8) Pf isolates.
NPEs were chosen for further study in the available isolates if
they were found across the original six malaria isolates tested to
have curve classes 1a, 1b, or 2a with efficacies exceeding a 75%
reduction in SYBR Green signal and potencies of >7.5 μg NPE/L
(Table S3a; also see Experimental Section for additional details).
This focused our candidate NPE follow-up to 231 samples
derived from 130 culturable species showing pan-activity across
the six Pf isolates with high-quality CRCs. Some NPE samples
were obtained from the same culture but extracted with a
different solvent, for example, acetone vs MeOH, therefore
resulting in several NPE samples per marine microorganism (see
Experimental Section, Isolation of Steptomyces sp., for an example
of the multiple-solvent extraction procedure). We reacquired
these samples from our 384-well source plates and retested these
231 NPEs in the following Pf isolates: cp250, Dd2, HB3, 7G8,
and GB4. The reacquired samples were tested in duplicate as 11
point titrations to increase the precision over the primary 7 point
qHTS (Figure 2c, Table S2, and Data Table S1). The EC50
Figure 1. Overview of natural product extract qHTS. A prefractionated
library of Amberlite XAD16 polymer resin enriched natural product
extracts (NPEs) is evaluated as a dilution series using a 1536-well assay
format to accommodate the added samples, here in a Plasmodium
falciparum (Pf) intraerythrocytic SYBR Green viability assay (A). qHTS-
derived concentration−response curve (CRC) profiles for each NPE
tested are assessed using parameters derived from each CRC, for
example, EC50, maximal effect, and curve class (CC), i.e., 1a, 1b, 2, 3, and
4 (inactive at all concentrations), and selected for follow-up studies (B).
NPE microbial strains highest ranking in project profile criteria, here
pan-activity on several Pf laboratory stains and low mammalian cell
toxicity, are recultured, confirmed, and fractionated for retesting (C).
Active fractions meeting criteria are subjected to structure elucidation
(D).
facilitate reacquisition of samples for further evaluation, we
employ minimally enriched natural product extracts from
culturable microorganisms. The extracts used in this study are,
by virtue of several solvent extractions from Amberlite XAD16
absorbent polymer, largely free of the resins (e.g., oligomeric
tannins) and salts that can accompany crude extracts, which
create significant interferences with sensitive optical outputs of
highly miniaturized HTS assays.6 By applying quantitative high
throughput screening (qHTS)7 across a series of assays where
B
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 2. Intraerythrocytic Pf proliferation qHTS of a marine NPE library. (A) Representative 3-axis plot for 15 500 microbial NPEs tested against the
Dd2 laboratory strain of P. falciparum. NPE numbers increase by alphanumeric sorting of the microbial extract names. Blue lines, high-quality
concentration−response curves selected for retesting. (B) Bar chart summarizing active NPEs in SYBR Green DNA staining viability assay for five
malaria lines. The number of active NPEs per assay is indicated on the y-axis, where the white bars represent class 1a CRC,7 while 3 SD and 6 SD cutoffs
for activity at a single concentration of 43 μg/mL are given by the gray and black bars. (C) EC50 correlation between activities determined in the primary
qHTS vs retests of 231 NPEs. R2(cp250) = 0.73, R2(Dd2) = 0.48, R2(HB3) = 0.78, R2(7G8) = 0.76, R2(GB4) = 0.73. NPEs not confirming do not appear in plots;
see Data Table S1.
correlation between these 231 NPEs and the original qHTS data
for the five strains was highly reproducible, as determined by the
generally high R2 values (Figure 2c) despite several outliers that
lowered the R2 in the Dd2 tests. From this analysis 221 NPEs
displayed well-fit CRCs across all Pf parasite lines with apparent
EC50 potencies ranging from ∼75 to ∼7500 ng/mL (Table S3b),
confirming the activity of >95% of the NPEs selected for retest
from the primary qHTS. These NPEs were further tested for
mammalian cell toxicity using HEK293 and HepG2 lines to
generate the activity profile in Figure 3a (also see Table S3b),
allowing the separation of NPEs with parasite-restricted vs broad
cellular toxicity (Figure S1a,b). From these we selected 18 NPEs
that displayed an activity profile having apparent high potency
across all five Pf lines, while showing minimal apparent toxicity in
HEK293 and HepG2 cells (see Figure 3b for example CRC
profiles).
Culture and Isolation of Active NPE Components.
Strains corresponding to 14 of the 18 active NPEs that met our
selection criteria were recultured and subjected to XAD-resin
extraction and solvent elution as originally employed to prepare
the NPE library.5 Stepwise bioassay-guided fractionation tradi-
tionally used to resolve the NPE active constituents through a
series of chromatographic stationary phases (e.g., silica gel, C4-
reversed phase) was bypassed in favor of a more direct approach,
reducing the assays needed and allowing more rapid acquisition
of MS and NMR signatures for preliminary structure assign-
ments. Accordingly, NPEs were subjected to reversed-phase
HPLC fractionation beginning with 5 mg of the extract for each
sample (Figure 4a). Forty-four fractions were collected for each
NPE sample and retested in the five Pf and human cell lines,
shown for Dd2 in Figure 4B.
Structure Elucidation of Actinoramides as Antimalarial
Agents. On the basis of the HPLC fractionation described
above, two phylogenetically related marine Streptomyces species,
S. ballenaensis and S. bangulaensis (Figure S2), showed similar
bioassay results (Figure 4c). After separating the similar
compounds from the two species, we focused our efforts on S.
bangulaensis, which produced significantly higher levels of an
identical suite of active compounds. Further analysis revealed
that the major metabolite was the recently described actino-
ramide A (1)/padanamide A,15,16 whose biosynthetic pathway
has recently been characterized.17 Actinoramide B (2), also
isolated among the active components, was identified by
comparison with published spectroscopic data (1H and 13C
NMR, HRMS, specific rotation).15 During the NMR experi-
ments, conducted in deuterated pyridine, 1 epimerized. The two
epimers were subsequently separated to the known actinoramide
A and previously undescribed 25-epi-actinoramide A.
Conversion of Actinoramide A (1) to 25-epi-Actino-
ramide A (3) and Marfey’s Analysis. 25-epi-Actinoramide A
C
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 3. NPEs selected from primary qHTS and toxicity assessment. (A) Spotfire clustering where NPE activity was calculated as the average EC50
value for all reacquired extracts (MeOH, acetone, and/or EtOAc) against the respective parasite isolate or mammalian cell line. EC50 values determined
where curve classes were not equal to 1a, 1b, 2a, 2b, or 3 were excluded, as were EC50 values where the p-value of the titration was greater than 0.05.
Darker color indicates greater activity in either the SYBR Green or Cell TiterGlo assay. Arrows indicate NPE CRCs shown in B. Dotted box frame:
mammalian cell line response values. The UPGMA clustering method was used for both row and column dimensions using Euclidean distances and
ordering weight set to average value. (B) Representative 11-point CRC confirmation and mammalian cell line toxicity evaluation for selected NPEs. Data
shown are for NPE 07-13-H1I (1), 7714-H2I (2), 276-1 (3), 22278-N3 (4), and 07-234-A1I (5) and broadly toxic NPEs, 7756-H5I (6) and 20755-1I
(7); numbers in parentheses following the NPE number refer to heat map positions. The two data sets per plot are replicate tests (n = 2) for a single
solvent extract, except for HepG2, which is a single test. Cytotoxicity was measured by the CellTiter-Glo assay on HepG2 and HEK293 cell lines. See
Figure S1A,B for additional examples.
(3) was isolated as a colorless powder with a molecular formula
of C31H48N7O9 based on the HRMS protonated molecule peak
at m/z 662.3519 [M + H]+ (Figure S5a). Comparison of the 1D
and 2D NMR data from actinoramide A (1) (Figure S3b,c) and
25-epi-actinoramide A (3) (Figure S5a−e) suggested only one
potential difference, at position 25, with a 1H signal shifted from
4.38 ppm for actinoramide A to 4.61 ppm for 25-epi-
actinoramide A (Figure S5a,b; Table 1). This prediction was
confirmed by Marfey’s analysis using 2,4-diaminobutyric acid
with both compounds, where actinoramide A was shown to have
the S configuration, whereas 25-epi-actinoramide A had the R
configuration (see Scheme S1).
Actinoramide D (4) was isolated as a colorless powder
corresponding to a molecular formula of C31H46N7O9 based on
HRMS (Figure S6a). Comparing the 1H NMR of actinoramide
D (Figure S6b) to actinoramide A (Figure S3b) revealed a loss of
an NH signal (δH 4.52 ppm) and substitution of the methylene
signals (δH 2.79, δC 46.0 ppm) with those for a methine (δH 6.94
D
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 4. Fractionation and testing of selected NPEs. (A) Diagram of NPE fractionation (44 fractions) and preliminary characterization scheme. (B)
Sixteen recultured microbial fractionated extracts tested against the Dd2 isolate. Vertical grids separate 44 fractions and original lead parent NPE
(magenta) from individual microbial strain growths. Blue curves denote high-quality concentration−response active extract fractions. (C) Activity
chromatograms for NPE 07-234-A1I from S. bangulaensis and 22278-N3 from S. ballenaensis where the following are indicated: UV absorbance at 210
nm (black line trace); pEC50 from the average of the five P. falciparum lines tested (blue diamond); viability of human RBCs (purple cross), HEK293
cells (red square), and HepG2 cells (green triangle); assay interference due to light absorption by fraction (light blue cross).
ppm, δC 143.8 ppm). Analysis of gCOSY, gHSQC, and gHMBC
NMR spectroscopic data (Figure S6c−e) revealed the
unsaturated piperazic acid with a double bond linking C-13 to
the adjacent nitrogen (Figure S6a). A previous study on the
kutzneride biosynthetic cluster18 indicates that the unsaturated
piperazic acid moiety is a probable precursor to the saturated ring
system. Another confirmation for the actinoramide D structure
came from a small-scale hydrogenation experiment where
actinoramide D was reduced to actinoramide A. The reduced
product had identical retention time, UV spectrum, and
molecular weight compared to actinoramide A. HPLC co-
injection was also performed to confirm the identification of
actinoramide A (Scheme S2).
C30H45N7O9 based on HRMS (Figure S8a). Comparison of
the 1H and 13C NMR spectra of actinoramide F and actinoramide
A showed no difference for the hydroxy leucine bearing the
methoxy acetate substitution, piperazic acid, and 4-amino-3-
hydroxy-2-methyl-5-phenylpentanoic acid backbone (Figure
S8b,c). The 4-amino-3-hydroxy-2-methyl-5-phenylpentanoic
acid carbonyl 13C signal at δC 174.4 ppm showed an HMBC
correlation from an amidic proton resonating at δH 8.04 ppm.
This amidic proton showed a COSY correlation with a methine
proton at δH 4.36 ppm (δC 47.3 ppm), which correlates to
methylene protons at δH 3.18 ppm (δC 40.0 ppm) and to an
amidic proton at δH 7.60 ppm (Figures S8d−f). All of these
correlations were confirmed by HMBC correlations (Figure
S8e). The amidic proton resonating at δH 7.60 ppm shows
another HMBC correlation to a carbonyl carbon resonating at δC
153.2 ppm, which shows a correlation from another amidic
proton at δH 10.20 ppm. The carbonyl carbon resonating at δC
170 ppm shows two HMBC correlations, the first to the amidic
carbon at δH 10.20 ppm and the second to the previously
described methine at δH 4.36 ppm, which forms the dihydrouracil
ring system (Figure 5).
Actinoramide E (5) was isolated as a colorless powder
corresponding to a molecular formula of C30H46N6O8 based on
the HRMS protonated molecule peak at m/z 619.3459 [M + H]+
1
(Figure S7a). The H and 13C NMR (Table 2) spectra of
actinoramide E are almost identical to those reported for
actinoramide A except that the actinoramide E spectra show the
loss of a carbonyl (δC 152.7 ppm) and NH (δH 7.73 and 7.42
ppm) signals and gain of an amide signal (δH 8.06 ppm). On the
basis of all of the NMR data actinoramide E should have the 3-
aminopyrrolidin-2-one moiety instead of the 2-amino-4-
ureidobutanoic acid in actinoramide A (Figure S7b−f).
During our study another dihydrouracil analogue was isolated
where the piperazic acid is unsaturated, similar to actinoramide
D. Due to the lack of material and purity issues (this compound
was isolated as a mixture with actinoramide F), this compound
was incompletely characterized primarily by NMR and HRMS.
A new natural product, actinoramide F (6), was isolated as a
colorless powder corresponding to the molecular formula
E
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
a
b
1
Table 1. H NMR Spectroscopic Data for Compounds 3−6 (DMSO-d6)
δH (J in Hz)
δH (J in Hz)
c
c
fragment
MAA
no.
3
4
5
6
fragment
no.
19
3
4
5
6
2
3
3.80, m
3.28, s
3.84, m
3.31, s
3.78, m 3.82, d (7.4)
7.21, m
7.13, m
3.45, m
7.22, m
7.13, m
3.41, m
4.97, m
7.23, m 7.24, m
7.14, m 7.15, m
3.49, m 3.47, m
3.27, s
3.30, s
20
Hle (Hleu) 4-NH 7.48, d
(9.6)
7.69, d (8.4) 7.30, d
(9.0)
7.29, m
21
21-
OH
4
5.44, t (8.8) 5.57, t (8.0) 5.48, t
(9.0)
5.52, t (8.7)
22
23
2.22, m
2.23, m
2.26, m 2.29, m
5
3.45, m
3.53, m
3.39, m 3.38, m
4.82, d (6.0)
1.06, d
(7.2)
1.02, d (7.2) 1.05, d
(6.6)
1.05, d (6.8)
5-OH
6
4.71, d (5.6)
1.68, m
1.63, m
1.58, m 1.65, dq
C-terminal
moiety
Auba
Auba
(Aopc)
Apd
adhU
(12.9, 6.5)
(Aopc)
7
0.87, d
(7.2)
0.89, d (6.8) 0.86, d
(6.8)
0.86, d (6.7)
25-
NH
8.03, d
(7.8)
8.09, d (8.0) 8.05, d
(7.6)
8.04, d (7.8)
8
0.82, d
(6.6)
0.82, d (6.8) 0.82, d
(6.7)
0.82, d (6.8)
(2.4, 6.0)
25
26
4.61, m
1.81, m
2.16, m
4.35, m
1.89, m
2.10, m
4.44, m 4.36, m
1.87, m 3.00, m
2.27, m 3.18, m
7.60, d (4.4)
Pip
10
11
4.92, dd
(2.4, 6.0)
5.00, m
4.90, m 4.97, dd
26-
NH
1.60, m
2.06, m
1.57, m
2.17, m
1.60, m
1.58, m 1.67, m
2.08, m 2.13, m
1.37, m 1.37, m
27
3.43, m
3.69, m
3.44, m
3.21, m
12
13
1.26, m;
1.33, m
3.69, t (9.2)
27-
NH
8.03, d
(7.8)
8.09 d (8.0) 8.10 d
(8.5)
10.20, s
2.01, m
2.75, m
6.94, d (4.0) 2.81, m 2.83, m
4.73, m 4.52, t (8.4)
28-
NH
8.06 s
13-
4.39, dd
(6.0, 8.4)
NH
29-
NH2
7.38, br s
7.71, br s
7.40, br s
Ahmppa
(Ahmpp)
15-
7.49, d
(8.4)
7.37, d (9.2) 8.14, d
(8.6)
7.65, d (9.0)
NH
7.72, br s
15
16
4.10, q
(8.4)
4.03, q (8.4) 4.08, m 4.18, br q
(8.1)
a
b
c
δ (ppm) 600 MHz. In DMSO-d6. Fragment nomenclature based on
refs 13 and 14: 2-methoxyacetic acid, MAA; 3-hydroxyleucine, Hle
(Hleu); piperazic acid, Pip; 4-amino-3-hydroxy-2-methyl-5-phenyl-
pentanoic acid, Ahmppa (Ahmpp); 3-amino-2-oxopyrrolidine-1-
carboxamide, Auba (Aopc); cyclic 4-acetoamide-2-aminobutanoic
acid (Aaba). 3-Aminopiperidine-2,6-dione (Apd); 5-aminodihydrour-
acil (adhU) this study.
2.75, m
2.65, dd
(8.0, 13.6)
2.77, m 2.78, m
2.75, dd,
(6.8,
2.87, m
13.6)
18
7.19, m
7.16, m
7.21, m 7.21, m
Antimalarial Activity of the Actinoramides. The natural
products isolated from S. bangulaensis include members of the
recently discovered actinoramide/padanamide family of modi-
fied tetrapeptides illustrated in Figure 6a.15,16 Once purified and
structurally characterized, the actinoramides were re-examined
against in vitro cultured P. falciparum parasites and human cell
lines. The results show a clear separation in activity between
actinoramide A and 25-epi-actinoramide A of approximately 20-
fold, with the highly active actinoramide A displaying an EC50 of
approximately 200 nM in all five parasite lines retested (Figure
6b; Figure S9a and Table S4). Strikingly these compounds show
virtually no toxicity toward the HepG2 and HEK293 human cell
lines tested in this study (Figure 6C and Figure S9b).
One possible P. falciparum candidate target class of actino-
ramide A is suggested by the presence of statine-like amino acid
moieties in the molecule (see Figure 6a, red substructures). By
mimicking the tetrahedral transition state in aspartyl protease-
catalyzed peptide bond hydrolysis statine-containing peptides
such as pepstatin are potent class-specific protease inhibitors.19
We therefore examined if actinoramide A would be capable of
inhibiting the aspartyl protease activity of P. falciparum
plasmepsins found in the digestive vacuole of the parasite.20
Using freshly prepared P. falciparum extracts we demonstrated
the presence of plasmepsin activity using a pro-fluorescent FRET
peptide substrate (Figure S10a) based on internal FRET
quenching.21 While pepstatin potently inhibited the aspartyl
protease activity of our Pf extracts, actinoramide A had no effect
up to 100 μM, nor did the S. ballenaensis or S. bangulaensis NPEs
(Figure S10b,c), leaving the Pf target of actinoramide A in
question.
The actinoramides/padanamides were recently isolated by
two independent laboratories from marine Streptomyces sp.
originating from the coastal regions of Papua New Guinea16 and
southern California.15 In the latter case the compounds isolated
from this halophilic Streptomyces species (CNQ-027) failed to
demonstrate activity against an HCT-116 human colon
carcinoma cell line and methicillin-resistant Staphylococcus
aureus. In the former, two compounds were isolated, padanamide
A, which is identical to actinoramide A, and padanamide B.
Padanamide A and padanamide B differ at their terminal
functional groups, bearing either the 2-oxopyrrolidine-1-
carboxamide or piperidine-2,6-dione, respectively. Padanamide
A (actinoramide A) and padanamide B were shown to be
somewhat cytotoxic to Jurkat cells (EC50 ≈ 60 μg/mL (90 μM)
and 20 μg/mL (30 μM), respectively), although such
concentrations can be considered relatively benign in cell culture
studies. In this study actinoramide A/padanamide A was
discovered independently from NPEs prepared from cultures
of Streptomyces sp. isolated from the Costa Rican coast by qHTS
in an array of P. falciparum viability assays. We observed a
consistent 20−50-fold separation in potency between the 25R
and 25S diastereomers of actinoramide A (padanamide A) with
no toxicity to HEK293 or HepG2 cells at the highest tested
concentrations (Figure S9; Table S4). This activity is differ-
F
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
a
Table 2. 13C NMR Spectroscopic Data for Compounds 1−6
entiated by the configuration at the terminal tertiary carbon of
the 2-oxopyrrolidine-1-carboxamide ring (Figure 6a). We further
observed that the apparent oxidation product of the piperazic
acid moiety (dehydro actinoramide A, characterized herein as
new congener actinoramide D; Figure 6a, 4) had an observable
effect on its activity in our malaria assays, leading to a ∼6−20-fold
decrease in potency depending on the parasite strain (Figure S9a;
Table S4). In addition to the actinoramide A congeners, we
isolated a structurally related metabolite. Actinoramide F (Figure
6a, 6) differs from actinoramide A at the terminus of the
tetrapeptide where the 3-amino-2-oxopyrrolidine-1-carboxamide
(Aopc; also referred to as cyclic 2-amino-4-ureidobutanoic acid
(Auba))15 is replaced by a 5-amino-5,6-dihydrouracil (adhU).
Actinoramide F showed no antimalarial activity (Figure 6b),
which strongly implicates the terminal moiety of actinoramide A
as a key active substituent of the molecule.
To gain additional insight into how, and under what
conditions, the Pf SYBR Green viability assay would best
perform in a single test concentration experiment with NPEs, we
reanalyzed the primary qHTS data obtained from the Dd2 Pf
isolate. Using the qHTS curve classification7 (Figure 1), we
defined the NPEs with active curve classes (i.e., 1a, 1b, and 2) as
true positives (TPs) and, for simplicity, all other curve classes
(i.e., 3 and 4) as true negatives. In a scenario with few limits on
follow-up capacity of candidate NPEs, a 1σ (i.e., 1 standard
deviation from the mean of the DMSO-treated control) and
more likely 3σ activity threshold were examined. In the 1σ
threshold case, the number of false positives (FPs) would likely
be untenable, numbering in the thousands and rising with
increasing tested concentrations. It is interesting to note that not
all the TPs with the 1σ hit cutoff would be found unless tested at
the second highest concentration of 10 μg/mL NPE as illustrated
in Figure 7. That is, 15 TP NPEs would be missed at the highest
concentration of 43 μg/mL NPE.
b
(DMSO-d6)
δC, type
c
fragment
MAA
no.
3
4
5
6
1
2
3
4
5
6
7
8
9
168.5, C
168.8, C
168.1, C
168.4, C
70.9, CH2
58.3, CH3
49.9, CH
74.7, CH
29.0, CH
19.8, CH3
15.3, CH3
172.3, C
70.8, CH2
58.1, CH3
50.8, CH
75.1, CH
29,0, CH
19.7, CH3
15.5, CH3
172.0, C
71.2, CH2 71.2, CH2
58.5, CH3 58.5, CH3
49.4, CH 49.4, CH
74.9, CH 75.1, CH
28.8, CH 29.2, CH
19.8, CH3 20.1, CH3
15.2, CH3 15.6, CH3
Hle (Hleu)
172.0, C
172.1, C
Pip
10 50.9, CH
11 25.1, CH2
12 20.5, CH2
13 45.7, CH2
14 170.0, C
15 52.1, CH
51.5, CH
18.3, CH2
19.8, CH2
140.8, CH
169.6 C
50.5, CH 50.9, CH
25.5, CH2 25.9, CH2
20.8, CH2 20.8, CH2
46.0, CH2 46.4, CH2
173.8 C
170.4 C
Ahmppa
(Ahmpp)
52.8, CH
52.7, CH 52.9, CH
16 37.7, CH2
17 138.8, C
18 128.6, CH
19 127.8, CH
20 125.6, CH
21 71.6, CH
22 43.1, CH
23 15.0, CH3
24 174.3, C
37.3, CH2
138.7, C
37.0, CH2 37.3, CH2
139.6, C
139.0, C
128.6, CH
127.7, CH
125.4, CH
71.4, CH
42.6, CH
14.5, CH3
174.0 C
129.1, CH 129.1, CH
128.0, CH 128.1, CH
125.9
125.9
71.7, CH 71.8, CH
42.9, CH 43.2, CH
15.3, CH3 15.0, CH3
173.8, C
Apd
174.4, C
adhU
C-terminal
moiety
Auba
(Aopc)
Auba
(Aopc)
25 51.3, CH
26 23.8, CH2
27 41.7, CH2
28 174.7, C
51.8, CH
23.0, CH2
41.1, CH2
174.4, C
154.2, C
49.2, CH 47.3, CH
23.7, CH2 40.0, CH2
41.2, CH2 153.2, C
Using a 3σ cutoff, most of the qHTS actives (807 of 840)
would be found if tested at 10 μg/mL, but would be accompanied
by 392 FPs. At the third and fourth highest concentrations, the
false negatives (FNs) would exceed the TPs and yield
comparable numbers of FPs to TPs (137 TPs, 167 FPs at
0.344 μg/mL NPE; 338 TPs, 228 FPs at 1.73 μg/mL NPE).
Again, at the highest tested concentration, only 15 NPEs
characterized as qHTS actives would be missed, but an
unreasonable 2843 hits are FPs, inactive in qHTS.
175.8, C
170.0, C
29 152.7 C
a
b
c
δ (ppm) 150 MHz. In DMSO-d6. Fragment nomenclature based on
refs 13 and 14: 2-methoxyacetic acid, MAA; 3-hydroxyleucine, Hle
(Hleu); piperazic acid, Pip; 4-amino-3-hydroxy-2-methyl-5-phenyl-
pentanoic acid, Ahmppa (Ahmpp); 3-amino-2-oxopyrrolidine-1-
carboxamide, Auba (Aopc); cyclic 4-acetoamide-2-aminobutanoic
acid (Aaba). 3-Aminopiperidine-2,6-dione (Apd); 5-aminodihydrour-
acil (adhU) this study.
At more stringent hit thresholds of 20σ and 6σ, the lowest
tested concentration results generally in higher FNs with an
accompanying reduction in FPs. As expected, using a 20σ
threshold yields fewer actives and more FNs at any tested
concentration. Examining the highest tested concentration, 692
hits are actives according to qHTS criteria and 148 FNs result,
the fewest of any of the tested concentrations. Of the 895 overall
hits in this scenario, 203 NPEs would be characterized as negative
by qHTS, supporting an overall trend of FPs tracking with TPs in
any single-concentration screening scenario and suggesting an
intrinsic FP rate related to the threshold. Of the four threshold
scenarios, the best case might use a 6σ cutoff at the second
highest concentration tested. Compared to the 20σ cutoff, a
slightly more modest 21% (176 of 840) of the actives found using
qHTS are not found using a threshold of 6σ, while
simultaneously keeping the false positive hits down to 74,
although the split between TPs and FNs is slightly worse at 664
and 176, respectively.
Figure 5. Dihydrouracil ring system. The novel terminal end group of
actinoramide F (6) differs from actinoramide A (1) by virtue of 5,6-
dihydrouracil replacing the 2-oxopyrrolidine-1-carboxamide.
In this study we examined the effect of a library of
prefractionated marine natural product extracts5,8 on the viability
G
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 6. Activity of the actinoramides on blood stage parasite (Dd2 line) and a hepatocyte cell line. (A) Structures of the actinoramides. (2R,3R)-2-
Amino-3-hydroxy-4-methylpentanoic acid and (2S,3S,4S)-4-amino-3-hydroxy-2-methyl-5-phenylpentanoic acid amino acids are shown in red. 1H and
13C NMR, HSQC, HMBC, and COSY data for compounds 1−6 and ROESY and TOCSY data for compounds 3−6 are provided in Tables 1 and 2 and
Figures S3−S8. (B) Actinoramide congener activity on 3D7 showing stereoselective antimalarial activity of the 25S (solid circles, actinoramide A) and
25R (open circles, 25-epi-actinoramide A) isomers of actinoramide A. The activity of these compounds on all five blood stage parasites (Dd2, HB3, 7G8,
GB4, and cp250) are provided in Figure S9a. (C) Effect of the actinoramides on viability of HEK293 cells (for effect on viability of HepG2 cells see
Figure S9b). Digitonin control (×). n = 2 or greater for all experiments; error bars are SD of the mean.
obtained on a Bruker Alpha-P UV/visible spectrophotometer. All
NMR experiments were carried out on a Varian INOVA 600 MHz
spectrometer. Compounds 1 and 3 were purified on an Agilent 1100
series HPLC (Agilent Technologies) using a preparative Phenomenex
Luna Phenyl-hexyl column. University of Michigan (compounds 1, 2, 5,
and 6): Optical rotations were determined on a Jasco P2000
polarimeter. UV spectra were measured on a UV−visible Molecular
Devices SpectraMax M5 spectrophotometer using 1 mL cuvettes with
1.0 cm path lengths at room temperature in CH3OH solvent. IR spectra
were obtained on a PerkinElmer FTIR spectrometer equipped with a
Miracle ATR accessory. The NMR spectra were acquired at 600 MHz
for 1H and 150 MHz for 13C on a Varian Inova 600 MHz spectrometer.
High-resolution MS spectra were measured at the University of
Michigan core facility in the Department of Chemistry using an Agilent
6520 Q-TOF mass spectrometer equipped with an Agilent 1290 HPLC
system. The LCMS analysis of HPLC fractions was performed on a
Shimadzu 2010 EV APCI spectrometer. HPLC separations were
performed on an Agilent 1100 HPLC system consisting of a G1311A
pump and DAD detector G1315B equipped with Clarity software using
Merck HiBar LiChrospher 60 RP-Select B, 5 μm, 250 × 25 mm, and
of P. falciparum in its RBC host using qHTS. The resulting CRCs
displaying potent antimalarial activity were clearly separable from
those having accompanying mammalian cellular toxicity as
indicated by a parallel HEK293 and HepG2 viability screen,
facilitating selection of NPEs with optimal profiles for
subsequent reversed-phase fractionation and structure elucida-
tion. The screening methodology described here focuses the
prioritization of NPEs to a limited number of microbial species
for reculture. We anticipate the paradigm’s capacity to resolve
primary activity by EC50 with minimal library consumption will
prove useful for the evaluation of other forms of complex or
multicomponent chemical libraries that may exist or be prepared
in initial limited quantities.22−26
EXPERIMENTAL SECTION
■
General Experimental Procedures. Harvard Medical School
(compounds 3 and 4): Optical rotations were obtained using a Jasco
P1010 polarimeter. UV spectra were measured on an Amersham
Biosciences Ultrospec 5300 spectrophotometer. IR spectra were
H
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 7. Analysis of qHTS data at a single NPE concentration with various hit threshold scenarios. Solid circles, true positives (TPs); open triangles,
false positives (FPs); open down triangles, false negatives (FNs). See discussion for description. σ = standard deviation.
Phenomenex, Cosmosil C8, 4.6 μm, 250 × 20 mm, Waters HPLC
columns using a solvent system of CH3CN and H2O.
in parentheses), of which 231 were available in sufficient amounts from
our 384-well archive for retesting. The relatively few samples displaying
color interference (11 NPEs) or RBC toxicity (1 NPE) were flagged.
The 231 NPEs retrieved from the archive and tested using 11-point
titrations were then selected for regrowth using the more stringent
conditions of only curve classes 1a or 2a with the IC50 of all five Pf
isolates < 7500 ng/mL and no measurable or weak effects in HEK293
and HepG2 cell viability assays. Using these criteria, 29 NPEs
representing 18 microbial strains were selected for regrowth. Fourteen
of the 18 microbial strains were available for regrowth and fractionation.
Mammalian Cell Viability Assays. Cell viability assays for
HEK293 and HepG2 lines and RBCs were carried out using Cell
TiterGlo (Promega) according to the manufacturer’s instructions.
Assays were performed in solid white 1536-well plates (Nexus
Biosystems), and measurements made on a ViewLux (PerkinElmer).
See Brown et al.14 for additional details.
Optical Interference Assay. Potential compound interference with
the SYBR Green viability assay was determined by adding test
compound following a 72 h incubation time and reading fluorescence
intensity at 485/535 nm EX/EM on an EnVision (PerkinElmer). See
Brown et al.14 for additional details.
Preparation of Actinoramides for Follow-up Testing. From ∼2
mg of actinoramide A (MW: 661) in 1 mL of CH2Cl2/CH3OH (50/50
v/v), 400 μg in that solution was taken (and then dried), and 30 μL of
molecular biology grade DMSO was added to make a 20 mM DMSO
solution. The other samples were prepared similarly.
Plasmepsin Protease Assay. A P. falciprarum parasite extract
(saponin treated and RBC removed) was prepared as previously
described22 to enable a plasmepsin protease assay using the EDANS-
DABCYL Plasmepsin II fluorogenic substrate, EDANS-
COCH2CH2CO-ALERMFLSFP-Dap(DABCYL)OH (AnaSpec). An
estimated KM value was calculated for the Pf preparation and compared
to the aspartic protease pepsin. Briefly, 2 μL of assay buffer (100 mM
sodium acetate, pH 5.0, and 20% glycerol) was dispensed into black,
clear bottom 1536-well plates (Greiner Bio-One) followed by 1 μL of
the respective protease or protease preparation across 3−5 different
concentrations using a BioRaptr Flying Reagent dispenser (Beckman
Coulter). Plates were sealed, centrifuged for 1 min, and incubated at
room temperature, protected from light, for 15 min. Baseline
fluorescence was read on the Tecan Infinite M1000. A 1 μL amount
of a 4 pt titration of the EDANS-DABCYL Plasmepsin II fluorogenic
substrate was added to respective wells as above. Plates were sealed and
centrifuged for 1 min, and fluorescence at EX 336 nm, 5 nm bandwidth,
and EM 490 nm, 10 nm bandwidth, was measured for 3 h. Actinoramide
SYBR Green DNA-Staining Parasite Viability Assay. Briefly, 3
μL of culture medium was dispensed into 1536-well black, clear-bottom
plates (Aurora Biotechnologies) using a Multidrop Combi (Thermo
Fisher Scientific Inc.) followed by 23 nL of NPE via pin tool transfer
(Kalypsys), then 5 μL of P. falciparum-infected RBCs (0.3% parasitemia,
2.5% hematocrit final concentration). The plates were incubated at 37
°C in a humidified incubator in 5% CO2 for 72 h, after which 2 μL of lysis
buffer (20 mM Tris-HCl, 10 mM EDTA, 0.16% saponin (w/v), 1.6%
Triton-X (v/v), 10× SYBR Green I (supplied as 10 000× concentration
by Invitrogen)) was added to each well. The plates were incubated
overnight at 22−24 °C in the dark. The following morning, fluorescence
intensity at 485 nm excitation and 535 nm emission wavelengths was
measured on an EnVision (PerkinElmer) plate reader. The compound
library was screened against each parasite line at seven 5-fold dilutions
beginning at 15 mg/mL DMSO stock. The antimalarial drugs,
artemisinin and mefloquine, and DMSO were included as positive and
negative controls for each plate, respectively. Additional details can be
found in Table S1 and refs 13 and 14.
Analysis of qHTS Data and Selection of Follow-up NPEs. NPE
concentration−response curves from the qHTS were determined from
the normalized and corrected data as previously described.13 CRCs for
each NPE were curve-fitted and categorized into four major curve classes
(CCs) as described7 using freely available NCATS software (http://
ncgc.nih.gov/pub/openhts/curvefit/). Briefly, NPEs in class 1 produce
a complete response curve containing two asymptotes. Subclassification
of CC = 1 CRCs into either 1a or 1b is based on the magnitude of the
response relative to control, where 1a is equal to ≥80% of control
inhibition and 1b < 80%. NPEs constituting class 2 have incomplete
response curves containing one asymptote and can be likewise
subcategorized into a or b. NPEs fitting class 3 curves display activity
only at the highest tested concentration or poorly fit class 1 or 2 curves
with r2 less than 0.9. NPEs with class 4 curves are inactive, and either
they do not have a curve fit or the curve fit is below 3 SD of the mean
basal activity (about 30% of the maximal control response).
Preliminary selection criteria for NPEs with panisolate antimalarial
activity used qHTS data from six Pf assays, two using chloroquine-
sensitive (HB3 and GB7) and four using chloroquine-resistant (cp250,
Dd2, GB4, 7G8) laboratory isolates. An NPE was considered active and
promoted for follow-up testing if it displayed a high-quality CC of either
1 or 2a in each of the six assays, with measured IC50 values less than 7500
ng/mL and efficacies exceeding a 75% reduction in SYBR Green signal.
This identified 235 pan-active NPEs (Table S3a, pan-active NPE counts
I
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
congeners and three fractions of NPE 07-234-A1I and 7714-H2I were
tested under KM conditions for inhibitory activity on the plasmepsin-
containing extracts. Assay plates were set up as above, and 23 nL of each
compound was transferred via pin tool (Wako) followed by addition of 1
μL of pepsin (50 nM final concentration) or plasmepsin pellet (P)
extract (0.16% final concentration). Following a 15 min incubation 1 μL
of EDANS-DABCYL Plasmepsin II fluorogenic substrate (5 or 10 μM
final concentration) was added to each well. Plates were read for 3 h as
described above (Table S5). Data were plotted, and KM and IC50 values
were calculated in GraphPad Prism 5.
Marine Microbes. Marine sediments were collected by scuba
diving; the source material resulting in S. ballenaensis (Sherman lab
22278-N3) was collected near the “whale tail” isthmus close to Uvita,
Costa Rica (9°8′37.5″, −83°45′21.7008″). The second strain, S.
bangulaensis (Sherman Lab 07-234-A1), was derived from marine
sediments collected at Bangula Bay, New Britain Island, Papua New
Guinea (−5°25′40.5600″, 150°46′10.0200″).
Promega pGEM-T vector kit. The resulting pGEM-16S rDNA construct
was transformed into chemically competent E. coli XL1-Blue cells and
cultured on LB-ampicillin plates. Plasmid DNA was isolated using the
Zymo Plasmid miniprep kit (cat. no. D4037) and sequenced using the
primers T7 and SP6. 16S rRNA sequences were deposited at GenBank
with the following accession numbers: Streptomyces ballenaensis
(KT333053) and Streptomyces bangulaensis (KT333054).
Phylogenetic Analysis of Marine Streptomyces Species. The
genetic analysis of S. ballenaensis and S. bangulaensis was performed using
a 1.5 kb sequence of the 16S rRNA gene. A Blast search of the 16S rRNA
sequence was employed to retrieve sequence data for the phylogenetic
analysis. Only sequences of at least 1.3 kb were retrieved from GenBank
as well as the two other actinoramide-like producing strains Streptomyces
sp. RJA2928 (padanamide producer) and Streptomyces sp. CNQ-
027_SD01. The 16S rRNA sequence of Micromonospora aurantiaca
ATCC 27029 was used as an out group. Phylogenetic analyses were
conducted with Geneious Pro 4.8.4. The evolutionary history was
inferred using the neighbor-joining method.30 S. ballenaensis and S.
bangulaensis show high phylogenetic similarity (d = 1.16 × 10−4). The
closest member was Streptomyces sp. VTT E-062996 (d = 4.21 × 10−4).
It is also worth noting that the padanamide A-producing Streptomyces sp.
RJA2928 (d = 0.0124) and the actinoramides-producing strain CNQ-
027_SD01 (d = 0.0147) were more distant than the bacteria of the S.
albiaxailis/paraguayensis clade (Figure S2).
Cultivation and Extraction. Two marine-derived microorganisms,
S. ballenaensis and S. bangulaensis, were each cultured in two 6 L
Erlenmeyer flasks each containing 2.5 L of ISP2 medium and shaken at
200 rpm at 27 °C. After 21 days of cultivation for S. ballenaensis strain
and 19 days of cultivation for S. bangulaensis strain the cells were
separated from the broth by centrifugation (5500 rpm for 30 min).
Washed XAD-16 resin (15 g/L) was added to the broth using 30 g resin
bags in order to adsorb the organic substances, and the culture and resin
were agitated at 200 rpm overnight. The XAD resin was washed with
Milli-Q H2O and extracted with three different solvents, MeOH,
acetone, and EtOAc. The biological activity of these two extracts was
evaluated using gradient HPLC separation (5−100% CH3CN over 30
min). Then 5 mg of the extract was injected onto the HPLC column, and
60 fractions were collected (every 30 s), evaporated, and transferred for
biological activity evaluation. Results of the biological activity assay were
plotted on top of the 210 nm UV chromatogram to identify the active
HPLC peaks (Figure 4).
Isolation of Streptomyces spp. S. ballenaensis and S. bangulaensis
isolation was based on a modified protocol of Imura and Yamamura27
where 500 mg of wet sediment was diluted in 10 mL of sterile water and
vortexed for 10 min. A 1 mL portion of this suspension was then applied
directly to the top of the discontinuous sucrose gradient and centrifuged
for 30 min at 300g. A 500 μL amount of the 20%, 30%, and 40% layers
was then plated to HVA agar supplemented with 10 μg/mL
chlortetracycline, 25 μg/mL cyclohexamide, and 25 μg/mL nalidixic
acid. The plates were then incubated at 28 °C for one month. The
colony was picked off the plate and streaked onto International
Streptomyces Project media 2 (ISP2) agar supplemented with 3% NaCl
to mimic seawater salinity until pure. Seed cultures were grown in 17 mL
dual position cap tubes containing 2 mL of ISP2 and grown for 4 days on
a rotary shaker at 200 rpm. The seed culture was then poured into a 250
mL baffled flask containing 100 mL of ISP2 and grown for 21 days for S.
ballenaensis sp. and 19 days for S. bangulaensis sp. on a rotary shaker at
200 rpm. The culture was centrifuged at 4000 rpm for 10 min to remove
the cells, and 2 g of XAD16N resin (Sigma-Aldrich) contained within a
polypropylene mesh bag was added to the broth and incubated
overnight on the rotary shaker. The resin bag was removed and placed
into 10 mL of CH3OH followed by 10 mL of acetone and 10 mL of
EtOAc. Each of the three fractions was dried in vacuo and reconstituted
to a final concentration of 15 mg/mL in DMSO.
DNA Extraction, PCR Amplification of 16S rRNA Genes,
Cloning, and Sequencing. Prior to DNA extraction, 1 mL of a liquid
culture of Streptomyces sp. (S. ballenaensis and S. bangulaensis) was
combined with 100 μL of 10 mg/mL lysozyme (Sigma catalog no.
L6876-5G), then incubated in a 35 °C water bath for 1 h. After
incubation, the cell material was centrifuged and the supernatant
discarded. DNA was extracted from the resulting cell material using the
Wizard Genomic DNA purification kit (Promega cat. A1120). The
manufacturer’s protocol was modified to include mechanical disruption
after the addition of the nuclei lysis solution (step 6). A portion of the
16S rDNA genes were amplified by PCR from the DNA using the
primers FC27 and RC1492 previously described.27−29 Reaction
mixtures consisted of 2.5 μL of DNA (250 ng), 10 μL of 5× Phusion
GC buffer, 1.0 μL of dNTP mix (10 μM of dATP, dCTP, dGTP, and
dTTP), 2.5 μL of each primer (10 μM), 1.5 μL of DMSO, 0.5 μL of
Phusion high-fidelity DNA polymerase (New England Biolabs cat. F-
530S), and 28.5 μL of H2O for a total volume of 50 μL. The reaction was
performed using a Bio-Rad iCycler thermal cycler with the following
reaction program: initial denaturation for 30 s at 98 °C, 30 cycles of
amplification with 10 s at 95 °C, 20 s at 50 °C, and 45 s at 72 °C, and a
final elongation of 5 min at 72 °C. PCR products were purified using the
Zymo Gel DNA recovery kit. The manufacturer’s instructions were
modified by combining 250 μL of the binding buffer with 50 μL of PCR
products, then skipping to step 4 of the instructions. The purified PCR
products were modified to add a 3′-A overhang using Taq DNA
polymerase. Reaction mixtures consisted of 3 μL of PCR product, 1 μL
of 10× Taq buffer, 2 μL of 1 mM dATP, 0.5 μL of Taq DNA polymerase,
and 3.5 μL of H2O for a total volume of 10 μL. The reaction was
incubated for 30 min at 72 °C using a Bio-Rad iCycler thermal cycler.
The modified PCR products were ligated into a pGEM vector using the
On the basis of the enhanced metabolite production and biological
activity results, S. bangulaensis was pursued for longer-term metabolite
analysis. However, all of the known and novel molecules described in
this work were isolated from the strains obtained from both Costa Rica
and Papua New Guinea.
Isolation Procedure. For the initial fractionation of S. bangulaensis
species, the XAD resin was washed and extracted as described above to
give three extracts: MeOH (402.9 mg), acetone (170 mg), and EtOAc
(46 mg). The MeOH extract showed the expected antimalarial activity
and was further purified using similar chromatography procedures. The
MeOH extract was separated on an ODS (YMC-GEL, 120 Å, 4.4 × 10
cm) flash column with 10% increasing amounts of MeOH in H2O and
washed using EtOAc to obtain 11 fractions labeled A−K. On the basis of
the NMR spectra and the biological activity assay fractions D (79.7 mg)
and E (18.1 mg) were further separated. Fraction D was subjected to a
reversed-phase HPLC column (Merck HiBar LiChrospher 60 RP-Select
B, 5 μm, 250 × 25 mm, Agilent 1100 series HPLC, DAD at 210 and 238
nm, flow rate 5.0 mL/min) eluted with 4:6 CH3CN/H2O to give two
semipure fractions, D4 (8.6 mg) and D6 (32 mg). Fraction D6 (32 mg)
was further separated on a reversed-phase HPLC column (Phenomenex,
Cosmosil C8, 4.6 μm, 250 × 20 mm, Agilent 1100 series HPLC, DAD at
210 and 238 nm, flow rate 5.0 mL/min) eluted with 4:6 CH3OH/H2O.
Seven fractions were collected; two of them were pure compounds.
Fraction D6b (1.2 mg, tR = 29.8 min, MH+ 660.37 g/mol) was
established to be the new compound actinoramide D (4). Fraction D6c
(24.1 mg, tR = 34.1 min, MH+ 662.47, MNa+ 684.37 g/mol) was
established to be the known compound padanamide A16/actinoramide
A (1),15 and this compound was also isolated from the crude isolate E.
Fraction D6e (2.4 mg, tR = 39.3 min, MH+ 661.3564, MNa+ 683.3373 g/
J
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
mol) was established to be the known compound actinoramide B.13
Fraction D4 (8.6 mg) was further separated on a reversed-phase HPLC
column (Phenomenex, Cosmosil C8, 4.6 μm, 250 × 20 mm, Jasco PU
2080 Plus HPLC, DAD at MaxAbs, flow rate 5.0 mL/min) eluted with
45:55 CH3OH/H2O. Five fractions were collected; one of them was a
pure compound. Fraction D4a (1.3 mg, tR = 24.3 min, MH+ 648.47 g/
mol) was established to be actinoramide F (6), while fraction D4c (2.2
mg, tR = 27.4 min, MH+ 646.43 g/mol) was established to be a mixture of
actinoramide F and its unsaturated piperazic analogue. Fraction 2D5
(1.1 mg, tR = 32.3 min, MH+ 619.37 g/mol) contains actinoramide E
(5).
S. bangulaensis Strain Secondary Cultivation and Fractiona-
tion. Seed cultures were generated by inoculating 10 mL of ISP2 growth
media in a 50 mL glass culture tube containing a stainless steel spring for
aeration with 10 μL of S. bangulaensis from spore preparation glycerol
stock. At this point, 1 mL of seed culture was used to inoculate 100 mL of
ISP2 growth media in a 250 mL Erlenmeyer flask containing a circular
stainless steel spring at the base for aeration. This culture was grown at
28 °C with shaking at 180 rpm until the culture reached the stationary
phase of growth. Finally, 20 mL of this culture was used to inoculate 2 L
of modified ISP2 media containing high C:N (contained 15 g/L malt
extract instead of 10 g/L that is used in standard ISP2 media) in a 6 L
flask containing a large stainless steel circular spring at the base for
aeration (15 flasks resulting in 30 L total culture). The process was
repeated as described above to isolate actinoramides A, E, and F.
Actinoramide D was detected as a mixture with actinoramide A in a 5:1
ratio, respectively.
Conversion of Actinoramide A (1) to 25-epi-Actinoramide A
(3) and the Separation. A 3 mg amount of actinoramide A (1) was
dissolved in 500 μL of pyridine-d5 for overnight NMR. 1H NMR
spectrum showed that it was a mixture of two components (4:1,
actinoramide A/25-epi-actinoramide A). The mixture was separated by
using semipreparative HPLC (Phenomenex, Luna, Phenyl-hexyl, 250 ×
10 mm, 5 μm, 2 mL/min, 33:67 CH3CN/H2O with 0.1% formic acid).
Actinoramide A (1) and 25-epi-actinoramide A (3, 0.6 mg) had
retention times of 21.2 and 23.0 min, respectively.
Small-Scale Reduction of Actinoramide D to Actinoramide A.
Actinoramide D (100 μg, 0.15 μmol), NaBH3CN (12 μg, 0.19 μmol),
and a trace amount of acetic acid in about 100 μL of dry CH3CN were
magnetically stirred at room temperature. After 1 h, the reaction mixture
was analyzed by LC/MS. The reduced product had the same retention
time, same UV spectrum, and same molecular weight as actinoramide A.
HPLC co-injection analysis was also performed to confirm the identity
of actinoramide A.
NMR (150 MHz, DMSO-d6), Table 2; HRMS m/z 660.3367 [M + H]+
(calcd for C31H46N7O9, 660.3357).
Actinoramide E (5): colorless powder; [α]23D −5.1 (c 0.09, CH3OH);
UV (CH3OH) λmax (log ε) 214(2.19), 232 (1.06, sh) nm; IR ν 3294,
1
2954, 2923, 2851,1701, 1649, 1529, 1454, 1349 cm−1; H NMR (600
MHz, DMSO-d6), Table 1; 13C NMR (150 MHz, DMSO-d6), Table 2;
HRMS m/z 619.3459 [M + H]+ (calcd for C30H46N6O8, 619.3505).
Actinoramide F (6): colorless powder; [α]23 −9.87 (c 0.15,
D
CH3OH); UV (CH3OH) λmax (log ε) 216 (2.31), 232 (1.26, sh), 292
1
(0.27) nm; IR ν 3292, 2958, 2931, 2875, 1708, 1649, 1529 cm−1; H
NMR (600 MHz, DMSO-d6), Table 1; 13C NMR (150 MHz, DMSO-
d6), Table 2; HRMS m/z 648.3353 [M + H]+ (calcd for C30H45N7O9,
648.3357).
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jnat-
prod.5b00489.
P. falciparum assay 1536-well protocols, performance
metrics, NPE retest, and EC50 (compounds 1−6) data,
Tables S1−S5, Data Table S1, Figures S1, S9, and S10.
Phylogenetic tree relating active S. sp. to other actino-
ramide-producing strains, Figure S2. Degradation and
Marfey reaction of compounds 1 and 3 and reduction of 4
to 1, Schemes S1 and S2. NMR and MS data for
compounds 1−6, and Figures S3−S8, S11−S15 (PDF)
Excel table (XLSX)
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail (D. H. Sherman): davidhs@umich.edu.
*E-mail (J. Inglese): jinglese@mail.nih.gov.
Author Contributions
□K. Chih-Chien Cheng, S. Cao, and A. Raveh contributed
equally to this work.
Notes
The authors declare no competing financial interest.
Marfey’s Analysis. Actinoramide A (1), 25-epi-actinoramide A (3),
and actinoramide E (5) (0.2 mg) were hydrolyzed in 170 μL of 6 N HCl
at 108 °C for 20 h, and the HCl was removed under vacuum. The dry
material was resuspended in 100 μL of H2O and dried three times to
remove residual acid. The residue was dissolved in 1 N NaHCO3 (100
μL). Marfey’s reagent (50 μL of 10 mg/mL 1-fluoro-2,4-dinitrophenyl-
5-L-alanine amide, in acetone) was added to each solution, and the
reaction mixtures were heated at 80 °C for 5 min. The reactions were
quenched by neutralization with 50 μL of 2 N HCl. Aqueous 50%
CH3CN (50 μL) was added to each solution to dissolve the products,
which were analyzed by LC/MS (Figures S11−14).
ACKNOWLEDGMENTS
■
This work was supported by NIH grant U01 TW007404 as part
of the International Cooperative Biodiversity Group Initiative at
the Fogarty International Center (D.H.S., G.T-C., and J.C.) and
the Hans W. Vahlteich Professorship (D.H.S.) and at the NIH
through the Divisions of Intramural Research at the National
Institute of Allergy and Infectious Diseases (X.-z.S.) and the
National Center for Advancing Translational Sciences (J.I.). The
authors are grateful to PNG BioNet and the University of Papua
New Guinea, Department of Environment and Conservation, for
permission to collect research samples. We thank P. Shinn
(NCATS) for compound management. We also thank S.
Carmeli (Tel Aviv University), A. Lowell (UMICH LSI) for
his help with the spectroscopic analysis, and R. L. Johnson (NCI)
for early support of this work. The Costa Rican marine sediment
sample was obtained under the permit R-CM-INBio-82-2009-
OT.
Actinoramide A (1): [α]23D −8.8 (c 0.26, MeOH); [α]21D −18 (c 0.1,
MeOH);15 [α]25D −10.7(c 5.2, MeOH).16
Actinoramide B (2): [α]23D −12 (c 0.21, MeOH); [α]21D −27 (c 0.1,
MeOH).15
25-epi-Actinoramide A (3): colorless powder; [α]23 −6.7 (c 0.15,
D
CH3OH); UV (CH3OH) λmax (log ε) 204 (4.20), 260 (sh), 268 (sh)
nm; IR ν 3377, 2842, 2356, 2349, 1725, 1645, 1385, 1354, 1023 cm−1;
1H NMR (600 MHz, DMSO-d6), Table 1; 13C NMR (150 MHz,
DMSO-d6), Table 2; HRMS m/z 662.3519 [M + H]+ (calcd for
C31H48N7O9, 662.3514).
REFERENCES
Actinoramide D (4): colorless powder; [α]23 −11.5 (c 0.20,
■
220.
D
CH3OH); UV (CH3OH) λmax (log ε) 207 (4.28), 237 (3.81, sh); 260
(sh), 268 (sh) nm; IR ν 3418, 2829, 2362, 2335, 2214, 1723, 1651, 1594,
(1) Koehn, F. E.; Carter, G. T. Nat. Rev. Drug Discovery 2005, 4, 206−
1381, 1351, 1026 cm−1; 1H NMR (600 MHz, DMSO-d6), Table 1; 13
C
(2) Butler, M. S. Nat. Prod. Rep. 2008, 25, 475−516.
K
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(3) Mayer, A. M.; Glaser, K. B.; Cuevas, C.; Jacobs, R. S.; Kem, W.;
Little, R. D.; McIntosh, J. M.; Newman, D. J.; Potts, B. C.; Shuster, D. E.
Trends Pharmacol. Sci. 2010, 31, 255−265.
(4) Carlson, E. E. ACS Chem. Biol. 2010, 5, 639−653.
(5) Cruz, P. G.; Auld, D. S.; Schultz, P. J.; Lovell, S.; Battaile, K. P.;
MacArthur, R.; Shen, M.; Tamayo-Castillo, G.; Inglese, J.; Sherman, D.
H. Chem. Biol. 2011, 18, 1442−1452.
(6) Inglese, J.; Johnson, R. L.; Simeonov, A.; Xia, M.; Zheng, W.;
Austin, C. P.; Auld, D. S. Nat. Chem. Biol. 2007, 3, 466−479.
(7) Inglese, J.; Auld, D. S.; Jadhav, A.; Johnson, R. L.; Simeonov, A.;
Yasgar, A.; Zheng, W.; Austin, C. P. Proc. Natl. Acad. Sci. U. S. A. 2006,
103, 11473−11478.
(8) Raveh, A.; Schultz, P. J.; Aschermann, L.; Carpenter, C.; Tamayo-
Castillo, G.; Cao, S.; Clardy, J.; Neubig, R. R.; Sherman, D. H.; Sjogren,
B. Mol. Pharmacol. 2014, 86, 406−416.
(9) Bennett, T. N.; Paguio, M.; Gligorijevic, B.; Seudieu, C.; Kosar, A.
D.; Davidson, E.; Roepe, P. D. Antimicrob. Agents Chemother. 2004, 48,
1807−1810.
(10) Corbett, Y.; Herrera, L.; Gonzalez, J.; Cubilla, L.; Capson, T. L.;
Coley, P. D.; Kursar, T. A.; Romero, L. I.; Ortega-Barria, E. Am. J. Trop.
Med. Hyg. 2004, 70, 119−124.
(11) Smilkstein, M.; Sriwilaijaroen, N.; Kelly, J. X.; Wilairat, P.; Riscoe,
M. Antimicrob. Agents Chemother. 2004, 48, 1803−1806.
(12) Yuan, J.; Cheng, K. C.; Johnson, R. L.; Huang, R.; Pattaradilokrat,
S.; Liu, A.; Guha, R.; Fidock, D. A.; Inglese, J.; Wellems, T. E.; Austin, C.
P.; Su, X. Z. Science 2011, 333, 724−729.
(13) Yuan, J.; Johnson, R. L.; Huang, R.; Wichterman, J.; Jiang, H.;
Hayton, K.; Fidock, D. A.; Wellems, T. E.; Inglese, J.; Austin, C. P.; Su, X.
Z. Nat. Chem. Biol. 2009, 5, 765−771.
(14) Brown, L. E.; Chih-Chien Cheng, K.; Wei, W. G.; Yuan, P.; Dai,
P.; Trilles, R.; Ni, F.; Yuan, J.; MacArthur, R.; Guha, R.; Johnson, R. L.;
Su, X. Z.; Dominguez, M. M.; Snyder, J. K.; Beeler, A. B.; Schaus, S. E.;
Inglese, J.; Porco, J. A., Jr. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 6775−
6780.
(15) Nam, S. J.; Kauffman, C. A.; Jensen, P. R.; Fenical, W. Tetrahedron
2011, 67, 6707−6712.
(16) Williams, D. E.; Dalisay, D. S.; Patrick, B. O.; Matainaho, T.;
Andrusiak, K.; Deshpande, R.; Myers, C. L.; Piotrowski, J. S.; Boone, C.;
Yoshida, M.; Andersen, R. J. Org. Lett. 2011, 13, 3936−3939.
(17) Du, Y. L.; Dalisay, D. S.; Andersen, R. J.; Ryan, K. S. Chem. Biol.
2013, 20, 1002−1011.
(18) Neumann, C. S.; Jiang, W.; Heemstra, J. R., Jr.; Gontang, E. A.;
Kolter, R.; Walsh, C. T. ChemBioChem 2012, 13, 972−976.
(19) Turk, B. Nat. Rev. Drug Discovery 2006, 5, 785−799.
(20) Goldberg, D. E.; Slater, A. F.; Cerami, A.; Henderson, G. B. Proc.
Natl. Acad. Sci. U. S. A. 1990, 87, 2931−2935.
(21) Matayoshi, E. D.; Wang, G. T.; Krafft, G. A.; Erickson, J. Science
1990, 247, 954−958.
(22) Asai, T.; Yamamoto, T.; Shirata, N.; Taniguchi, T.; Monde, K.;
Fujii, I.; Gomi, K.; Oshima, Y. Org. Lett. 2013, 15, 3346−3349.
(23) Balthaser, B. R.; Maloney, M. C.; Beeler, A. B.; Porco, J. A., Jr.;
Snyder, J. K. Nat. Chem. 2011, 3, 969−973.
(24) Charlop-Powers, Z.; Milshteyn, A.; Brady, S. F. Curr. Opin.
Microbiol. 2014, 19, 70−75.
(25) Kawamura, T.; Matsubara, K.; Otaka, H.; Tashiro, E.; Shindo, K.;
Yanagita, R. C.; Irie, K.; Imoto, M. Bioorg. Med. Chem. 2011, 19, 4377−
4385.
(26) McNally, A.; Prier, C. K.; MacMillan, D. W. Science 2011, 334,
1114−1117.
(27) Yamamura, H.; Hayakawa, M.; Iimura, Y. J. Appl. Microbiol. 2003,
95, 677−685.
(28) Magarvey, N. A.; Keller, J. M.; Bernan, V.; Dworkin, M.; Sherman,
D. H. Appl. Environ. Microbiol. 2004, 70, 7520−7529.
(29) Mincer, T. J.; Jensen, P. R.; Kauffman, C. A.; Fenical, W. Appl.
Environ. Microbiol. 2002, 68, 5005−5011.
(30) Kearse, M.; Moir, R.; Wilson, A.; Stones-Havas, S.; Cheung, M.;
Sturrock, S.; Buxton, S.; Cooper, A.; Markowitz, S.; Duran, C.; Thierer,
T.; Ashton, B.; Meintjes, P.; Drummond, A. Bioinformatics 2012, 28,
1647−1649.
L
DOI: 10.1021/acs.jnatprod.5b00489
J. Nat. Prod. XXXX, XXX, XXX−XXX