Enzymatic basis of "hybridity" in thiomarinol biosynthesis

Article abstract of DOI:10.1002/anie.201411667

Thiomarinol is a naturally occurring double-headed antibiotic that is highly potent against methicillin-resistant Staphylococcus aureus. Its structure comprises two antimicrobial subcomponents, pseudomonic acid analogue and holothin, linked by an amide bond. TmlU was thought to be the sole enzyme responsible for this amide-bond formation. In contrast to this idea, we show that TmlU acts as a CoA ligase that activates pseudomonic acid as a thioester that is processed by the acetyltransferase HolE to catalyze the amidation. TmlU prefers complex acyl acids as substrates, whereas HolE is relatively promiscuous, accepting a range of acyl-CoA and amine substrates. Our results provide detailed biochemical information on thiomarinol biosynthesis, and evolutionary insight regarding how the pseudomonic acid and holothin pathways converge to generate this potent hybrid antibiotic. This work also demonstrates the potential of TmlU/HolE enzymes as engineering tools to generate new "hybrid" molecules. The biosynthetic mechanism responsible for generating the antibiotic thiomarinol was elucidated. In contrast to previous hypotheses, TmlU acts as a CoA-ligase and works in tandem with a second enzyme, acyltransferase HolE, to link two antimicrobial warheads pseudomonic acid and holothin, creating a hybrid antibiotic (see scheme).

Full text of DOI:10.1002/anie.201411667

Angewandte
Chemie
DOI: 10.1002/anie.201411667
Biosynthesis
Enzymatic Basis of “Hybridity” in Thiomarinol Biosynthesis**
Zachary D. Dunn, Walter J. Wever, Nicoleta J. Economou, Albert A. Bowers, and Bo Li*
Abstract: Thiomarinol is a naturally occurring double-headed
antibiotic that is highly potent against methicillin-resistant
Staphylococcus aureus. Its structure comprises two antimicro-
bial subcomponents, pseudomonic acid analogue and holo-
thin, linked by an amide bond. TmlU was thought to be the sole
enzyme responsible for this amide-bond formation. In contrast
to this idea, we show that TmlU acts as a CoA ligase that
activates pseudomonic acid as a thioester that is processed by
the acetyltransferase HolE to catalyze the amidation. TmlU
prefers complex acyl acids as substrates, whereas HolE is
relatively promiscuous, accepting a range of acyl-CoA and
amine substrates. Our results provide detailed biochemical
information on thiomarinol biosynthesis, and evolutionary
insight regarding how the pseudomonic acid and holothin
pathways converge to generate this potent hybrid antibiotic.
This work also demonstrates the potential of TmlU/HolE
enzymes as engineering tools to generate new “hybrid”
molecules.
natural product thiomarinols (1 and 2) from Pseudoaltero-
monas spp. SANK73390,[3] have already been honed by
nature for selectivity and biological activity. Thus, these
natural hybrids can provide valuable insight into useful
therapeutic combinations and linker strategies.
Thiomarinol combines the monic acid warhead of the
FDA-approved agent mupirocin (pseudomonic acids, e.g. 3,
Bactroban—GlaxoSmithKline)[4] and the compact holothin
(4) core of the dithiolopyrrolones (DTPs), such as holomycin
(5)[5] and thiolutin (6; Figure 1A).[6] These fragments, which
in thiomarinol are linked by a fatty acyl amide bridge, exhibit
broadly different antibiotic activities: mupirocin exhibits high
specificity for the bacterial isoleucyl-tRNA synthetase and
subsequent inhibition of protein synthesis,[7] whereas the
T
he rapid rise of antibiotic resistance creates an urgent need
for new antibiotics. One strategy to meet the diminishing
returns on traditional antibiotics is to covalently link combi-
nations of existing antibiotics to produce novel hybrids. These
hybrid antibiotics exhibit enhanced bioactivity and pharma-
cology compared to the parent compounds. This strategy has
a synergistic effect because it improves activity against drug-
resistant bacteria, expands the spectrum of the individual
compounds, and reduces the potential for new resistance.[1]
A
drawback of synthetic hybrid antibiotics is that the partner
compound or the covalent linker may hinder target binding.
Furthermore, the effective concentration of both species may
be reduced if each compound targets disparate cellular sites.[2]
Naturally occurring hybrid antibiotics, such as the marine
[*] Z. D. Dunn, Prof. B. Li
Department of Chemistry
University of North Carolina at Chapel Hill
Carolina Center for Genome Sciences
Chapel Hill, NC, 27599 (USA)
E-mail: boli@email.unc.edu
W. J. Wever, Dr. N. J. Economou, Prof. A. A. Bowers
Division of Chemical Biology and Medicinal Chemistry
Eshelman School of Pharmacy
University of North Carolina at Chapel Hill
Chapel Hill, NC, 27599 (USA)
[**] We acknowledge Dr. Wilfred van der Donk and Dr. Christopher T.
Walsh for helpful discussions. We thank Dr. Gary Pielak, Andrew
Chan, Kevin Santa Maria, and Dr. Christopher Neumann for careful
reading of the manuscript. Financial support was provided by the
University of North Carolina at Chapel Hill and NIH grant
5R00GM099904-03 (B.L.).
Figure 1. A) Structures of thiomarinols, pseudomonic acids, and
dithiolopyrrolones. B) Gene cluster for thiomarinol. Open arrows
indicate ORFs with homology to the mupirocin pathway; blue ORFs
are homologous to DTP biosynthetic genes; black ORFs are unique to
the thiomarinol pathway; red ORFs, TmlU and HolE, the targets of this
study, have counterparts in the mupirocin and holomycin pathway,
respectively. They are represented as red arrows instead of open and
blue arrows for clarity.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201411667.
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
proposed holomycin mechanism of action involves inhibition
of bacterial transcription.[8] Importantly, despite these differ-
ences, the hybrid is more potent than its constituents, with
enhanced activity against many drug-resistant pathogens
including methicillin-resistant Staphylococcus aureus
(MRSA).[3a]
Thiomarinolꢀs hybrid structure generates at least two
thought-provoking questions. How did nature come to couple
these two distinct moieties? What are the mechanistic
benefits of combining these seemingly unrelated antibiotic
motifs? Here, we begin to answer the first question by
detailed characterization (and functional reassignment) of the
tmlU and holE gene products.
The thiomarinol gene cluster was discovered by whole
genome
sequencing
of
Pseudoalteromonas
spp.
SANK73390.[9] The genes are present on a 97 kb plasmid
comprising a polyketide synthase (PKS) portion and a non-
ribosomal peptide synthetase (NRPS) portion, which are
responsible for synthesizing the marinolic acid and holothin
segments, respectively (Figure 1B). Subsequent genetic dele-
tions by Simpson and co-workers identified TmlU as a key
enzyme responsible for coupling the holothin and marinolic
acid moieties.[9,10] TmlU was initially assigned as an amide
ligase on the basis of its significant sequence identity with the
amide ligases NovL (20.7%), CouL (21.1%), CloL (19.7%),
and SimL (18.5%), which catalyze amide bond formation in
the biosynthesis of the aminocoumarin antibiotics novobio-
cin, coumermycin, chlorobiocin, and simocyclinone, respec-
tively (Scheme 1A, see Figure S1 in the Supporting Informa-
tion).[11]
We tested whether TmlU could similarly act as a stand-
alone amide ligase by generating recombinant TmlU in
E. coli. The proposed amine donor for TmlU-mediated amide
coupling, holothin, could be readily accessed by total syn-
thesis in five steps with 13% overall yield.[13] The C7 fatty acyl
monic acid, marinolic acid (8), was not readily available. The
C8 fatty acyl monic acid, pseudomonic acid A (PAA), which is
commercially available (Sigma), was used instead. The epoxy
group in PAAwas reduced in three steps to give pseudomonic
acid C (PAC, 9) with a C-10,11 trans-olefin, similar to
marinolic acid.
Scheme 1. A) Reported mechanism of amide formation in the biosyn-
thesis of simocyclinone catalyzed by SimL.[12] B) Our proposed mecha-
nism of thiomarinol formation catalyzed by TmlU/HolE. ATP=adeno-
sine triphosphate, AMP=adenosine monophosphate, PPi=inorganic
pyrophosphate, CoASH=coenzyme A.
Despite investigating a large number of conditions, TmlU
failed to yield the anticipated product with the substrates
holothin and either PAA or PAC. This observation led us to
reexamine the assignment of TmlU. A Phyre2 homology
model (Figure S2) indicated close structural similarity to the
SrfA-C termination module of the nonribosomal peptide
synthetase responsible for surfactin biosynthesis.[14] More-
over, we found that TmlU has 14% sequence identity to
MupU, a putative acyl-CoA ligase in the mupirocin biosyn-
thetic pathway (Figures S1 and S3). As a result, we wondered
whether TmlU might activate marinolic acid by linking to
CoA or an acyl carrier protein, which could then be
transferred onto the acceptor holothin through tandem
action of another enzyme from the cluster, and speculated
that the putative acyltransferase HolE might play this
auxiliary role.
ing the terminal acylation of holothin with acetyl-CoA.[15] It
was hypothesized that HolE was solely responsible for the
“background” acylation observed in isolates from Pseudo-
monas spp. SANK73390, which gives rise to a series of short-
chain xenorhabdin-like molecules.[16]
HolE was expressed in E. coli and purified to homoge-
neity. For in vitro reconstitution of enzyme activities, we
treated PAC and holothin with 1 mm purified TmlU and 1 mm
HolE in the presence of CoASH, MgCl2, and ATP. We
detected a significant peak for the pseudomonic acid C-
holothin (PAC-holothin, 10) conjugate, which displayed the
same molecular weight and retention time as a synthetic
standard (Figure 2A, B, and G; see the Supporting Informa-
tion for synthesis). Omitting CoA abolished PAC-holothin
production, which suggests that CoA is necessary for efficient
conversion (Figure 2C). Assays carried out in the absence of
HolE is a homologue of HlmA, which is present in the
holomycin pathway from Streptomyces clavuligerus, catalyz-
2
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
Figure 3. Kinetic measurement of TmlU with different substrates.
A) PAC, KM =6Æ1 mm and kcat =3.2Æ0.1 sÀ1. B) PAA,
Figure 2. Enzymatic production of PAC-holothin, a thiomarinol ana-
logue, in vitro by TmlU and HolE, in the presence of 1 mm ATP, 2 mm
MgCl2, and 1 mm CoASH at pH 7.5. A) Synthetic PAC-holothin stan-
dard, B) in vitro reconstitution of TmlU and HolE activity generating
PAC-holothin, C) control lacking CoA, D) control lacking TmlU, E) con-
trol lacking HolE, F) mass spectrum of PAC-CoA generated by TmlU
(calculated [M+H]+, 1234.4155), G) mass spectrum of PAC-holothin
generated enzymatically by TmlU and HolE (calculated [M+H]+,
639.2768).
KM =5.2Æ0.5 mm and kcat =3.0Æ0.1 sÀ1. C) Octanoic acid,
KM =0.5Æ0.1 mm and kcat =(5.0Æ0.3)ꢀ10À3 sÀ1
.
a lesser extent (Figure 4A). In particular, we measured the
kcat/KM value of TmlU for octanoic acid and found it to be
50000 fold less than those for PAC or PAA (Figure 3C).
Overall, TmlU appears selective for long and relatively
complex fatty acyl carboxylates.
TmlU or ATP failed to yield the expected product (Figure 2D
and Figure S6). Furthermore, when HolE was omitted, the
substrate PAC was consumed, but the final PAC-holothin
product was not observed (Figure 2E). Instead, a pseudo-
monic acid C CoA (PAC-CoA) adduct accumulated (Fig-
ure 2Fand Figures S7 and S8). These results demonstrate that
TmlU is an acyl-CoA ligase and that HolE catalyzes the
subsequent acyl-transfer step required for thiomarinol bio-
synthesis. With this understanding, we assessed the kinetics
and promiscuity of this two-step enzymatic process.
Kinetic parameters for TmlU were measured using
saturating concentrations of the cosubstrates CoA and ATP
with PAC as a substrate. The formation of the PAC-CoA
product was measured by a coupled assay with saturating
concentrations of HolE and 3-aminocoumarin. 3-Aminocou-
marin was used instead of holothin, because we found it to be
well accepted by HolE and more stable than holothin. Under
these conditions, TmlU displays a KM value of (6 Æ 1) mm for
PAC and a kcat value of (3.2 Æ 0.1) sÀ1 (Figure 3A). These
parameters are consistent with those reported for other acyl-
CoA ligases, such as 4-chlorobenzoate-CoA ligase.[17] Inter-
estingly, the use of PAA as substrate yielded similar values to
PAC (Figure 3B), thus suggesting that the presence of the
epoxy group does not affect the TmlU activity. Thus, the lack
of the epoxide moiety in thiomarinols is likely due to the
absence of epoxide-forming enzymes in thiomarinol biosyn-
thesis rather than the substrate selectivity of TmlU or HolE.
We additionally investigated the substrate scope against
a number of other carboxylic acids. Under high enzyme
concentration, TmlU was capable of activating acetic, octa-
noic, 2,4-dodecadienoic, and 2,4-decadienoic acids, albeit to
To assess the kinetics of HolE, the substrate PAC-CoA
was generated using TmlU under the conditions of full
conversion. HolE and holothin were subsequently added and
the conversion of PAC-CoA into PAC-holothin was observed.
For a fixed concentration of PAC-CoA (100 mm), holothin
displayed inhibitory activity against HolE at concentrations
above 20 mm (Figure S5). This prevented determination of the
kcat/KM value. Efforts instead turned to assessing its promis-
cuity at a fixed concentration of holothin and in the presence
of several different, commercially available acyl-CoA sub-
strates. HolE accepted linear CoA substrates of different
lengths, including propionyl-, hexanoyl-, octanoyl-, oleoyl-,
and dodecanoyl-CoA, readily converting all into the corre-
sponding acyl-holothin adducts (Figure 4B and Figure S12).
This finding is consistent with our observation regarding the
substrate tolerance of HlmA, the acetyltransferase in the
holomycin biosynthetic pathway.[15] The promiscuity of HolE
with respect to fatty acyl CoA derivatives suggests that it is
likely responsible for the formation of the xenorhabdins (7),
which were seen as pathway by-products by Simpson et al.[10]
Given the potential for new and useful hybrid antibiotics
from this pathway, we explored the promiscuity of the HolE/
TmlU pair by supplying the reaction with various amine
donors and measuring the conversion to PAC-amine products
over a fixed time (Figure 4C and Figure S10). The HolE/
TmlU pair could readily attach PAC to a variety of primary
amines including 3-aminocoumarins, but was less effective
with a series of aryl amines. Overall, an adjacent substrate
carbonyl group appears useful or important for recognition of
the amine donor. This promiscuity stands in contrast to the
related enzyme systems, SimL and CouL from simocyclinone
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Figure 5. Mechanism of thiomarinol production by TmlU and HolE
compared to assembly-line tethered mechanisms to create PKS/NRPS
hybrids. A) The condensation domain of a NRPS directly adds a PKS
product to a growing peptide chain. Reactions occur on the assembly line.
B) A stand-alone transglutaminase-like domain (TGH) catalyzes the trans-
fer between tethered PKS and NRPS products in the biosynthesis of
andrimid.[21] C) Tailoring enzymes TmlU and HolE create a NRPS/PKS
hybrid with released products of discrete PKS and NRPS pathways. KS:
ketosynthase, AT: acyltransferase, ACP: acyl carrier protein, C: condensa-
tion domain, A: adenylation domain, PCP: peptidyl carrier protein.
amenable to combinatorial biosynthesis. Creation of new
natural products by combinatorial biosynthesis has been
successfully explored in glycosyltransferases and several
other natural product systems.[20]
Figure 4. Substrate promiscuity of TmlU and HolE. A) Incorporation of
carboxylic acids into thiomarinol by 5 mm TmlU in the presence of
saturating concentrations of HolE and holothin. Activity was measured
by the formation of the acyl-holothin products at l=360nm and
normalized to activity with PAC as substrate. [a] 1 mm TmlU was used
in this assay. [b] 2,4-DDA: (2E,4E)-2,4-dodecadienoic acid. [c] 2,4-DA:
(2E,4E)-2,4-decadienoic acid. B) Activity of HolE with acyl-CoA deriva-
tives as substrates to generate acyl-holothin products, measured as
described in (A). SNAC: N-acetyl cysteamine thioester. C) Activity of
HolE with different amines as substrates. Cosubstrate PAC-CoA was
generated by TmlU. Activity was measured by integration of ion
intensities in mass spectra.
It is intriguing to consider the implications for the
evolution of the pathway and the future of thiomarinol or
similar hybrid molecules in light of this new understanding.
Both the biosynthetic logic and the antibiotic mechanism
must have dictated the role of the TmlU/HolE pair. MupU,
the TmlU homologue in mupirocin biosynthesis, loads the
fatty acyl carboxylate in pseudomonic acid B onto the stand-
alone acyl carrier protein (ACP) MacpE for additional
tailoring.[22] In contrast, thiomarinol biosynthesis lacks
a MacpE-like stand-alone ACP. Instead, an additional ACP
and ketosynthase domain appear to have been incorporated
into the large PKS subunits, possibly replacing MacpE and
MupU, respectively. TmlU would then be free to facilitate
alternative post-PKS tailoring (i.e. ligation to holothin via
HolE), by taking advantage of the promiscuity of HolE, which
provides a branch point for further molecular diversity.
Although there is no clear evidence that the mupirocin
pathway preceded the thiomarinol pathway in terms of
evolution, the observations that these late-stage modifications
have been subsumed by larger pathway enzymes and that
other enzymes (TmlU) are repurposed suggests that thiomar-
inol is the more recent of the two. The pronounced biological
activity and reduced antibiotic resistance of the hybrid
molecule would further substantiate this idea. It remains to
be determined how adding holothin to a monic acid aids its
and coumermycin biosynthesis, respectively. Both SimL and
CouL reportedly show strong preference for native amino-
coumarin derived substrates.[11b,12,18] The biological activities
of PAC-aminocoumarin and other analogues generated in this
study will be the subject of future studies.
TmlU/HolE points to an intriguing convergent strategy
for creating PKS/NRPS hybrid molecules. While many known
hybrid linkages are installed on the PKS/NRPS assembly line
(Figure 5A,B),[19] TmlU/HolE makes use of a fully mature
PKS product, which is reactivated by a stand-alone acyl-CoA
ligase (TmlU) and transferred to a fully mature NRPS
product by a stand-alone acyltransferase (HolE; Figure 5C).
This strategy does not require direct participation of assembly
proteins or attachment to a carrier protein and is highly
4
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
[10] A. C. Murphy, D. Fukuda, Z. S. Song, J. Hothersall, R. J. Cox,
mechanism of action: the crystal structure of Ile-tRNA
synthetase bound pseudomonic acid A shows the carboxylate
group jutting from the active site, uninvolved in the key
inhibitory binding event.[23] Although the cyclic disulfide in
holomycin was shown to be important for the antimicrobial
action,[24] additional mechanistic studies are needed to reveal
the role that holothin could play in the Ile-tRNA synthetase
steric space. The combination of the two molecules is
a potentially fortuitous evolutionary event, not easily pre-
dicted by a modern structure-based approach. We anticipate
that our characterization of TmlU and HolE will aid efforts to
gain insight into the evolution and confluence of biosynthetic
pathways.
C. L. Willis, C. M. Thomas, T. J. Simpson, Angew. Chem. Int. Ed.
2011, 50, 3271 – 3274; Angew. Chem. 2011, 123, 3329 – 3332.
[11] a) E. Schmutz, M. Steffensky, J. Schmidt, A. Porzel, S. M. Li, L.
Heide, Eur. J. Biochem. 2003, 270, 4413 – 4419; b) C. L. Free-
l Meyers, M. Oberthur, L. Heide, D. Kahne, C. T. Walsh,
Biochemistry 2004, 43, 15022 – 15036; c) M. Steffensky, S. M.
Li, L. Heide, J. Biol. Chem. 2000, 275, 21754 – 21760.
[12] M. Pacholec, C. L. F. Meyers, M. Oberthur, D. Kahne, C. T.
Walsh, Biochemistry 2005, 44, 4949 – 4956.
[13] T. Hjelmgaard, M. Givskov, J. Nielsen, Org. Biomol. Chem. 2007,
5, 344 – 348.
[14] a) A. Tanovic, S. A. Samel, L. O. Essen, M. A. Marahiel, Science
2008, 321, 659 – 663; b) A. M. Gulick, ACS Chem. Biol. 2009, 4,
811 – 827.
[15] B. Li, C. T. Walsh, Proc. Natl. Acad. Sci. USA 2010, 107, 19731 –
19735.
Received: December 4, 2014
[16] A. C. Murphy, S. S. Gao, L. C. Han, S. Carobene, D. Fukuda,
Z. S. Song, J. Hothersall, R. J. Cox, J. Crosby, M. P. Crump, C. M.
Thomas, C. L. Willis, T. J. Simpson, Chem. Sci. 2014, 5, 397 – 402.
[17] R. Wu, J. Cao, X. Lu, A. S. Reger, A. M. Gulick, D. Dunaway-
Mariano, Biochemistry 2008, 47, 8026 – 8039.
[18] a) U. Galm, M. A. Dessoy, J. Schmidt, L. A. Wessjohann, L.
Heide, Chem. Biol. 2004, 11, 173 – 183; b) U. Galm, S. Heller, S.
Shapiro, M. Page, S. M. Li, L. Heide, Antimicrob. Agents
Chemother. 2004, 48, 1307 – 1312.
Published online: && &&, &&&&
Keywords: antibiotics · biosynthesis · dithiolopyrrolone ·
.
enzymes · evolution
[1] a) C. K. Chu, T. J. Bardos, J. Med. Chem. 1977, 20, 312 – 314;
b) D. D. Long, D. G. Marquess, Future Med. Chem. 2009, 1,
1037 – 1050; c) T. Karoli, S. K. Mamidyala, J. Zuegg, S. R. Fry,
E. H. L. Tee, T. A. Bradford, P. K. Madala, J. X. Huang, S.
Ramu, M. S. Butler, M. A. Cooper, Bioorg. Med. Chem. Lett.
2012, 22, 2428 – 2433.
[2] a) V. Pokrovskaya, T. Baasov, Expert Opin. Drug Discovery
2010, 5, 883 – 902; b) S. Shapiro, J. Antibiot. 2013, 66, 371 – 386.
[3] a) H. Shiozawa, T. Kagasaki, T. Kinoshita, H. Haruyama, H.
Domon, Y. Utsui, K. Kodama, S. Takahashi, J. Antibiot. 1993, 46,
1834 – 1842; b) H. Shiozawa, T. Kagasaki, A. Torikata, N.
Tanaka, K. Fujimoto, T. Hata, Y. Furukawa, S. Takahashi, J.
Antibiot. 1995, 48, 907 – 909; c) H. Shiozawa, A. Shimada, S.
Takahashi, J. Antibiot. 1997, 50, 449 – 452.
[4] a) A. T. Fuller, G. Mellows, M. Woolford, G. T. Banks, K. D.
Barrow, E. B. Chain, Nature 1971, 234, 416 – 417; b) E. B. Chain,
G. Mellows, J. Chem. Soc. Perkin Trans. 1 1977, 318 – 322; c) E. B.
Chain, G. Mellows, J. Chem. Soc. Perkin Trans. 1 1977, 294 – 309;
d) J. P. Clayton, P. J. Ohanlon, N. H. Rogers, T. J. King, J. Chem.
Soc. Perkin Trans. 1 1982, 2827 – 2833.
[5] L. Ettlinger, E. Gaumann, R. Hutter, W. Kellerschierlein, F.
Kradolfer, L. Neipp, V. Prelog, H. Zahner, Helv. Chim. Acta
1959, 42, 563 – 569.
[19] a) L. Du, C. Sanchez, B. Shen, Metab. Eng. 2001, 3, 78 – 95;
b) A. C. Mercer, M. D. Burkart, Nat. Prod. Rep. 2007, 24, 750 –
773; c) C. M. Rath, J. B. Scaglione, J. D. Kittendorf, D. H.
Sherman in Comprehensive Natural Products II: Chemistry and
Biology, Vol. 1 (Eds.: L. Mander, H.-w. Liu), Elsevier, Amster-
dam, 2010, pp. 453 – 492; d) D. Boettger, C. Hertweck, Chem-
BioChem 2013, 14, 28 – 42.
[20] a) C. Fischer, L. Rodriguez, E. P. Patallo, F. Lipata, A. F. Brana,
C. Mendez, J. A. Salas, J. Rohr, J. Nat. Prod. 2002, 65, 1685 –
1689; b) M. Yang, M. R. Proctor, D. N. Bolam, J. C. Errey, R. A.
Field, H. J. Gilbert, B. G. Davis, J. Am. Chem. Soc. 2005, 127,
9336 – 9337; c) M. Perez, F. Lombo, I. Baig, A. F. Brana, J. Rohr,
J. A. Salas, C. Mendez, Appl. Environ. Microbiol. 2006, 72,
6644 – 6652; d) G. J. Williams, C. Zhang, J. S. Thorson, Nat.
Chem. Biol. 2007, 3, 657 – 662; e) R. W. Gantt, R. D. Goff, G. J.
Williams, J. S. Thorson, Angew. Chem. Int. Ed. 2008, 47, 8889 –
8892; Angew. Chem. 2008, 120, 9021 – 9024; f) R. W. Gantt, P.
Peltier-Pain, W. J. Cournoyer, J. S. Thorson, Nat. Chem. Biol.
2011, 7, 685 – 691; g) M. D. Shepherd, T. Liu, C. Mendez, J. A.
Salas, J. Rohr, Appl. Environ. Microbiol. 2011, 77, 435 – 441;
h) G. Wang, P. Pahari, M. K. Kharel, J. Chen, H. Zhu, S. G.
Van Lanen, J. Rohr, Angew. Chem. Int. Ed. 2012, 51, 10638 –
10642; Angew. Chem. 2012, 124, 10790 – 10794.
[6] W. D. Celmer, I. A. Solomons, J. Am. Chem. Soc. 1955, 77, 2861 –
2865.
[7] a) J. Hughes, G. Mellows, Biochem. J. 1978, 176, 305 – 318;
b) C. M. Thomas, J. Hothersall, C. L. Willis, T. J. Simpson, Nat.
Rev. Microbiol. 2010, 8, 281 – 289.
[8] a) B. Oliva, A. OꢀNeill, J. M. Wilson, P. J. OꢀHanlon, I. Chopra,
Antimicrob. Agents Chemother. 2001, 45, 532 – 539; b) B. Li, W. J.
Wever, C. T. Walsh, A. A. Bowers, Nat. Prod. Rep. 2014, 31, 905 –
923.
[21] P. D. Fortin, C. T. Walsh, N. A. Magarvey, Nature 2007, 448, 824 –
827.
[22] S. S. Gao, J. Hothersall, J. Wu, A. C. Murphy, Z. Song, E. R.
Stephens, C. M. Thomas, M. P. Crump, R. J. Cox, T. J. Simpson,
C. L. Willis, J. Am. Chem. Soc. 2014, 136, 5501 – 5507.
[23] L. F. Silvian, J. M. Wang, T. A. Steitz, Science 1999, 285, 1074 –
1077.
[9] D. Fukuda, A. S. Haines, Z. S. Song, A. C. Murphy, J. Hothersall,
E. R. Stephens, R. Gurney, R. J. Cox, J. Crosby, C. L. Willis, T. J.
Simpson, C. M. Thomas, Plos ONE 2011, 6, e18031.
[24] a) B. Li, C. T. Walsh, Biochemistry 2011, 50, 4615 – 4622; b) B. Li,
R. R. Forseth, A. A. Bowers, F. C. Schroeder, C. T. Walsh,
ChemBioChem 2012, 13, 2521 – 2526.
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Biosynthesis
The biosynthetic mechanism responsible
for generating the antibiotic thiomarinol
was elucidated. In contrast to previous
hypotheses, TmlU acts as a CoA-ligase
and works in tandem with a second
enzyme, acyltransferase HolE, to link two
antimicrobial warheads pseudomonic
acid and holothin, creating a hybrid anti-
biotic (see scheme).
Z. D. Dunn, W. J. Wever, N. J. Economou,
A. A. Bowers, B. Li*
&&&& — &&&&
Enzymatic Basis of “Hybridity” in
Thiomarinol Biosynthesis
6
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
These are not the final page numbers!