Biosynthetic products from a nearshore-derived gram-negative bacterium enable reassessment of the kailuin depsipeptides

Article abstract of DOI:10.1021/np500840n

Sampling of California nearshore sediments resulted in the isolation of a Gram-negative bacterium, Photobacterium halotolerans, capable of producing unusual biosynthetic products. Liquid culture in artificial seawater-based media provided cyclic depsipeptides including four known compounds, kailuins B-E (2-5), and two new analogues, kailuins G and H (7 and 8). The structures of the new and known compounds were confirmed through extensive spectroscopic and Marfeys analyses. During the course of these studies, a correction was made to the previously reported double-bond geometry of kailuin D (4). Additionally, through the application of a combination of derivatization with Moshers reagent and extensive ¹³C NMR shift analysis, the previously unassigned chiral center at position C-3 of the β-acyloxy group of all compounds was determined. To evaluate bioactivity and structure-activity relationships, the kailuin core (13) and kailuin lactam (14) were prepared by chiral synthesis using an Fmoc solid-phase peptide strategy followed by solution-phase cyclization. All isolated compounds and synthetic cores were assayed for solid tumor cell cytotoxicity and showed only minimal activity, contrary to other published reports. Additional phenotypic screenings were done on 4 and 5, with little evidence of activity.

Full text of DOI:10.1021/np500840n

Article
pubs.acs.org/jnp
Biosynthetic Products from a Nearshore-Derived Gram-Negative
Bacterium Enable Reassessment of the Kailuin Depsipeptides
Christine M. Theodore,† Nicholas Lorig-Roach,† Patrick C. Still,† Tyler A. Johnson,† Marija Draskovic,†
̌
́
Joshua A. Schwochert,† Cassandra N. Naphen,† Mitchell S. Crews,† Simone A. Barker,†
Frederick A. Valeriote,‡ R. Scott Lokey,† and Phillip Crews*,†
†Department of Chemistry and Biochemistry, University of California, Santa Cruz, California 95064, United States
‡Josephine Ford Cancer Center, Henry Ford Health System, Detroit, Michigan 48202, United States
S
* Supporting Information
ABSTRACT: Sampling of California nearshore sediments
resulted in the isolation of a Gram-negative bacterium,
Photobacterium halotolerans, capable of producing unusual
biosynthetic products. Liquid culture in artificial seawater-
based media provided cyclic depsipeptides including four known
compounds, kailuins B−E (2−5), and two new analogues,
kailuins G and H (7 and 8). The structures of the new and
known compounds were confirmed through extensive spectro-
scopic and Marfey’s analyses. During the course of these studies,
a correction was made to the previously reported double-bond
geometry of kailuin D (4). Additionally, through the application
of a combination of derivatization with Mosher’s reagent and
extensive 13C NMR shift analysis, the previously unassigned
chiral center at position C-3 of the β-acyloxy group of all compounds was determined. To evaluate bioactivity and structure−
activity relationships, the kailuin core (13) and kailuin lactam (14) were prepared by chiral synthesis using an Fmoc solid-phase
peptide strategy followed by solution-phase cyclization. All isolated compounds and synthetic cores were assayed for solid tumor
cell cytotoxicity and showed only minimal activity, contrary to other published reports. Additional phenotypic screenings were
done on 4 and 5, with little evidence of activity.
shift in the focus for the study of marine biosynthetic
products appears to be under way. This change involves
Vibrio, Pseudoalteromonas, and Pseudomonas. Yet, the tally of
natural products from this assemblage is relatively small.
Additionally, while members of the α-Proteobacteria seem to
be the most abundant, it is the γ-Proteobacteria that appear to
be most successfully cultured.7 Gaining command of the
literature linking biosynthetic products to their various Gram-
negative sources was accomplished at the outset of our project.
This involved building a small database of 63 lead molecular
structures (shown in Figure S1, Supporting Information) using
results from four of the six major classes of the Proteobacteria
(the other two remain unstudied). Also evident to others and
ourselves is that many sponge metabolite scaffolds have parallel
or identical features to those isolated from terrestrial-derived
Gram-negative bacteria, especially myxobacteria (δ-Proteobac-
teria).8−10
A
an emphasis on compounds isolated from culturable hetero-
trophic microorganisms rather than from marine invertebrates
(sponges, cnidarians, and tunicates). It is clear that the ocean
environment represents a vast resource for obtaining new
chemical entities from bacteria and fungi. In fact, recent
estimates indicate more than 105 microorganisms could exist.1
Foreshadowing this swing to marine microorganisms was a
1993 review published by Fenical2 entitled “Chemical Studies
of Marine Bacteria: Developing a New Resource”. This work
summarized some 50 biosynthetic products derived from
oceanic bacteria and has been highly cited by others [ca. 270
times, Web of Science]. As a measure of recent progress, 119
novel natural products from the culturing of marine bacteria
were reported during the 2012 calendar year.3−5 A majority of
the microorganisms were Gram-positive species.
One motivation for the project described herein was our
conclusion that the study of products from marine-derived
Gram-negative bacteria remains underexplored. The diversity of
culturable heterotrophic bacteria in seawater, while incom-
pletely understood,6 is thought to be dominated by Gram-
negative rods, mainly of the phylum Proteobacteria. Some
examples of well-known genera within this group include
The initial goal of this project was to develop a library of
California coastal zone-derived Gram-negative strains to seed
subsequent metabolite studies. Our hypothesis was that
processing fresh environmental samples using baiting techni-
Special Issue: Special Issue in Honor of William Fenical
Received: October 24, 2014
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 1. A short list of natural products isolated via culturing of marine Gram-negative Photobacterium (family Vibronaceae) strains. Only lead
structures are shown, and each is categorized by biosynthetic type as a(n) alkaloid (ALK), polyketide (PK), nonribosomal peptide (NRP), or
polyketide−nonribosomal peptide (PK−NRP). Further remarks include additional marine bacteria (coded with N for Gram-negative or P for Gram-
positive) that also produce these exact metabolites in culture.
ques designed to maximize isolation of δ-Proteobacteria would
yield enhanced strain diversity. However, large numbers of
rapidly growing Gram-negative cultures were observed that
were γ-Proteobacteria, and only a few had the characteristics
expected from the δ-class. Specific details of these biological
outcomes, including an overview of the strains and their
diversity, will be published elsewhere. This account reports on
an initial success in the saltwater culturing of a nearshore γ-
Proteobacteria species characterized as Photobacterium haloto-
lerans (M128SB283Ax), which produces six cyclic depsipep-
tides of the kailuin family (2−5, 7, and 8) including two new
analogues (7 and 8).
L) to metabolite profiling via LC/MS-ELSD. Since this
operation was done prior to the completion of molecular
taxonomic analysis, the initial prioritizations for advanced
chemical study relied on mass spectrometric results. Thus,
attention was drawn to one sample displaying intense m/z ions
(726, 740, 752, 754 [M + H]+) for several chromatographic
peaks that were not seen in other cultures. This strain, coded
M128SB283Ax, was derived from sediment obtained at
Manresa State Beach, CA, and subsequently identified as
Photobacterium halotolerans.
The next step involved cross referencing the initial mass
spectrometric data described above to the formulas and
structures known from prior study of the Photobacterium
cultures. Even though Photobacterium is often associated with
disease states in a number of marine organisms,13 the literature
on its biosynthetic products is relatively scant. Representative
examples are shown in Figure 1. Some highlights from this
compilation are (a) only four defined species of the genera have
been explored chemically including P. angustum, P. damselae, P.
halotolerans, and P. phosphoreum; (b) no compound isolated
from Photobacterium cultures matched the MS m/z data
observed; and (c) cyclic depsipeptides dominate the list,
indicating dereplication efforts should emphasize this class.
As the project progressed, strain M128SB283Ax was firmly
characterized through repeated 16S rRNA analysis. It showed
98% likeness to the P. halotolerans type specimen (MACL01),
which is above the cutoff value for bacterial species similarity
identification.23 This identification was further validated by
follow-up examination using the recA and rpoA genes24 (P.
Dunlap, unpublished results using our sample). As an
interesting biological circumstance, P. halotolerans was first
reported in 2006 from water samples gathered in Lake Martel, a
RESULTS AND DISCUSSION
■
The early reports of metabolites from Gram-negative marine
bacterial strains involved the pursuit of samples obtained from
seawater.2 This strategy continues to be successful. For
example, the antitumor cyclic depsipeptide didemnin B
(originally from Caribbean tunicates) was recently reisolated
through the culturing of Tistrella sp. (α-Proteobacteria, Figure
S1, Supporting Information), obtained from an Indo-Pacific
water column sample.5,11 The current work began by collecting
sediments from 41 beaches in the Monterey Bay, CA, region. A
total of 820 distinct samples yielded 54 cultivable strains thus
far. The strain isolation step involved baiting with Escherichia
coli streaks on nutrient-deficient agar.12 Each strain was later
examined using 16S rRNA sequences, and the majority (59%)
could be classified as members of the phylum Proteobacteria;
most proved to be γ-Proteobacteria.
The campaign to identify the top chemically prolific strains
from the isolates described above involved subjecting dichloro-
methane (CH2Cl2) extracts of selected small-scale cultures (5
B
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
subterranean saline environment in Mallorca, Spain.24 At the
time of the 2006 report, there were 12 defined Photobacterium
species. Subsequently, this group has now grown to include 29
members, including a few previously assigned to the genus
Vibrio. Figure 2 depicts strain M128SB283Ax phylogeny in a
molecular formula yielded only one natural product hit,
(+)-kailuin D (4),26 shown in Figure 3. The H and 13C
1
NMR data (shown in Table 1; see also Figures S2−S8,
Supporting Information) confirmed the match. A 1997 report
on the structure of kailuin D employed liquid culture of a
bacterium identified provisionally as in the genus Vibrio and
obtained from driftwood in Hawaii. Also, three other analogues,
(+)-kailuins A−C (1−3), were reported from this culture
broth. Added to this list, shown in Figure 3, are two other
analogues as follows. Kailuin E (5) was isolated from a culture
sample identified as a Vibrio sp. (marine environmental source
not stated)27 and kailuin F (6), obtained from a Vibrio sp.
cultured from the surface of an artificial cellulose sponge
deployed for 1 h at −10 m in Okinawa.28
On close inspection, uncertainties arose pertaining to the
properties of the compound from fraction H4 compared to
those published for kailuin D (4). First, the NMR data shown
here were consistent with the presence and sequence of the five
amino acids and acyloxy substructures in 4, but the optical
rotation measured ([α]D −15.0, c 0.8, MeOH) was not in
agreement with the published value ([α]D +9.5, c 1.0
MeOH).26 Second, the Marfey’s analysis published on 4 did
not distinguish between the possibilities of (2S,3R)-Thr, vs
(2S,3S)-Thr.29 Third, the 13C NMR data listed for the allylic
carbons at C-6 (δC 26.6) and C-9 (δC 27.6) were consistent
with a Z rather than the E geometry reported.30 Finally, the
absolute configuration at the β-acyloxy position was not
defined, but R configurations had been reported for this same
residue in related compounds, solonamide A and unnarmicin A
(Figure 1). Despite lacking original samples from previously
published reports, the immediate next steps were to resolve all
of these issues.
A multistep process was used to reinvestigate the total
structure of kailuin D (4), and this also facilitated work on the
five additional compounds. The optical properties of 4 were
reassessed, and the negative rotation data obtained were
invariant. The published configurations of each of the four
amino acid residues were verified as shown in Figure 3. This
involved acid-mediated hydrolysis of 4 (0.5 mg in 6 M HCL at
120 °C) followed by derivatization with Marfey’s reagent (1-
fluoro-2,4-dinitrophenyl-5-L-alanine amide, FDAA).31 The
initial separation scheme, using a standard gradient of
acetonitrile (CH3CN) and water with 0.1% formic acid, did
not provide sufficient separation between (2S,3R)-Thr and
(2S,3S)-Thr or between (S)-Ser and (R)-Ser. These observa-
tions mirror those difficulties seen by others29,32 (Figures S25−
S28, Supporting Information). Thus, gradient optimization
steps were developed to achieve baseline resolution between
these amino acid-FDAA diastereomers (Figures S25−S28,
Supporting Information). A solvent system of 90:10
CH3CN−MeOH as the organic solvent and 5% acetic acid in
water as the aqueous solvent was found to provide accurate
results.32 The Marfey’s analysis confirmed the amino acid
configurations previously published for kailuin D (4) as
(2S,3S)-Ile, (R)-Leu (2), (2S,3R)-Thr, (S)-Ser, and (R)-Leu
(1). Marfey’s analysis was also completed on kailuin E (5), with
the same absolute configuration deduced for 4 (Figures S25−
28, Supporting Information).
Figure 2. Phylogeny depicted by a maximum likelihood tree
constructed in MEGA6 (Molecular Evolutionary Genetics Analysis)
from the alignments of 16S rRNA gene sequences. Compared here are
(a) UCSC strain M128SB283Ax, which produced compounds 2−5, 7,
and 8, (b) 29 representative and/or type strains currently deposited in
the NCBI database (accession numbers shown in parentheses), and
(c) Vibrio cholera used as an outgroup. Bootstrap values >50% are
listed for nodes with the scale bar representing 0.005 substitution per
position.
maximum likelihood tree, with P. halotolerans forming a branch
alongside P. ganghwense, which has not yet been a subject of
natural products study (Figure 1). Many Photobacterium species
live in association through mutual and pathogenic relationships
with marine organisms. Consequently, they can be difficult to
isolate. The present finding represents the first example of the
isolation and chemical study of any Photobacterium species from
California coastal zone sediment and suggests that its members
could be more widespread than previously believed.
The gateway to the isolation and analysis of the secondary
metabolites from strain M128SB283Ax involved the scale-up
culture and purification outlined in Scheme 1. Exhaustive
extraction of the culture broths (16 L) provided an oil, which
was then subjected to solvent partitioning. This afforded 0.45 g
of a CH2Cl2-soluble fraction and 1.2 g of water-soluble material.
The CH2Cl2 material was subjected to a preparative HPLC step
gradient25 to provide six prefractions (F0 was discarded). The
fractions F3−F5 contained several compounds with mass
spectrometric [M + H]+ m/z values ranging from 726 to 754.
These were pooled, and their constituents were further
separated by HPLC. Six pure compounds were isolated from
the 12 fractions collected, and six were pursued further
including H2 (0.8 mg), H3 (2.0 mg), H4 (3.2 mg), H6 (1.7
mg), H9 (1.7 mg), and H10 (4.0 mg).
Additional scrutiny of the NMR data previously reported for
kailuin D (4), as noted above, revealed discrepancies in the
lipid portion of the structure. This led to the conclusion that
both the geometry of the double bond and the assignment of
the allylic resonances were incorrect. Davidson et al.26 reported
A metabolite in fraction H4 served as the lead for
dereplication. The molecular formula was established from
high-accuracy mass spectrometry (HAMS) as C39H69N5O9,
based on m/z [M + K]+ 790.4739. A literature search on this
C
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Scheme 1. Isolation Scheme of the Six Kailuins (2−5, 7, and 8) from Photobacterium halotolerans (M128SB283Ax)
an E geometry using a coupling-based argument and assigned
the vinyl signals at δH 5.36 (δC 130.9) and δH 5.26 (δC 128.5) at
positions 7 and 8 of the acyl chain, respectively. Closer
examination revealed Z geometry by 13C NMR data and J-
coupling values and also indicated a reversal of the assignments
of 7 and 8 via 2D NMR experiments (Figures S2 and S3,
Supporting Information). The acyl chain spin system and the
position of the double bond were verified by HMBC and
COSY experiments. These results confirmed that the
diastereotopic proton resonances of position 2 (δ 2.47 and
2.54) correlated to a carbonyl resonance at δ 173.1, to which
the amide of Leu-1 (δ 6.36) also correlated, thereby
establishing the position of C-2 as adjacent to the N-terminal
amide of Leu-1. The COSY correlations from position C-2 to
C-9 showed C-6 (δ 2.02) is allylic to C-7 (δ 5.26), while C-9 (δ
1.97) is allylic to C-8 (δ 5.36). The 13C NMR resonances of
these allylic positions (δ 26.6 and 27.2) are in agreement with
13C NMR shift analysis for the Z configuration, using
benchmark data of δC 29 for Z and δC 35 for E.30 Additionally,
the resolution of the vinyl multiplets at 600 MHz in this work
was a significant improvement over that reported at 500 MHz
(Figure S2, Supporting Information), enabling the correct
coupling constants to be extracted. Poor resolution in the vinyl
multiplet region appears to have misled previous elucidation
efforts. The vicinal coupling between C-7 and C-8 (J = 10.8
Hz) also supported the Z double-bond configuration, where J =
11 Hz is typical of Z, and J = 17 Hz is indicative of an E
configuration.30
To date, no attempts have been made to deduce the absolute
configuration at the β-acyloxy position of the known kailuin
analogues A−F (1−6). Also relevant is that more than 12
microbial-derived cyclic depsipeptides have been described
possessing a β-acyloxy substructure (see Figure 1 and Figure
S33, Supporting Information). An R configuration at the β-
acyloxy carbon has been deduced previously for six members of
this collection including (a) solonamide A,19 unnarmicin A,21
and kahalide V33 by Mosher’s method; (b) tripropeptin C by
chiroptical data of a β-hydroxy acid product;34 and (c)
peptidolipin B34 and serratiomycin36 by NMR NOE measure-
3
ments, JH−H values, and 13C NMR calculated δ’s (Figure 4).
However, similar to the situation with the kailuins, no
assignments have been made for other members of this list
including ngercheumicins D and E,20 halobacillin,37 and
kahalalides Y and W (W was assigned only by biosynthetic
analogy).33 To date, no natural products of this type have been
reported with an S configuration at the β-acyloxy carbon,
although NMR properties for the peptidolipin B have been
predicted via DFT calculations (see Figure 4).35
Deducing the assignment of the β-acyloxy carbinol carbon
for kailuins A−E (1−5) and new compounds kailuins G (7)
and H (8), whose structures are all discussed below, was
eventually accomplished, although it proved to be more
challenging than initially expected. An obvious first step
D
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 3. Overview of the kailuin family: known analogues A−F (1−6) [with D (4) revised] and new members G and H (7 and 8). Compounds 2−
5, 7, and 8 were isolated from Photobacterium halotolerans (M128SB283Ax). The coupling constant between protons 7 and 8 is shown to justify
double-bond revision for 4. The R/S configurations for the amino acids were determined through Marfey’s analysis starting with compounds 4 and 5
(see Figures S25−28, Supporting Information). The configuration at position 3 was determined using two methods: (a) compound 2 as the starting
material to prepare compounds 12a/12b (see Scheme 2), which were evaluated by the Mosher’s method 1H NMR shift analysis (see Figure 4, panel
A); (b) analysis of 13C NMR chemical shifts (shown in Figure 4, panel B).
b
1
Table 1. Partial 13C and H NMR Data (125/600 MHz, CDCl3, 25 °C) for the 3-Hydroxy Acyl Chain for Kailuin D (4) and
b
b
Kailuins G (7) and H (8)
4
7
8
position
δC
type
δH, mult (J in Hz)
δC
type
δH, mult (J in Hz)
δC
type
δH, mult (J in Hz)
1
171.5
43.0
C
171.7
43.0
C
171.7
43.1
C
2a
2b
3
CH2
2.47, dd (2.3, 16.4)
2.54, dd (5.1, 16.4)
CH2
2.47, dd (2.1, 16.4)
2.54, dd (5.1, 16.4)
CH2
2.47, dd (2.2, 16.4)
2.54, dd (5.1, 16.4)
c
c
c
73.1
35.1
CH
5.06, m
73.2
35.6
CH
5.05, m
73.2
35.6
CH
5.05, m
c
c
c
4a
4b
5
CH2
1.58, m
CH2
1.56, m
CH2
1.57, m
c
c
c
1.69, m
1.66, m
1.68, m
c
c
c
25.7
26.9
128.6
131.1
27.5
29.9
29.2
32.0
23.0
14.3
CH2
CH2
1.32−1.38, m
25.8
29.8
29.7
29.7
29.5
29.5
32.1
23.0
14.3
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH3
1.66, m
25.8
30.1
29.8
29.7
29.4
27.5
39.2
28.2
22.9
22.9
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH
1.68, m
c
c
6
2.02, q (7.2)
1.45, m
1.46, m
a
c
c
7
CH
CH
5.26, dt (10.8, 7.2)
5.36, dt (10.8, 7.2)
1.97, q (7.2)
1.19−1.26, m
1.22−1.26, m
1.22−1.26, m
1.22−1.26, m
a
c
c
8
1.19−1.26, m
c
c
9
CH2
CH2
CH2
CH2
CH2
CH3
1.19−1.26, m
c
c
c
10
11
12
13
14
1.21−1.31, m
1.19−1.26, m
1.22−1.26, m
c
c
c
1.24−1.28, m
1.19−1.26, m
1.11, m
c
c
c
1.24, m
1.19−1.26, m
1.51, m
c
1.27, m
0.85, t (7.2)
CH3
CH3
0.84, d (6.5)
0.84, d (6.5)
0.86, t (6.9)
a
b
Data reported from Harrington et al.26 incorrectly transposed these two assignments. Measured at 125 MHz in CDCl3. For complete NMR data
c
for kailuin D, G, and H (4, 7, 8) see Tables S1 and S2, Supporting Information. Overlapping signals.
involved Mosher’s analysis by preparing and examining R/S-
MTPA (α-methoxy-α-trifluoromethylphenylacetic acid) esters
of the β-hydroxyl group, and kailuin B (2) was chosen for this
process. As noted above, this strategy has been successfully
used by others, although such an analysis failed during the
evaluation of the peptidolipins.35 As described below, Mosher’s
method was tenuous for the kailuins and motivated the quest
for back-up analysis. This was achieved by assembling a mini-
database of 13C NMR chemical shift values. Shown below are
that these 13C δ data can provide a fast, general approach to
assessing R/S configurations for almost any metabolite having a
β-acyloxy moiety.
The Mosher’s tactic began by ester cleavage on kailuin B (2)
using lithium hydroxide. Two products, outlined in Scheme 2,
were obtained and consisted of depsipeptide alcohol I (m/z
748.4856 [M + Na]+) (9) and a linear alcohol (m/z 758.8 [M +
Na]+) (10). The major product 9 arose from intramolecular
attack by the (S)-serine hydroxy group at the ester carbon. The
hypothetical product 11, potentially formed from attack by the
(2S,3R)-Thr hydroxy group, was not observed. Alcohol I (9)
E
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 4. Strategies to assign the β-acyloxy configuration. Panel A shows the ΔδS−R chemical shift differences measured for (R)/(S)-MTPA esters
(12a and 12b) obtained via the transformations shown in Scheme 2, which employed kailuin B (2) as the starting material. The subsequent ΔδS−R
data shown here were used to assign the 3R-side chain and the 3R-Thr absolute configurations, respectively. Panel B provides, in the upper portion of
the table, a database of 13C NMR shifts at three sites proximal to the oxygen center in compounds whose β-acyloxy configuration was assigned in
previous studies. The trends in the δC at the β and γ are diagnostic of a variation between R and S and can be used to assign a final configuration at
the β-acyloxy position. The lower portion of the table includes compounds whose configurations at the β-acyloxy position were assigned in this study
by comparison of the 13C NMR shifts at the β and γ positions to those collected in the database.
was converted in an NMR tube to the bis esters R-MTPA (12a)
and S-MTPA (12b) (Scheme 2)38,39 (both exhibited nominal
m/z = 599 [M + H + K]2+ for molecular formula
C57H81F6N5O13). However, there were complications in
measuring and interpreting the ΔδS−R patterns. Anomalous
1H δ values can arise for protons in proximity to multiple
MTPA residues,40 and this is apparent in the strange data
in the literature for carbons associated with a β-acyloxy
functionality, but these data have not yet been compiled or
analyzed. The collection in Figure 4 panel B provides values for
β-acyloxy 13C NMR chemical shifts of 14 compounds at the key
α, β, and γ positions. Unfortunately, the chemical shift trends
for the α position are not useful, whereas the data from the
other two positions provide a pattern for assigning the
configuration at Cβ as R or S. Important benchmark
information is contained in the first three entries, which
present experimental R-peptiolipin B 13C shifts (δCβ 72.6, δCγ
32.5)35 alongside those from DFT calculations for R-peptiolipin
B (δCβ 74.1, δCγ 35.7) and S-peptiolipin B (δCβ 77.5, δCγ
40.6).35 Other diagnostic insights can be derived from the data
available for an additional five compounds definitively assigned
with an R-β-acyloxy configuration (solonamide A,19 unnarmicin
A,21 serratiomycin,36 triopeptin,34 and kahalalide V33), whose
shift values are δCβ 71−73 and δCγ 34−36. Thus, our proposed
analysis framework consists of (a) assigning an R configuration
when δCβ ≤ 74/δCγ ≤ 36 and (b) assigning an S configuration
when δCβ ≥77/δCγ ≥ 40.
The patterns shown in Figure 4 panel B, interpreted using
the rules for R/S configuration assignment at the β-acyloxy
summarized above, were applied to reexamine the kailuins.
Using the experimental data collected (Tables S2 and S4,
Supporting Information) allows an R configuration to be
assigned for kailuin B (2) (δCβ 73.2 δCγ 35.5), kailuin C (3)
(δCβ 73.2 δCγ 35/6), kailuin D (4) (δCβ 73.1, δCγ 35.1), and
kailuin E (5) (δCβ 73.0 δCγ 35.4), as well as for the new
compounds described below: kailuin G (7) (δCβ 73.2, δCγ 35.6)
and kailuin H (8) (δCβ 73.2, δCγ 35.6). Extending this analysis
pattern shown in Figure 4 (panel A) at Ser: NH ΔδS−R
=
+0.076, Hα ΔδS−R = −0.036, Hβ ΔδS−R = +0.018, +0.111).
Unfortunately, no useful Δδ data were observed for the nine-
carbon side chain flanking the CO-MPTA group because the
resonances of 12a/12b appeared as unresolved multiplets. This
same situation also confounded the Mosher approach to the
peptidolipins, possessing a 23-carbon side chain,35 but it was
not a problem for the three five-carbon side chain containing
compounds19,21,33 (Figure S32, Supporting Information). The
only remaining useful data were minimal and consisted of
ΔδS−R = −0.059 at C-2, suggesting an R β-acyloxy
configuration. Correspondingly, this was identical in sign and
magnitude to data reported at this same position (see Figure
S32, Supporting Information) for the R β-acyloxy of solo-
namide A (ΔδS−R = −0.04),19 solonamide B (ΔδS−R
=
−0.04),19 unnarmicin A (ΔδS−R = −0.02),21 and kahalide V
(ΔδS−R = −0.16/−0.18).33 Finally, the ΔδS−R = +0.030 at Thr
supported the 3R configuration at this amino acid group, in
agreement with the assignment obtained (discussed above)
from Marfey’s analysis.
The power and speed in using 13C NMR δ trends for
configuration analysis is well accepted; however to do so
requires a set of diagnostic data.30 There is relevant information
F
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Scheme 2. Summary of the Reaction Products of Kailuin B (2) Used in the Determination of the Absolute Configurations
Justified by the Data Shown in Figure 4, Panel A
allows an R configuration to also be assigned, based on
literature data (δCβ 73.0, δCγ 35.3),24 for kailuin A (1).
Employing this same logic allows four other compounds to now
be assigned as R including kahalalides W and Y,33 halobacillin,37
and ngercheumicin D20 (Figure 4).
only difference in this compound was the length of the alkyl
chain. Taken together, these data represent a new member of
the kailuin class of compounds, and it was given the name
kailuin G (7). Fraction H9 contained a compound with an m/z
of 754 [M + H]+. HAMS confirmed a molecular formula of
C39H69N5O9, which was the same as kailuin E (5). The 1H and
13C NMR data again confirmed that its tetrapeptide core
remained unchanged. The H−13C HMBC and H−1H COSY
analysis confirmed the presence of a terminal isopropyl group
on the alkyl chain. This suggested the occurrence of a new
kailuin analogue, which has been named kailuin H (8).
Biosynthetically, the same gene cluster likely produces each
of the kailuin compounds isolated, so it is reasonable to assume
that all of the isolated kailuins have the same configuration at
each of the amino acid subunits as well as at the C-3 position.
Therefore, these compounds were assigned the same absolute
configuration as the previously isolated kailuins. Finally, all of
the kailuins isolated here exhibited negative [α]D values.
In addition to kailuin D (4), the other five pure compounds
(2, 3, 5, 7, and 8) obtained from the HPLC fractions shown in
Scheme 1 all had negative optical rotations. They could be best
tracked using the ESIMS m/z [M + H]+ values consisting of
726, 740, and 754. Structures for kailuins B, C, and E (2, 3, and
5) were dereplicated by comparing HAMS and NMR data to
literature values.26−28 Fraction H6 contained a compound with
an m/z of 740 [M + H]+ and a molecular formula calculated
from HAMS data as C38H69N5O9. This molecular formula was
unique compared to previously isolated kailuins and
represented the loss of a methyl group from kailuin E (5).
1
1
1
The H and 13C NMR data (shown in Table 1) indicated that
its tetrapetide core remained unchanged compared to the other
known kailuins. When combined with the 2D NMR data, the
G
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
a
Scheme 3. Concise Total Chiral Synthesis of Two Macrolide Analogues: the Kailuin Core (13) and Kailuin Lactam (14)
a
Details on the total synthesis are described in Schemes S1 and S2, Supporting Information.
Although the kailuins have shown antimicroalgal activity27,28
and low micromolar cytotoxicity against several tumor cell
lines,26 no mechanism of action has been reported for these
activities. To assess the biological activity of the kailuins, the
small library of compounds on hand was screened against the
tumor cell lines described in previous publications to provide
insights into their structure−activity relationships and further
explore their mechanistic target(s).26 The role of the
depsipeptide core contribution to the cytotoxicity represented
an additional, important issue, because this could lead to next-
generation simplified bioactive analogues. To explore these
ideas, we employed a standard synthetic methodology to
prepare two reduced complexity compounds for screening. The
strategies outlined in Scheme 3 utilized fluoroenylmethyloxy-
carbonyl (Fmoc) solid-phase chiral peptide synthesis followed
by solution-phase cyclization and provided two compounds,
which both showed negative optical rotation, including the
kailuin core (13) ([α]24D = −45.8) and the kailuin lactam (14)
high dilution (1 mg crude peptide/mL solvent) provided a
cyclization solvent system that was easily removed via rotary
evaporation. The coupling reagent COMU ((1-cyano-2-ethoxy-
2-oxoethylidenaminooxy)dimethylamino-morpholino-carbe-
nium hexafluorophosphate) was used as the cyclization reagent,
which afforded complete cyclization of the peptide within 20
min. This represented a tremendous improvement over the
commonly reported reaction times of 16 h.45 It has been
reported previously that COMU may not provide an efficient
cyclization reagent,46 but in this system, however, it enabled
expedient macrolactamization. Moreover, global side chain
deprotection using 50% TFA in CH2Cl2 in the presence of
COMU leads to many undesired side reactions (unpublished
data). This problem was avoided by removing the COMU
byproducts via flash chromatography with high recovery prior
to deprotection. Final reversed-phase HPLC purification
yielded the kailuin core (13) in 40% yield, based on resin
loading, over 14 steps. An overview of this sequence is shown in
Scheme S1, Supporting Information.
([α]24 = −1.6).
D
The detailed synthetic strategy employed in the synthesis of
the kailuin core (13) is shown in Scheme 3. It was centered on
microwave-assisted couplings and deprotections,41 along with
an expedient solution-phase cyclization protocol. The linear
peptide was synthesized on 2-chlorotrityl polystyrene resin
using the Fmoc protection group tactic. Microwave couplings
were carried out using HBTU (N,N,N′,N′-tetramethyl-O-(1H-
benzotriazol-1-yl)uronium hexafluorophosphate) and 1-hy-
droxy-7-azabenzotriazole (HOAt) (10 min, 50 °C, 12 W),
while deprotections were carried out with 2% DBU (1,8-
diazabicyclo[5.4.0]undec-7-ene) and 2% piperidine in DMF (5
min, 50 °C). Synthesis of the protected 3-hydroxypropionic
acid monomer was accomplished via monoprotection of 1,3-
propanediol42 followed by oxidation of the alcohol directly to
the carboxylic acid with Confer’s reagent that was generated in
situ.43 Coupling of this monomer was achieved using N,N′-
diisopropylcarbodiimide (DIC) (DMF, 1 h, twice). Removal of
the TBDPS (tert-butyldiphenylsilyl) group was accomplished
using TBAF (tetra-n-butylammonium fluoride) in THF over-
night. The last amino acid, Fmoc-(2S,3S)-Ile, was coupled onto
the unprotected alcohol using 5 equiv of Fmoc-(2S,3S)-Ile, 5
equiv of DIC, 0.25 equiv of DMAP (4-dimethylaminopyridine),
and 1 equiv of DIPEA (N,N-diisopropylethylamine) in DMF
for 45 min twice.44 In order to prevent nucleophilic
displacement of the ester bond by piperidine, the Fmoc
group of this amino acid was instead removed using the non-
nucleophilic base DBU (5% v/v in DMF, RT, 5 min).
The kailuin lactam (14), devoid of the lipid chain, was
envisioned as a key compound to test side-by-side with the
lactone for their potential to modulate the bioactivity. The
lactam was concisely prepared by a multistep chiral synthesis
sequence (Scheme S2, Supporting Information), using a
peptide synthesizer with room-temperature couplings without
HOAt. The same 20 min CH3CN/THF COMU cyclization
conditions were utilized, and this peptide was obtained in
26.4% overall yield.
The mini-library consisting of compounds 2−5, 7, 8, 13, and
14 was screened against a human colon tumor (HCT-116) cell
line to investigate their cytotoxicity and preliminary structure−
activity relationships.47 The kailuins (2−5, 7, 8) displayed weak
activity in the assay used against HCT-116, a human colon
cancer cell line, with IC50 values of 17−32 μM, as shown in
Table 2. Previous publications have reported activity against
human lung (A-549), breast (MCF-7), and colon (HT-29)
tumor cell lines in the 3−6 μM range,26 representing an order
of magnitude difference compared to results found in this work.
To determine if the observed differences were due to selectivity
against certain cell lines, kailuin D (4) was run against the
MCF-7 breast cancer cell line reportedly used in previous
studies. The IC50 in our assay was 39 μM and did not compare
to the 4 μM IC50 in the previous report.26 The synthetic kailuin
depsipeptide core (13) and lactam (14) were inactive against
HCT-116 cells at >100 μM IC50 values, and Table 2
summarizes these overall differences in assay results.
The deprotected linear peptide was cleaved off resin, and
residual TFA was removed in vacuo overnight to prevent
activation of the acid and subsequent N-terminal capping. It
was found that a mixture of THF and CH3CN (1:1 v/v) under
Due to the disparities between the current and published
IC50 values described above for kailuins D and E (4, 5), they
were submitted for additional screening. First, the compounds
were examined by cytological profiling (CP).48 This screening
H
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
publication date of initial work on the kailuins (in 1997 and
2000), which preceded the description (2006) of the type
strain, P. halotolerans sp. nov.24 There are some important next
steps to be suggested. Further study and bioactivity assessment
are warranted for metabolites from the 25 unstudied Photo-
bacterium species. Strains of note include the fully sequenced P.
leiognathi subsp. mandapamensis51 and P. gangwense, which was
originally isolated from seawater.52 Little is known about the
secondary metabolism of these organisms, and minimal work
has been done to discover quorum-sensing compounds
potentially produced by these organisms. For example,
preliminary secondary metabolite gene cluster analysis of P.
leiognathi subsp. mandapamensis by antiSMASH53 did not
reveal any clusters that appear to produce typical quorum-
sensing compounds (unpublished data). Currently, our culture
collection contains 10 different Photobacterium strains, and
studies on each are currently under way.
Table 2. Summary of Cytotoxicity of the Kailuins against
Tumor Cell Lines
a
MCF-7
b
b
c
compound
HCT-116
this study
literature
kailuin A
NT
22
NT
NT
NT
39
4
3
6
4
kailuin B
kailuin C
kailuin D
kailuin E
50
28
18
NT
NT
NT
NT
NT
kailuin G
kailuin H
kailuin core
kailuin lactam
32
17
>1000
>1000
a
b
Human cancer cell lines: HCT-116 (colon); MCF-7 (breast). IC50
values reported in μM, calculated based on Valeriote et al.47 IC50
values reported in μM, from Harrington et al.26
c
EXPERIMENTAL SECTION
General Experimental Procedures. Optical rotations were
performed on a JASCO P-2000 polarimeter. All NMR experiments
■
tool uses automated microscopy and image analysis to gain
insights into phenotypic features. The CP assay develops
“fingerprints” obtained by comparing the cytotoxic phenotypic
effects on HeLa cells to a library of 500 reference compounds
with known biological mechanisms of action. Similar to the
initial screens against solid tumor cell lines, 4 and 5 showed
only weak cytotoxicity in the CP assay. The phenotypic
“fingerprint” of this weak cytotoxicity did not cluster to any
compounds with known mechanisms of cancer cell cytotoxicity.
In addition, compounds 4 and 5 were submitted for screening
in the BioMAP assay.49 This assay tests compounds against 15
bacterial strains and compares the phenotypic response to 72
antibiotics with known mechanisms of action. The BioMAP
assessment showed only weak effects with no clustering to any
antibiotics in the compound library.
In summary, six kailuins (2−5, 7, and 8), a rare class of cyclic
depsipeptides, were isolated and characterized from a marine-
derived Gram-negative bacterium, P. halotolerans. New insights
have been obtained on the kailuins and two synthetic
derivatives, and several outcomes contained in previous
publications have been re-examined and clarified. Two new
analogues have been isolated (7 and 8), and the double-bond
geometry of 4 has been revised. The basis for absolute
configuration assignments in each of the constituent depsipep-
tide subunits was re-examined. An NMR 13C shift database was
created and used to determine the previously unassigned
configuration as R at the β-acyloxy position of all but one of the
previously isolated kailuins. This strategy was also extended in
the R/S configuration analysis of other β-acyloxy-containing
natural products. Additionally, it appears that the bioactivity
properties of the kailuins have been overstated in the past. The
screening responses measured here, including the cytotoxic data
of Table 2, cytological profiling, and BioMAP antimicrobial
evaluations (data not shown), revealed weak to no activity. The
biological activity properties reported in previous publications
were not reproduced. Such an outcome is surprising because
other members of this class have been reported to exhibit
activity, including bacteriostatic,36 quorum-sensing interfer-
ence,22 antifungal effects,50 and mild cytotoxicity to HCT-116
cells.40
1
were run on a Varian Unity spectrometer (500 and 125 MHz for H
and 13C, respectively) or Varian INOVA spectrometer (600 and 150
1
MHz for H and 13C, respectively) equipped with a cryoprobe. High-
accuracy mass spectrometry measurements were recorded for all
compounds (except 14) using an Applied Biosystems Mariner
electrospray ionization time-of-flight (ESITOF) mass spectrometer.
HAMS measurements for 14 were obtained using a Thermo Velos Pro
electrospray ionization hybrid ion trap−orbitrap mass spectrometer.
Prefractionation HPLC was performed using a Phenomenex Gemini-
NX C18 (50 × 21.1 mm, 5 μm) column. Semipreparative HPLC
fractions were generated using a Phenomenex Luna 5 μm C18 (250 ×
10 mm, 5 μm) column. Analytical LC-MS analysis was performed on
samples at a concentration of approximately 15−20 mg/mL, using a
reversed-phase 150 × 4.60 mm 5 μm C18 Phenomenex Luna column
in conjunction with a 4.0 × 3.0 mm C18 (octadecyl) guard column and
cartridge (holder part number: KJ0-4282; cartridge part number: AJ0-
4287, Phenomenex, Inc., Torrance, CA, USA). An injection volume of
15 μL and a solvent flow rate of 1 mL/min were used. Injections were
monitored using a Waters model 996 photodiode array UV detector.
The elution was subsequently split (1:1) between a SEDERE model 55
evaporative light scattering detector (ELSD) and an Applied
Biosystems Mariner electrospray ionization time-of-flight mass
spectrometer. Peptides were purified using an HPLC equipped with
an XBridge BEH130 5 μm 19 × 150 C18 column. Synthetic
preparation of intermediates, kailuin core (13) and kailuin lactam
(14), involved the purchase of 2-chlorotrityl chloride resin from
Aapptec. Fmoc-protected amino acids were purchased from Aapptec,
Novabiochem, or P3 Biosystems. HOAt was also purchased from
Aapptec. HBTU and TBDPSCl were purchased from Oakwood
Chemical. TFA and CH2Cl2 were purchased from Fisher. DMF,
EtOAc, and hexanes were purchased from Macron, and 1,3-
propanediol was purchased from Acros. All reagents were used
without further purification.
Biological Material and Taxonomic Identification. Strain
M128SB283Ax was isolated from intertidal sediment from Manresa
State Beach, Watsonville, CA. After drying in air, a small amount of the
sediment was placed on E. coli-streaked AWCX agar plates12 made
with SWS54 and 25 mg of cycloheximide. A nearly transparent
swarming bacterium was transferred to VY2 agar plates39 with SWS
and Zobell Marine Broth (DIFCO) plates until a pure culture was
obtained. Genomic DNA from strain M128SB283Ax was isolated from
the cell mass that was scraped from CY-SWS agar plates12 using the
Wizard Genomic DNA purification kit (Promega). The eubacterial
primers 27F (5′ AGA GTT TGA TCC TGG CTC AG 3′) and 1492R
(5′ ACG GCT ACC TTG TTA CGA CTT 3′) were used to amplify
16S rRNA by PCR using QIAGEN 2X master mix, with 25 pmol of
each primer, and yielded approximately 300 ng of genomic DNA. PCR
Moving forward, it would appear that the provisional
taxonomic identification of Vibrio sp. described in each of the
previous reports on the kailuins should be revisited and possibly
revised to Photobacterium. The discrepancy in taxonomic
identification of these organisms is most likely due to the
I
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
was performed with an Eppendorf Mastercycler personal thermocycler
under the following conditions: denaturation at 94 °C for 2 min,
followed by 25 cycles of denaturation at 94 °C for 1 min, annealing at
51 °C for 1 min, extension at 72 °C for 1.5 min, terminated by a final
extension at 72 °C for 7 min. PCR cleanup and sequencing was
performed by the University of California Berkeley Sequencing
Facility. Forward and reverse reads were analyzed using FinchTV
and assembled into a 1379 bp fragment of 16S rRNA. Initial analysis
via the BLAST server of the National Center for Biotechnology
Information (NCBI) revealed a match to the Photobacterium genus.55
The sequence was aligned against all Photobacterium species in the
NCBI’s nonredundant nucleotide database (RefSeq) using MAFFT
(G-INS-i strategy with 1PAM/K=2).56 A phylogenetic tree was built
using the neighbor-joining method (total of 5000 bootstraps) on the
conserved regions of the alignment, indicating that strain
M128SB284Ax is a variant of P. halotolerans.57 The DNA sequence
was deposited in GenBank with accession number KJ755052. Cultures
are maintained as a cryopreserved glycerol stock at University of
California, Santa Cruz.
Culture Conditions. A starter culture of bacterium strain
M128SB283Ax was prepared in MD1 medium12 in 75% seawater54
and rotary shaken at 200 rpm at 30 °C for 7 days. For scale-up
preparation, eight 4 L baffled Delong flasks, each containing 2 L of
sterile MD1 medium in 75% seawater with 2% w/v XAD-7 adsorbent
resin, were inoculated with 10 mL of starter culture. The flasks were
shaken at 200 rpm at 30 °C for 20 days.
Biological Assays. Data from the cytological profiling screen
involved an analysis of the cytotoxic phenotypic effects observed for
kailuins D (4) and E (5) and the kailuin core (13) and kailuin lactam
(14) against HeLa cells. The profiles of 4, 5, 13, and 14 were
compared against a panel of 500 biologically active molecular
structures with diagnostic cytotoxic phenotypic effects against HeLa
cells, as is described in detail elsewhere.48 The IC50 values reported
against HCT-116 cell lines were obtained using methods described
previously.47
Kailuin G (7): white film; [α]24 −20.0 (c 0.45, MeOH);
D
HAESITOFMS m/z [M
+
Na]+ 762.4963 (calcd for
C38H69N5O9Na, 762.4988); for 1H, 13C, and 2D NMR data, see
Figures S9−S11 and Table S2, Supporting Information.
Kailuin H (8): white film; [α]24 −21.7 (c 0.6, MeOH);
D
HAESITOFMS m/z [M + K]+ 792.4875 (calcd for C39H71N5O9K,
1
792.4883); for H, 13C, and 2D NMR data, see Figures S12−S14 and
Table S2, Supporting Information.
Marfey’s Analysis of Kailuin D (4) and Kailuin E (5). Marfey’s
analysis31 was performed to determine the absolute configuration of
each amino acid contained in 4 and 5 (Figures S25 and S26,
Supporting Information). Briefly, 0.5 mg of each compound was
dissolved in 6 M HCl and heated at 100 °C for approximately 18 h.
After heating, the sample was dried under reduced pressure,
redissolved in 50 μL of water, and transferred to a 1.5 mL centrifuge
tube. Aliquots consisting of 50 μL of 50 mM of each enantiomer of
each standard amino acid were treated independently in 1.5 mL
centrifuge tubes. To each standard and sample were added 100 μL of
1% FDAA and 20 μL of 1 M NaHCO3. Tubes were capped and heated
in a 40 °C water bath with periodic mixing for 1 h. After cooling to
room temperature, 10 μL of 2 M HCl was added to each tube. The
mixtures were dried under reduced pressure and redissolved in 200 μL
of 9:1 MeOH−H2O. Samples were diluted further to one-tenth of
their original concentration with 9:1 MeOH−H2O prior to LC-MS
analysis. The configuration of the Leu and Ile amino acids was
confirmed using a standard HPLC gradient using CH3CN and water
spiked with 0.1% formic acid as the mobile phase. To achieve
separation of the FDAA derivatives of (R)- and (S)-Ser and (2S,3R)-
and (2S,3S)-Thr, a mixture of 90:10 CH3CN−MeOH was used as the
organic solvent and 5% acetic acid in water was used as the aqueous
solvent. Details of the gradient profiles, columns, flow rates, and
chromatographic method development are described elsewhere
(Figures S25−S28, Supporting Information).
Synthesis and Purification of Kailuin B Depsipetide Alcohol
1 (9). A 2.9 mg (3.99 μmol) sample of kailuin B (2) was dissolved in
2175 μL of MeOH in a 20 mL scintillation tube. Then, 725 μL of a
0.025 M LiOH solution was added to the dissolved depsipeptide to
make a 1 mg/mL reaction solution. The solution was briefly sonicated
and placed in a 30 °C incubator overnight. The reaction was then
quenched with equimolar (0.025 M) HCl and dried on a SpeedVac.
The dried residue was washed with anhydrous THF three times to
solubilize the product and the resulting LiCl salt, to yield 1.2 mg of a
mixture of 9 (m/z 725) and 10 (m/z 758). Separation with a
CH3CN−H2O gradient from 50% to 100% acetonitrile over 20 min (1
mL/min) using a Kinetex C18 column (2.6 μm, 100 × 4.6 mm,
Phenomenex) yielded 0.6 mg of 9 and 0.2 mg of 10.
Extraction and Isolation. Resin and cell mass were vacuum-
filtered from the fermentation broth and exhaustively extracted with
methanol, then acetone. The crude extract was evaporated to dryness
and then partitioned between water and dichloromethane to yield 450
mg of organic extract. Half of the organic extract was prefractionated
via HPLC using CH3CN−H2O, 10 mL/min flow rate, and the
following step gradient: 0% CH3CN for 0.5 min, linear gradient to
40% CH3CN for 0.6 min, hold at 40% CH3CN for 0.4 min, linear
gradient to 70% CH3CN over 1 min, hold at 70% CH3CN for 0.4 min,
linear gradient to 100% CH3CN over 0.5 min, hold at 100% CH3CN
for 10 min. Prefraction collection times were as follows: F0: 0.0−2.0
min, F1: 2.0−3.5 min, F2: 3.5−5.0 min, F3: 5.0−6.5 min, F4: 6.5−8
min, F5: 8−9.5 min, F6: 9.5−14.0 min (see Scheme 1). Prefractions
3−5 (47 mg) were combined and subjected to semipreparative HPLC
using 80% CH3CN−20% H2O (with 0.1% formic acid) at 4 mL/min,
UV absorbance detection at 210 nm, to yield 0.8 mg of 3 (tR =13.5
min), 2.0 mg of 2 (tR = 14.4 min), 3.2 mg of 4 (tR = 17.6 min), 1.7 mg
of 7 (tR = 19.4 min), 1.7 mg of 8 (tR = 24.8 min), and 4.0 mg of 5 (tR =
26.9 min).
Kailuin B depsipeptide alcohol I (9): white solid; HAESIMS m/z
748.48310 [M + Na]+ (calcd for C37H67N5O9Na, 748.48562). See
1
Figure S29−S31 (Supporting Information) for relevant H and 2D
NMR data.
Kailuin B linear alcohol (10): white solid; nominal m/z 758.8 [M +
Na]+.
Preparation of the (R)- and (S)-MTPA Ester Derivatives.
Compound 9 (0.6 mg) was dried in a vacuum desiccator overnight.
The dried residue was dissolved in deuterated pyridine (666 μL), and
equal volumes (333 μL) were transferred into two 5 mm CHCl3
Shigemi NMR tubes under a N2 gas stream. (S)-(+)-α-Methoxy-α-
(trifluoromethyl)phenylacetyl ((S)-MTPA) chloride (2 μL) was added
into the NMR tube and lightly shaken to mix 9 evenly with the
reagent. In a separate NMR tube, (R)-(−)-α-methoxy-α-
(trifluoromethyl)phenylacetyl chloride ((R)-MTPA) (2 μL) was
added and shaken to mix 9 with the reagent. After 4 h the reaction
mixture showed a nominal mass spectrometric signal (m/z 599.49)
representing Mosher esterification to C-3 and the (2S,3R)-Thr Cβ
hydroxy group. Relevant 1H NMR data used for Mosher’s analysis are
described below (see also Figure S32, Supporting Information)
(R)-MTPA ester (12a): white solid; 1H NMR of 12 (600 MHz,
pyridine-d5) δ 9.69 (1H, d, J = 6.96 Hz, Ser NH), 5.28 (1H, m, Ser
Hα), 4.55 (1H, dd, J = 10.68, 5.4 Hz, Ser Hβ, Hb), 4.33 (1H, dd, J =
Kailuin B (2): white film; [α]24 −20.1 (c 0.9, MeOH);
D
HAESITOFMS m/z [M + K]+ 764.4591 (calcd for C37H67N5O9K,
764.4570); for 1H and 13C NMR data, see Table S2, Supporting
Information.
Kailuin C (3): white film; [α]24 −14.5 (c 0.55, MeOH);
D
HAESITOFMS m/z [M + K]+ 764.4573 (calcd for C37H67N5O9K,
764.4570); for 1H and 13C NMR data, see Table S3, Supporting
Information.
Kailuin D (4): white film; [α]24 −15.0 (c 0.8, MeOH);
D
HAESITOFMS m/z [M + K]+ 790.4739 (calcd for C39H69N5O9K,
1
790.4727); for H, 13C, and 2D NMR data, see Figures S4−S8 and
Table S2, Supporting Information.
Kailuin E (5): white powder; [α]24 −19.4 (c 0.1, MeOH);
D
HAESITOFMS m/z 776.5134 (calcd for C39H71N5O9Na, 776.5144
[M + Na]+); for 1H, 13C, and 2D NMR data, see Table S3, Supporting
Information.
J
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
10.98, 5.7 Hz, Ser Hβ, Hb), 2.19 (2H, m, H-2), 1.47 (3H, d, J = 6.06,
Thr Hγ).
30% EtOAc−CH2Cl2) and concentrated to yield the target compound
(0.4 g, 24% yield) as a white solid. The observed TLC Rf = 0.43 30%
EtOAc−Hex 1% acetic acid, silica. 1H NMR (500 MHz, chloroform-d)
δ 1.04 (s, 9H), 2.61 (t, J = 6.2 Hz, 2H), 3.95 (t, J = 6.2 Hz, 2H), 7.35−
7.48 (m, 6H), 7.67 (dd, J = 7.6, 1.4 Hz, 4H); 13C NMR (125 MHz,
CDCl3) δ 19.3, 26.9, 37.7, 59.7, 127.9, 129.9, 133.5, 135.7, 178.0;
LCESIMS m/z (relative intensity) 351 (M + Na) (6), 292 (17), 251
(M − Ph+) (100), 172.7 (7), 118 (14), 83 (17).
(S)-MTPA ester (12b): white solid; 1H NMR of 12 (600 MHz,
pyridine-d5) δ 9.76 (1H, d, J = 5.58 Hz, Ser NH), 5.25 (1H, m, Ser
Hα), 4.56 (1H, dd, J = 10.68, 5.4 Hz, Ser Hβ, Hb), 4.35 (1H, dd, J =
10.68, 5.3 Hz, Ser Hβ, Hb), 2.13 (2H, m, H-2), 1.50 (3H, d, J = 6.18,
Thr Hγ).
Peptide Purification. Peptides were purified after solution-phase
cyclization on a Biotage Isolera Prime system equipped with a KP-
C18-HS 12 g column eluting with H2O−CH3CN modified with 0.1%
TFA. Peptides were purified after global deprotection on a Waters
mass-directed preparative HPLC system eluting with H2O−CH3CN
modified with 0.1% formic acid.
Kailuin core (13): white powder; [α]24 −45.8 (c 1.55, MeOH);
D
HAESITOFMS m/z 600.3603 [M + H]+ (calcd for C28H50N5O9,
600.3602 [M + H]+); for 1H and 13C NMR data, see Figures S16−S19
and Table S5, Supporting Information.
Kailuin lactam (14): white powder; [α]24 −1.6 (c 1.25, MeOH);
HAESITOFMS m/z 599.37705 [M + H]+ D(calcd for C28H51N6O8,
Loading Fmoc-S-Leucine-OH onto 2-Chlorotrityl Chloride
Resin. A 3.0 g aliquot of Fmoc-S-Leu-OH was dried under a vacuum
for 24 h. A 5 g bottle of 2-chlorotrityl chloride resin (200−400 mesh;
1% DVB) was stored in a desiccator under vacuum prior to use.
Anhydrous CH2Cl2 was added to the Fmoc-S-Leu-OH, followed by
dry DIPEA (4 equiv, 5.6 mL). The solution was transferred by syringe
to a flame-dried round-bottom flask containing 5.0 g of resin and a stir
bar. The resin and solution were stirred gently at room temperature
for 3 h. After 3 h, the resin and solution were transferred to a
polypropylene synthesis vial containing a fritted disk and fitted with a
Teflon stopcock. The resin was drained and rinsed with CH2Cl2 (3 ×
30 mL), DMF (3 × 30 mL), and again with CH2Cl2 (3 × 30 mL). The
remaining resin was treated with a solution of CH2Cl2, methanol, and
DIPEA (17:2:1 respectively, 2 × 30 min), in order to cap any
unreacted linker. The resin was dried thoroughly, and a 10 mg aliquot
was taken for use in determining the loading value.
Determination of Resin Loading by the Fmoc Cleavage
Method. A sample of dry resin (10 mg) was transferred to a
microcentrifuge tube and allowed to swell in 800 μL of DMF. After 15
min, 200 μL of piperidine was added. The resin and solution were
vortexed to effect good mixing and allowed to sit for an additional 15
min. After the Fmoc cleavage was complete, 10 μL of the solution was
transferred to a quartz cuvette and diluted with 990 μL of DMF. The
absorbance was measured at λ = 301 nm. Using an extinction
coefficient of 7800 L·mol−1·cm−1, the loading value was determined to
be 1.6 mmol of Fmoc-S-Leu-OH per gram of resin.
Preparation of 3-[(tert-Butyldiphenylsilyl)oxy)]propan-1-ol.
To a round-bottom flask charged with CH2Cl2 (10 mL) were added
1,3-propanediol (0.54 mL, 7.5 mmol) and DIPEA (1.96 mL, 11.25
mmol). TBDPSCl (2.04 mL, 7.88 mmol) was slowly added dropwise,
and the reaction was allowed to proceed under N2 for 15 h at room
temperature. The reaction was reduced in vacuo, redissolved in ethyl
acetate (100 mL), and washed with saturated citric acid (3 × 50 mL),
followed by 0.5 M HCl (50 mL). The combined aqueous layers were
back-extracted with ethyl acetate (50 mL). The combined organic
layers were washed with brine (50 mL) and dried over sodium sulfate.
The product was purified via flash chromatography (3% EtOAc−
hexanes to 30% EtOAc−hexanes) and concentrated to yield the title
compound (1.69 g, 70% yield) as a translucent, colorless solid. The
observed TLC Rf = 0.5 30% EtOAc−hexanes, silica gel. 1H NMR (500
MHz, CDCl3) δ 1.06 (16H, s), 1.82 (3H, p J = 5.6 Hz), 2.06 (1H, br
s), 3.85 (6H, td, J = 5.6, 2.6 Hz), 7.37−7.48 (6H, m), 7.68 (4H, dd, J =
8.0, 1.6 Hz); 13C NMR (125 MHz, CDCl3) δ 19.3, 27.1, 34.4, 62.1,
63.4, 127.9, 130.0, 133.4, 135.7; LC-ESIMS m/z 278 (8), 237 (M −
Ph+) (12), 199 (30), 158 (96), 119 (8), 118 (29), 117 (100), 83 (13).
Preparation of 3-[(tert-Butyldiphenylsilyl)oxy]propanoic
Acid. A 1.6 g (5.1 mmol) amount of 3-[(tert-butyldiphenylsilyl)oxy)]
was dissolved in 15 mL of DMF and added to a round-bottom flask.
To this was added pyridinium HCl (4.13 g, 35.7 mmol) followed by a
slow addition of K2Cr2O7 (5.25 g, 17.6 mmol), and the reaction was
allowed to stir overnight. The reaction was cooled to 4 °C, and 3 g of
Celite was added, followed by 40 mL of H2O and 20 mL of diethyl
ether. This mixture was stirred at 4 °C for 45 min, then filtered over
Celite. The aqueous layer was extracted with diethyl ether (3 × 30
mL). The combined organic layers were washed with 1 M HCl (50
mL) and saturated brine (50 mL), then dried over sodium sulfate. The
product was purified via flash chromatography (3% EtOAc−CH2Cl2 to
1
599.37684 [M + H]+); for H and 13C NMR data, see Figures S20−
S24 and Table S5, Supporting Information.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information contains a selective view of
biosynthetic products from marine-derived strains of Gram-
negative bacteria, NMR spectra, detailed explanations of
Marfey’s analysis and chromatographic optimizations, and
information on synthetic methodologies. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Tel: (831) 459-2603. Fax: (831) 459-2935. E-mail: pcrews@
ucsc.edu.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by grants from the NSF (CHE
1214065) (P.C.) and NIH (R01 CA 47135) (P.C.). We also
acknowledge additional funding from NIH (R01GM084530)
(R.S.L.) and an NSF award (CHE 1427922) for MS
instrumentation. The authors would like to acknowledge W.
Bray at the UCSC Chemical Screening Center for assistance
with cytological profiling and K. Tenney for collection of the
sediment sample. Also, P. Dunlap at the University of Michigan
is acknowledged for sample exchanges, guidance, and
confirmation of bacterial taxonomy. The authors also wish to
acknowledge S. Loveridge and H. Lee for general technical help
especially in NMR data processing.
DEDICATION
■
Dedicated to Dr. William Fenical of Scripps Institution of
Oceanography, University of California−San Diego, for his
pioneering work on bioactive natural products.
REFERENCES
■
(1) Whitman, W. B.; Coleman, D. C.; Wiebe, W. J. Proc. Natl. Acad.
Sci. U.S.A. 1998, 95, 6578−6583.
(2) Fenical, W. Chem. Rev. 1993, 93, 1673−1683.
(3) Blunt, J. W.; Copp, B. R.; Keyzers, R. A.; Munro, M. H. G.;
Prinsep, M. R. Nat. Prod. Rep. 2014, 31, 160−258.
(4) Souto, A.; Montaos, M. A.; Rivas, A. J.; Balado, M.; Osorio, C. R.;
Rodriguez, J.; Lemos, M. L.; Jimenez, C. Eur. J. Org. Chem. 2012, 29,
5693−5700.
(5) Xu, Y.; Kersten, R. D.; Nam, S.-J.; Lu, L.; Al-Suwailem, A.;
Zhenge, H.; Fenical, W.; Dorrestein, P. C.; Moore, B. S.; Qian, P.-Y. J.
Am. Chem. Soc. 2012, 134, 8625−8632.
K
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(6) Pinhassi, J.; Zweifel, U. L.; Hagstrom, A. Appl. Environ. Microbiol.
1997, 63, 3359−3366.
(39) Su, B.-N.; Park, E. J.; Mbwambo, Z. H.; Santarsiero, B. D.;
Mesecar, A. D.; Fong, H. H. S.; Pezzuto, J. M.; Kinghorn, A. D. J. Nat.
Prod. 2002, 65, 1278−1282.
(7) Joint, I.; Muhling, M.; Querellou, J. Microb. Biotechnol. 2010, 3,
564−575.
́
(40) Seco, J. M.; Quinoa, E.; Riguera, R. Chem. Rev. 2004, 104, 17−
̃
117.
(8) Still, P. C.; Johnson, T. A.; Theodore, C. M.; Loveridge, S. T.;
Crews, P. J. Nat. Prod. 2014, 77, 690−702.
(41) Yu, H. M.; Chen, S. T.; Wang, K. T. J. Org. Chem. 1992, 57,
4781−4784.
(9) Weissmann, K. J.; Muller, R. Nat. Prod. Rep. 2010, 27, 1276−
̈
(42) Druais, V.; Hall, M. J.; Corsi, C.; Wendeborn, S. V.; Meyer, C.;
Cossy, J. Org. Lett. 2009, 11, 935−938.
1295.
(10) Konig, G. M.; Kehraus, S.; Seibert, S. F.; Abdel-Lateff, A.;
̈
(43) Salit, A.-F.; Meyer, C.; Cossy, J.; Delouvrie, B.; Hennequin, L.
Tetrahedron 2008, 64, 6684−6697.
Muller, D. ChemBioChem 2006, 7, 229−238.
̈
(11) Tsukimoto, M.; Nagaoka, M.; Shishido, Y.; Fujimoto, J.;
Nishisaka, F.; Matsumoto, S.; Harunari, E.; Imada, C.; Matsuzaki, T. J.
Nat. Prod. 2011, 74, 2329−2331.
(44) Gracia, C.; Isidro-Llobet, A.; Cruz, L. J.; Acosta, G. A.; Alvarez,
M.; Cuevas, C.; Giralt, E.; Albericio, F. J. Org. Chem. 2006, 71, 7196−
7204.
(12) Reichenbach, H.; Dworkin, M. In The Prokaryotes, 2nd ed.;
(45) Chatterjee, J.; Laufer, B.; Kessler, H. Nat. Protoc. 2012, 7, 432−
444.
Balows, A.; Trueper, H. G.; Dworkin, M.; Harder, W.; Schleifer, K.-H.,
̈
Eds.; Springer-Verlag: Berlin, 1992; pp 3416−3487.
(13) Berrue, F.; Withers, S. T.; Haltli, B.; Withers, J.; Kerr, R. G. Mar.
Drugs 2011, 9, 369−381.
(46) Subiros-Funosas, R.; Nieto-Rodriguez, L.; Jensen, K. J.;
Albericio, F. J. Pept. Sci. 2013, 19, 408−414.
(47) Valeriote, F. A.; Tenney, K.; Media, J.; Pietraszkiewicz, H.;
Edelstein, M.; Johnson, T. A.; Amagata, T.; Crews, P. J. Exp. Ther.
Oncol. 2012, 10, 119−134.
(48) Woehrmann, M. H.; Bray, W. M.; Durbin, J. K.; Nisam, S. C.;
Michael, A. K.; Glassey, E.; Stuart, J. M.; Lokey, R. S. Mol. Biosyst.
2013, 9, 2604−2617.
(49) Wong, W. R.; Oliver, A. G.; Linington, R. G. Chem. Biol. 2012,
16, 1583−1495.
(50) Enticknap, J.; Hamann, M. T.; Hill, R. T.; Rao, K. V. U.S. patent
WO2005042720A2, 2005.
(51) Urbanczyk, H.; Ogura, Y.; Hendry, T. A.; Gould, A. L.; Kiwaki,
N.; Atkinson, J. T.; Hayashi, T.; Dunlap, P. V. J. Bacteriol. 2011, 193,
3144−3145.
(14) Wietz, M.; Mansson, M.; Gotfredsen, C. H.; Larsen, T. O.;
Gram, L. Mar. Drugs 2010, 8, 2946−2960.
(15) Qin, Z.; Baker, A. T.; Raab, A.; Huang, S.; Wang, T.; Yu, Y.;
Jaspars, M.; Secombes, C. J.; Deng, H. J. Biol. Chem. 2013, 288,
14688−14697.
(16) Kenig, M.; Reading, C. J. Antibiot. 1979, 32, 549−554.
(17) de Nys, R.; Kumar, N.; Sharara, K. A.; Srinivasan, S.; Ball, G.;
Kjelleberg, S. J. Nat. Prod. 2001, 64, 531−532.
(18) Suzuki, A.; Goto, M. Biochim. Biophys. Acta 1973, 313, 229−234.
(19) Mansson, M.; Nielsen, A.; Kjærulff, L.; Gotfredsen, C. H.; Wietz,
M.; Ingmer, H.; Gram, L.; Larsen, T. O. Mar. Drugs 2011, 9, 2537−
2552.
(52) Park, Y. D.; Baik, K. S.; Seong, C. N.; Bae, K. S.; Kim, S.; Chun,
J. Int. J. Syst. Evol. Microbiolol. 2006, 56, 745−749.
(53) Blin, K.; Medema, M. H.; Kazempour, D.; Fischbach, M. A.;
Breitling, R.; Takano, E.; Weber, T. Nucleic Acids Res. 2013, 41,
W240−W212.
(20) Adachi, K., Kawabata, Y.; Kasai, H.; Katsuta, M.; Shizuri, Y.
Japanese patent JP 2007230911, 2007.
(21) Oku, N.; Kawabata, K.; Adachi, K.; Katsuta, A.; Shizuri, Y. J.
Antibiot. 2008, 61, 11−17.
(22) Kjaerullf, L.; Nielsen, A.; Mansson, M.; Gram, L.; Larsen, T. O.;
Ingmer, H.; Gotfredsen, C. H. Mar. Drugs 2013, 11, 5051−5062.
(23) Stackebrandt, E.; Goebel, B. M. Int. J. Syst. Bacteriol. 1994, 44,
846−849.
(24) Rivas, R.; Garcia-Fraile, P.; Mateos, P. F.; Martinez-Molina, E.;
Velazquez, E. Int. J. Sys. Evol. Microbiol. 2006, 56, 1067−1071.
(25) Appleton, D. R.; Buss, A. D.; Butler, M. S. Chim. Int. J. Chem.
2007, 61, 327−331.
(54) Iizuka, T.; Jojima, Y.; Fudou, R.; Yamanaka, S. FEMS Microbiol.
Lett. 1998, 169, 317−322.
(55) Altschul, S. F.; Gish, W.; Miller, W.; Myers, E. W.; Lipman, D. J.
J. Mol. Biol. 1990, 215, 403−410.
(56) Katoh, S. Mol. Biol. Evol. 2013, 30, 772−780.
(57) Saitou, N.; Nei, M. Mol. Biol. Evol. 1987, 4, 406−425.
(26) Harrigan, G. G.; Harrigan, B. L.; Davidson, B. S. Tetrahedron
1997, 53, 1577−1582.
(27) Yoshitaka, K.; Kenichi, M.; Miyuki, N.; Masato, S. Japanese
patent JP2000245497-A 20000912, 2000.
(28) Raju, R.; Kawabata, K.; Nishijima, M.; Aalbersberg, W. G. L.
Tetrahedron Lett. 2012, 53, 6905−6907.
(29) Fuji, K.; Ikai, Y.; Mayumi, T.; Oka, H.; Suzuki, M.; Harada, K. I.
Anal. Chem. 1997, 69, 3346−3352.
(30) Crews, P.; Rodríguez, J.; Jaspars, M. Organic Structure Analysis,
2nd ed.; Oxford University Press: New York, 2010; pp 82, 152.
(31) Marfy, P. Carlsberg. Res. Commun. 1984, 49, 591−596.
(32) Hess, S.; Gustafson, K. R.; Milanowski, D. J.; Alvira, E.; Lipton,
M. A.; Pannell, L. K. J. Chromatogr. A 2004, 1035, 211−219.
(33) Rao, K. V.; Na, M.; Cook, J. C.; Peng, J.; Matsumoto, R.;
Hamann, M. T. J. Nat. Prod. 2008, 71, 772−778.
(34) Hashizume, H.; Hirosawa, S.; Sawa, R.; Muraoka, Y.; Ikeda, D.;
Naganawa, H.; Igarashi, M. J. Antibiot. 2004, 57, 52−58.
(35) Wyche, T. P.; Hou, Y.; Vazquez-Rivera, E.; Braun, D.; Bugni, T.
S. J. Nat. Prod. 2012, 75, 735−740.
(36) Lam, T.; Salvatore, M. J.; Hirsch, C.; Hensens, O. D.; Zink, D.;
Bernard-King, A.; Lee, N.; Graham, A.; Pelak, B. Tetrahedron 1998, 54,
4755−4760.
(37) Trischman, J. A.; Jensen, P. R.; Fenical, W. Tetrahedron Lett.
1994, 35, 5571−5574.
(38) Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512−519.
L
DOI: 10.1021/np500840n
J. Nat. Prod. XXXX, XXX, XXX−XXX