Dapdiamides, tripeptide antibiotics formed by unconventional amide ligases

Article abstract of DOI:10.1021/np900685z

Construction of a genomic DNA library from Pantoea agglomerans strain CU0119 and screening against the plant pathogen Erwinia amylovora yielded a new family of antibiotics, dapdiamides A-E (1-5). The structures were established through 2D-NMR experiments and mass spectrometry, as well as the synthesis of dapdiamide A (1). Transposon mutagenesis of the active cosmid allowed identification of the biosynthetic gene cluster. The dapdiamide family's promiscuous biosynthetic pathway contains two unconventional amide ligases that are predicted to couple its constituent monomers.

Full text of DOI:10.1021/np900685z

J. Nat. Prod. 2010, 73, 441–446
441
Dapdiamides, Tripeptide Antibiotics Formed by Unconventional Amide Ligases†
Jessica Dawlaty,‡ Xiaorong Zhang,‡ Michael A. Fischbach,§ and Jon Clardy*,‡
Department of Biological Chemistry and Molecular Pharmacology, HarVard Medical School, 240 Longwood AVenue,
Boston, Massachusetts 02115, and Department of Bioengineering and Therapeutic Sciences and California Institute
for QuantitatiVe Biology, UniVersity of California, San Francisco, 1700 4th Street, San Francisco, California 94158
ReceiVed October 27, 2009
Construction of a genomic DNA library from Pantoea agglomerans strain CU0119 and screening against the plant
pathogen Erwinia amyloVora yielded a new family of antibiotics, dapdiamides A-E (1-5). The structures were established
through 2D-NMR experiments and mass spectrometry, as well as the synthesis of dapdiamide A (1). Transposon
mutagenesis of the active cosmid allowed identification of the biosynthetic gene cluster. The dapdiamide family’s
promiscuous biosynthetic pathway contains two unconventional amide ligases that are predicted to couple its constituent
monomers.
Condensation reactions, the joining of two molecules with the
formal loss of water, dominate the biosynthesis of large and small
biological molecules. In nonribosomal peptide synthetase (NRPS)
pathways, condensation domains that are evolutionarily related to
the coenzyme A-dependent acyltransferases used in primary
metabolism carry out repetitive condensations between amino,
hydroxy, and aryl acid monomers.1 A recent study on the andrimid
biosynthetic gene cluster from Pantoea agglomerans and other
bacteria led to the realization that transglutaminase-like enzymes
also carry out biosynthetic condensations.2,3 In this report, we
describe another antibiotic family from P. agglomerans (1-5) in
which an L-2,3-diaminopropionic acid fragment (DAP) is linked
to two variable units through two amide bonds. The family’s
promiscuous biosynthetic pathway contains no NRPS condensation
domains.
pUC19 and transformed into E. coli. From this plasmid subclone
(pUC19 A10A) members of the dapdiamide family were isolated
using bioassay-guided fractionation and linked to their biosynthetic
gene cluster.
From the 1H NMR spectra of the active fractions it was apparent
that several structurally related compounds were responsible for
the antibiotic activity against E. amyloVora. The high-resolution
MS spectrum of the most abundant compound (1) exhibited an
[M + H]+ ion at m/z 301.1517, indicating the molecular formula
C12H20N4O5 + H. In a DQCOSY spectrum of 1, three proton-proton
spin systems corresponding to fumaramic acid, L-2,3-diaminopro-
pionate (DAP), and L-valine units were evident (Figure 1). HMQC
and HMBC 2D NMR experiments (Table 1) demonstrated that
fumaramic acid is joined by an amide bond to the ꢀ-amino group
of DAP, while Val is linked through an amide bond to the DAP
carboxyl. Several less abundant structural variants of dapdiamide
A were also isolated: dapdiamides B (2) and C (3) contain isoleucine
and leucine instead of valine (Table 1); dapdiamide D (4) is identical
to 1 except the linkage between fumaramic acid and DAP is through
the R-amino group of DAP; and dapdiamide E (5) is identical to 4
but has an epoxide in place of the fumaramic acid double bond
(Table 2).
In order to determine the absolute configuration of the natural
product, dapdiamide A (1) was synthesized with the L configuration
1
at both DAP and Val alpha carbons (Figure 2). The H and 2D
NMR data of synthetic dapdiamide A were identical to dapdiamide
A isolated from A10A. The specific rotation [R]23 for isolated
D
dapdiamide A was +5.9 and that for synthetic dapdiamide A was
+9.5 in water. Circular dichroism spectra of synthetic and isolated
dapdiamide A indicate that the configuration of the two compounds
is the same (Figure 3).
The most likely annotation of the gene cluster for dapdiamide
biosynthesis (Figure 4) identifies two genes responsible for produc-
ing DAP (blue), four genes involved in the biosynthesis of
fumaramic acid (red), two genes responsible for amide bond
formation (green), and a gene for self-resistance (yellow). DdaAB
are homologous to enzymes previously implicated in DAP biosyn-
thesis6 and will not be discussed further here.
Biosynthesis of Fumaramic Acid (DdaCDEH). The pathway
for fumaramic acid biosynthesis likely begins with succinate. Like
other NRPS pathways that generate oxidatively modified building
blocks,7thedapdiamidegeneclusterencodesanadenylation-thiolation
(A-T) enzyme, DdaD, that sequesters the monomer to be oxidized.
The A domain of DdaD is predicted to activate succinate by
adenylation and then tether it to the adjacent DdaD thiolation
domain as a thioester. This sequestered monomer would be oxidized
by DdaC, a putative Fe(II)/R-ketoglutarate-dependent dioxygenase.
Results and Discussion
Both the dapdiamide family members and their biosynthetic gene
cluster were captured by constructing a genomic library of P.
agglomerans strain CU0119 and screening for cosmid clones active
against Erwinia amyloVora. Strains of P. agglomerans (formerly
E. herbicola) have previously been shown to produce numerous
antibiotics active against E. amyloVora, the causative agent of the
devastating plant disease fire blight.4,5 A subclone of a dapdiamide-
producing cosmid was transposon mutagenized to identify the
biosynthetic genes. In this way, a cluster of nine ORFs from the
cosmid insert was identified, and it was further subcloned into
† Dedicated to the late Dr. John W. Daly of NIDDK, NIH, Bethesda,
Maryland, and to the late Dr. Richard E. Moore of the University of Hawaii
at Manoa for their pioneering work on bioactive natural products.
* To whom correspondence should be addressed. Tel: (617) 432-2845.
Fax: (617) 432-6424. E-mail: jon_clardy@hms.harvard.edu.
‡ Harvard Medical School.
§ University of California, San Francisco.
10.1021/np900685z  2010 American Chemical Society and American Society of Pharmacognosy
Published on Web 12/30/2009

442 Journal of Natural Products, 2010, Vol. 73, No. 3
Dawlaty et al.
Figure 1. DQCOSY correlations for the dapdiamides are indicated with bold lines.
Table 1. NMR Spectroscopic Data (600 MHz, DMSO-d6) for Dapdiamides A-C (1-3)
dapdiamide A (1)
δCa, mult. δH (J in Hz)
dapdiamide B (2)
δCa, mult. δH (J in Hz)
dapdiamide C (3)
δH (J in Hz)
HMBCb
position
Fum
HMBCb
HMBCb
δCa, mult.
C1
164.3, qC
164.6, qC
164.3, qC
C2
C3
C4
132.3, CH 6.78, d (15.2) 1, 3, 4
133.1, CH 6.85, d (15.2) 1, 2, 4
164.9, qC
132.5, CH 6.78, d (15.2) 1,3,4
133.3, CH 6.84, d (15.3) 1,2,4
165.2, qC
132.2, CH 6.78, d (15.2) 1, 3, 4
133.1, CH 6.84, d (15.2
164.9, qC
1, 2, 4
4a-NH
4b-NH
DAP
C1′
7.34, s
7.80, s
3, 4
4
7.34, s
7.80, s
3,4
4
7.34, br s
7.81, br s
166.7, qC
166.8, qC
C2′
51.5, CH 4.06, t (5.9)
8.22, br s
39.5, CH2 3.52, m
3.56, m
1′, 3′
51.8, CH 4.03, t (5.9)
8.13, br s
39.8, CH2 3.53 (m)
1′, 3′
51.6, CH 3.96, m
39.7, CH2 3.55, t
2′-NH2
C3a′
1, 1′, 2′
1, 1′, 2′
1, 3′
1, 1′, 2′
1, 3′
C3b′
3′-NH
Val/Ile/Leu
C1′′
8.44, t (5.9)
8.44, t (5.8)
172.5, qC
172.2, qC
1′′-OH
C2′′
12.95, br s
57.2, CH 4.24, dd
(5.2, 8.3)
1′, 1′′, 3′′, 4a′′, 4b′′ 56.6, CH 4.28, dd
(5.2, 8.1)
8.68, d (7.9) 1′, 2′′
1′, 1′′, 3′′, 4′′, 6′′ 50.3, CH 4.28, dd
(7.8, 14.7)
8.79, br d (7.6)
39.5, CH2 1.56, m
23.9, CH 1.67, m
2′′-NH
C3′′
8.68, d (8.4) 1′′, 2′′
29.7, CH 2.12, m 1′′, 2′′, 4a′′, 4b′′
36.6, CH 1.82, m
2′′, 4′′, 5′′, 6′′
2′′, 3′′, 5′′, 6′′
2′′, 3′′, 5′′, 6′′
3′′, 4′′,
5a′′
C4a′′
17.5, CH3 0.91, d (7.2) 2′′, 3′′, 4b′′
18.8, CH3 0.92, d (7.2) 2′′, 3′′, 4a′′
24.5, CH 1.21, m
24.5, CH 1.43, m
11.2, CH3 0.87, t
C4b′′
C5a′′
20.9, CH3 0.87, d (6.5)
22.6, CH3 0.91, d (6.6)
3′′, 4a′′, 5b′′
3′′, 4a′′, 5a′′
C5b′′ or C6′′
a
15.4, CH3 0.89, d (7.0) 2′′, 3′′, 4′′
13C NMR chemical shifts are assigned from 1H-13C HMQC and HMBC correlations. b HMBC correlations are from proton(s) stated to the
indicated carbon.
Enzymes from this family catalyze oxidative modifications of
Therefore, DdaF likely catalyzes one of the two condensation
reactions needed to join the dapdiamide monomers. Related ATP-
grasp enzymes have been shown to form intramolecular ester and
amide cross-links in microviridin and related cyanobacterial natural
products, hinting at a wider role for this enzyme superfamily in
natural product biosynthesis.13,14
The other amide bond is likely formed by DdaG, a distant
homologue of phenylacetate-CoA ligase, which activates the
carboxylic acid of phenylacetate by adenylation for the subsequent
attack of the terminal thiol of coenzyme A. DdaG may similarly
activate a carboxylic acid by adenylation, perhaps for direct amide
bond formation without the intermediacy of a CoA-linked thioester
species. While acyl-CoA ligase enzymes are not known to catalyze
subsequent condensation reactions, DdaG has an unassigned ∼100
amino acid domain at its C-terminus that may be involved in
positioning an amine nucleophile for attack.
Self-Resistance (DdaI). Antibiotic biosynthetic pathways typi-
cally include a mechanism for resistance to the antibiotic being
produced. Plasmid-borne DdaI, a predicted transmembrane efflux
pump, conferred resistance to 1 in an otherwise sensitive E. coli
strain, confirming its role as a dapdiamide resistance gene.
The target of ꢀ-linked N-fumaramoyl-DAP has previously been
described as glucosamine-6-phosphate synthase.15 This enzyme
unactivated carbon centers, including hydroxylation, desaturation,
and epoxide formation.8 In the case of the dapdiamides, DdaC
typically catalyzes the formation of a double bond, but may
occasionally catalyze an additional two-electron oxidation to form
an epoxide instead. There is precedent for the formation of an
epoxide from a hydroxy group by another enzyme from this class,
HppE.9 Finally, the resulting oxidized succinate derivative is likely
released from DdaD by DdaE, a homologue of the type II
thioesterase domains that cleave acyl groups from thiolation
domains.10 DdaH, a homologue of asparagine synthetase, should
transaminate one of the carboxylic acids of the succinate-based
monomer, but the timing of its action cannot be predicted
bioinformatically.11,12
Amide Bond Formation (DdaFG). The dapdiamide gene
cluster does not encode any enzymes known to form amide bonds
in secondary metabolic pathways. However, it encodes two proteins
with sequence homology to enzymes that activate carboxylate
monomers in primary metabolic pathways: DdaF, a member of the
ATP-grasp superfamily, and DdaG, a homologue of phenylacetate-
CoA ligase. DdaF is homologous to biotin carboxylase and
phosphoribosylglycinamide synthetase, both of which catalyze the
formation of amide bonds via an acyl phosphate intermediate.

Dapdiamides: Tripeptide Antibiotics
Journal of Natural Products, 2010, Vol. 73, No. 3 443
Table 2. NMR Spectroscopic Data for Dapdiamides D (4) (600 MHz, DMSO-d6) and Dapdiamide E (5) 600 MHz, D2O)
dapdiamide D (4)
δH (J in Hz)
dapdiamide E (5)
δH (J in Hz)
position
δCa, mult.
HMBCb
δCa, mult.
HMBCb
Fum
C1
C2
C3
C4
164.5, qC
132.4, CH
133.3, CH
165.1, qC
169.1, qC
53.7, CH
53.6, CH
170.9, qC
6.85, d (15.4)
6.86, d (15.3)
1, 3, 4
1, 2, 4
3.82, d (2.0)
3.75, d (2.0)
1
4
4a-NH
4b-NH
DAP
C1′
7.36, br s
7.83, br s
3
168.5, qC
50.4, CH
169.4, qC
50.8, CH
C2′
4.72, d (4.7, 8.5, 8.5)
8.83, d (8.2)
2.97, dd (9.7, 12.4)
3.23, dd (8.6, 13.1
1, 1′, 3′
1, 2′
4.89, dd (4.9, 8.5)
1
2′-NH
C3a′
C3b′
3′-NH2
Val
39.9, CH2
40.0, CH2
3.37, dd (4.5, 9.0)
3.55, dd (4.9, 13.5)
1′, 2′
1′, 2′
C1′′
C2′′
2′′-NH
C3′′
C4a′′
C4b′′
a
172.5, qC
57.3, CH
175.9, qC
59.4, CH
4.16, dd (5.6, 8.5)
8.14, d (8.5)
2.08, m
0.86, d (6.8)
0.87, d (6.8)
1′, 1′′, 3′′, 4a′′, 4b′′
1′
1′, 2′′, 4a′′, 4b′′
2′′, 3′′, 4b′′
2′′, 3′′, 4a′′
4.35, d (5.7)
1′, 1′′, 3′′, 4a′′
29.7, CH
17.8, CH3
18.9, CH3
30.0, CH
17.2, CH3
18.6, CH3
2.27, m
0.98, d (7.0)
0.99, d (7.0)
4a′′, 4b′′
2′′, 3′′, 4b′′
2′′, 3′′, 4a′′
13C NMR chemical shifts are assigned from 1H-13C HMQC and HMBC correlations. b HMBC correlations are from proton(s) stated to the
indicated carbon.
Figure 2. Synthesis of dapdiamide A (1). Details of yields and procedures are given in the Experimental Section.
of the glutamine binding domain of glucosamine-6-phosphate
synthase becomes inactivated when it attacks the double bond of
the fumaramoyl moiety and becomes irreversibly attached.17
There are a few natural products structurally similar to the
dapdiamides. CB-25-I, isolated from Serratia plymuthica, has a
structure nearly identical to 1 but has an epoxide in place of the
olefin.18 While the biosynthetic gene cluster for CB-25-I has never
been identified, a pathway nearly identical to the biosynthetic
pathway for the dapdiamide compounds in GenBank from the
Serratia proteamaculans 568 sequencing project (locus tags
Spro_0339-Spro_0347) suggests that this pathway has been shared
between these enterobacteria. Two actinomycete-derived com-
pounds, A19009 and Sch37137 (the epoxide analogue of A19009),
contain the N-fumaramoyl-DAP core but have L-alanine linked to
DAP via the R-amino group.19,20 These compounds could be made
by the same pathway that uses an enzyme with different chemo-
and regioselectivity for the formation of the second amide bond,
but no sequence data are available. The isolation of such similar
molecules from unrelated Gram-negative and Gram-positive bacteria
suggests that the dapdiamide gene cluster is highly mobile, and
the divergent structures of the dapdiamides A19009 and Sch37137
Figure 3. Circular dichroism of isolated (natural) and synthetic
dapdiamide A (1).
converts D-fructose-6-phosphate to D-glucosamine-6-phosphate, a
precursor to UDP-N-acetyl-D-glucosamine, an essential building
block for bacterial cell wall biosynthesis.16 The catalytic cysteine

444 Journal of Natural Products, 2010, Vol. 73, No. 3
Dawlaty et al.
Figure 4. Gene cluster and proposed scheme for dapdiamide biosynthesis. The order of steps shown here is one of several possibilities.
into electrocompetent DH10B E. coli cells (Invitrogen) and plated on
LB plates containing cm, X-gal, and IPTG. White colonies were
screened for antibiotic production in E. amyloVora bioassays. An active
subclone, p119C1.1, was identified, transformed into E. coli DH10B,
and rescreened against E. amyloVora to confirm the activity.
are a testament to the plasticity of a pathway based on atypical
condensation catalysts.
Experimental Section
General Experimental Procedures. CD spectra were obtained using
an Aviv circular dichroism spectrometer model 202. NMR spectra were
obtained using a Varian Inova 600 MHz NMR spectrometer. LC-MS
spectra were obtained with a Micromass LCZ quadrupole mass
spectrometer. High-resolution mass spectra were obtained with a
Micromass Q-Tof2 mass spectrometer using reserpine as a lock mass.
Samples were injected through the HPLC system, and the LC flow
was combined with the flow of a reserpine solution from a syringe
pump. Restriction enzymes were purchased from New England Biolabs.
Plasmids were isolated using Qiaprep Spin Miniprep kits. All other
reagents were purchased from Sigma-Aldrich unless otherwise indicated.
Culture Media and Strains. Pantoea agglomerans strain CU0119
and Erwinia. amyloVora strain 273 were obtained from Professor Steven
Beer at Cornell University. Cultures of E. coli were grown in
Luria-Bertani (LB) media or on LB-agar plates. Antibiotics were used
at the following concentrations: ampicillin (amp) 100 µg/mL; chloram-
phenicol (cm) 25 µg/mL; kanamycin (kan) 30 µg/mL. For blue/white
screening 5-bromo-4-chloro-3-indolyl-ꢀ-D-galactoside (X-gal) was
added to a final concentration of 80 µg/mL, and isopropyl ꢀ-D-1-
thiogalactopyranoside (IPTG) was added to a concentration of 20 mM.
Bioassays with E. amyloVora were conducted using glucose-asparagine
(GA) medium containing 20 g glucose, 0.3 g L-asparagine, 0.05 g
nicotinic acid, 11.5 g K2HPO4, 4.5 g KH2PO4, and 0.12 g MgSO4 per
liter.21 Cultures of E. coli with cosmids or plasmids containing the
dapdiamide biosynthetic pathway were grown in E. coli minimal
medium (EcMM) containing per liter 0.25 g yeast extract, 20 mL
glycerol, 4.0 g K2HPO4, 1.72 g KH2PO4, 0.5 g NaCl, 2.0 g (NH4)2SO4,
0.2 g sodium citrate, and 0.02 g MgSO4 ·7H2O. Both GA and EcMM
were sterilized by filtration through a 0.2 µm filter.
Erwinia amyloWora Bioassays. Bioassay plates were made by
diluting an overnight culture of E. amyloVora grown in LB into warm,
sterile top agar (GA with 0.75% agar). To test E. coli transformants
for activity, a small portion of a colony was patched onto the surface
of the agar using a wooden stick. To test liquid samples for activity, 3
µL of each sample was pipetted directly onto the surface of the agar.
Assay plates were incubated overnight at room temperature and then
examined for zones of inhibition.
Cosmid Library Construction and Screening. P. agglomerans
strain CU0119 exhibited a large zone of inhibition in E. amyloVora
bioassays, so it was selected for further study. A cosmid library of P.
agglomerans strain CU0119 genomic DNA was constructed using the
pWEB cosmid cloning kit from Epicenter according to the manufac-
turer’s instructions. The cosmid clones were screened against E.
amyloVora, and one active cosmid clone, 119C1, was identified. The
cosmid was isolated and retransformed into E. coli, and the antibiotic
activity was confirmed by rescreening against E. amyloVora on a GA
agar plate.
Transposon Mutagenesis of p119C1.1. To identify genes respon-
sible for the biosynthesis of the antibiotics, the active plasmid
(p119C1.1) was transposon mutagenized using the GPS-1 Genome
Priming System from New England Biolabs according to the manu-
facturer’s instructions. The mutagenized plasmids were transformed
into DH10B cells, and transformants were screened against E. amylo-
Vora for loss of antibiotic production. Transposon insertion mutants
that eliminated antibiotic production were mapped by sequencing out
from the ends of the transposons using primers S and N, which are
supplied with the GPS kit. DNA sequencing was performed at the Dana-
Farber/Harvard Cancer Center DNA Resource Core in Boston, MA.
The sequences obtained from the knockouts were assembled into a
contiguous DNA sequence spanning roughly 14 kilobases. Using the
ORF Finder on MacVector (Accelrys), a cluster of nine open reading
frames was identified that had transposon insertions that knocked out
antibiotic production.
Subcloning p119C1.1. To confirm that the cluster of nine open
reading frames identified by transposon mutagenesis was indeed
responsible for antibiotic production, this portion of p119C1.1 was
subcloned. p119C1.1 was digested with DrdI, and the ∼12 kb fragment
containing the biosynthetic cluster was gel purified using a Qiaquick
gel extraction kit (Qiagen). To form blunt ends, the End-It DNA end-
repair kit (Epicenter) was used according to the manufacturer’s
instructions. The pUC19 vector was prepared by digestion with SmaI,
dephosphorylation with calf intestinal alkaline phosphatase (CIP), and
gel purification using the Qiaquick gel extraction kit. The DrdI fragment
and SmaI-digested pUC19 were ligated using the Fast-Link DNA
ligation kit according to the manufacturer’s instructions, and the ligation
reaction was transformed into electrocompetent EC100 E. coli cells
(Epicenter). An active subclone, named A10A, was identified and
verified by double restriction digest. A10A was checked for stability
by growing cultures from several different colonies in EcMM and
testing them for activity on GA plates seeded with E. amyloVora. The
construct consistently inhibited growth.
Isolation of Dapdiamides. A starter culture of E. coli subclone in
pUC19 (A10A) was grown in 10 mL of LB amp overnight at 30 °C.
Twenty 500 mL Erlenmeyer flasks containing 50 mL of EcMM media
were inoculated with 150 µL of starter culture. The cultures were
incubated at 30 °C for 6 h at 250 rpm and then incubated at room
temperature for 16 h at 200 rpm. The cultures were combined and
centrifuged for 25 min at 12000g. The culture supernatant was applied
to a cation exchange column (AG 50W-X8 resin from Bio-Rad). Twenty
50 mL fractions were eluted with 500 mM ammonium hydroxide at a
flow rate of 6 mL/min. The pH of each fraction was adjusted to 6.5-8.5
using acetic acid or ammonium hydroxide. Fractions were dried and
assayed for activity against E. amyloVora. The active fractions were
combined, and this material (∼1 g) was further purified by reversed-
phase C-18 chromatography. Thirty 125 mL fractions were eluted with
a MeOH/H2O step gradient. The fractions active against E. amyloVora
Subcloning the Active Cosmid. The active cosmid and plasmid
vector pBC were digested with NotI and ligated using the Fast-Link
DNA ligation kit from Epicenter. The ligation mixture was transformed

Dapdiamides: Tripeptide Antibiotics
Journal of Natural Products, 2010, Vol. 73, No. 3 445
were combined except for the first active fraction, fraction 3, which
was high in salt and was purified separately. Fraction 3 (∼450 mg)
was purified by HPLC using a reversed-phase C-30 semipreparative
column with a flow rate of 2 mL/min and a H2O/MeOH gradient
(0-100% MeOH over 25 min). The remaining active fractions from
the C-18 column, F4 to F16, were purified by HPLC on a semiprepara-
tive C-18 column with a flow rate of 4 mL/min using H2O/MeOH
acidified with 0.01% trifluoroacetic acid as the eluent (2.5-25% MeOH
over 20 min). Approximately 3 mg of dapdiamide A, 1.4 mg of
dapdiamide B, 1.6 mg of dapdiamide C, 1 mg of dapdiamide D, 1.5
mg of dapdiamide E, and 1.5 mg of N-fumaramoyl-DAP were isolated.
Synthesis of Dapdiamide A. (E)-Ethyl 4-(4-methoxybenzylamino)-
4-oxobut-2-enoate (6). To form the fumaramic acid portion of
dapdiamide A, we chose to make the N-(methoxybenzyl)fumaramic
acid (6) derivative. Monoethyl fumaric acid (40 mmol) and 4-dim-
ethylaminopyridine (DMAP) (44 mmol) were dissolved in CH2Cl2 (400
mL), and 4-methoxybenzylamine (4-MBA) (40 mmol) was added. N,N’-
Dicyclohexylcarbodiimide (DCC) (44 mmol) dissolved in CH2Cl2 (100
mL) was added, and the mixture was stirred overnight. The precipitate
from the reaction was removed by filtering and washed with CH2Cl2.
The CH2Cl2 was extracted with sulfuric acid (0.5 M, 500 mL), then
H2O (500 mL), and then saturated aqueous NaHCO3 (500 mL) and
dried with sodium sulfate. The CH2Cl2 was evaporated and the product
was further purified by silica flash chromatography eluting with EtOAc/
hexanes (3:2). The fractions containing the product were combined and
evaporated to afford 6 (7.83 g, 74%): TLC Rf ) 0.53, EtOAc/hexanes
(1H, dd, J ) 4.6, 8.8 Hz), 5.65 (1H, br s), 7.69 (1H, br s); LC(ESI)MS
m/z ) 360 [M + H]+.
(S)-tert-Butyl 2-((S)-2-(tert-butoxycarbonylamino)-3-((E)-4-(4-
methoxybenzylamino)-4-oxobut-2-enamido)propanamido)-3-meth-
ylbutanoate (10). The dipeptide 9 was coupled with 7 using PyBop
and DIEA. The product 9 from the previous reaction (approximately
0.86 mmol) was dissolved in CH2Cl2 (8 mL) along with 7 (0.86 mmol),
PyBop (0.95 mmol), and DIEA (2.6 mmol). After stirring the reaction
mixture for 2 h the CH2Cl2 was washed with H2O acidified with 5%
citric acid (pH 4), H2O, and then saturated aqueous NaHCO3. The
CH2Cl2 was dried and purified by silica flash chromatography with
CH2Cl2/MeOH (9:1) as an eluent to afford 10 (0.82 mmol, 95%): TLC
Rf ) 0.31, EtOAc; 1H NMR (500 MHz, CDCl3) δ 0.87 (3H, d, J ) 6.5
Hz), 0.88 (3H, d, J ) 5.7 Hz), 1.41 (9H, s), 1.43 (9H, s), 2.13 (1H,
m), 3.26 (1H, br s), 3.68 (2H, br s), 3.75 (3H, s), 4.32 (2H, dd, J )
4.6, 8.5 Hz), 4.38 (2H, brs), 5.88 (1H, br s), 6.54 (1H, br s), 6.82 (2H,
d, J ) 8.5 Hz), 6.95 (2H, br s), 7.16 (2H, d, J ) 8.3 Hz); LC(ESI)MS
m/z ) 577 [M + H]+.
(S)-tert-Butyl 2-((S)-3-((E)-4-amino-4-oxobut-2-enamido)-2-(tert-
butoxycarbonylamino)propanamido)-3-methylbutanoate (11). The
methoxybenzyl group was removed from 10 to form the amide by
dissolving 10 (0.82 mmol) and 2,3-dichloro-5,6-dicyano-1,4-benzo-
quinone (DDQ)22 (1.2 mmol) in CH2Cl2/H2O (25:1), and the reaction
was stirred overnight. CH2Cl2 (200 mL) was added to the reaction,
washed twice with H2O (200 mL), and then washed with saturated
aqueous NaHCO3 (200 mL). The CH2Cl2 was dried and the product
was purified by silica flash chromatography using CH2Cl2/MeOH
(5:1) as an eluent to yield 11 (0.37 mmol, 45%): TLC Rf ) 0.53,
1
(3:2); H NMR (500 MHz, CD3OD) δ 1.30 (3H, t, J ) 7.1 Hz), 3.77
(3H, s), 4.24 (2H, q, J ) 7.1 Hz), 4.39 (2H, s), 6.73 (1H, d, J ) 15.5
Hz), 6.88 (2H, d, J ) 8.6 Hz), 7.02 (1H, d, J ) 15.5 Hz), 7.22 (2H, d,
J ) 8.6 Hz); LC(ESI)MS m/z ) 264 [M + H]+.
1
CH2Cl2/MeOH (5:1); H NMR (500 MHz, CD3OD) δ 0.96 (6H, d,
J ) 6.8 Hz), 1.44 (9H, s), 1.48 (9H, s), 2.16 (1H, m), 3.49 (1H, dd, J
) 8.2, 13.5 Hz), 3.64 (1H, dd, J ) 5.1, 13.7 Hz), 4.22 (1H, d, J ) 5.4
Hz), 4.34 (1H, dd, J ) 5.3, 7.8 Hz), 6.88 (1H, d, J ) 15.2 Hz), 6.93
(1H, d, J ) 15.3 Hz); LC(ESI)MS m/z ) 457 [M + H]+.
(E)-4-(4-Methoxybenzylamino)-4-oxobut-2-enoic acid (7). To form
the acid of 6, the ethyl group of 6 was hydrolyzed with lithium
hydroxide by dissolving 6 (4 g, 15 mmol) in MeOH (50 mL) and adding
3 equiv of LiOH (46 mL of 1 M solution in water). After stirring at
room temperature for 2 h the MeOH was evaporated using a rotary
evaporator, 100 mL of water was added, the pH was adjusted to roughly
4 with citric acid (5% in water), and the water was extracted with EtOAc
(450 mL) three times. The EtOAc extract was dried to yield 7 (3.66 g,
102%): TLC Rf ) 0.41, CH2Cl2/MeOH/CH3COOH (5:1:0.001); 1H
NMR (500 MHz, CD3OD) δ 3.77 (3H, s), 4.39 (2H, s), 6.71 (1H, d,
J ) 15.5 Hz), 6.88 (2H, d, J ) 8.6 Hz), 6.98 (1H, d, J ) 15.5 Hz),
7.22 (2H, d, J ) 8.6 Hz); LC(ESI)MS m/z ) 236 [M + H]+.
(S)-tert-Butyl 2-((S)-3-(((9H-fluoren-9-yl)methoxy)carbonylamino)-
2-(tert-butoxycarbonylamino)propanamido)-3-methylbutanoate (8).
To couple DAP and L-valine, Boc-DAP(Fmoc)-OH (1 mmol, Bachem),
Val-tBu (1 mmol, Bachem), benzotriazol-1-yloxytripyrrolidinophos-
phonium hexafluorophosphate (PyBOP) (1.1 mmol), and N,N-diiso-
propylethylamine (DIEA) (3 mmol) were dissolved in 10 mL of CH2Cl2
and stirred for 2 h. After the reaction had gone to completion (based
on the disappearance of the DAP and L-valine starting materials in the
LC-MS profile) CH2Cl2 was added to 100 mL and extracted with H2O
(100 mL), H2O acidified with citric acid (100 mL), H2O (100 mL),
and then aqueous NaHCO3 (100 mL). The CH2Cl2 was dried and then
purified by flash chromatography on a silica column using EtOAc/
hexanes (3:2) as an eluent to afford 8 (0.87 mmol, 87%): TLC Rf )
0.75, EtOAc/hexanes (3:2); 1H NMR (500 MHz, CD2Cl2) δ 0.89 (3H,
d, J ) 6.9 Hz), 0.91 (3H, d, J ) 6.9 Hz), 1.44 (9H, s), 1.46 (9H, s),
2.15 (1H, m), 3.54 (2H, t, J ) 5.3 Hz), 4.22 (1H, t, J ) 6.8 Hz), 4.33
(1H, dd, J ) 4.6, 8.5 Hz), 4.39 (2H, t, 7.4 Hz), 5.60 (1H, br s), 5.68
(1H, br s), 6.89 (1H, br s), 7.32 (2H, t, J ) 7.4 Hz), 7.40 (2H, t, J )
7.4 Hz), 7.61 (2H, d, J ) 7.5 Hz), 7.78 (2H, d, J ) 7.5 Hz); LC(ESI)MS
m/z ) 582 [M + H]+.
Dapdiamide A (1). The final step in the synthesis of dapdiamide A
is the removal of the Boc and tBu protecting groups. 11 (0.37 mmol)
was dissolved in CH2Cl2 (3 mL), TFA was added to 20%, and the
reaction was stirred at room temperature for 24 h. After addition of
H2O (40 mL), the reaction was extracted twice with CH2Cl2 (40 mL).
The aqueous layer was dried to yield dapdiamide A (0.34 mmol, 91%).
1H and 2D NMR data of synthetic dapdiamide A were identical to
dapdiamide A isolated from A10A. Dapdiamide A was tested against
E. amyloVora, and it was active in a similar concentration range as the
isolated dapdiamide A.
Dapdiamide A (1): [R]23 +5.9 (isolated) and +9.5 (synthetic)
D
1
(H2O); H and 13C NMR spectroscopic data, see Table 1; HRMS m/z
301.1499 [M + H]+ (isolated) and 301.1504 [M + H]+ (synthetic)
(calcd for C12H21N4O5, 301.1512).
Dapdiamide B (2): 1H and 13C NMR spectroscopic data, see Table 1;
HRMS m/z 315.1656 [M + H]+ (calcd for C13H23N4O5, 315.1668).
Dapdiamide C (3): 1H and 13C NMR spectroscopic data, see Table 1;
HRMS m/z 315.1666 [M + H]+ (calcd for C13H23N4O5, 315.1668).
Dapdiamide D (4): 1H and 13C NMR spectroscopic data, see Table 2;
HRMS m/z 301.1509 [M + H]+ (calcd for C12H21N4O5, 301.1512).
Dapdiamide E (5): 1H and 13C NMR spectroscopic data, see Table 2;
HRMS m/z 317.1461 [M + H]+ (calcd for C12H21N4O6, 317.1461).
Acknowledgment. This work was supported by NIH grant CA24487
(J.C.) and a HHMI Fellowship (J.D.). We thank Prof. S. Beer (Cornell,
Plant Pathology) for sharing his P. agglomerans strains, and M.
Hollenhorst, L. Blasiak, and C. Walsh (HMS, BCMP) for helpful
discussions.
1
Supporting Information Available: H and 2D NMR spectra for
(S)-tert-Butyl 2-((S)-3-amino-2-(tert-butoxycarbonylamino)pro-
panamido)-3-methylbutanoate (9). The Fmoc protecting group of 8
was removed to enable coupling with 7 by dissolving 8 (0.87 mmol)
in a solution of 10% piperidine in dimethylformamide (5 mL) and
stirring at room temperature for 2 h. CH2Cl2 was added to the reaction
mixture, and it was extracted with saturated aqueous NaHCO3 (75 mL).
The organic layer was dried and then purified by silica flash chroma-
tography with CH2Cl2/MeOH (7:1) as an eluent. The fractions contain-
ing 9 were dried to yield an oil: TLC Rf ) 0.36, CH2Cl2/MeOH (7:1);
1H NMR (500 MHz, CD2Cl2) δ 0.89 (3H, d, J ) 6.9 Hz), 0.92 (3H, d,
J ) 6.9 Hz), 1.44 (9H, s), 1.45 (9H, s), 2.16 (1H, m), 2.75 (1H, dd,
J ) 7.8, 12.1 Hz), 3.21 (1H, dd, J ) 2.8, 12.2 Hz), 3.99 (1H, m), 4.33
1-5, map of transposon insertions, and table of sequence homologues.
This information is available free of charge via the Internet at http://
pubs.acs.org.
References and Notes
(1) Keating, T. A.; Marshall, C. G.; Walsh, C. T.; Keating, A. E. Nat.
Struct. Biol. 2002, 9, 522–526.
(2) Fortin, P. D.; Walsh, C. T.; Magarvey, N. A. Nature 2007, 448, 824–
827.
(3) Jin, M.; Fischbach, M. A.; Clardy, J. J. Am. Chem. Soc. 2006, 128,
10660–10661.

446 Journal of Natural Products, 2010, Vol. 73, No. 3
Dawlaty et al.
(4) Jin, M.; Wright, S. A.; Beer, S. V.; Clardy, J. Angew. Chem., Int. Ed.
2003, 42, 2902–2905.
(5) Sutton, A. E.; Clardy, J. J. Am. Chem. Soc. 2001, 123, 9935–9946.
(6) Thomas, M. G.; Chan, Y. A.; Ozanick, S. G. Antimicrob. Agents
Chemother. 2003, 47, 2823–2830.
(7) Vaillancourt, F. H.; Yeh, E.; Vosburg, D. A.; O’Connor, S. E.; Walsh,
C. T. Nature 2005, 436, 1191–1194.
(8) Hausinger, R. P. Crit. ReV. Biochem. Mol. Biol. 2004, 39, 21–68.
(9) Higgins, L. J.; Yan, F.; Liu, P.; Liu, H. W.; Drennan, C. L. Nature
2005, 437, 838–844.
(10) Yeh, E.; Kohli, R. M.; Bruner, S. D.; Walsh, C. T. ChemBioChem
2004, 5, 1290–1293.
(11) Humbert, R.; Simoni, R. D. J. Bacteriol. 1980, 142, 212–220.
(12) Scofield, M. A.; Lewis, W. S.; Schuster, S. M. J. Biol. Chem. 1990,
265, 12895–12902.
(13) Ziemert, N.; Ishida, K.; Liaimer, A.; Hertweck, C.; Dittmann, E.
Angew. Chem., Int. Ed. 2008, 47, 7756–7759.
(14) Philmus, B.; Christiansen, G.; Yoshida, W. Y.; Hemscheidt, T. K.
ChemBioChem 2008, 9, 3066–3073.
(15) Chmara, H.; Andruszkiewicz, R.; Borowski, E. Biochem. Biophys. Res.
Commun. 1984, 120, 865–872.
(16) Ghosh, S.; Blumenthal, H. J.; Davidson, E.; Roseman, S. J. Biol. Chem.
1960, 235, 1265–1273.
(17) Kucharczyk, N.; Denisot, M. A.; Le Goffic, F.; Badet, B. Biochemistry
1990, 29, 3668–3676.
(18) Shoji, J.; Hinoo, H.; Sakazaki, R.; Kato, T.; Hattori, T.; Matsumoto,
K.; Tawara, K.; Kikuchi, J.; Terui, Y. J. Antibiot. (Tokyo) 1989, 42,
869–874.
(19) van der Baan, J. L.; Barnick, J. W.; Bickelhaupt, F. J. Antibiot. (Tokyo)
1983, 36, 784–792.
(20) Cooper, R.; Horan, A. C.; Gentile, F.; Gullo, V.; Loebenberg, D.;
Marquez, J.; Patel, M.; Puar, M. S.; Truumees, I. J. Antibiot. (Tokyo)
1988, 41, 13–19.
(21) Wright, S. A. I. Ph.D. dissertation, Cornell University, Ithaca, NY, 1997.
(22) Oikawa, Y.; Tanaka, T.; Horita, K.; Yoshioka, T.; Yonemitsu, O.
Tetrahedron Lett. B 1984, 393–5396.
NP900685Z