Chloramphenicol Derivatives with Antibacterial Activity Identified by Functional Metagenomics

Article abstract of DOI:10.1021/acs.jnatprod.7b00903

A functional metagenomic approach identified novel and diverse soil-derived DNAs encoding inhibitors to methicillin-resistant Staphylococcus aureus (MRSA). A metagenomic DNA soil library containing 19 200 recombinant Escherichia coli BAC clones with 100 Kb average insert size was screened for antibiotic activity. Twenty-seven clones inhibited MRSA, seven of which were found by LC-MS to possess modified chloramphenicol (Cm) derivatives, including three new compounds whose structures were established as 1-acetyl-3-propanoylchloramphenicol, 1-acetyl-3-butanoylchloramphenicol, and 3-butanoyl-1-propanoylchloramphenicol. Cm was used as the selectable antibiotic for cloning, suggesting that heterologously expressed enzymes resulted in derivatization of Cm into new chemical entities with biological activity. An esterase was found to be responsible for the enzymatic regeneration of Cm, and the gene trfA responsible for plasmid copy induction was found to be responsible for inducing antibacterial activity in some clones. Six additional acylchloramphenicols were synthesized for structure and antibacterial activity relationship studies, with 1-p-nitrobenzoylchloramphenicol the most active against Mycobacterium intracellulare and Mycobacterium tuberculosis, with MICs of 12.5 and 50.0 μg/mL, respectively.

Full text of DOI:10.1021/acs.jnatprod.7b00903

Article
pubs.acs.org/jnp
Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX
Chloramphenicol Derivatives with Antibacterial Activity Identified by
Functional Metagenomics
Shamima Nasrin,†,# Suresh Ganji,‡,# Kavita S. Kakirde,† Melissa R. Jacob,‡ Mei Wang,‡
Ranga Rao Ravu,‡ Paul A. Cobine,† Ikhlas A. Khan,‡,§ Cheng-Cang Wu,⊥ David A. Mead,∥
Xing-Cong Li,*,‡,§ and Mark R. Liles*,†,∥
†Department of Biological Sciences, Auburn University, Auburn, Alabama 36849, United States
‡National Center for Natural Products Research, Research Institute of Pharmaceutical Sciences, School of Pharmacy, The University
of Mississippi, University, Mississippi 38677, United States
§Department of Biomolecular Sciences, School of Pharmacy, The University of Mississippi, University, Mississippi 38677, United
States
⊥Lucigen Corporation, Middleton, Wisconsin 53562, United States
∥Varigen Biosciences Corporation, Madison, Wisconsin 53719, United States
S
* Supporting Information
ABSTRACT: A functional metagenomic approach identified novel
and diverse soil-derived DNAs encoding inhibitors to methicillin-
resistant Staphylococcus aureus (MRSA). A metagenomic DNA soil
library containing 19 200 recombinant Escherichia coli BAC clones
with 100 Kb average insert size was screened for antibiotic activity.
Twenty-seven clones inhibited MRSA, seven of which were found
by LC-MS to possess modified chloramphenicol (Cm) derivatives,
including three new compounds whose structures were established
as 1-acetyl-3-propanoylchloramphenicol, 1-acetyl-3-butanoyl-
chloramphenicol, and 3-butanoyl-1-propanoylchloramphenicol. Cm
was used as the selectable antibiotic for cloning, suggesting that
heterologously expressed enzymes resulted in derivatization of Cm into new chemical entities with biological activity. An esterase
was found to be responsible for the enzymatic regeneration of Cm, and the gene trfA responsible for plasmid copy induction was
found to be responsible for inducing antibacterial activity in some clones. Six additional acylchloramphenicols were synthesized
for structure and antibacterial activity relationship studies, with 1-p-nitrobenzoylchloramphenicol the most active against
Mycobacterium intracellulare and Mycobacterium tuberculosis, with MICs of 12.5 and 50.0 μg/mL, respectively.
he need for novel antimicrobial compounds is increasing
to discover novel antimicrobial compounds to combat MRSA
Twith the emergence of hypervirulent and multidrug- and other MDR pathogens.1,2
resistant (MDR) bacterial pathogens, such as MRSA.1,2 The
increasing frequency of antimicrobial resistance observed
among Staphylococcus aureus clinical isolates and their ability
to produce biofilms in tissues and medical devices limit
treatment options.3 The glycopeptide antibiotic vancomycin is
the first choice of drug for the treatment of methicillin-resistant
S. aureus (MRSA) infections; however, its application is limited
by the emergence of strains with reduced antimicrobial
susceptibility4 and the occurrence of vancomycin treatment
failure and mortality in patients with S. aureus bacteremia.5,6
Several new alternative antibiotics such as daptomycin,
linezolid, and tigecycline are used to treat MRSA-infected
patients.7 However, their application efficacies are limited by
the different varieties of infection caused by MRSA.8 The three
newest drugs oritavancin, dalbavancin, and tedizolid have
recently been approved by the FDA for treating MRSA
infections,9 yet these new drugs can only be used for skin
infection and nosocomial pneumonia. There is therefore a need
Functional metagenomics has opened a new avenue for
natural products drug discovery.10−12 By sampling DNA
directly from a natural environment it is possible to access a
greater proportion of microbial genomes13,14 and screen the
resultant metagenomic libraries for recombinant clones that
express a specific activity. Fosmid-based metagenomic
approaches15 have been used in other studies to identify
novel small molecules with antimicrobial properties,16 such as
the isolation of terragine E and related compounds,17 long-
chain N-acyl antibiotics,18 and the antibiotic palmitoylputres-
cine.19 Screening larger-insert metagenomic libraries for
antimicrobial activity is advantageous due to the higher
probability that a recombinant clone will encode an intact
biosynthetic gene cluster or antimicrobial peptides. In this
study, we screened a soil metagenomic library containing 110
Received: October 27, 2017
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Kb average size inserts constructed in a copy-inducible BAC
vector17,20 for recombinant clones that inhibited MRSA
metabolic activity. We identified seven metagenomic clones
that were capable of modifying chloramphenicol (Cm) to
biologically active Cm derivatives. The structures of the three
new Cm derivatives were determined, and these compounds
were found to have antibacterial activity against MRSA,
Mycobacterium intracellulare, and M. tuberculosis. These findings
demonstrate that large-insert soil metagenomic libraries can be
functionally screened to access previously undescribed environ-
mental DNA and explore their biochemical diversity for
antibacterial discovery.
the addition of arabinose to the growth medium.23 Enhanced
expression of cloned genes due to gene dosage amplification
may increase downstream concentrations of antibiotic products,
and therefore the likelihood of their detection. This effect was
observed when representative clones with the highest activity
(P11K11, P15G24, P16O8, P17L5, P29A3, P30A5, and P30L8)
were grown with and without 0.01% arabinose, with arabinose
addition showing increased antibiotic activity ranging from 3.5-
to 13.2-fold relative to the clone cultures without arabinose
(Figure S2).
Preliminary Characterization of Antibiotic-Expressing
Clones. The MRSA strain EAMC30 inhibition activity for each
clone or clone fraction was calculated as the % reduction of
resazurin fluorescence as compared to the respective negative
control. For the negative control, both the cell-free supernatant
and the cell lysate were tested so that a true comparison could
be made with the respective clone samples. MRSA inhibition
ranged from 8% to 75% (Table S1) among the 27 validated
clones tested. Seventeen clones showed cell-free supernatant
inhibition activity equivalent to that of their cognate cell lysates
(Table S1). This suggests that the active compound(s) is most
likely extracellular or readily secreted out of the cell. For 10
clones the activity in the cell lysate was significantly higher than
the activity detected in the cell-free supernatant. This
presumably indicates that the active compound(s) is intra-
cellular or membrane-associated, is not readily secreted out of
the cell, and/or is freely soluble in the supernatant. For all of
the subsequent tests for these 10 clones with higher activity in
the cell lysate, the cell lysate was used in bioassays, whereas the
cell-free supernatant was used for the other clones.
Cell-free supernatants or lysates that were incubated for 10
min in a boiling water bath resulted in up to 91% inhibition
activity against MRSA strain EAMC30 (Table S1), with four
clones showing complete loss of activity, suggesting that the
active compound(s) is heat sensitive. Seven clones showed
partial loss in activity, while 16 clones showed a gain in activity,
indicating that the active compound(s) is heat stable and likely
to be nonproteinaceous. Twelve clones showed a 5−12-fold
increase in inhibitory activity after boiling, perhaps because the
active compounds are sequestered in some way in the crude
lysate.
In the fractionation assay, the filtrate obtained after
separation using a 3 kDa molecular weight cutoff (MWCO)
membrane was tested. For 11 of the 27 clones the filtrate (<3
kDa fraction) did not show a loss of activity. For 14 clones,
even though there was a decrease in the % activity the overall
MRSA inhibition was still more than 50%. For seven clones,
there was a significant drop in the activity with less than 50%
inhibition. Taken together, these data indicated that for many
of the clones the activity was attributable to a heat-resistant
nonproteinaceous small-molecule compound with molecular
weight less than 3 kDa.
RESULTS AND DISCUSSION
■
The metagenomic library used in this study was obtained using
randomly sheared metagenomic DNA that resulted in large
insert sizes that increased the number of genes per each clone.
An in situ lysis method was developed for high-throughput
library screening for clones expressing an antibacterial activity
that allowed for growth of the metagenomic cultures, lysis, and
screening in the same 96-well plates and detected both extra-
and intracellular compounds. An inducible copy number BAC
vector was used to enhance natural product expression. The
identified clones with antibacterial activity were analyzed in
terms of biosynthetic gene clusters, active metabolites, and
relevant genes and enzymes associated with the activity.
Identification of Antibiotic-Producing Clones. To
enhance expression of the cloned genes, cultures were
incubated for a prolonged time in stationary phase.21,22 Each
E. coli clone was grown for 48 h at 37 °C, and cells were lysed
by freezing at −80 °C followed by rapid thawing at 55 °C, prior
to assaying supernatants for inhibition of MRSA metabolic
activity. The addition of nalidixic acid along with the diluted
log-phase MRSA inhibited E. coli growth but did not interfere
with MRSA growth. The use of the metabolic indicator dye
resazurin allowed visualization of the wells in which MRSA cells
were metabolically active and induced a color change from blue
to pink as well as the generation of fluorescence. A total of 136
metagenomic clones that resulted in at least a 20% reduction in
fluorescence compared to an empty vector control were
selected for subsequent validation. The 136 clones identified
from the primary screen out of 19 200 clones in the library
reflects a hit rate of 0.7%.
Validation of Recombinant Clones with Antimicrobial
Activity. Clones that were only weakly positive compared to
the empty vector negative control and/or did not show
consistent activity were eliminated, as well as clones that were
potential cross-well contamination, resulting in 27 candidates
selected for additional biochemical and genetic characterization.
Retransformation of Antibiotic-Producing Clones. The
27 clones that expressed an antibacterial activity were subjected
to a restriction fragment length analysis using pulsed field gel
electrophoresis (PFGE), and each had large and unique DNA
inserts (Figure S1). Each clone’s DNA was used to transform a
Sequence Analysis of Anti-MRSA Clone Inserts. DNA
insert sequences were obtained for all 27 clones using bar-
coded next-generation sequencing. After reads were trimmed
the sequences were used for assembly de novo. Most of the
clones contained large inserts with more than 100 predicted
open reading frames (ORFs) (Supplemental Data 1). Based on
an analysis of the nearest relatives to the top BLAST hit for the
majority of ORFs in each respective clone, the distribution of
clone origins at a phylum level was 37% Proteobacteria, 33%
Acidobacteria, 7% Planctomycetes, 7% Bacteriodetes, and 4%
Gemmatimonadetes and Firmicutes each (Supplementary Data
̈
naive E. coli, and the resulting recombinants all showed
significant anti-MRSA activity when assayed with MRSA
EAMC30 (Table S1). These experiments indicated that the
cloned DNA was necessary and sufficient to confer bioactivity
on the E. coli heterologous host for all 27 of the positive hits
that were validated.
Effect of Arabinose Induction on Antibiotic Activity.
The copy-control BAC vector used for this library can be
amplified from single copy up to ∼50 copies per E. coli cell by
B
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
1). Analysis of clone insert DNAs using antiSMASH4.0
predicted that eight clones (P5C24, P22C4, P23K15,
P27M10, P28H1, P28I7, P37O10, and P46O24) encode
biosynthetic gene clusters. Among these eight clones, five
clones (P5C24, P22C4, P23K15, P27M10, and P28I7) were
predicted to contain a cluster, but in each clone less than 10%
of the genes in the putative cluster had a significant ClusterBlast
hit. The other three clones were predicted to contain a
biosynthetic gene cluster with significant homology to a known
cluster. The clone P28H1 was predicted to encode a 30.5 Kb
gene cluster with limited homology (at most 13% of genes were
similar) to a fatty acid−ladderane biosynthetic cluster
(Supplemental Data 1). While the predicted core biosynthetic
genes had closest GenBank matches to hypothetical proteins
identified in Armatimonadetes taxa (range of 45−54%
similarity), other predicted ORFs on this clone had relatively
low % similarities and diverse putative origins (including from
the phyla Acidobacteria, Firmicutes, and Proteobacteria),
leaving the encoded functions and origin of this cloned DNA
in doubt. The clone P37O10 was predicted to encode four
ORFs required for biosynthesis of a bacteriocin (Supplemental
Data 1), and all four ORFs had a top BLAST match as a gene
product from Paracoccus denitrificans (range of 61−68%
similarity). While to our knowledge no bacteriocins have
been reported from a Paracoccus spp., the related bacterium
Rhodobacter capsulatus has been reported to produce
bacteriocins;24 however, given the phylogenetic distance
between Paracoccus spp. and S. aureus and the identification
of Cm derivatives from clone P37O10 supernatants (Figure 1),
it is unlikely that this predicted cluster is involved in the
observed anti-MRSA activity. The clone P46O24 was predicted
to contain a biosynthetic gene cluster ∼77.4 Kb in size that
encodes a resorcinol−arylpolyene-like cluster, with 88% of the
genes showing similarity to the flexirubin cluster from
Chitinophaga pinensis (Supplemental Data 1),25 which is
responsible for biosynthesis of orange flexirubin pigments
that are characteristic of many bacteria within the Bacteriodetes
phylum.26 Like the flexirubin cluster, clone P46O24 was
predicted to contain several β-ketoacyl synthases that suggests a
type II fatty acid-like synthesis of the polyene component;25
however, no pigmentation was observed for E. coli containing
clone P46O24, and to our knowledge no antibacterial activity
has been associated with flexirubin pigments.
In the case of clone P6L4, in silico analysis indicated the
presence of three ORFs predicted to encode esterase-like gene
products, which were subcloned by targeted PCR and cloning.
In contrast, with the seven clones that expressed a Cm
derivative (see sections below), sequence analysis did not
indicate any obvious candidates (e.g., gene products that would
catalyze Cm acetyl-, propanol-, or butanolation). Since these
clones contained many predicted ORFs with no significant
BLAST hit or ORFs with a significant hit annotated as a
hypothetical protein of unknown function, a random
subcloning approach was adopted in order to identify gene
products contributing to Cm derivative formation. At least 96
subclones derived from each clone were tested for anti-MRSA
activity, but only two of the seven clones (P6B5 and P35B14,
see below) were found to have subclones that expressed anti-
MRSA activity. It is possible that Cm derivative formation in
the other five clones required additional genetic loci that were
not captured on the 4−8 Kb subcloned fragments or that
differences in plasmid copy number or other regulatory factors
affected gene expression.
Figure 1. LC-MS chromatograms of seven metagenomic clones
showing Cm (1) and its derivatives (2−9). HPLC conditions: Acquity
UPLC BEH C18 column (2.1 × 150 mm, 1.7 μm); gradient elution
from 10% to 100% CH3CN in H2O.
Characterization of Cm Derivatives from Metagenom-
ic Clones. For the majority of clones that expressed an anti-
MRSA activity, no unique metabolite was observed by LC-MS
analysis of E. coli cultures in comparison with E. coli containing
the empty pSmartBAC-S vector (data not shown). LC-MS
analysis revealed that seven metagenomic clones (P35B14,
P22E10, P28I7, P37O10, P27M10, P6B5, and P5A4) active
against MRSA had similar metabolic profiles (Figure 1). Cm
(1) was identified in all seven clones with a retention time (tR)
of 6.37 min that produced typical pseudomolecular ion peaks at
m/z 321.2 and 323.2 in a ratio of 5:3 in the negative ESIMS
spectra. Similar isotopic patterns and UV spectra were observed
for several compounds with longer tR (e.g., compounds 2−9 in
C
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
clone P35B14), indicating they were chlorine-containing Cm
derivatives. These compounds gave pseudomolecular ion peaks
at m/z 363.2/365.2, 377.2/379.2, 391.2/393.2, 405.2/405.2,
419.2/421.2, 433.2/435.2, and 447.4/449.4 in the negative
ESIMS spectra, with a difference of 14 corresponding to the
mass of a methylene group (CH2). Although Cm derivatives
have been reported,27−30 analogues producing some of these
molecular ions were not present in current literature reports.
Thus, clone P35B14 was selected for a scale-up of fermentation
and chemical isolation due to the presence of typical Cm
derivatives (Figure 1).
with the carbonyl carbon at δ 173.6. Therefore, the structure of
7 was determined as 1-acetyl-3-propanoylchloramphenicol.
Compound 8 was also isolated as a colorless oil and showed
close similarities to compound 7 in terms of 1H and 13C NMR
spectra. Its HRESIMS indicated a molecular formula of
C17H20N2Cl2O7, which was one CH2 more than compound
1
7. The H and 13C NMR data suggested that 8 had one acetyl
group and one butanoyl group in the molecule (Table 1). The
key was to determine the linkage positions of the two acyl
groups to the 1,3-dihydroxy groups of the Cm skeleton, which
was achieved by the HMBC correlation of H-1 at δ 6.06 (1H, d,
J = 6.3 Hz) with the acetyl carbonyl C-1″ at δ 169.5 (Figure 2).
In addition, the two methylene protons of H-3 at δ 4.20 (dd, J
= 11.7, 5.5 Hz) and 4.05 (dd, J = 11.7, 6.3 Hz) showed HMBC
correlations with the butanoyl carbonyl C-1‴ at δ 173.3.
Therefore, the structure of 8 was determined as 1-acetyl-3-
butanoylchloramphenicol.
An ethyl acetate extract of the P35B14 clone fermentation
broth was subjected to column chromatography on normal-
phase and reversed-phase silica gel to yield compounds 2−9,
along with Cm (1). Compounds 2−6 were identified as 3-
acetylchoramphenicol,29 3-propanoylchloramphenicol,29 3-
butanoylchloramphenicol,31 1,3-diacetylchloramphenicol,29
and 1,3-dipropanoylchloramphenicol,29 respectively, by com-
Compound 9 had the least polarity among the eight
compounds produced by this metagenomic clone. Its molecular
formula was determined as C18H22N2Cl2O7 by the HRESIMS
and 13C NMR spectra, which was one more CH2 than
1
parison of their H and 13C NMR data and optical rotations
with those reported in the literature. The structures of the new
compounds 7−9 were established as follows.
1
Compound 7 was isolated as a colorless oil and
corresponded to the peak at tR 11.39 min in the LC-MS
chromatograms (Figure 1) and the pseudomolecular ions at m/
z 419.2/421.2 in the ESIMS spectrum. The high-resolution
ESIMS (negative mode) gave, in conjunction with its 13C NMR
data, the molecular formula C16H18N2Cl2O7. Comparison of
the 1H and 13C NMR data of 7 with those of 5 and 6 indicated
that it was a 1,3-diacylchloramphenicol. It was evident that one
acetyl group at δH 2.18 (3H, s, COCH3)/δC 20.8 (COCH3)
and 169.4 (COCH3) was present in the molecule. The
presence of one propanoyl group at δH 1.15 (3H, t, J = 7.8
Hz, COCH2CH3) and 2.37 (2H, q, J = 7.8 Hz, COCH2CH3)/
δC 173.6 (COCH2CH3), 27.3 (COCH2CH3), and 9.0
(COCH2CH3) was also established by 2D NMR COSY,
HSQC, and HMBC correlations (Figure 2). Since H-1 at δ 6.06
compound 8. The H NMR data of 9 were more similar to
those of 1,3-dipropanoylchloramphenicol (6), and careful
1
analysis of the H NMR data concluded that a propanoyl and
a butanoyl group were present in the molecule. However, the
minute quantity of 9 obtained from the isolation prevented
acquisition of reasonable 13C NMR and 2D NMR spectra to
determine the linkage positions of the two acyl groups to the
Cm skeleton. Two regioisomers, 3-butanoyl-1-propanoyl-
chloramphenicol (11) and 1-butanoyl-3-propanoyl-
chloramphenicol (12), were thus synthesized for structure
elucidation (see Scheme 1 and the Experimental Section for
details). Comparison of the 1H NMR spectrum of 9 with those
of the two regioisomers indicated that it should possess the
1
same structure as 11. The differences in the H NMR spectra
between 11 and 12 were in the range of δ 2.3−2.5 due to the
two methylene protons near the carbonyl groups in the two
1
acyl side chains (see their H NMR spectra in the Supporting
Information). In addition, acquisition and analysis of the 13C
NMR and 2D NMR spectra of 11 allowed unequivocal
assignment of the 1H and 13C NMR resonances for this
compound, including the two acyl carbonyl carbons with close
chemical shifts at δ 173.1 and 173.2 assigned to C-1‴ and C-1″,
respectively (Table 1, Figure 2).
The available purified natural and synthetic Cm derivatives as
reference compounds for LC-MS analysis allowed unequivocal
chromatographic assignments of compounds 1−9 in the seven
metagenomic clones (Figure 1). Monoacylated compounds 2−
4 were major metabolites present in all seven clones, while
diacylated compounds 5−9 were expressed from clones
P35B14 and P6B5 that showed similar metabolic profiles. It
was noted that compounds 6 and 8 coeluted at 12.40 min.
Compound 9 (as a minor metabolite) was present only in
clones P35B14, P6B5, and P5A4 and coeluted with an
unidentified compound at tR 13.19 min. The above analyses
suggest that all seven metagenomic clones contain highly active
acyltransferases that are able to acylate Cm. It should be
pointed out that intramolecular rearrangement of acyl groups
involved in the mono- and diacylated product generation may
occur during biosynthesis32 or chemical extraction and
isolation.33 However, the lack of corresponding rearranged
compounds for 1-acyl Cm derivatives 1−3 and 1,3-diacyl Cm
derivatives 7−9 in the LC-MS profiles suggests that all these
Figure 2. 2D NMR correlations of compounds 7−9.
(1H, d, J = 6.3 Hz) showed HMBC correlations with the
carbonyl carbon of the acetyl group as well as C-2 (δ 52.8), C-3
(δ 66.0), and the aromatic C-1′ (δ 143.3), the acetyl group
should be attached to the C-1 hydroxy group of the Cm
skeleton. The linkage of the propanoyl group to the C-3
hydroxy group was confirmed by the HMBC correlations of H-
3 at δ 4.20 (dd, J = 11.7, 5.5 Hz) and 4.06 (dd, J = 11.7, 6.3 Hz)
D
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Table 1. NMR Spectroscopic Data (400 MHz, CDCl3) for Compounds 7−9 and 11
a
7
8
9/11
position
δC, type
δH (J in Hz)
6.06, d (6.3)
δC, type
δH (J in Hz)
6.06, d (6.3)
δC, type
72.7
δH (J in Hz)
1
72.7, CH
52.8, CH
66.0, CH2
72.7
52.8
66.0
6.08, d (6.3)
4.62, m
2
3
4.62, m
4.61, m
53.0
66.1
4.20, dd (11.7, 5.5)
4.06, dd (11.7, 6.3)
4.20, dd (11.7, 5.5)
4.05, dd (11.7, 6.3)
4.20, dd (11.7, 4.7)
4.05, dd (11.7, 6.3)
1′
2′
3′
4′
5′
6′
143.3, C
143.3
127.6
124.0
148.0
124.0
127.6
143.6
127.7
124.1
148.1
124.1
127.7
127.6, CH
124.0, CH
148.1, C
7.52, d (8.6)
8.21, d (8.6)
7.52, d (8.6)
8.23, d (9.4)
7.53, d (8.6)
8.24, d (8.6)
124.0, CH
127.6, CH
8.21, d (8.6)
7.52, d (8.6)
6.76, d (9.4)
8.23, d (9.4)
7.52 (d, 8.6)
6.74 (d, 9.4)
8.24, d (8.6)
7.53, d (8.6)
6.78, d (9.4)
−NH
−CONH
−CHCl2
1″
2″
3″
1‴
2‴
3‴
4‴
164.1, C
62.0, CH
169.4, C
20.8, CH3
164.1
61.9
164.2
5.85, s
2.18, s
5.85, s
2.18, s
62.1
5.86, s
169.5
20.8
173.2, C
27.6, CH2
9.0, CH3
173.1, C
35.9, CH2
18.4, CH2
13.8, CH3
2.47, qd (7.8, 3.1)
1.17, t (7.8)
173.6, C
27.3, CH2
9.0, CH3
173.3, C
2.37, q (7.8)
1.15, t (7.8)
35.8, CH2
18.3, CH2
13.6, CH3
2.33, t (7.8)
1.65, m
2.34, t (7.1)
1.65, m
0.96, t (7.8)
0.96, t (7.1)
a
Both have the same structure, with 9 an isolated natural product and 11 a synthetic compound.
only marginal growth inhibition of 11%, whereas compounds
17 and 18 (see below) were observed to inhibit S. aureus
EAMC30 growth by 35% and 59%, respectively. Interestingly,
compounds 2, 4, and 10 showed activities in an initial testing
against Mycobacterium intracellulare comparable to Cm.
Considering that the three active compounds were 3-acyl- or
1,3-diacylchloramphenicols, additional 1-acetylchlorampheni-
cols were prepared for SAR studies. Thus, protection of the
C-3 primary hydroxy group of 1 with TBSCl using imidazole as
a base in dichloromethane gave compound 13. Acylation of 13
with acyl chloride using pyridine as a base in dichloromethane
afforded compound 14, which was deprotected with p-
toluenesulfonic acid in H2O−THF (1:4) to furnish the final
products 1-acetylchloramphenicol (15), 1-propanoyl-
chloramphenicol (16), 1-cinnamoylchloramphenicol (17), and
1-p-nitrobenzoylchloramphenicol (18) (Scheme 2; see detailed
synthetic procedures in the Supporting Information).
Scheme 1. Synthesis of Regioisomers 11 and 12
Scheme 2. Synthesis of 1-Acylchloramphenicols 15−18
Cm derivatives are natural products biosynthesized by the
clones.
Acylation of Cm is well known by E. coli34 and Streptomyces
spp.16,28 and has previously demonstrated to be associated with
drug resistance. This is consistent with biological testing in
which the isolated acylchloramphenicols (2−7) and the three
synthetic analogues (10−12) were inactive at 200 μg/mL
against MRSA (strain ATCC 1708), E. coli, and Pseudomonas
aeruginosa. In contrast, the S. aureus EAMC30 strain used in the
primary screening was observed to have some susceptibility at a
concentration of 62.5 μg/mL to purified acylchloramphenicols,
exhibiting growth inhibition of 15%, 72%, 36%, and 40% for
compounds 2, 3, 4, and 6, respectively. At this same
concentration the synthetic analogue compound 10 showed
E
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 3. Growth inhibition of E. coli expressing trfA-containing clones at high Cm concentrations. The trfA gene was cloned into pSmartBAC-S
vector and transformed into E. coli strain DH10B (dashed line) and compared to the pSmartBAC-S empty vector (solid line). The growth of each
culture in the presence of Cm concentrations ranging from 12.5 to 200 μg/mL was determined by the optical density at 600 nm and expressed as the
percent growth relative to growth at the standard 12.5 μg/mL Cm added to the E. coli pSmartBAC-S culture for plasmid maintenance. Cultures in
which there was a statstically significant increase (P < 0.05) in the percent inhibition of the culture containing trfA relative to the empty vector at the
same Cm concentration are marked with an asterisk.
Compounds 2−7, 10−12, and 15−18 were tested for in vitro
anti-Mycobacterium activity against M. intracellulare and the
tuberculosis-causing strain M. tuberculosis in comparison with
the positive controls Cm (1), ciprofloxacin, streptomycin, amd
rifampin. The data are shown in Table S2, indicating that these
compounds possessed moderate or marginal activity compared
to ciprofloxacin and streptomycin. However, compound 18,
with an aromatic-containing acyl substituent, was more active
than Cm (1), giving minimum inhibitory concentrations
(MICs) of 12.5 and 50.0 μg/mL against M. intracellulare and
M. tuberculosis, respectively, versus 25.0 and >100 μg/mL,
respectively, for Cm (1). The limited SAR information
indicated that monoacyl substitutions on either C-1 or C-3
hydroxy groups of the Cm skeleton did not affect the activity
against M. intracellulare, as evident by the equivalent potency of
2 and 15 (MIC, 25.0 μg/mL).
Characterization of the Gene Responsible for Anti-
microbial Activity in Clones P6B5 and P35B14. Subclones
of P6B5 and P35B14 with anti-MRSA activity were found to
contain the trfA gene sequence derived from the BAC-
optimized E. coli host strain (catalog no. 60215-1, Lucigen
Corp.). The TrfA replication protein together with the oriV
replication origin is considered the minireplicon for RK2
plasmids.35 When oriV is present on a plasmid and trfA is under
control of an arabinose-inducible promoter, the addition of
arabinose induces BAC copy number to approximately 50
copies per cell.26 To determine the effect of copy induction on
antibacterial activity, the trfA gene from clone P6B5 was PCR
amplified and ligated into pSmartBAC-S, and the resulting
plasmid pSmartBAC-S:trfA was transformed into E. coli strain
DH10B. We observed that E. coli expressing pSmartBAC-S:trfA
was inhibited in its growth at the highest concentrations (>30
μg/mL) of Cm as compared to the empty vector control
(Figure 3). This finding suggests that Cm derivatives produced
in E. coli clones expressing trfA attenuated E. coli growth by an
as-yet-undescribed mechanism.
These data therefore indicate that trfA was the single gene
necessary for antibacterial activity against MRSA in these two
metagenomic clones. TrfA is known to be responsible for
replication initiation and copy number control for RK2
plasmids;36 therefore, having trfA present on the BAC vector
results in greater levels of BAC copy induction, and this is
hypothesized to impact the generation of Cm derivatives
presumably through a gene dosage mechanism in which the
plasmid-borne chloramphenicol acetyltransferase is expressed at
higher levels. TrfA contains a single-stranded DNA binding
domain that activates the origin of vegetative replication in
diverse bacterial species.37 The increased susceptibility of E. coli
to Cm in the presence of the cloned trfA suggests that TrfA
levels are directly proportional to the formation of Cm
derivatives. Further research will be required to understand
the mechanism(s) by which increased TrfA levels result in Cm
derivative formation.
Characterization of the Esterase Expressed by Clone
P6L4 That Regenerates Cm. Enzymatic reactivation of Cm
by chloramphenicol acetate esterases that counteract the
chloramphenicol acetyl transferase (CAT) activity has been
reported previously.38−41 A similar mechanism is probably
responsible for the activity of P6L4 since the inset sequence
had multiple ORFs with esterases as the predicted gene product
(Supplemental Data 1). Three different esterases were
predicted, including an esterase, a carboxylesterase, and a
metallophosphoesterase (ORFs 12, 35, and 77, respectively;
GenBank accession MF620030.1). These predicted esterases
had low homology with each other, with the esterase amino
acid sequence showing 30% identity to the carboxylesterase and
38% to the metallophosphoesterase, whereas the carboxylester-
ase and metallophosphoesterase amino acid sequences had 31%
F
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
identity. A multiple alignment of all three esterases and their
nearest GenBank matches did not reveal any conserved amino
acid positions, whereas separate multiple alignments of each
esterase and its respective homologous amino acid sequences
obtained from GenBank did reveal conserved positions
(Figures S3−S5, see Supporting Information). A phylogenetic
analysis of each of these predicted esterases with a database of
their nearest respective GenBank matches revealed that the
predicted esterase and metallophosphoesterase were in clades
well supported by bootstrapping and were affiliated with
esterases originating from members of the phylum Chloroflexi,
whereas there was poor bootstrap support for the clade that
contained the predicted carboxylesterase (Figures S6−S8, see
Supporting Information).
predicted for each respective esterase as compared to that of
the uninduced sample.
The role of the esterases in the activity of clone P6L4 was
evaluated by testing the subclones in a bioassay against the
MRSA strain EAMC30. Transformation of the subclones (or
the empty vector control) with pGNS-BAC enabled growth in
LB broth containing Cm. Inhibition of MRSA growth was
observed for the esterase subclone but not for the
carboxylesterase. The inhibition of MRSA growth was much
higher when rhamnose was added to induce expression (Figure
5), as expected. These results suggest that the putative esterase
The esterase activity expressed by clone P6L4 was
determined by adding a fluorescent BODIPY FL Cm substrate
(BCAM) to the cultures. TLC results (Figure 4) indicated that
Figure 5. Comparison of anti-MRSA activity of the P6L4 subclones
pRham-esterase (dark bars) and pRham-carboxylesterase (light bars)
in the presence and absence of rhamnose-induced expression. The
graph represents the % growth inhibition of MRSA strain EAMC 30
by the subclones relative to the empty vector negative control,
considered to have no inhibitory effect and calculated by measuring
the fluorescence of reduced resazurin.
Figure 4. Comparison of negative control and clone P6L4 culture
extract with BCAM as a substrate. Aliquots withdrawn every 12 h were
processed for extraction with ethyl acetate and analyzed by TLC.
Lanes were normalized for direct comparison of the negative control
with P6L4.
(clone P6L4, ORF 12) is at least partially responsible for the
restoration of nonacetylated Cm responsible for the anti-MRSA
activity observed in clone P6L4 supernatants. This esterase
activity counteracts the mechanism encoded by the cat gene,
thereby resulting in active Cm that was capable of inhibiting
MRSA growth. This result is similar to that obtained from a
functional metagenomic screen that resulted in the isolation of
an esterase that reactivated chloramphenicol.43
the concentration of BCAM in the negative control decreased
(T0−T48) and that of the acetylated forms increased over time.
Two different acetylated forms of BCAM (1-acetyl Cm and 3-
acetyl Cm) were observed in the negative control, whereas only
1-acetyl Cm was seen in the presence of clone P6L4.42
A
reverse trend was seen in clone P6L4, wherein the BCAM
concentration increased and the acetylated form decreased over
time, suggesting reactivation of Cm by esterase activity. The
genes encoding three different predicted esterases in P6L4 were
subcloned to investigate this and ascertain the role of these
enzymes in Cm reactivation.
The carboxylesterase- and esterase-encoding genes, respec-
tively, were successfully amplified using gene-specific primer
sets, while the putative metallophosphoesterase gene failed to
amplify under these same conditions despite multiple attempts.
Further validation of the successful cloning and expression of
the esterases was seen from the results of SDS-PAGE analysis
(Figure S9). For the esterase and the carboxylesterase the use
of rhamnose-inducible expression showed a greater concen-
tration of the protein at an approximate molecular mass
In conclusion, chemical analysis of the metabolites expressed
from metagenomic clones with anti-MRSA activities identified
multiple Cm modifications. The supernatants of E. coli hosting
these clones contained multiple Cm derivatives, three of which
are novel Cm derivatives. The mechanism by which these Cm
derivatives are synthesized is believed to be by enzymatic
acylation of the Cm skeleton. This is supported by the data
obtained from these metagenomic clones, indicating that
expression of metagenomic-encoded enzymes (i.e., esterases)
or a cloned gene that induced BAC vector copy number (i.e.,
trfA) resulted in modification of the exogenously supplied Cm
to produce these novel Cm derivatives. This methodology
could be applied to other natural product scaffolds to generate
novel derivatives of other chemical entities with biological
activity.
G
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
BAC clones were determined to find conditions that contained the
desired insert size. Once the suitability of the trial ligation reaction was
confirmed, large-scale ligations and transformations were conducted to
achieve 19 200 clones for the BAC library (50 × 384-well plates
arrayed).
The progress made in these studies toward generation of
large-insert metagenomic libraries in shuttle BAC vectors will
be applied in the future for generation of larger-scale libraries
that can encompass a greater diversity of soil microbial
metagenomes and be expressed in multiple heterologous
hosts. In particular, the use of other heterologous hosts
engineered for expression of polyketide synthases as well as
other biosynthetic gene clusters will take further advantage of
these libraries for natural product discovery.
Construction of Shuttle BAC Cloning Vector pSmartBAC-S.
Construction of a BAC vector that would enable conjugal transfer into
other bacterial hosts and integration into the host chromosome
followed a method to construct a similar shuttle BAC vector
(pMDB14) reported by Martinez and colleagues.46 We first restriction
digested pSmartBAC (EU101022.1) with ApaI, end-repaired the DNA
to make it blunt, and dephosphorylated the digested pSmartBAC
vector ends. A synthesized and blunt-ended gene cassette that
contained the ϕC31 integrase (int), attachment site (attP), and
aac3(IV) that encodes for apramycin resistance (4793 bp) was ligated
to the linearized pSmartBAC-S, and apramycin-resistant transformants
in E. coli 10G were selected. The other phenotype (Cm resistance) and
the ability of the pSmartBAC-S to integrate into the ϕC31 attachment
site of Streptomyces coelicolor were also confirmed (data not shown).
The primary difference between pSmartBAC-S and pMDB14 is that
pSmartBAC-S contains the inducible high-copy replication origin oriV.
Plasmid pSmartBAC-S has been deposited at Addgene (ID# 107268),
and this and all other plasmids used in this study are listed in Table S3.
Metagenomic Library Screening for Antimicrobial Activity.
For high-throughput screening of the library, an in situ lysis method
was developed. Using a pin replicator, the BAC library containing E.
coli cells was inoculated into deep 96-well plates containing 1.0 mL of
LB supplemented with 12.5 μg/mL Cm for selection of the vector plus
a final concentration of 0.01% arabinose, followed by incubation at 37
°C for 48 h while shaking at 200 rpm. The E. coli cells were lysed by
freezing at −80 °C, followed by rapid thawing at 55 °C. Then 100 μL
of a 1:1000-diluted, log-phase MRSA strain EAMC30 culture in LB
containing 30 μg/mL nalidixic acid was added to each well followed by
incubation at 37 °C for 24 h. Finally, 165 μL of the viability indicator
resazurin (0.02% solution) was added to each well, and the plates were
incubated at 37 °C for 1 h until a color change from blue to pink was
observed for the majority of wells.47,48
Validation of Recombinant Clones with Antimicrobial
Activity. Each recombinant clone that inhibited the growth and/or
viability of the tester strain was regrown from the library that was
cryopreserved in 384-well plate format into a 2 mL LB broth culture
containing 12.5 μg/mL Cm, incubated overnight at 37 °C, and a
separate glycerol stock was stored at −80 °C. Each positive clone was
then inoculated into replicate wells (n = 4) of a 96-well plate and
grown as above to retest the clone for inhibitory activity. Every positive
clone that demonstrated reproducible antibiotic activity was tested for
its ability to inhibit MRSA growth by removal of E. coli cells by
centrifugation and transfer of supernatants to another microtiter plate.
Results for each positive clone were noted according to the degree and
consistency of inhibitory activity observed. In the resazurin-based
bioassays, fluorescence readings of reduced resazurin (resorufin) were
recorded (530 nm excitation and 590 nm emission) and used for
calculating the % inhibition of MRSA growth relative to the empty
vector negative control. These results were used to narrow down the
list of positive clones to the top candidates for additional genetic and
biochemical characterization.
EXPERIMENTAL SECTION
■
General Experimental Procedures. Specific rotations were
measured on an Autopol IV polarimeter at room temperature. 1H
and 13C NMR spectra were recorded on a Bruker DRX NMR
spectrometer operating at 400 (1H) and 100 (13C) MHz. Chemical
shifts are expressed in ppm relative to the solvent residue signals. 2D
NMR spectra (gradient DQF-COSY, HMQC, gradient HMBC, and
NOESY) were run using standard Bruker pulse programs. High-
resolution ESIMS data were obtained on an Agilent Series 1100 SL
mass spectrometer. Column chromatography (CC) was performed on
silica gel (40 μm, J. T. Baker) and reversed-phase silica gel (C18, 40
μm, J. T. Baker). Silica gel 60 F254 TLC plates (Merck, Darmstadt,
Germany) and reversed-phase TLC plates (C18, Merck, Darmstadt,
Germany) were used for analytical TLC. LC-MS was run on an
Agilent 1290 Infinity series chromatograph with a dual pump,
autosampler, thermostated column compartment, and a diode array
detector. The chromatograph was coupled with an Agilent 6120 single
quadruple mass spectrometer with a dual APCI/ESI source operated
in both the positive and negative modes. The system was controlled by
ChemStation software. A Waters Acquity UPLC BEH C18 column (2.1
× 150 mm, 1.7 μm) was used. The experiments were carried out at a
gradient elution from 10% to 100% CH3CN in H2O in 20 min and
then held for 6 min. The quadrupole mass analyzer was operated in
the scan mode with the mass range from 100 to 1500. The drying gas
flow was 10 L/min at 300 °C, the nebulizer pressure was 35 psi, and
the vaporizer temperature was 200 °C. The capillary voltage used was
3 kV, the corona current was 10 μA, and the charging voltage was 4
kV. The fragmentor was set to 120.
Bacterial Strains and Plasmids. Metagenomic libraries were
transformed into the BAC-optimized E. coli strain DH10B BAC
Replicator V2.0 (Lucigen Corp., Middleton, WI, USA) that contains
the arabinose-inducible PBAD:trfA in the chromosome to support copy
induction of BAC vectors. A clinical isolate from the East Alabama
Medical Center (EAMC) designated as MRSA strain EAMC30 was
used as a tester strain for antimicrobial activity screens.
Metagenomic Library Construction. High molecular weight
(HMW) metagenomic DNA was isolated from a Cullars Rotation
(Auburn, AL, USA) soil plot that had not been amended with
fertilizers for the past 100 years.44 The isolation and purification of soil
HMW DNA was conducted by isolating soil microorganisms that were
embedded in low melting point agarose, treated with proteinase K, and
washed extensively as in a previously published protocol.45 The
agarose was melted and the DNA was sheared by pipetting up to five
times to generate DNA in the >150 Kb range based on pulsed field gel
electrophoresis. The agar was allowed to solidify again, and the DNA
was end-repaired with the DNATerminator kit (Lucigen) in a total
volume of 500 μL with 10 μL of enzymes and then heat killed at 70 °C
for 15 min. The end-repaired DNA was ligated with BstXI adaptors
(10 μL of 100 μM each) in a total volume of 700 μL consisting of
ligation reaction containing 10 μL of ligase (2 U/μL, Epicenter),
followed by gel fractionation of large DNA fragments ranging from
100 to 200 Kb purified by pulse-field gel electrophoresis. Purified large
DNA fragments (about 100 μL, 1−3 ng/μL) were ligated into the
cloning-ready BstXI shuttle vector pSmartBAC-S (16 °C for ∼18 h).
The ligated DNA mixture was electroporated into competent E. coli
cells (BAC-Optimized E. coli 10G Replicator Cells, Lucigen). Small-
scale ligations and transformations (1 μL of DNA per 20 μL of cells)
were used to judge the cloning efficiency. The insert sizes of about 50
Retransformation of Antibiotic-Producing Clones. The
validated antibiotic-producing clones were grown in 3 mL of LB
containing Cm, and after 16 h at 37 °C the BAC DNA was extracted
using an alkaline lysis method. A restriction digestion of each BAC
clone with BamHI or EcoRI was resolved by PFGE at 6 V/cm, 1 to 15
s switch time for 12 h at 15 °C. The insert size for each BAC clone was
estimated by comparison with DNA size standards. BAC DNA was
̈
retransformed into a naive E. coli strain and selected on LB containing
12.5 μg/mL Cm for the presence of the plasmid. Transformation was
done by electroporation (1 mm gap cuvette, 1.8 kV, 600 ohms, 10 μF)
into commercially available electrocompetent BAC replicator V2.0
cells (Lucigen). Each retransformed clone was tested as described
above for antibiotic activity. Only recombinant clones showing
H
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
evidence of a metagenomic insert and consistent and retransformable
antibiotic activity were selected for further studies.
identified using a GeneMark heuristic model for prediction of
prokaryotic genes.49 The ORF sequences were compared against the
GenBank nr/nt database using BLASTx to predict the function of
putative gene products. Sequences for each BAC clone were deposited
in GenBank under accession numbers MF620010 to MF620035 and
MF797910 for clone P35B14.
Testing for the Effect of Arabinose Induction. The
antimicrobial expressing clones were evaluated for their antibiotic
activity when grown in the presence and absence of arabinose
induction. These clones were grown in LB containing 12.5 μg/mL Cm
with and without a final concentration of 0.01% arabinose, and the
antibacterial assay was conducted as described above. The fluorescent
intensities were recorded, and the data were analyzed to determine
whether the addition of arabinose to the growth medium affected
antibiotic activity.
Subcloning ORFs Responsible for Antibiosis Activity. Clones
that produced a Cm derivative were characterized by subcloning to
identify the ORF(s) responsible for producing the anti-MRSA
compound. BAC DNA was isolated from 500 mL cultures of each
clone according to methods described previously.48 Approximately 20
μg of BAC DNA from each clone was sheared using a g-TUBE
(Covaris, MA, USA) to obtain fragmented DNA within the range of 4
to 8 kb. Fragmented DNA was separated using gel electrophoresis, and
DNA of the targeted size was excised for purification. The ends of the
fragmented DNA obtained by shearing were repaired using a DNA
Terminator kit (Lucigen). End-repaired DNA from each clone was
ligated into the pSMART vector (Lucigen) and then electroporated
into electrocompetent E. coli 10G cells containing pGNS-BAC to
provide Cm resistance.49 Subclones were selected on LB agar
supplemented with 200 μg/mL ampicillin and 12.5 μg/mL Cm. The
transformants were picked robotically using a QPix2 (Molecular
Devices, Sunnyvale, CA, USA) in 96-well plates and were grown
overnight with shaking at 200 rpm at 37 °C. All the cultures grown in
shallow 96-well plates were transferred to deep-well plates, and their
anti-MRSA activity was determined according to the method described
above. Subclones with anti-MRSA activity were grown for plasmid
extraction, and the ORF(s) contained within active subclones were
identified by primer walking PCR (Table S4) followed by sequencing.
The size of the inset within positive subclones was determined by PCR
using vector-specific primers.
Clone P6L4 ORFs Responsible for Antibacterial Activity. To
assess the acylation of Cm, 2 mL broth cultures (containing Cm and
arabinose) of clone P6L4 and an empty vector negative control were
grown at 37 °C for 12 h at 200 rpm. A portion (600 μL) of the culture
was transferred to fresh medium (6 mL of LB broth containing 0.01%
arabinose), and 30 μL of BODIPY FL Cm (FAST CAT kit,
ThermoFisher Scientific, Waltham, MA, USA) was added to each
tube. Tubes were covered with aluminum foil and incubated in the
dark at 37 °C for 48 h at 200 rpm. A 1 mL aliquot was withdrawn
every 12 h (T0 until T48) and extracted with an equal volume of ethyl
acetate. The organic layer was transferred and stored in brown
microcentrifuge tubes at −20 °C. At the end of 48 h all of the extracts
were dried in the Centrivap and resuspended in 15 μL of ethyl acetate,
and 5 μL of each sample was spotted on a silica gel TLC plate. A
chloroform and methanol mixture (87:13) was used as the mobile
phase, and the plate was visualized under UV light.
Amplification, cloning, and expression of each putative esterase gene
from clone P6L4 were conducted using the Expresso Rhamnose
SUMO cloning and expression system (Lucigen). Custom primers
were designed for amplification of the respective gene sequence
encoding each of the three different predicted esterases (Table S4).
The thermal cycler was preheated to 94 °C, and the reactions were
incubated at 94 °C for 2 min. A total of 25 cycles were performed with
denaturation at 94 °C for 15 s, annealing at 55 °C for 15 s, and
extension at 72 °C for 1 min, with a final extension step at 72 °C for
10 min. A 10 μL amount of the PCR product was loaded onto an
agarose gel (SB gel run for 2 h at 165 V) for analysis. The PCR
amplicon was cloned into the pRham vector using E. coli DH10B
chemically competent cells (Lucigen) as per the kit protocol. A 100 μL
amount of the transformed cells was plated on YT agar plates
containing 30 μg/mL kanamycin (Kan), and plates were incubated
overnight at 37 °C. Transformants from each plate were grown in YT
broth with Kan at 37 °C for 16 h at 200 rpm followed by extraction of
plasmid DNA. PCR (design and conditions as before) was used to
confirm the presence of the cloned insert by using two primer sets, the
esterase-specific primers and the SUMO forward and pETite reverse
primers specific to the vector. The PCR product was analyzed as
before by agarose gel electrophoresis and purified by the Wizard PCR
Clean-up System (Promega, Madison, WI, USA). The amplified and
Preliminary Characterization of Active Clones. A total of 27
clones that were reproducibly found to express antimicrobial activity
̈
when their BAC clone DNA was used to retransform a naive E. coli
were selected for further testing to determine if the active
compound(s) was extra- or intracellular. For each clone, 2 mL of
LB broth containing 12.5 μg/mL Cm and 0.01% arabinose was
inoculated with an E. coli glycerol stock stored at −80 °C. The culture
tubes were incubated at 37 °C for 48 h while shaking at 200 rpm.
Cultures were then divided into two sets (1 mL each), with one set
subjected to a freeze−thaw process (freezing at −80 °C followed by
rapid thawing at 55 °C) and the other set subjected to centrifugation
at 10000g, and then the supernatant was passed through a 0.2 μm filter
to obtain cell-free supernatants. These samples were then tested in 96-
well microtiter plates with three replicates (200 μL in each well) for
each clone and treatment. Appropriate negative controls (empty
vector with no insert) were used in the bioassay, and 20 μL of diluted
log phase MRSA strain EAMC30 (approximately 5.0 × 107 CFU) was
added to each well. The addition of 30 μg/mL nalidixic acid was used
to inhibit any residual E. coli cells in the cell lysates from the set
subjected to the freeze and thaw treatments. Plates were incubated at
37 °C for 24 h while shaking at 200 rpm. After incubation, 30 μL of a
0.02% (w/v) resazurin solution was then added to each well, and
plates were further incubated at 37 °C for 4−5 h while shaking at 200
rpm. Fluorescence readings were recorded using a microtiter plate
reader (excitation at 530 nm and emission at 590 nm), and the percent
reduction of resazurin fluorescence for each clone was determined by
comparison with the respective negative controls.
To evaluate the heat stability and molecular weight of the
heterologously expressed antibiotic activity, cultures were inoculated
for the same set of 27 clones and the negative controls as before. At
the end of 48 h, the cultures were processed using either the freeze−
thaw or cell-free supernatant method. For two of the clones (P6B5 and
P37O10) the freeze−thaw treatment was used, and all of the
remaining cell-free supernatants were divided into three sets to test
the heat stability and estimate the molecular weight of the active
fractions. Antibiotic activity due to a heat-labile and/or proteinaceous
product was tested by boiling the cell-free supernatant or the lysate for
10 min. Tubes were cooled to room temperature, and a bioassay was
set up as described above. To determine if the activity was due to a
compound less than 3 kDa in size, 1 mL of the sample was fractionated
using a centrifugal filter (VWR) with a modified poly(ether sulfone)
membrane with 3 kDa MWCO. The spin time was 15 min at 14000g,
and the concentrate (reconstituted with half-strength LB broth) and
the filtrate were tested in the bioassay format as described before.
DNA Sequence Generation and Analysis. Each of the 27
bioactive clones was selected for complete insert sequencing using
either 454 pyrosequencing (454 Life Sciences, Branford, CT, USA) or
an Ion Torrent PGM (Life Technologies, Grand Island, NY, USA).
BAC DNA was isolated from 100 mL cultures of each respective clone
using an alkaline lysis method.48 The purified DNA was used to
generate bar-coded shotgun subclone libraries that were sequenced at
Lucigen (in the case of Ion Torrent PGM) or at EnGenCore at the
Univeristy of South Carolina (Columbia, SC, USA) using a Genome
Sequencer FLX system as per the manufacturer’s instructions. The
sequences were trimmed for quality using a Q score of >30 and
assembled into contiguous fragments (contigs) using the CLC
genomics workbench (Cambridge, MA, USA) de novo assembler.
The contig(s) that represented the complete (or nearly complete)
clone insert DNA was exported in FASTA format. ORFs were
I
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
purified DNA was sequenced and compared with the original sequence
using the CLC Genomics Workbench and also analyzed by BLASTx
comparison with the GenBank nr/nt database. A standard induction
protocol was used for protein expression followed by SDS-PAGE
analysis.50
The esterase subclones were selected for growth on LB agar
containing 50 μg/mL kanamycin, but in order to evaluate their
respective ability to modify Cm, it was necessary to introduce an
additional vector that confers Cm resistance. E. coli strains containing
the esterase subclones were made competent and transformed by
electroporation (using conditions as described previously) with the
pGNS-BAC vector DNA.49 An aliquot (100 μL) of the transformed
cells was plated on YT agar plates containing 12.5 μg/mL Cm, and
plates were incubated overnight at 37 °C. An appropriate negative
control was also conducted by using an E. coli strain containing the
empty pRham vector and processing it similarly for electroporation
with the pGNS-BAC vector.
1,3-Dipropanoylchloramphenicol (6): colorless liquid; [α]25D −6.3
(c 0.7, DCM); 1H and 13C NMR data matching reported values in the
literature;29 HRESIMS m/z 433.0567/435.0543 [M − H]− (calcd for
[C17H20N2Cl2O7 − H]−, 433.0575/435.0545).
1-Acetyl-3-propanoylchloramphenicol (7): colorless liquid; [α]25
D
1
−4.9 (c 0.1, EtOH); H and 13C NMR data, see Table 1; HRESIMS
m/z 419.0413/421.0389 [M − H]− (calcd for [C16H18N2Cl2O7
−
H]−, 419.0418/421.0389).
1-Acetyl-3-butanoylchloramphenicol (8): colorless liquid; [α]25
D
−8.9 (c 0.1, EtOH); H and 13C NMR data, see Table 1; HRESIMS
1
m/z 433.0570/435.0541 [M − H]− (calcd for [C17H20N2Cl2O7
−
H]−, 433.0575/435.0545).
1-Propanoyl-3-butanoylchloramphenicol (9): colorless liquid;
1
[α]25 −10.9 (c 0.5, EtOH); H NMR data, see Table 1; HRESIMS
D
m/z 447.0735/449.0714 [M − H]− (calcd for [C18H22N2Cl2O7
−
H]−, 447.0731/449.0702).
Synthesis of 3-Acylchloramphenicols 3 and 4 and 1,3-
Diacylchloramphenicols 6 and 10. To a solution of 1 (100 mg) in
dry DCM (5 mL) were added pyridine (0.03 mL, 1.1 equiv) and
dimethylaminopyridine (DMAP) (3.7 mg, 0.1 equiv) with a 2 min
interval. After stirring for 5 min, acyl chloride (butanoyl chloride or
propionic anhydride, 1.1 equiv) was added. The reaction was
monitored by TLC and stopped in 5 min. Removal of solvents from
the reaction mixture by evaporation afforded a residue, which was
chromatographed on silica gel using 30% hexanes in EtOAc to yield
monoacylated Cms 3 and 4 (85%) and 1,3-diacylchloramphenicols 6
and 10 (15%). The 1H NMR data of synthetic 3, 4, and 6 were
identical to those of the isolated compounds.
Each of the transformants was inoculated into 5 mL of LB broth
containing 12.5 μg/mL Cm, 30 μg/mL Kan, and 0.2% rhamnose, and
additional inoculations were made into 5 mL of LB broth containing
12.5 μg/mL Cm and 30 μg/mL Kan without any added rhamnose.
Cultures were incubated at 37 °C for 48 h at 200 rpm, and cell-free
supernatants were collected followed by testing against MRSA strain
EAMC30 using the 96-well microtiter plate bioassay as described
above. Results for induced expression with and without the addition of
rhamnose were compared for the inhibition of MRSA growth.
Fermentation of Metagenomic Clones. Fermentation was
conducted using LB broth, 0.01% arabinose, and 12.5 μg/mL of Cm
at 37 °C for 48 h with shaking at 200 rpm. LB broth was prepared
using 10 g of tryptone, 5 g of yeast extract, and 10 g of NaCl. For each
of the E. coli BAC clones expressing antimicrobial activity, 200 mL of
culture broth was placed in a 1 L flask with shaking at 250 rpm. After
48 h, the supernatant was separated by centrifugation for 15 min and
extracted with EtOAc. A one-liter culture was prepared for clones
P35B14, P22E10, P28I7, P37I0, P27M10, P6B5, and P5A4, and the
resultant EtOAc extracts were used for LC-MS analysis. For scale-up of
fermentation of clone P35B14, a total of 40 L of culture was prepared
to yield 1.614 g of EtOAc extract for follow-up isolation work.
Isolation of Chloramphenicol Derivatives from Clone
P35B14. The EtOAc (1.35 g) extract was adsorbed over 4.0 g of
normal-phase silica gel and loaded on a column using 120 g of normal-
phase silica gel. The column was eluted with a gradient solvent system
consisting of acetone in hexanes (0−100%) to afford 15 combined
fractions (F1−F15). Fraction F3 (135 mg) was chromatographed on
reversed-phase silica gel C18 using 75% MeOH in H2O to obtain
subfractions A1−A7. Fractions A3 (8.5 mg) and A4 (3.5 mg) were
combined and purified by normal-phase silica gel column chromatog-
raphy using 30% acetone in hexanes to afford fractions B1−B12.
Fractions B3, B5, and B7 were identified as pure compounds 9 (0.9
mg), 6 (2.5 mg), and 8 (2.5 mg), respectively. Purification of fraction
F4 (184.2 mg) on reversed-phase silica gel C18 with 75% MeOH in
H2O yielded 64 fractions. Similar fractions were combined based on
TLC to give compounds 1 (20.0 mg), 2 (20.0 mg), 3 (8.2 mg), 4 (6.0
mg), 5 (1.2 mg), and 7 (1.1 mg).
1,3-Dibutanoylchloramphenicol (10): colorless liquid; [α]25D −6.6
1
(c 0.58, EtOH); H NMR (CDCl3, 400 MHz) δ 8.23 (2H, d, J = 9.4
Hz), 7.53 (2H, d, J = 8.6 Hz), 6.86 (1H, d, J = 9.4 Hz, NH), 6.08 (1H,
d, J = 5.5 Hz, H-1), 5.89 (1H, s, CHCl2), 4.64 (1H, m, H-2), 4.20 (1H,
dd, J = 11.7, 5.5 Hz, H-3a), 4.06 (1H, dd, J = 11.7, 6.3 Hz, H-3b), 2.42
(2H, t, J = 7.8 Hz), 2.34 (2H, t, J = 7.8 Hz), 1.65 (4H, m), 0.95 (6H,
dt, J = 3.1, 7.1 Hz); 13C NMR (CDCl3, 100 MHz) δ 173.2, 172.3,
164.2, 148.1, 143.6, 127.6, 124.1, 72.6, 66.1, 62.0, 52.9, 35.9, 35.8, 18.4,
18.3, 13.7, 13.6; HRESIMS m/z 461.0902/463.0886 [M − H]− (calcd
for [C19H24N2Cl2O7 − H]−, 461.0888/463.0858); 485.0841/487.0812
[M + Na]+ (calcd for [C19H24N2Cl2O7 + Na]+, 485.0853/487.0823).
Synthesis of Regioisomers 11 and 12. To a solution of 3 or 4
(40 mg) in dry dichloromethane (DCM) (5 mL) were added pyridine
(0.015 mL, 1.1 equiv) and DMAP (0.1 equiv) sequentially with a 2
min interval. After stirring for 5 min, propanoyl chloride or butanoyl
chloride (1.1 equiv) was added. The reaction was monitored by TLC
and continued for 15 min. Removal of solvents from the reaction
mixture by evaporation afforded a residue, which was chromato-
graphed on silica gel using 30% hexanes in EtOAc to yield 11 or 12 in
a yield of approximately 95%.
3-Butanoyl-1-propanoylchloramphenicol (11): colorless liquid;
1
[α]25 −10.9 (c 0.5, EtOH); H NMR data are identical to those of
D
compound 9; 13C NMR data, see Table 1; HRESIMS m/z 447.0774/
449.0746 [M − H]− (calcd for [C18H22N2Cl2O7 − H]−, 447.0731/
449.0702).
1-Butanoyl-3-propanoylchloramphenicol (12): colorless liquid;
[α]25D −7.9 (c 0.5, EtOH); 1H NMR (CDCl3, 400 MHz) δ 8.23 (2H,
d, J = 8.6 Hz), 7.54 (2H, d, J = 8.6 Hz), 6.89 (1H, d, J = 9.4 Hz, NH),
6.09 (1H, d, J = 5.5 Hz, H-1), 5.89 (1H, s, CHCl2), 4.64 (1H, m, H-2),
4.20 (1H, dd, J = 11.7, 5.5 Hz, H-3a), 4.09 (1H, dd, J = 11.7, 6.3 Hz,
H-3b), 2.42 (2H, t, J = 7.1 Hz), 2.36 (2H, pentet, J = 7.1 Hz), 1.67
(2H, m), 1.15 (3H, t, J = 7.8 Hz), 0.96 (3H, m); 13C NMR (CDCl3,
100 MHz) δ 173.9, 172.3, 164.3, 148.0, 143.6, 127.6, 123.9, 72.6, 66.1,
62.1, 52.9, 35.9, 27.3 18.2, 13.6, 8.9; HRESIMS m/z 447.0813/
449.0787 (calcd for [C18H22N2Cl2 O7 − H]−, 447.0731/449.0702).
In Vitro Antibacterial Assay. Organisms were obtained from the
American Type Culture Collection (Manassas, VA, USA) and include
methicillin-resistant S. aureus ATCC 1708, E. coli ATCC 2452, P.
aeruginosa ATCC 2108, M. intracellulare ATCC 23068, and M.
tuberculosis ATCC 25177. All organisms were tested using modified
versions of the CLSI (formerly NCCLS) methods.51 Samples
(dissolved in DMSO) were serially diluted in 20% DMSO/saline
3-Acetylchloramphenicol (2): white solid; [α]25 +29.4 (c 0.5,
D
EtOH); 1H and 13C NMR data matching reported values in the
literature;29 HRESIMS m/z 363.0149/365.0119 [M − H]− (calcd for
[C13H14N2Cl2O6 − H]−, 363.0156/365.0127).
3-Propanoylchloramphenicol (3): colorless liquid; [α]25 +34.5 (c
D
1
0.5, EtOH); H and 13C NMR data matching reported values in the
literature;29 HRESIMS m/z 401.0276/403.0245 [M + Na]+ (calcd for
[C14H16N2Cl2O6 + Na]+, 401.0278/403.0248).
3-Butanoylchloramphenicol (4): colorless liquid; [α]25 +25.4 (c
D
0.45, EtOH); 1H and 13C NMR matching reported values in the
literature;29 HRESIMS m/z 415.0428/417.0396 [M + Na]+ (calcd for
[C15H18N2Cl2O6 + Na]+, 415.0434/417.0405).
1,3-Diacetylchloramphenicol (5): colorless liquid; [α]25 −9.4 (c
D
0.16, EtOH); 1H and 13C NMR matching reported values in the
literature;29 HRESIMS m/z 405.0259/407.0231 [M − H]− (calcd for
[C15H16N2Cl2O7 − H]−, 405.0262/407.0232).
J
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
and transferred (10 μL) in duplicate to 96-well flat-bottom
microplates. Inocula were prepared by correcting the OD630 of
microbe suspensions in 5% Alamar Blue (BioSource International,
Camarillo, CA, USA) in Middlebrook 7H9 broth with OADC
enrichment, pH = 7.0. Final sample test concentrations were 1/
100th the DMSO stock concentration. Mycobacterium cultures were
read at 544ex/590em using the Polarstar Galaxy plate reader (BMG
LabTechnologies, Germany) prior to and after incubation at 37 °C
and 10% CO2 for 70−74 h for M. intracellulare and 5 days for M.
tuberculosis. IC50 values (concentrations that afford 50% inhibition
relative to controls) were calculated using XLfit 4.2 software (IDBS,
Alameda, CA, USA) using fit model 201. The MIC is defined as the
lowest test concentration that allows no detectable bacterial growth.
Minimum bactericidal concentrations were determined by removing 5
μL from each clear (or blue) well, transferring to fresh media, and
incubating as previously mentioned. The MBC is defined as the lowest
test concentration that kills the organism (allows no growth).
(2) Fischbach, M. A.; Walsh, C. T. Science 2009, 325, 1089−1093.
(3) Archer, N. K.; Mazaitis, M. J.; Costerton, J. W.; Leid, J. G.;
Powers, M. E.; Shirtliff, M. E. Virulence 2011, 2, 445−59.
(4) Howden, B. P.; Davies, J. K.; Johnson, P. D.; Stinear, T. P.;
Grayson, M. L. Clin. Microbiol. Rev. 2010, 23, 99−139.
(5) Lodise, T. P.; Graves, J.; Evans, A.; Graffunder, E.; Helmecke, M.;
Lomaestro, B. M.; Stellrecht, K. Antimicrob. Agents Chemother. 2008,
52, 3315−3320.
(6) Soriano, A.; Marco, F.; Martinez, J. A.; Pisos, E.; Almela, M.;
Dimova, V. P.; Alamo, D.; Ortega, M.; Lopez, J.; Mensa, J. Clin. Infect.
Dis. 2008, 46, 193−200.
(7) Micek, S. T. Clin. Infect. Dis. 2007, 15, S184−S190.
(8) Liu, C.; Bayer, A.; Cosgrove, S. E.; Daum, R. S.; Fridkin, S. K.;
Gorwitz, R. J.; Kaplan, S. L.; Karchmer, A. W.; Levine, D. P.; Murray,
B. E.; M, J. R.; Talan, D. A.; Chambers, H. F. Clin. Infect. Dis. 2011, 52,
285−292.
(9) Tatarkiewicz, J.; Staniszewska, A.; Bujalska-Zadrony, M. Arch.
Med. Sci. 2016, 12, 1327−1336.
(10) Handelsman, J.; Rondon, M. R.; Brady, S. F.; Clardy, J.;
Goodman, R. M. Chem. Biol. 1998, 5, R245−R249.
(11) Iqbal, H. A.; Craig, J. W.; Brady, S. F. FEMS Microbiol. Lett.
2014, 354, 19−26.
(12) Hover, B. M.; Kim, S. H.; Katz, M.; Charlop-Powers, Z.; Owen,
J. G.; Ternei, M. A.; Maniko, J.; Estrela, A. B.; Molina, H.; Park, S.;
Perlin, D. S.; Brady, S. F. Nature Microbiol. 2018, 3, 1.
(13) Santana-Pereira, A. L. R.; Liles, M. R. In Functional
Metagenomics: Tools and Applications; Charles, T. C.; Liles, M. R.;
Sessitch, A., Eds.; Springer: Cham, Switzerland, 2017; pp 119−138.
(14) Rondon, M. R.; August, P. R.; Bettermann, A. D.; Brady, S. F.;
Grossman, T. H.; Liles, M. R.; Loiacono, K. A.; Lynch, B. A.; MacNeil,
I. A.; Minor, C.; Tiong, C. L.; Gilman, M.; Osburne, M. S.; Clardy, J.;
Handelsman, J.; Goodman, R. M. Appl. Environ. Microbiol. 2000, 66,
2541−2547.
(15) Felczykowska, A.; Dydecka, A.; Bohdanowicz, M. L. G.; Gasior,
T.; Sobon, M.; Kobos, J.; Bloch, S.; Nejman-Falenczyk, B.; Wegrzyn,
G. Microb. Cell Fact. 2014, 13, 105.
(16) Gillespie, D. E.; Brady, S. F.; Bettermann, A. D.; Cianciotto, N.
P.; Liles, M. R.; Rondon, M. R.; Clardy, J.; Goodman, R. M.;
Handelsman, J. Appl. Environ. Microbiol. 2002, 68, 4301−4306.
(17) Wang, X.; Liu, Q.; Wang, H.; Luo, C. X.; Wang, G.; Luo, M.
BMC Genomics 2013, 14, 1471−2164.
(18) Brady, S. F.; Chao, C. J.; Clardy, J. Appl. Environ. Microbiol.
2004, 70, 6865−6870.
(19) Banik, J. J.; Brady, S. F. Curr. Opin. Microbiol. 2010, 13, 603−
609.
(20) Rine, J.; Hansen, W.; Hardeman, E.; Davis, R. W. Proc. Natl.
Acad. Sci. U. S. A. 1983, 80, 6750−6754.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jnat-
prod.7b00903.
Tables indicating the anti-MRSA activity of metagenom-
ics clones, antibacterial activity of Cm derivatives, and the
plasmids and oligonucleotides used in this study;
synthetic procedures for compounds 15−18; NMR and
MS spectra of compounds 1−18; multiple alignments of
cloned and GenBank-obtained esterases, and phyloge-
netic analysis of each esterase (PDF)
Excel file (XLSX)
AUTHOR INFORMATION
Corresponding Authors
*(X.-C. Li) Tel: 662-915-6742. Fax: 662-915-7989. E-mail:
xcli7@olemiss.edu.
■
*(M. R. Liles) Tel: 334-844-1656. Fax: 334-844-1645. E-mail:
lilesma@auburn.edu.
ORCID
Mei Wang: 0000-0003-3686-4292
Mark R. Liles: 0000-0002-9313-8150
Author Contributions
#S. Nasrin and S. Ganji contributed to this work equally and
should be considered as cofirst authors.
(21) Strobel, R. J.; Sullivan, G. R. Manual of Industrial Microbiology
and Biotechnology, 2nd ed.; ASM Press: UK, 1999.
(22) Nystrom, T. Mol. Microbiol. 2004, 54, 855−862.
(23) Wild, J.; Hradecna, Z.; Szybalski, W. Genome Res. 2002, 12,
1434−1444.
Notes
The authors declare the following competing financial
interest(s): M.R.L. and D.A.M. are cofounders of the Varigen
Biosciences Corporation, which has financial interests in the
metagenomic library.
(24) Willison, J. C.; Magnin, J. P.; Vignais, P. M. Arch. Microbiol.
1987, 147, 134−142.
ACKNOWLEDGMENTS
(25) Schoner, T. A.; Fuchs, S. W.; Schonau, C.; Bode, H. B. Microb.
Biotechnol. 2014, 7, 232−241.
(26) Achenbach, H.; Kohl, W.; Reichenbach, H. Chem. Ber. 1976,
109, 2490−2502.
(27) Elkersh, T. A.; Plourde, J. R. J. Antibiot. 1976, 29, 292−302.
(28) Argoudel, A. D.; Coats, J. H. J. Antibiot. 1971, 24, 206−208.
(29) Gross, F.; Lewis, E. A.; Piraee, M.; van Pee, K. H.; Vining, L. C.;
White, R. L. Bioorg. Med. Chem. Lett. 2002, 12, 283−286.
(30) Zhang, W.; Huffman, J.; Li, S. Y.; Shen, Y. M.; Du, L. C. BMC
Biotechnol. 2017, 17, 59.
(31) Ottolina, G.; Carrea, G.; Riva, S. J. Org. Chem. 1990, 55, 2366−
2369.
(32) Nakagawa, Y.; Nitahara, Y.; Miyamura, S. Antimicrob. Agents
Chemother. 1979, 16, 719−723.
■
This work was supported by the National Institutes of Health
NIAID (#2R44AI085840-02 and #5R44AI085840-03) and the
USDA Agricultural Research Service Specific Cooperative
Agreement No. 58-6408-2-0009. The authors wish to thank
Ms. N. Capps and M. Wright for biological testing, Dr. B. Avula
for recording high-resolution ESIMS spectra, R. Ye for technical
help with library construction, and S. Jasinovica, M. Warner,
and A. Kerowicz for technical help with next-generation
sequencing.
REFERENCES
(1) Levy, S. B.; Marshall, B. Nat. Med. 2004, 10, S122−S119.
■
K
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(33) Kutrzeba, L. M.; Li, X.-C.; Ding, Y.; Ferreira, D.; Zjawiony, J. K.
J. Nat. Prod. 2010, 73, 707−708.
(34) Suzuki, Y.; Okamoto, S. J. Biol. Chem. 1967, 242, 4722−4730.
(35) Fang, F. C.; Helinski, D. R. J. Bacteriol. 1991, 173, 5861−5868.
(36) Perri, S.; Helinski, D. R.; Toukdarian, A. J. Biol. Chem. 1991,
266, 12536−12543.
(37) Wegrzyn, K.; Fuentes-Perez, M. E.; Bury, K.; Rajewska, M.;
Moreno-Herrero, F.; Konieczny, I. Nucleic Acids Res. 2014, 42, 7807−
7818.
(38) Nakano, H.; Tomita, F.; Yamaguchi, K.; Nagashima, M.; Suzuki,
T. Biotechnol. Bioeng. 1977, 19, 1009−18.
(39) Sohaskey, C. D.; Barbour, A. G. Antimicrob. Agents Chemother.
1999, 43, 655−660.
(40) Sohaskey, C. D.; Barbour, A. G. J. Bacteriol. 2000, 182, 1930−
1934.
(41) Sohaskey, C. D. FEMS Microbiol. Lett. 2004, 240, 187−192.
(42) Young, S. L.; Barbera, L.; Kaynard, A. H.; Haugland, R. P.; Kang,
H. C.; Brinkley, M.; Melner, M. H. Anal. Biochem. 1991, 197, 401−
407.
(43) Tao, W. X.; Lee, M. H.; Yoon, M. Y.; Kim, J. C.; Malhotra, S.;
Wu, J.; Hwang, E. C.; Lee, S. W. J. Microbiol. Biotechnol. 2011, 21,
1203−1210.
(44) Mitchell, C. C.; Delaney, D. Beltwide Cotton Conferences 2012,
1335−1344.
(45) Liles, M. R.; Williamson, L. L.; Rodbumrer, J.; Torsvik, V.;
Goodman, R. M.; Handelsman, J. Appl. Environ. Microbiol. 2008, 74,
3302−3305.
(46) Martinez, A.; Kolvek, S. J.; Yip, C. L.; Hopke, J.; Brown, K. A.;
MacNeil, I. A.; Osburne, M. S. Appl. Environ. Microbiol. 2004, 70,
2452−2463.
(47) Martin, A.; Camacho, M.; Portaels, F.; Palomino, J. C.
Antimicrob. Agents Chemother. 2003, 47, 3616−3619.
(48) Drummond, A. J.; Waigh, R. D. Recent Res. Dev. Phytochem.
2000, 4, 143−152.
(49) Besemer, J.; Borodovsky, M. Nucleic Acids Res. 1999, 27, 3911−
3920.
(50) Sambrook, J.; Russell, D. W. Molecular Cloning: A Laboratory
Manual; Cold Spring Harbor Laboratory Press: Cold Spring Harbor,
New York, NY, 2001; Vol. 1.
(51) National Committee on Clinical Laboratory Standards.
Susceptibility testing of Mycobacteria, Nocardia, and other aerobic
Actinomycetes. Approved Standard, M24-A, 2003, 23.
L
DOI: 10.1021/acs.jnatprod.7b00903
J. Nat. Prod. XXXX, XXX, XXX−XXX