Metabolomics-Guided Discovery of Microginin Peptides from Cultures of the Cyanobacterium Microcystis aeruginosa

Article abstract of DOI:10.1021/acs.jnatprod.7b00829

We report a mass-spectrometry-based metabolomics study of a laboratory-cultured strain of Microcystis aeruginosa (UTEX LB2385), which has led to the discovery of five peptides (1-5) belonging to the microginin class of linear cyanopeptides. The structures and configurations of these peptides were determined by spectroscopic analyses and chemical derivitization. The microginin peptides described herein are the first reported derivatives containing N-methyl methionine (1, 5) and N-methyl methionine sulfoxide (2-4). The two tripeptide microginin analogues (4, 5) identified represent the smallest members of this peptide family. Their angiotensin-converting enzyme (ACE) inhibitory activity was also investigated. Microginin 527 (4) was the most potent of the group, with an IC₅₀ of 31 μM.

Full text of DOI:10.1021/acs.jnatprod.7b00829

Article
Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX
pubs.acs.org/jnp
Metabolomics-Guided Discovery of Microginin Peptides from
Cultures of the Cyanobacterium Microcystis aeruginosa
Allison K. Stewart, Rudravajhala Ravindra, Ryan M. Van Wagoner, and Jeffrey L. C. Wright*
Center for Marine Science, Department of Chemistry and Biochemistry, University of North Carolina Wilmington, 5600 Marvin K.
Moss Lane, Wilmington, North Carolina 28409, United States
*
S Supporting Information
ABSTRACT: We report a mass-spectrometry-based metabolomics study of a
laboratory-cultured strain of Microcystis aeruginosa (UTEX LB2385), which has led
to the discovery of five peptides (1−5) belonging to the microginin class of linear
cyanopeptides. The structures and configurations of these peptides were
determined by spectroscopic analyses and chemical derivitization. The microginin
peptides described herein are the first reported derivatives containing N-methyl
methionine (1, 5) and N-methyl methionine sulfoxide (2−4). The two tripeptide
microginin analogues (4, 5) identified represent the smallest members of this
peptide family. Their angiotensin-converting enzyme (ACE) inhibitory activity was also investigated. Microginin 527 (4) was the
most potent of the group, with an IC50 of 31 μM.
yanobacteria are prokaryotic organisms also known as
blue-green algae. Blooms of cyanobacteria are a natural
derivatives, with few exceptions, contain 4−6 residues including
one or more tyrosines at the C-terminus and the characteristic
N-terminal side chain, 3-amino-2-hydroxy decanoic acid
(Ahda). Both mono- and dichlorinated variants of this side
chain have been reported as well as some derivatives containing
C
phenomenon and can occur in both inland and coastal
1
−3
waters.
Freshwater cyanobacterial blooms are on the rise
worldwide, resulting in increased attention to monitoring and
regulation of the hepatotoxic microcystins often produced in
24
a modified octanoic acid moiety. Congeners of this peptide
class contain various amino acids between the N-terminal β-
amino acid and the C-terminal tyrosine residues including Ile,
Leu, Val, Ala, Pro, and Ser, some of which contain a methylated
amino group. Many of the microginins reported to date exhibit
angiotensin-converting enzyme (ACE) inhibition and thus may
serve as lead compounds in the discovery of novel
antihypertensive agents. The microginin peptides described
herein include five new variants (1−5), which are the first
reported microginin derivatives containing N-methyl methio-
nine (1, 5) and N-methyl methionine sulfoxide (2−4). The two
tripeptide microginin analogues (4, 5) identified represent the
smallest members of this peptide family. The identification,
isolation, and structure determination of these five microginin
congeners are described as well as their ACE inhibitory activity.
4
these events. Microcystis sp. is the most common bloom-
5
forming genus of cyanobacteria and, to date, has been reported
6
on every continent except Antarctica. Several classes of
bioactive linear and cyclic peptides, in addition to the
heptapeptide microcystins, have been isolated from Microcystis
7
−10
aeruginosa strains,
and recent genomic studies have
highlighted the potential of this species in drug discovery
efforts by uncovering further biosynthetic gene clusters
11,12
encoding unidentified chemistry.
Application of secondary metabolomics to guide isolation of
new chemistry has been investigated in marine cyanobac-
1
3,14
15
16
teria
as well as plants, nonphotosynthetic bacteria, and
7
1
fungi, but has hardly been applied to discover new chemistry
18
in freshwater cyanobacteria, where published metabolomics
studies have largely focused on untargeted analysis of primary
metabolites.
1
9,20
The identification of strains of M. aeruginosa as
RESULTS AND DISCUSSION
■
an untapped resource for the discovery of novel chemistry
coupled with its cosmopolitan distribution and implication in
toxic bloom events globally make it an attractive and relevant
candidate for secondary metabolomics analysis.
Metabolomics of Microcystis aeruginosa (UTEX
LB2385). Methanolic extracts of harvested M. aeruginosa cells
were loaded onto SPE cartridges for removal of matrix
interferences, such as salts and phospholipids, prior to UPLC-
QTof MS analysis. Six replicate cell extractions were processed
in an identical manner in order to account for natural variance
within biological systems and meet proposed minimum
A rapid and robust UPLC-QTof MS metabolomics method
for secondary metabolite profiling of laboratory cultures of M.
aeruginosa (UTEX LB2385) resulted in the identification of
new congeners of the microginin class of linear nonribosomal
cyanopeptides. The first microginin congener was reported
from M. aeruginosa in 1993 by Okino and co-workers, and
further studies have described the isolation of numerous
microginin congeners from several cyanobacteria genera
25
standards for metabolomics analysis. Raw data files were
imported into Progenesis QI software for processing.
Received: September 28, 2017
21−23
including Microcystis, Nostoc, and Oscillatoria.
Microginin
©
XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.7b00829
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 1. S-plot of M. aeruginosa UTEX LB2385 cell extracts (6 replicates, 3 injections of each) versus B3N media blank (1 replicate, 3 injections).
The blue box indicates compound ions with loading scores of >0.03 from the cell extracts, which were selected for manual data mining and
dereplication. Compound ions of microginin derivatives discovered using this approach are designated with green circles.
Compound ions of interest after application of various filters
were exported to EZInfo 3.0 software for multivariate analysis,
including principle component analysis (PCA) and orthogonal
projections to latent structures discriminant analysis (OPLS-
DA). The OPLS-DA model was used to construct an S-plot
comparing compound ions from M. aeruginosa UTEX LB2385
cell extracts to those in the media extract (Figure 1).
Compound ions with loading values of >0.03 from the cell
extracts were selected for further analysis, resulting in 88
compound ions of interest. Manual data mining of the resulting
compound list identified extraneous compound ions that were
falsely identified by the software algorithm, most often high-
intensity fragment ions from more abundant metabolites in the
samples. This is a trade-off in selecting a sample concentration
to maximize the number of secondary metabolites that are
above the limit of detection and a caveat of natural product
metabolomics data sets that contain diverse chemistry
produced over a wide range of concentrations. After exclusion
of compound ions of abundant fragment ions during manual
data mining, a total of 41 compounds were identified in the
intracellular secondary metabolome of M. aeruginosa (UTEX
LB2385). Eight of these compounds could further be classified
into three cyanopeptide classes based on characteristic fragment
tolerance), isotope ratio, and fragmentation scores within
selected Progenesis databases (ChemSpider, MassBank, Metlin,
KEGG, HMDB). Compound identification is a recognized
bottleneck in metabolomics experiments due to lack of database
27
coverage. Acquisition of MS/MS data for these compound
ions in order to query fragment databases such as GNPS,
MoNA, and CSIFingerID is one approach to improve
compound identification and is currently under investigation
28
in our metabolomics workflow.
Microginin Derivatives from Microcystis aeruginosa
(1−5). Mass-spectrometry-guided isolation of three microginin
congeners (m/z 675.3427, m/z 691.3378, m/z 528.2739;
fragment ion of m/z 128.1438 in all high-energy spectra)
identified by metabolomics analysis (Figure 1) using scaled-up
cultures of M. aeruginosa UTEX LB2385 successfully resulted in
the isolation of sufficient material for full structure elucidation.
Futhermore, an additional two minor related compounds (m/z
512.2785 and m/z 705.3530; fragment ion of m/z 128.1438)
were identified in the scaled-up material, likely not found by
metabolomics due to intensities below the limit of detection in
the small-scale analysis. The isolation and structure elucidation
of these five related microginin congeners (1−5) is described.
Microginin 674 (1) was isolated as a glassy yellow solid. The
HRMS and NMR data revealed that 1 had a molecular formula
26
ions observed in the high-energy spectra. The presence of
three microcystins (Adda fragment at m/z 135, [M + H −
of C H N O S. The UV data (λ
224, 276 nm) were
34
50
4
8
max
+
1
34] neutral loss), two cyanopeptolins (Ahp-Phe substructures
consistent with the presence of an aromatic ring, and
preliminary inspection of the NMR data confirmed the
presence of four ortho-coupled aromatic proton signals, an
N-Me group (δ 2.65), and 16 aliphatic protons in the range δ
indicated by ions at m/z 215 and m/z 243), and three
microginins (Ahda fragment at m/z 128) were identified using
this method, in addition to a number of compound ions that
could not be further classified by high-energy fragmentation
data. This metabolite profile was consistent with genome
sequencing of 10 representative M. aeruginosa strains, which
revealed each strain contained two to nine biosynthetic gene
H
H
13
1.19−1.43 ppm. The C spectrum contained four carbonyl
resonances (δ 172.8, 170.7, 171.6, and 169.1) suggesting that
C
this compound was a tetrapeptide. A doubling of many of the
resonances was observed in both the H and C spectra in a
1
13
1
2
clusters. Aside from the microcystin congeners, all other
metabolites identified were concluded to represent new
structural variants after dereplication using both in-house and
commercial databases (ChemSpider, Metlin, MassBank, Anti-
base, MarinLit). Compound ions not classified as cyanopep-
tides by high-energy fragmentation data were not able to be
confidently identified using high-resolution mass (2 ppm
ratio of ca. 3:1, consistent with the presence of an N-methylated
2
9
amide linkage in a linear peptide. The presence of two
aromatic residues was confirmed by the ESIMS data, which
displayed a strong fragment ion at [M + H − 181] , indicating
loss of a terminal tyrosine residue, and a fragment ion at m/z
343 corresponded to a dityrosine fragment (Table 1, Figure 2).
This was supported by extensive NMR analyses including
+
B
DOI: 10.1021/acs.jnatprod.7b00829
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Table 1. ESIMS Fragment Ions for Microginins 1−5
at m/z 128 and substantiated by extensive NMR data. The
NMR data easily established the presence of two tyrosine units
ion structure
1
2
3
4
5
1
(
Table 2), but many of the resonances observed in both the H
[
[
[
[
[
[
[
[
[
[
[
[
M + Na]+
697
675
713
691
627
510
727
705
641
550
528
464
534
512
13
and C spectra of 2 and tentatively assigned to an N-methyl
methionine residue appeared as doublets in a ratio of ca. 3:1.
Furthermore, compared with 1, the S-methyl resonances in the
NMR spectra of 2 now appeared deshielded at δ 2.47 and δ
M + H]+
+
M + H − CH SOH]
M + H − Tyr]
M + H − O-Me Tyr]
M + H − Tyr − CO]
M + H − O-Me Tyr − CO]
H + Tyr + O-Me Tyr]+
H + Tyr + Tyr]+
N-Me Met(O)]+
N-Me Met]+
3
+
494
H
C
+
510
3
7.8 (Table 3). Additionally, ESIMS of 2 revealed a strong
+
482
+
fragment ion of [M + H − 64] not seen in 1, corresponding to
the loss of CH SOH (Table 1). Collectively these data
+
482
357
3
established 2 contained an N-methyl methionine sulfoxide
amino acid residue. The presence of the sulfoxide group in 2 as
a mixture of R and S configurations resulted in the doubling of
several NMR signals for this residue, making spectral
interpretation more difficult. The doubling of signals due to
R and S sulfoxides has been described in other methionine
sulfoxide-containing cyanopeptides, and, as has been suggested
for those metabolites, the sulfoxide in 2 is most likely an artifact
of aerial oxidation of the N-methyl methionine natural product
343
343
180
180
128
180
128
164
128
164
128
+
CH (CH ) CHNH ]
128
3
2
6
2
31,32
during isolation.
The high-resolution MS and NMR data for microginin 704
(
3) (λmax 224, 276 nm) established a molecular formula of
C H N O S, which differs from 2 by 14 Da, suggesting the
35
52
4
9
1
13
Figure 2. Schematic of ESIMS fragmentation of microginin 674 (1).
presence of an additional methyl group. Indeed, the H and C
NMR spectra of 3 closely resembled those of 2, except for new
signals in the NMR spectra of 3 suggesting the presence of a
TOCSY, COSY, HSQC, and HMBC spectra. A characteristic
fragment ion at m/z 128 indicated the presence of an N-
methoxy group at δ 3.60 ppm and δ 53 ppm (Table 2). The
H C
1
13
terminal Ahda residue, which was supported by the H and C
spectra (Table 2). The fragment at m/z 128 results by scission
of the C-2 and C-3 bond in the Ahda moiety (Figure 2).
location of the methoxy group was determined by ESIMS
fragment ions at m/z 510 and 357, corresponding to loss of a
terminal methoxy tyrosine residue and a MeOTyr-Tyr fragment
(Table 1). The presence of a methionine sulfoxide residue was
1
1
The presence of Ahda was further verified by H− H
TOCSY data, which displayed correlations from the proton on
C-2 to the terminal methyl group, which was consistent with
+
supported by a strong fragment ion at [M + H − 64] in the
ESIMS of 3, corresponding to a loss of CH SOH. An ESIMS
3
30
reported data for Ahda-containing microginins. The NMR
features of the final peptide residue contained a characteristic
set of signals (δ 2.02 and δ 14.3) corresponding to an S-
fragment ion of m/z 128 and the 2D NMR data indicated the
Ahda residue was also unchanged.
H
C
The high-resolution MS spectrum as well as NMR data
revealed that microginin 527 (4) had a molecular formula of
C H N O S, corresponding to seven double-bond equiv-
methyl group of a methionine residue. While other NMR
resonances were consistent with a methionine side chain, the
amide nitrogen was methylated as indicated by an N-methyl
shift at δH 2.65 and δC 30.3. These data led to the final
assignment of this residue as N-methyl methionine and the
amino acid sequence of 1 to be Ahda-N-MeMet-Tyr-Tyr.
25
41
3
7
1
13
alents. While the H and C NMR spectra of 4 resembled
those of 2, the difference of 163 Da and four double-bond
equivalents in 4 suggested that this compound lacked a tyrosine
residue. This was confirmed by the NMR data (Table 3), which
displayed resonances for only one tyrosine residue. Once again
the N-terminal Ahda was confirmed by a fragment ion at m/z
1
28 and by TOCSY data. The ESIMS of 4 also showed a strong
+
fragment ion at [M + H − 64] , indicating the loss of CH SOH
3
(
Table 1) which, coupled with the NMR data, established the
presence of an N-methyl methionine sulfoxide.
A molecular formula of C H N O S for microginin 511
25
41
3
6
(
5), established by high-resolution MS and NMR data,
suggested that this minor component was similar to 4, but
contained an N-methyl methionine residue rather than the
1
sulfoxide derivative. This residue was verified by the H NMR
1
1
and H− H TOCSY data, which was also used to identify a
single tyrosine residue and the N-terminal Ahda moiety, as
indicated by the fragment ion at m/z 128. The MW for 5 was
163 Da less than 1, and the similarity of the NMR data between
the two compounds established that 5 was similar to 1 but
contained a single C-terminal tyrosine.
While the spectrometric data for these linear peptides were
relatively straightforward to interpret, as noted earlier the
presence of epimeric sulfoxides in 2−4 and the 3:1 ratio of
The molecular formula of microginin 690 (2) was
determined to be C H N O S based on the high-resolution
3
4
50
4
9
MS and NMR data and denoted the presence of an additional
oxygen compared with 1. The spectroscopic data for this
second compound (UV data: λ
224, 276 nm; ESIMS
max
+
fragment ions at [M + H − 181] as well as m/z 343) were
consistent with a dityrosine C-terminus. Once again, the N-
terminal Ahda was confirmed by the characteristic fragment ion
C
DOI: 10.1021/acs.jnatprod.7b00829
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Table 2. H and 13C NMR Data for Tetrapeptide Microginins (1−3) in DMSO-d ( H 500 MHz, 13C 125 MHz)
1
1
6
microginin 674 (1)
microginin 690 (2)
δC, type δH (J in Hz)
170.7, C
microginin 704 (3)
δC, type δH (J in Hz)
170.7, C
position
δC, type
δH (J in Hz)
Ahda
1
2
3
4
5
6
7
8
9
1
170.7, C
69.6, CH
53.2, CH
28.4, CH2
24.6, CH2
28.6, CH2
29.1, CH2
31.0, CH2
22.0, CH2
13.9, CH3
4.01, br s
3.10, m
69.6, CH
53.2, CH
28.4, CH2
24.6, CH2
28.6, CH2
29.1, CH2
31.0, CH2
22.0, CH2
13.9, CH3
4.01, br s
3.10, m
69.6, CH
53.1, CH
28.4, CH2
24.6, CH2
28.6, CH2
29.1, CH2
31.0, CH2
22.0, CH2
13.9, CH3
4.01, br s
3.05, m
1.27, m
1.27, m
1.27, m; 1.43 (m)
1.19, m
1.19, m
1.19, m
1.22, m
1.22, m
1.22, m
1.21, m
1.21, m
1.21, m
1.23, m
1.23, m
1.23, m
1.24, m
1.24, m
1.24, m
0
0.85, t (6.7)
6.49, br s
7.78, br
0.85, t (6.8)
6.49, br s
7.79, br
0.85, t (6.1)
6.52, br s
7.78, br
OH
NH2
N-Me Met/N-Me Met (O)
1
169.1, C
169.1, C
169.1, C
a
a
a
a
2
55.1, CH
28.6, CH2
29.4, CH2
14.3, CH3
30.3, CH3
171.6, C
5.01, m
2.42, m
2.27, t (7.8)
2.02, s
54.9 (55.3 ), CH
5.06 (4.99 ), m
54.9 (55.3 ), CH
5.06 (4.99 ), m
a
a
3
20.9 (21.1 ), CH
2.10, m
20.9 (21.1 ), CH
2.09, m
2
2
3
2
2
3
a
a
a
a
4
49.4 (49.6 ), CH
2.54 (2.57 ), m
49.4 (49.6 ), CH
2.54 (2.57 ), m
a
a
a
a
S-Me
37.8 (38.0 ), CH
2.47 (2.49 ), s
37.8 (38.0 ), CH
2.45 (2.47 ), s
a
a
a
a
N-Me
2.65, s
31.1 (31.2 ), CH3
171.8, C
2.63 (2.66 ), s
32.0 (32.1 ), CH3
171.8, C
2.63 (2.66 ), s
Tyr
1
2
3
4
50.5, CH
35.8, CH2
127.0, C
4.36, m
2.81, m
50.9, CH
4.41, m
2.89, m
50.9, CH
4.41, m
2.89, m
35.4, CH2
127.0, C
35.4, CH2
127.0, C
5
6
7
7
, 9
, 8
130.0, CH
115.0, CH
156.0, C
6.99, m
6.64, m
130.0, CH
115.1, CH
156.0, C
7.03, m
6.64, m
130.2, CH
115.1, CH
156.0, C
7.01, m
6.63, m
-OH
9.25, br
9.25, br
9.25, br
NH
1
8.38, d (6.7)
8.05, d (7.5)
8.30, d (7.3)
Tyr/O-Me Tyr
172.8, C
172.8, C
172.8, C
2
53.8, CH
35.8, CH2
127.2, C
4.73, m
2.86, m
53.8, CH
35.8, CH2
127.6, C
4.72, m
2.82, m
53.8, CH
35.8, CH2
127.6, C
4.72, m
2.82, m
3
4
5
6
7
7
, 9
, 8
129.9, CH
114.9, CH
155.9, C
7.05, m
6.68, m
129.9, CH
114.9, CH
155.9, C
7.05, m
6.68, m
129.9, CH
114.9, CH
155.9, C
6.99, m
6.62, m
-OH
9.22, br
9.25, br
NH
-OMe
8.18, d (7.5)
8.21, d (7.5)
8.20, d (7.4)
3.60, s
7
53.0, CH3
a
Doubling of signals due to R and S sulfoxides.
conformational isomers as a result of N-methylation in 1−5
made assignment of the NMR resonances more difficult.
Configurations of the Microginins 1−5. The L-
configuration of the N-methyl methionine and tyrosine residues
in 2 was determined by oxidation of the sulfoxide to the
sulfone, then acid hydrolysis followed by derivatization with
Biological Activity. Four of the microginin derivatives (1−
4) reported in this study were produced in sufficient amounts
to allow testing for inhibition of angiotensin-converting
enzyme. The tripeptide microginin 527 (4) was the only
congener of the four tested in this study that displayed ACE
inhibition below 100 μM with an IC50 value of 31 μM. IC50
values above 100 μM were not determined due to scarcity of
available material. The LC/MS assay allowed observation of the
microginin peptides used for inhibition, confirming that 1 was
not oxidized to 2 during the course of the assay. Despite this,
there were no significant differences between 1 and 2 in
inhibitory potency by this assay, indicating that the presence of
the sulfoxide was not the cause for low levels of inhibition.
Previous structure−activity studies on this class of ACE
inhibitors indicated that the dityrosine C-terminus as well as
the free amino and hydroxy groups of the Ahda side chain play
33
Marfey’s reagent and LC/MS analysis. Due to lack of
commercially available standards, N-methyl methionine sulfone
required synthesis from L-methionine and D-methionine in
order to determine the configuration of this amino acid. The
relative stereochemistry of Ahda was determined by NMR
analysis of the oxazolidinone derivative of 2, which revealed a
coupling constant of 5.4 Hz between H-2 and H-3 (numbering
based on the Ahda chain prior to derivatization), which, upon
comparison to previous reports for similar Ahda derivatives in
34,35
other microginins,
suggests a trans relationship between the
35
alcohol and amine functionalities. Consistency in the specific
rotation values obtained for the five congeners suggests that the
configurations of 1 and 3−5 are identical to those determined
for 2.
a key role in ACE inhibition. Early studies on snake venom
peptide analogues revealed that the tripeptide residues at the C-
terminus of long-chain peptides were the most important for
36
ACE inhibition. More recent studies on hydrolyzed food
D
DOI: 10.1021/acs.jnatprod.7b00829
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Table 3. H and 13C NMR Data for Tripeptide Microginins
1
the biosynthetic machinery for more unidentified products.
Consequently, metabolomics evaluation of this genus will
undoubtedly provide new chemistry as well as new insight into
the biological functions of cyanobacterial secondary metabo-
lites. This preliminary metabolomics study serves as a baseline
of the secondary metabolome of M. aeruginosa (UTEX
LB2385) under standard culture conditions, and further studies
using changes in culture parameters or addition of chemical
elicitors may facilitate the expression of cryptic biosynthetic
clusters, allowing access to additional chemistry. Furthermore,
as the occurrence of freshwater cyanobacterial blooms persists
globally, an in-depth understanding of the complex metab-
olome of Microcystis spp. is essential to rapidly and accurately
assess the true threat of bloom events on human health.
1
13
(
4 and 5) in DMSO-d ( H 500 MHz, C 125 MHz)
6
microginin 527 (4)
, type (J in Hz)
microginin 511 (5)
δH (J in
Hz)
position
δ
C
δ
H
δ
C
, type
Ahda
1
2
3
4
5
6
7
8
9
1
171.3, C
69.6, CH
53.1, CH
28.4, CH
24.6, CH
28.6, CH
29.1, CH
31.0, CH
22.0, CH
13.9, CH
171.4, C
71.0, CH
53.5, CH
28.4, CH
24.6, CH
28.6, CH
29.1, CH
31.0, CH
22.0, CH
4.00, br s
3.01, m
1.27, m
1.19, m
1.22, m
1.21, m
1.23, m
1.24, m
4.00, br s
3.05, m
1.27, m
1.19, m
1.22, m
1.21, m
1.23, m
1.24, m
0.85, t
2
2
2
2
2
2
3
2
2
2
2
2
2
3
0
0.85, t (6.6) 14.3, CH
(
6.5)
EXPERIMENTAL SECTION
■
OH
NH
1
6.55, br s
7.78, br
6.51, br s
7.79, br
General Experimental Procedures. Optical rotations were
measured on a Rudolph Research Analytical Autopol III automatic
polarimeter. UV spectra were recorded on an HP1100 diode array
2
N-Me Met/ N-
Me Met(O)
170.4, C
170.5, C
1
detector. H and 2D homonuclear (TOCSY, COSY) NMR spectra
N-Me Met (O)
2
3
4
55.9
4.98
(4.91 ), m
58.1, CH
4.95, m
a
a
were measured on an Avance 500 MHz NMR spectrometer (Bruker)
(
55.6 ),
1
3
using 5.0 mm TXI and 1 mm TXI probe heads. C and 2D
heteronuclear (HSQC, HMBC) NMR spectra were also measured on
an Avance 500 MHz NMR spectrometer using a 5 mm BBO probe
CH
20.8
20.9 ),
CH2
49.5
49.6 ),
CH
2.21, m
28.8, CH2 2.35, m
29.2, CH2 2.28, m
a
(
head. All NMR spectra were acquired in DMSO-d , and the chemical
6
2.51
(2.59 ), m
shifts were referenced to DMSO-d (δH 2.50 for proton and δ 39.5
6
C
a
a
(
for carbon). Low-resolution ESI mass spectra were measured on a
Micromass ZQ with electrospray ionization (capillary voltage 3.1 kV;
cone voltages 30 V/50 V; source temperature 80 °C; desolvation
temperature 150 °C; cone gas flow 50 L/h; desolvation gas flow 300
L/h) (Waters Corporation) in positive ion mode with a mass range of
2
S-Me 37.9, CH
3
2.79
2.82 ), s
2.89
(2.92 ), s
14.9, CH
32.1, CH
3
3
2.01, s
2.90, s
a
(
N-Me 31.9
a
a
(
32.1 ),
1
00−1400 Da and a scan time of 0.5 s acquired using MassLynx v 4.1
CH
3
software. High-resolution data for accurate molecular weight, MS/MS
fragmentation, and metabolomics were measured using a Xevo G2-XS
QTof spectrometer controlled by MassLynx v 4.1 software (capillary
voltage 2.5 kV; sampling cone 80 V; source offset 80 V; source
temperature 120 °C; desolvation temperature 550 °C; cone gas flow
Tyr
1
2
3
4
5
171.5, C
50.5, CH
35.9, CH
126.6, C
171.0, C
52.0, CH
36.4, CH
127.0, C
4.85, m
2.71, m
4.82, m
2.71, m
2
2
, 9
, 8
130.2, CH 7.05, d (8.8) 131.8, CH 7.02, d
8.6)
115.0, CH 6.68, d (8.7) 115.0, CH 6.63, d
8.4)
1
00 L/h; desolvation gas flow 800 L/h) (Waters Corporation)
(
coupled with an Acquity iClass UPLC system (Waters Corporation)
with flow-through needle. HPLC separations were performed using a
Waters HPLC system (model 515 pumps and a model 680 automated
gradient controller) equipped with a Waters 2487 dual λ UV detector.
All MS solvents and modifiers were Fisher Optima MS grade, and all
HPLC solvents were BDH HiPerSolv Chromanorm HPLC grade.
Biological Material. Microcystis aeruginosa (UTEX LB2385) was
obtained from the University of Texas Culture Collection and cultured
in our laboratory in Bold 3N (B3N) media and a 14/10 h light−dark
6
(
7
7
156.0, C
156.0, C
-OH
9.25, br
9.25, br
8.22, d
NH
8.11, d (7.5)
(
7.4)
a
Doubling of signals due to R and S sulfoxide.
proteins have shown that ACE inhibition is enhanced in
peptides containing aromatic and branched chain amino acids
at the three C-terminal positions.
could explain the vast differences in inhibitory activity between
microginin 527 and the tripeptide degradation product of
microginin Ahda-Ala-Val, which display IC values of 31 and
39
cycle.
Metabolomics Sample Preparation. Cells from 1.0 L of M.
aeruginosa (UTEX LB2385) were harvested by vacuum filtration and
stored on glass fiber filters at −20 °C prior to extraction. Six replicate
harvests were processed individually. Cells were extracted with 80:20
37,38
This substrate preference
MeOH/H O (2×), sonicating for 15 min prior to filtration. Combined
2
5
0
3
5
and filtered extracts were taken to dryness in vacuo and resuspended in
460 μM respectively.
1
0 mL of 95:5 H O/MeOH for processing. An aliquot (1 mL) of the
2
Conclusion. The UPLC-QTof MS method for obtaining
extract was loaded onto an SPE column (Waters Oasis Prime HLB 6
cm /200 mg) and washed with 1 column volume (CV) of 95:5 H O/
secondary metabolite profiles of laboratory-cultured cyanobac-
teria described here allows rapid access to the cyanobacterial
secondary metabolome using minimal biomass. This method
successfully captured a large number of cyanopeptide
metabolites and is amenable to high-throughput processing.
The discovery of five congeners of the microginin class of linear
cyanopeptides using this method adds to the mounting
evidence for freshwater cyanobacteria as a source of new
chemistry. A large proportion of Microcystis spp. have yet to be
analyzed for secondary metabolite production, and genomic
analysis on a small number of Microcystis strains has revealed
3
2
MeOH in order to remove matrix interferences. The column was then
eluted with 1 CV of 90:10 MeCN/MeOH collected into a scintillation
vial. An aliquot (10%) of this fraction was filtered using a 0.2 μm 13
mm PTFE syringe filter (Pall Acrodisc) directly into an MS certified
autosampler vial with preslit cap (Waters TruView). B3N media (100
mL) was also processed using this protocol. A quality control (QC)
sample containing equal parts (50 μL) of each of the six replicate cell
extracts and media control was prepared from the filtered samples.
Metabolomics Data Acquisition. Chromatographic separation
for metabolomics analysis was achieved using a Waters Acquity BEH
C18 2.1 × 100 mm 1.7 μm column with gradient elution (mobile phase
E
DOI: 10.1021/acs.jnatprod.7b00829
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
A: H O + 0.1% formic acid (FA); mobile phase B: MeCN + 0.1% FA;
diluted with H O and extracted with CH Cl to afford N-Boc-L-
2
2
2
2
+
held at 5% B for 1 min, 5−65% B over 8 min, 65−100% B over 1 min,
hold at 100% B for 1.5 min, re-equilibrate at 5% B for 4.5 min; flow
rate: 0.5 mL/min). Samples were injected in triplicate (1 μL), and QC
samples were analyzed at the beginning, middle, and end of the data
acquisition in order to aid the software in alignment of the
chromatograms during processing. Continuum data were acquired in
methionine sulfone (m/z 282 [M + H] ). The sulfone was suspended
in dry THF (2 mL). NaH (50 mg) and MeI (330 μL) were added, and
the mixture was stirred at room temperature for 24 h. The reaction
mixture was diluted with H O and extracted with EtOAc. The aqueous
2
layer was then acidified with 20% citric acid and extracted with EtOAc.
The combined organic layers were concentrated to afford N-Boc-N-
e
+
resolution mode using the MS function (mass range 50−1800 Da;
methyl-L-methionine sulfone (m/z 296 [M + H] ), which was
scan time 0.1 scans/s; function 1 CE off; function 2 CE ramp 45−50
V; lockmass data acquired but not applied). Leucine enkephalin (200
pg/μL, infused at 20 μL/min) was used as the lockmass solution.
Metabolomics Data Processing Parameters. Continuum data
were uploaded into Progenesis QI software (Nonlinear Dynamics) for
alignment, lock mass calibration, and peak picking. QC samples were
chosen for alignment reference, and peak picking parameters were set
to default settings (automatic sensitivity threshold; no minimum
chromatographic peak width), with the exception of retention time,
which was restricted to 1−10 min. Progenesis identified 6287
compound ions under the selected threshold values. This data set
was filtered to remove compounds with ANOVA p-values of >0.05 as
well as compound ions with the highest mean intensity in the solvent
blank sample, resulting in 3724 compound ions, which were then
subjected to statistical analysis using an OPLS-DA model to construct
an S-plot comparing cell extracts to the media extract using EZInfo 3.0
software (UMetrics) (Figure 1).
dissolved in equal volumes of 1 M HCl (0.5 mL) and EtOAc (0.5 mL)
and stirred at room temperature for 2 h. The solution was
concentrated to give N-methyl-L-methionine sulfone (m/z 196 [M +
+
H] ). This material was used for the Marfey’s analysis. N-Methyl-D-
methionine sulfone was synthesized from N-Boc-D-methionine in the
same manner. To a solution of microginin 690 (2, 0.5 mg) in MeOH
(0.5 mL) was added oxone (5 mg in 100 μL of H O), and the mixture
2
was stirred at room temperature for 2 h. The solution was diluted with
3
H O and applied to a small SPE column (Waters Oasis Maxx 6 cm ),
2
washed with 10% MeOH, and eluted with 100% MeOH to afford
+
microginin 690 sulfone (m/z 707 [M + H] ). Microginin 690 sulfone
(0.5 mg) was dissolved in 6 N HCl (0.5 mL) and heated at 110 °C for
5 h. To the acid hydrolysate of a 500 μg portion of microginin 690
sulfone were added 100 μL of 1-fluoro-2,4-dinitrophenyl-5-L-alanine
amide (L-FDAA) (10 mg/mL) and 1 M NaHCO (100 μL), and the
3
reaction mixture heated at 40 °C for 1 h. After cooling to room
temperature, the reaction was quenched with 6 N HCl (5 μL) and
concentrated under a stream of nitrogen. D- and L-Amino acid
standards were treated with L-FDAA in the same manner. The L-FDAA
derivatives were resuspended in MeOH (100 μL) and filtered (0.2 μm
13 mm GHP syringe filter) for analysis by reversed-phase UPLC-MS
(Waters Acquity BEH C , 2.1 × 100 mm 1.7 μm column; 5−40%
Extraction and Isolation Procedure for Microginins. The 80%
aqueous MeOH extract of cells collected on filter papers by vacuum
filtration (from 256 L of culture, extracted in 32 L batches) was
washed with hexane. The aqueous layer was dried in vacuo and applied
to a reversed-phase column eluted with a MeOH gradient
18
(
Phenomenex Strata C18 20g/60 mL; step gradient 20−100%
MeCN + 0.1% FA over 9 min; flow rate 0.65 mL/min). Retention
MeOH), yielding fractions containing the masses of interest. Further
times (t ) of the standard amino acids: L-tyrosine (5.97 min), D-
R
purification of these fractions by gel permeation chromatography (LH-
tyrosine (6.70 min), N-Me-L-methionine sulfone (4.97 min), N-Me-D-
2
×
4
0), followed by C18 reversed-phase HPLC (Waters SunFire RPC18 19
250 mm; 40−100% MeOH + 0.05% trifluoroacetic acid (TFA) over
0 min; flow rate 10 mL/min, UV detection at 280 nm), provided five
methionine sulfone (5.05 min). The absolute configuration of the
amino acids in 2 were determined as L-tyrosine (t 5.97 min) and N-
R
Me-L-methionine sulfone (t 4.97 min).
R
new microginins (1−5). Final purification of these peptides was
accomplished by Amide C16 reversed-phase HPLC (Supelco Discovery
Amide C16 10 × 250 mm; 25−45% MeCN + 0.05% TFA over 20 min;
flow rate 2.5 mL/min, UV detection 280 nm) to yield 1 (18.2 min,
Formation of the Oxazolidinone Derivative. A vial containing
microginin 690 (2, 2.1 mg) was desiccated under vacuum prior to
addition of carbonyl diimidiazole (2 mg) and dry THF (700 μL)
under a blanket of nitrogen prior to sealing. The reaction was
subjected to magnetic stirring at room temperature for 24 h, after
which the sample was dried under a stream of nitrogen. The sample
was purified by HPLC (Waters XTerra MS C18 4.6 × 150 mm; 20−
45% MeCN + 0.05% TFA over 25 min; flow rate 1.0 mL/min) and
analyzed by NMR.
1
4.0 mg), 2 (12.1 min, 23.0 mg), 3 (13.5 min, 6.5 mg), 4 (9.7 min, 9.0
mg), and 5 (15.0 min, 2.2 mg).
2
6
Microginin 674 (1): [α] −41 (c 0.28, MeOH); UV (MeOH) λ
D
max
1
13
(
log ε) 224 (4.19), 276 (3.39); H and C NMR, Table 2; positive
ESIMS m/z (relative intensity) 675 (53.7), 494 (100), 343 (36.2), 164
25), 128 (18.7); HRMS m/z 675.3427 [M + H]+ (calcd for
C H N O S, 675.3428).
(
Angiotensin-Converting Enzyme Inhibitory Assay. ACE
inhibitory activity was determined by a modified method of Cushman
3
4
51
4
8
2
6
41
Microginin 690 (2): [α] −21 (c 0.175, MeOH); UV (MeOH) λ
and Cheung. The reaction mixture (100 μL) contained potassium
D
max
1
13
(
log ε) 224 (4.19), 276 (3.39); H and C NMR, Table 2; positive
phosphate buffer (50 mM; pH 8.3), NaCl (300 mM), ZnCl (10 μM),
2
ESIMS m/z (relative intensity) 691 (100), 627 (7.6), 510 (54.1), 343
hippuryl-L-histidyl-L-leucine (5 mM) (Sigma), and ACE (5 mU)
(Sigma). Test compounds were dissolved in DMSO for testing,
keeping the concentration of the DMSO in the reaction mixture at 5%.
At this concentration DMSO did not significantly affect enzyme
activity. Test compounds were preincubated with the substrate for 30
min at 37 °C, after which the ACE was added and the tubes were again
incubated at 37 °C for 30 min. At the end of the incubation, tubes
were placed on ice, and the reaction was terminated by the addition of
1 N HCl (100 μL). The hippuric acid (HA) was then extracted with
EtOAc (1 mL). An aliquot of the EtOAc layer (200 μL) was
transferred to Target DP vials (National Scientific) and evaporated
using a SpeedVac (Thermo). The HA was then dissolved in MeOH (1
mL), and the solution was analyzed by LC/MS (Waters XTerra MS
+
(
49.3), 180 (29.3), 128 (8.0); HRMS m/z 691.3378 [M + H] (calcd
for C H N O S, 691.3377).
3
4
51
4
9
2
6
Microginin 704 (3): [α] −16 (c 0.40, MeOH); UV (MeOH) λ
log ε) 224 (4.19), 276 (3.39); H and C NMR, Table 2; positive
D
max
1
13
(
ESIMS m/z (relative intensity) 705 (100), 641 (6.2), 510 (68.3), 357
+
(
40.6), 180 (31.9), 128 (8.5); HRMS m/z 705.3530 [M + H] (calcd
for C H N O S, 705.3533).
3
5
53
4
9
2
6
Microginin 527 (4): [α] −38 (c 0.16, MeOH); UV (MeOH) λ
log ε) 224 (4.18), 276 (3.34); H and C NMR, Table 3; positive
D
max
1
13
(
ESIMS m/z (relative intensity) 528 (37.4), 464 (3.3), 180 (100), 128
+
(
29.6); HRMS m/z 528.2739 [M + H] (calcd for C H N O S,
25
42
3
7
5
28.2743).
2
6
Microginin 511 (5): [α] −20 (c 0.10, MeOH); UV (MeOH) λ
C , 2.1 × 30 mm; 10−100% MeCN + 0.1% acetic acid over 6.1 min;
D
max
18
1
13
(log ε) 224 (4.18), 276 (3.34); H and C NMR, Table 3; positive
flow rate 0.5 mL/min). The LC/MS method was validated using pure
HA (Sigma). A standard curve of HA (1−250 μM) was generated and
found to be linear with the HA produced due to the hydrolysis of the
substrate falling within this range. Typically, about 1% of the substrate
was hydrolyzed with 5 mU of ACE, corresponding to about 60 μM
HA. To further validate this method, the effect of captopril and EDTA,
known inhibitors of ACE activity, was determined. The IC50 values
ESIMS m/z (relative intensity) 512 (100), 164 (6.8), 128 (17.5);
+
HRMS m/z 512.2785 [M + H] (calcd for C H N O S, 512.2794).
25
42
3
6
Determination of the Configuration of Microginin 690 (2).
40
Following reported methods, to a solution of N-Boc-L-methionine
100 mg) in MeOH (2.5 mL) was added oxone (500 mg), and the
solution was stirred at room temperature for 2 h. The mixture was
(
F
DOI: 10.1021/acs.jnatprod.7b00829
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
were 25 nM and 14 μM for captopril and EDTA, respectively, and are
in agreement with published values.
(19) Schwarz, D.; Orf, I.; Kopka, J.; Hagemann, M. Metabolites 2013,
3, 72−100.
42
(
20) Baran, R.; Ivanova, N. N.; Jose, N.; Garcia-Pichel, F.; Kyrpides,
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jnat-
prod.7b00829.
N. C.; Gugger, M.; Northen, T. R. Mar. Drugs 2013, 11, 3617−3631.
■
(
21) Okino, T.; Matsuda, H.; Murakami, M.; Yamaguchi, K.
*
S
Tetrahedron Lett. 1993, 34, 501−504.
(22) Ploutno, A.; Carmeli, S. Tetrahedron 2002, 58, 9949−9957.
(23) Sano, T.; Kaya, K. Phytochemistry 1997, 44, 1503−1505.
(24) Strangman, W. K.; Wright, J. L. Tetrahedron Lett. 2016, 57,
1
(
801−1803.
Spectra of microginins 1−5, including NMR and MS
25) Sumner, L. W.; Amberg, A.; Barrett, D.; Beale, M. H.; Beger, R.;
(
PDF)
Daykin, C. A.; Fan, T. W.-M.; Fiehn, O.; Goodacre, R.; Griffin, J. L.
Metabolomics 2007, 3, 211−221.
AUTHOR INFORMATION
Corresponding Author
E-mail: wrightj@uncw.edu.
̌
■
*
(26) Welker, M.; Marsa
̌
lek, B.; Sejnohova,
́
L.; Von Doehren, H.
Peptides 2006, 27, 2090−2103.
27) Vaniya, A.; Fiehn, O. TrAC, Trends Anal. Chem. 2015, 69, 52−
1.
̈
̈
28) Duhrkop, K.; Shen, H.; Meusel, M.; Rousu, J.; Bocker, S. Proc.
(
6
(
ORCID
Jeffrey L. C. Wright: 0000-0001-9299-9744
Notes
The authors declare no competing financial interest.
Natl. Acad. Sci. U. S. A. 2015, 112, 12580−12585.
(29) Luesch, H.; Yoshida, W. Y.; Moore, R. E.; Paul, V. J.; Mooberry,
S. L.; Corbett, T. H. J. Nat. Prod. 2002, 65, 16−20.
(30) Ishida, K.; Matsuda, H.; Murakami, M. Tetrahedron 1998, 54,
1
(
3475−13484.
ACKNOWLEDGMENTS
■
31) Harrigan, G. G.; Luesch, H.; Yoshida, W. Y.; Moore, R. E.;
We thank E. Wright for the culture of M. aeruginosa (UTEX
LB2385) for this project. We acknowledge the late T. Schuster
for his assistance with the acquisition of preliminary
spectroscopic data. This work was funded by CDC-NCDHHS
Nagle, D. G.; Paul, V. J. J. Nat. Prod. 1999, 62, 655−658.
(32) Ogino, J.; Moore, R. E.; Patterson, G. M.; Smith, C. D. J. Nat.
Prod. 1996, 59, 581−586.
(33) Marfey, P. Carlsberg Res. Commun. 1984, 49, 591−596.
(
34) Ishida, K.; Matsuda, H.; Murakami, M.; Yamaguchi, K.
(
01505-07) as well as NIGMS (GM076300-01).
Tetrahedron 1997, 53, 10281−10288.
35) Ishida, K.; Kato, T.; Murakami, M.; Watanabe, M.; Watanabe,
M. F. Tetrahedron 2000, 56, 8643−8656.
36) Cushman, D.; Plusc c, J.; Williams, N.; Weaver, E.; Sabo, E.;
Kocy, O.; Cheung, H.; Ondetti, M. Experientia 1973, 29, 1032−1035.
37) Vermeirssen, V.; Van Camp, J.; Verstraete, W. Br. J. Nutr. 2004,
2, 357−366.
38) Wu, J.; Aluko, R. E.; Nakai, S. J. Agric. Food Chem. 2006, 54,
(
REFERENCES
■
(
(
1) Reynolds, C.; Walsby, A. Biol. Rev. 1975, 50, 437−481.
2) O’Neil, J.; Davis, T.; Burford, M.; Gobler, C. Harmful Algae 2012,
(
̌ ̌
e
1
(
4, 313−334.
3) Buratti, F. M.; Manganelli, M.; Vichi, S.; Stefanelli, M.; Scardala,
S.; Testai, E.; Funari, E. Arch. Toxicol. 2017, 91, 1−82.
4) Harke, M. J.; Steffen, M. M.; Gobler, C. J.; Otten, T. G.; Wilhelm,
S. W.; Wood, S. A.; Paerl, H. W. Harmful Algae 2016, 54, 4−20.
5) Carmichael, W. W. Freshwater Cyanobacteria (Blue-Green Algae)
Toxins; Pergamon Press: Oxford, 1989; Vol. 3.
6) Zurawell, R. W.; Chen, H.; Burke, J. M.; Prepas, E. E. J. Toxicol.
(
9
(
(
7
(
(
32−738.
39) Starr, R. C.; Zeikus, J. A. J. Phycol. 1993, 29, 1−106.
(
40) Imae, Y.; Takada, K.; Okada, S.; Ise, Y.; Yoshimura, H.; Morii,
Y.; Matsunaga, S. J. Nat. Prod. 2013, 76, 755−758.
41) Cushman, D.; Cheung, H. Biochem. Pharmacol. 1971, 20, 1637−
648.
(
(
1
Environ. Health, Part B 2005, 8, 1−37.
(
(
7) Namikoshi, M.; Rinehart, K. J. Ind. Microbiol. 1996, 17, 373−384.
(42) Hooper, N. M.; Turner, A. J. Biochem. J. 1987, 241, 625−633.
8) Van Wagoner, R. M.; Drummond, A. K.; Wright, J. L. Adv. Appl.
Microbiol. 2007, 61, 89−217.
(
9) Chlipala, G. E.; Mo, S.; Orjala, J. Curr. Drug Targets 2011, 12,
1
(
(
654−1673.
10) Agha, R.; Quesada, A. Toxins 2014, 6, 1929−1950.
11) Frangeul, L.; Quillardet, P.; Castets, A.-M.; Humbert, J.-F.;
Matthijs, H. C.; Cortez, D.; Tolonen, A.; Zhang, C.-C.; Gribaldo, S.;
Kehr, J.-C. BMC Genomics 2008, 9, 274.
(
12) Humbert, J.-F.; Barbe, V.; Latifi, A.; Gugger, M.; Calteau, A.;
Coursin, T.; Lajus, A.; Castelli, V.; Oztas, S.; Samson, G. PLoS One
013, 8, e70747.
13) Boudreau, P. D.; Monroe, E. A.; Mehrotra, S.; Desfor, S.;
2
(
Korobeynikov, A.; Sherman, D. H.; Murray, T. F.; Gerwick, L.;
Dorrestein, P. C.; Gerwick, W. H. PLoS One 2015, 10, e0133297.
(
14) Kleigrewe, K.; Almaliti, J.; Tian, I. Y.; Kinnel, R. B.;
Korobeynikov, A.; Monroe, E. A.; Duggan, B. M.; Di Marzo, V.;
Sherman, D. H.; Dorrestein, P. C. J. Nat. Prod. 2015, 78, 1671−1682.
(
15) Li, P.; Anandhi Senthilkumar, H.; Figueroa, M.; Wu, S.-B.; Fata,
J. E.; Kennelly, E. J.; Long, C. J. Nat. Prod. 2016, 79, 1619−1627.
16) Gill, K. A.; Berrue, F.; Arens, J. C.; Kerr, R. G. J. Nat. Prod. 2014,
7, 1372−1376.
17) Klitgaard, A.; Iversen, A.; Andersen, M. R.; Larsen, T. O.;
Frisvad, J. C.; Nielsen, K. F. Anal. Bioanal. Chem. 2014, 406, 1933−
943.
18) Briand, E.; Bormans, M.; Gugger, M.; Dorrestein, P. C.;
Gerwick, W. H. Environ. Microbiol. 2016, 18, 384−400.
(
́
7
(
1
(
G
DOI: 10.1021/acs.jnatprod.7b00829
J. Nat. Prod. XXXX, XXX, XXX−XXX