SPECIAL FEA TURE: PERSPECTIVE

Biosynthetic origin of natural products isolated from

marine microorganism–invertebrate assemblages
T. Luke Simmons*, R. Cameron Coates*, Benjamin R. Clark*, Niclas Engene*, David Gonzalez†, Eduardo Esquenazi‡,
Pieter C. Dorrestein†§¶储, and William H. Gerwick*¶储
*Scripps Institution of Oceanography, Departments of †Chemistry and Biochemistry, §Pharmacology, and ‡Biology, and ¶Skaggs School
of Pharmacy and Pharmaceutical Sciences, University of California at San Diego, La Jolla, CA 92093

Edited by Jerrold Meinwald, Cornell University, Ithaca, NY, and approved December 10, 2007 (received for review October 16, 2007)

In all probability, natural selection began as ancient marine microorganisms were required to compete for limited resources. These
pressures resulted in the evolution of diverse genetically encoded small molecules with a variety of ecological and metabolic roles.
Remarkably, many of these same biologically active molecules have potential utility in modern medicine and biomedical research.
The most promising of these natural products often derive from organisms richly populated by associated microorganisms (e.g.,
marine sponges and ascidians), and often there is great uncertainty about which organism in these assemblages is making these in-
triguing metabolites. To use the molecular machinery responsible for the biosynthesis of potential drug-lead natural products, new
tools must be applied to delineate their genetic and enzymatic origins. The aim of this perspective is to highlight both traditional
and emerging techniques for the localization of metabolic pathways within complex marine environments. Examples are given from
the literature as well as recent proof-of-concept experiments from the authors’ laboratories.

E
xtensive drug discovery and small Of 20 compounds deriving from (or turing methods, and methods of chemical
molecule screening programs over inspired by) marine natural products detection are enabling the physical separa-
the past two decades have shown that are currently or were recently in tion of associated species for subsequent
marine organisms to be rich clinical trials for the treatment of cancer chemical analysis. Additionally, mass spec-
sources of structurally diverse and highly (5), 15 were isolated from sponges, tuni- tral imaging is proving exceptionally pow-
bioactive natural products (1). The molec- cates, and mollusks, with only 5 coming erful for identifying the location of natural
ular architectures of marine metabolites directly from a microorganism. How- products in mixed species assemblages.
are distinct from those of their terrestrial ever, based on biosynthetic parallels, Finally, gene probing techniques, such as
relatives in that the physicochemical re- distribution in taxonomically diverse or- catalyzed reporter deposition (CARD)–
quirements of adaptation to an aqueous ganisms, and, in a few cases, subsequent FISH or in situ hybridization, can identify
world, the biosynthetic pathways used, and direct isolation from a producing micro- those organisms with the genetic capacity
even the elements employed in crafting organism, 16 of these anticancer mole- to produce specific molecules of interest.
their arsenal of defensive molecules are cules actually derive from microbial This article is a review of advances made
quite different (2). As a consequence of sources and only 4 derive from macroor- in each of these three areas as reported in
their structural diversity and uniqueness, ganisms. For example, dolastatin 10 the literature and provides some new and
marine natural products are providing a (1; Fig. 1), originally isolated from the original results from our laboratories on
prominent share of the recent clinical and marine gastropod Dolabella auricularia, these subjects.
preclinical lead compounds for the treat- was later shown to originate from cya-
nobacteria of the genera Symploca and Results and Discussion
ment of various diseases, most promi-
Lyngbya upon which they feed (7–10). Microorganism Isolation Approaches. Con-
nently cancer (3).
The powerful PKC-activating cancer cell fronted with a mixture of species that
At first analysis, the sources of these
toxin bryostatin (2) was isolated from produces important secondary metabo-
important leads are taxonomically diverse
the Californian bryozoan Bugula lites, the simplest approach is to sepa-
and include sponges, tunicates, corals,
neritina, but the biosynthetic capacity to rate cell types and then analyze these
mollusks, fungi, and sediment-derived bac-
make this unique macrolide has been for the compound(s) of interest. It is
teria. However, there is growing recogni- localized to an associated bacterium En- plausible, however, that the natural
tion that the ‘‘collected source’’ for these dobugula sertula (11). Although the eco- product under study is produced in one
molecules is not necessarily the ‘‘meta- logical roles of most marine natural species, excreted, and then assimilated
bolic source.’’ That is, marine bacteria and products are not well understood, most by a second species. For example, ef-
cyanobacteria, either assimilated by an appear to be involved in defense against forts to localize peptides in the tunicate
invertebrate grazer (e.g., sea hare grazing predation or competition for space and Lissoclinum patella showed that they
on cyanobacteria) or growing in associa- nutrients (12–14). occur predominately in the ascidian
tion with an invertebrate host in a symbi- It has been surprisingly difficult to dem-
otic or commensal relationship, are onstrate that an invertebrate-associated
frequently the true origin of these mole- microorganism is unequivocally responsi- Author contributions: T.L.S., P.C.D., and W.H.G. designed
cules (4–6). The growing appreciation of ble for the production of a given second- research; T.L.S., R.C.C., B.R.C., N.E., D.G., E.E., and P.C.D.
performed research; T.L.S., R.C.C., B.R.C., N.E., E.E., P.C.D.,
microbial metabolism in the production of ary metabolite (6). However, recent and W.H.G. analyzed data; and T.L.S., R.C.C., P.C.D., and
many of the marine pharmaceutical leads advances on a number of fronts are pro- W.H.G. wrote the paper.
comes from evidence in several sectors. viding new and powerful methods for The authors declare no conflict of interest.
Ultimately, firm knowledge of the natural investigating these relationships. For ex- This article is a PNAS Direct Submission.
product-producing organism, the ‘‘holder ample, in the plant-endophyte arena, the 储To whom correspondence may be addressed. E-mail:
of the genes,’’ is essential to numerous anticancer compounds podophyllotoxin pdorrest@ucsd.edu or wgerwick@ucsd.edu.
branches of science and technology, in- and camptothecin were recently discov- This article contains supporting information online at www.
cluding ecology, biosynthesis, and natural ered as products of endophytic microbes pnas.org/cgi/content/full/0709851105/DC1.
product drug development. (15, 16). New advances in cell sorting, cul- © 2008 by The National Academy of Sciences of the USA

www.pnas.org兾cgi兾doi兾10.1073兾pnas.0709851105 PNAS 兩 March 25, 2008 兩 vol. 105 兩 no. 12 兩 4587– 4594
 surfaces of the sponge larvae, suggesting
O S vertical transmission of symbionts between
H H
N N N sponge generations. However, no natural
N N N
products have been reported from this
O OCH3O OCH3O bacterium to date (27).
In another recent example, the spatial
Dolastatin 10 (1)
H O distribution of bacteria within the
HO O
sponge Tethya aurantium was mapped by
OAc
H3COOC using denaturing gradient gel electro-
O O H phoresis (DGGE) and 16S rDNA clone
Herbadysidolide (3)
library analysis. These studies showed
O
HO that a new phylotype of Flexibacteria
O O Cl3C N CCl3 occurred in the sponge cortex whereas
HO
O
the endosome contained the cyanobac-
O OH O NH terium Synechococcus sp. Interestingly,
O COOCH3
N 16S rDNA sequences for a ␤-proteobac-
S Dysidenin (4) teria were found throughout the endo-
Bryostatin 1 (2) some and cortex (27).
OCH3 O Various gene probing methods (e.g.,
O N S FISH, PCR screening, Southern analysis,
N
H
N and genome sequencing) can demonstrate
O the occurrence of a biosynthetic gene in a
O
OH NH HN particular isolated cell type. For example,
OH OH O OH O the Haygood laboratory demonstrated a
O N H N reduction in the abundance of a ketosyn-
O N
S O thase gene from the putative bryostatin
OCH3
H3CO O gene cluster in B. neritina larvae that had
O
O
Patellamide A (6) been treated with antibiotics (28). In a
OH O OH OH second example, a genome sequence of
HO the cyanobacterium Prochloron sp., a sym-
O biont in the tunicate Lissoclinum patella,
O O
identified the cyanobacterium as the bio-
OH
OH
HO synthetic source of the ribosomally en-
OCH3 HO coded peptides patellamides A (6) and C
Swinholide A (5) Pseudopterosin A (7) (18, 19).
Eukaryotic dinoflagellates are common
Fig. 1. Structures 1–7. coral symbionts, and soft corals in the
Gorgonacea are particularly well studied
for their bioactive diterpenoids. The pseu-
tunic and not in the symbiotic cyanobac- ized to the associated cyanobacterium, dopterosins (e.g., 7) for example are po-
terium Prochloron sp. (17). However, Oscillatoria spongeliae (22). In studies of tent antiinflammatory terpenes obtained
subsequent heterologous expression and the cellular localization of several com- from the soft coral Pseudopterogorgia elisa-
genome sequence analysis indicated that plex metabolites isolated from the bethae and are used in commercially avail-
these peptides are indeed genetically sponge Theonella swinhoei, including able skin care products. Investigation of
encoded for and biosynthesized by the swinholide A (5) and the peptide theo- the site of biosynthesis of the pseudopt-
cyanobacterium (18, 19). In a few cases, palauamide, a combination of dissection erosins by using differential centrifugation
and despite considerable technological and differential centrifugation gave cell produced a 99% pure preparation of the
difficulties, successful segregation of cell preparations that were chemically ana- dinoflagellate Symbiodinium sp. This mac-
types and their cultivation has allowed lyzed. Curiously, theopalauamide was roalgal preparation was found to contain
isotope incorporation studies to demon- found in a filamentous bacterial popula- compound 7, and provision of either
strate new biosynthesis and, hence, iden- tion morphologically related to Beggia- [14C]NaHCO3 or 3H-geranylgeraniol pyro-
tification of the producing organism toa, whereas swinholide A (6) was phosphate led to radioactive HPLC peaks
(4, 6, 20). localized to a mixed unicellular bacterial corresponding to the pseudopterosins
Sponge tissues are extremely rich in fraction (23, 24). This latter finding is at (29). From these and other data, it was
bacteria that can comprise up to 40% of odds with our subsequent isolation of concluded that the dinoflagellate was re-
their biomass (21). Early efforts to lo- swinholide A from a free-living marine sponsible for pesudopterosin biosynthesis.
calize natural products to specific cell cyanobacterium, Geitlerinema sp. (25). However, a subsequent patent has ap-
types used formalin or glutaraldehyde- The microbial communities associated peared that identifies associated bacteria
fixed sponge-cyanobacterial tissues from with various sponges have been investi- as the biosynthetic source of the pseu-
the tropical sponge Dysidea herbacea gated by using both culture-dependent dopterosins, pointing out the danger of
(Fig. 2). The tissues were disrupted, and and culture-independent methodologies concluding too much from a cell separa-
individual cells were separated by using (4). Recently, a culturable ␣-proteobacte- tion followed by chemical analysis or undi-
a fluorescence-activated cell sorter rial symbiont was obtained from several luted radiolabeling experiments (30).
(FACS). Chemical analysis of the sorted sponges, including Mycale laxissima (26). Although most marine microorganisms
cell types showed sesquiterpenoid com- A fluorescently labeled 16S rRNA gene are not easily separated from complex
pounds (e.g., 3) to be physically associ- probe was prepared for FISH analysis to assemblages by direct manipulation, fila-
ated with the sponge cells, whereas visualize gene distribution. Intriguingly, mentous cyanobacteria grow to dimen-
chlorinated peptides (e.g., 4) were local- the bacteria were concentrated on the sions more amenable to manual isolation

4588 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0709851105 Simmons et al.

 interface with many types of chromato-
graphic supports such as reverse phase
(C18, C4, etc.), hydrophilic interaction
liquid chromatography (HILIC), cationic
exchange, or even combinations of multi-
ple resins [see supporting information (SI)
Fig. 8] (45, 46). For example, we detected
(and subsequently isolated and confirmed
by NMR) the anticancer lead compound
iejimalide A (8; Fig. 4) from a 0.1-␮g
crude cyanobacterial extract. This was sig-
nificant because this macrolide had only
Fig. 2. Photomicrographs of a thin section of sponge tissue showing embedded cyanobacterial cells. (A) previously been found from an Okinawan
Cross-section of the sponge Dysidea herbacea stained with O-toluidine showing composition of cells types. tunicate (47, 48). Detection was achieved
(B) Autofluorescence (522 nm) of cyanobacterial cells due to chlorophyll. [Reprinted with permission from by using a 100-␮m fused silica capillary
ref. 92 (Copyright 2005, Springer).] column loaded with 10 cm of C18 resin
and a 5-␮m tip. The estimated detection
limit of iejimalide A by this technique is
(e.g., Lyngbya majuscula filaments are up larly noteworthy improvements in mass in the picomole-to-femtomole range. By
to 100 ␮m wide and several centimeters spectrometry include the ability to ana- contrast, traditional LC-MS with an LCQ
long). However, the tough polysaccharide lyze small quantities of material while instrument and 4.6-mm-diameter HPLC
sheath surrounding cyanobacterial fila- increasing both resolution and mass ac- column required ⬎100 ␮g of iejimalide-
ments can harbor a multitude of hetero- curacy. These developments have been containing crude extract for detection
trophic bacteria that are difficult to largely driven by the proteomics com- (estimated detection limit in the micro-
remove (31–33) (Fig. 3). We and others munity, usually in the context of discov- mole-to-nanomole range).
find that the dominating heterotrophic ering disease biomarkers (40–43). These Nanospray Fourier transform (FT)–ion cyclotron
bacteria associated with these sheaths are efforts have spurred considerable devel- resonance MS (ICRMS). High-resolution MS
gram-negatives from the phyla Proteobac- opment of new pre-separation methods, is another extremely useful technique for
teria and Bacteroidetes. These include the new ionization approaches, improved analyzing crude mixtures for the presence
␣-proteobacteria (Caulobacter, Thalasso- mass filters, faster scan rates for im- of a specific natural product (49). We
bius, and Rhodospirillaceae), ␤-proteobac- proved time scale detection, and new used a complement of nanospray and FT-
teria (Achromobacter), ␥-proteobacteria detection methods, only some of which ICRMS to analyze the crude extract of
(Pseudomonas and Oceanospirillum), and have found their way into the routine the iejimalide A-producing cyanobacteria
Bacteroidetes (Flavobacterium and Flexi- analysis of natural products. These de- from Papua New Guinea. Despite the fact
bacter) (34, 35). Interestingly, many of velopments in MS forecast that soon we that the crude extract showed ⬎40 ions in
these same bacterial groups are also may be able to perform rigorous struc- the region m/z 675–705, the sodiated ieji-
commonly associated with sponges and ture elucidation on submicrogram quan- malide A peak at m/z 701.418 (calculated
tunicates (36, 37). tities within certain metabolite classes. m/z 701.414) was well resolved and firmly
Creating axenic cultures of cyanobacte- In the following sections, we present identified the presence of iejimalide A
ria is extraordinarily difficult, and rela- several examples from the authors’ labo- (8; see SI Fig. 9). In total, this extract dis-
tively few are reported in the literature. ratories that demonstrate how some of played several hundred unique molecular
Previous investigations proposed that ax- these very sensitive MS technologies can ions in a single spectrum (including sev-
enic cultures were produced by using irra- be used in the discovery of new marine eral consistent with novel hydroxylated
diation and antibiotics; however, proof of natural products, and how variations in analogs of 8), and these were only resolv-
their monoculture state used lack of bac- these methods can help in identifying able because of the high-resolution at-
terial growth on multiple media types as the true biosynthetic sources of these tributes of FT-ICRMS. It should be noted
evidence (31–33, 38). Because the media, compounds. that additional characterization of these
nutrient, and environmental combinations Electrospray ionization (ESI) and metabolites can be achieved by combining
were not exhaustive, doubt remains as to MALDI MS are revolutionary soft ioniza- the high resolution of FT-ICRMS with
whether these cultures were truly free of tion methods that enable MS measure- tandem MS (49). This ‘‘snapshot’’ of the
bacteria. Many oligotrophic organisms ments of chemically sensitive biomolecules secondary metabolome of a cyanobacte-
grow very slowly and to low cell densities; (44). However, for either of these tech- rium will surely become an important dis-
therefore, their presence can be easily niques to be effective, there is an absolute covery tool as natural product scientists
overlooked by classic culture-dependant requirement for ionization. ESI has be- learn to process and use such rich
tests for axenicity (39). Hence, it remains come the most common approach of ion- datasets.
uncertain whether sheath-associated bac- ization in the natural product chemist’s MALDI-TOF-MS. The species diversity asso-
teria are involved in the production or arsenal because it can be interfaced with ciated with complex marine assemblages
modification of natural products reported LC capabilities, thereby connecting molec- is well suited for analysis by another
from cyanobacteria. ular mass data with elution characteristics powerful MS technique, matrix-assisted
of a particular molecule. These data laser desorption/ionization–time-of-flight
Chemical Detection Methods. Microanalysis greatly expedite the assignment of a natu- MS (MALDI-TOF-MS). The utility of
by mass spectrometry (MS). In conjunction ral product as either novel or known MALDI-MS for the analysis of biologi-
with the physical and genetic approaches (dereplication). cal samples (e.g., single cells and tissues)
described above, chemical detection In addition to traditional HPLC-ESI- has been demonstrated for both biomed-
methods have improved significantly, in MS, capillary LC-ESI-MS has proven ex- ical and natural science applications.
particular MS, and these are becoming traordinarily useful in proteomic research. For example, MALDI-techniques were
highly useful in locating the biosynthetic Capillary LC requires extremely small applied to investigate entire colonies of
origin of specific metabolites. Particu- sample sizes (⬍1.0 ␮g) with the ability to the cyanobacterium Microcystis to under-

Simmons et al. PNAS 兩 March 25, 2008 兩 vol. 105 兩 no. 12 兩 4589
 pseudomolecular ion at m/z 840 (see SI
Fig. 10), a value corresponding to the
molecular formula of apratoxin A
[C45H69N5O8S ⫹ H].
We also used MALDI-MS in combina-
tion with ESI-FTMS, LC-ESI-MS, and
NMR to detect and characterize a novel
analog of curacin A (10) (54–56). Aside
from two double-bond isomers, curacins B
and C (57), and the 8-desmethyl analog
curacin D (58), no other naturally occur-
ring analogs have yet been isolated. The
biosynthesis of the distinctive cyclopropyl
ring has been the focus of detailed molec-
ular genetic and mechanistic biochemistry
investigations because it appears to in-
volve the intermediacy of a cryptic chlori-
nation reaction (59–61). MALDI analysis
of a single L. majuscula filament from
Curaçao revealed a pseudomolecular ion
Fig. 3. Trichome structure of the filamentous marine cyanobacterium Lyngbya majuscula-3L. Shown is
at m/z 374 [M⫹H]⫹, consistent with the
DAPI staining and epifluorescent imaging at ⫻1,000 magnification using a Zeiss Axioskop (filter set #2, presence of curacin A. Another peak at
emission 420 nm⫹). A, cyanobacterial phycoerythrin in filament tip (orange); B, filament sheath- m/z 372 [M⫹H]⫹ was also observed that
associated bacterial DNA (blue). did not correspond to any known cya-
nobacterial metabolite. FT-ICRMS of the
extract from this strain allowed character-
stand the diversity and distribution of for the identification of known and new ization of the 372 peak as exactly 2.015
hepatotoxic oligopeptides from these secondary metabolites. The analysis of Da less than curacin A, for a deduced
natural communities (50). Recently, intact filament assemblages enables the molecular formula of C23H33NOS; thus, it
‘‘intact cell MALDI-TOF’’ (ICM) was localization of specific secondary metab- represented a potentially intriguing struc-
used to assess the diversity of microbes olites within specific filament types. tural homolog of curacin A. The pure
from the tissues of marine invertebrates MALDI-TOF-MS analysis was per- compound was isolated and characterized
(e.g., sponges) by ‘‘proteometric cluster- formed on fresh single filaments of a by NMR and other spectroscopic tech-
ing,’’ allowing both the dereplication of strain of L. bouillonii known to produce niques and shown to be an oxidized form
known secondary metabolites and pro- apratoxin A (9), a potent mammalian of curacin A with a thiazole rather than a
viding a taxonomic characterization of cell cytotoxin (52, 53). In this example, thiazoline ring. Previously, we had pro-
the species present (51). a single filament was placed on a duced analog 11 via semisynthesis from
We have extended the utility of MALDI target plate and coated with an natural curacin A and given it the trivial
MALDI-TOF-MS to the analysis of sin- ␣-cyano-4-hydroxycinnamic acid matrix. name ‘‘curazole’’ (62). Based on detection
gle filaments of marine cyanobacteria Data analysis readily gave a strong of the [M⫹H] ion from intact cyanobacte-
rial filaments and through direct analysis
of unfractionated crude extracts, it seems
CH3 CH3 O
H certain that curazole (11) is a genuine
H3CO O N H
N natural product of this L. majuscula strain.
O H
CH3
CH3
CH3
OH
O O Contrary to literature precedent (60, 63),
OCH3
N O our sequence analysis revealed that the
Iejimalide A (8)
N
O oxidase required for the thiazoline-
CH3
N S
H
H OH to-thiazole transformation is not present
H HN
N O
N
S
in the curacin A gene cluster, nor is it
N
O NH
O O found immediately up- or downstream of
O
Dercitamide (13) Latrunculin (14) the cluster (60). This suggests that this
O O
N
O O
Cl
R transformation of the thiazoline to the
O
N
N thiazole is not encoded by a dedicated
O S H
N O
protein or protein domain.
O Jamaicamide A, R = Br (15) There are pitfalls in the direct
HO
Apratoxin A (9)
Jamaicamide B, R = H (16) MALDI analysis of biological tissues, as
Cl3C N CCl3
illustrated by our attempt to detect ieji-
S
O
malide A (8) from cyanobacterial fila-
H H
OCH3
Curacin A (10)
N
H
HN O ments. As noted, iejimalide A had been
S
previously detected by traditional or-
S N
N H 13-demethylisodysidenin (17)
ganic extraction of the collected biomass
OCH3
Curazole (11) H followed by NMR and FT-ICRMS ap-
N OMe
Br OH O O HO Br N proaches. MALDI-TOF-MS of single
S
O NH[CH2]4NH O
filaments from this sample failed to de-
O CCl3
O N N O tect any 8. Consistent with this finding,
Br
Aerothionin (12)
Br
Barbamide (18)
pure iejimalide A, once it became avail-
able, also failed to provide a detectable
Fig. 4. Structures 8–18. ion signal by MALDI-TOF, likely be-

4590 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0709851105 Simmons et al.

 Step 1 Step 2 Step 3 powerful method for directly observing
A Collect Mass both new and known secondary metab-
Define Raster Display the Spatial
Spectra at Each Distribution of an m/z olites. Can MALDI-TOF also provide
Raster Point. Value. insight into the identity of metabolite-
producing species found in complex
%
% environmental assemblages? As noted,
%
m/z this has been an extraordinarily diffi-
%
%
m/z
b Step 4 cult phenomenon to study because co-
m/z
%
% m/z Superposition existing strains are difficult to grow as
m/z
m/z
m/z
single pure strains, and many artifacts
a
b are introduced during sample prepara-
c c
= Laser raster point. tion. An emerging strategy showing
= Sample of mixed cultures. promise is MALDI-TOF imaging (75–
= MALDI target plate. 78). MALDI imaging allows the spatial
b
%
a
c
a localization of precise molecular
masses to particular cell types that are
m/z
then observed through assignment of
Average mass spectrum
specific colors to a given mass (Fig. 5).
Although this approach has not previ-
B ously been applied to the analysis of
natural products, it has been used to
identify biomarkers of disease and the
accumulation of therapeutics in whole-
animal histological preparations
(79, 80).
Here, we briefly demonstrate the
power and utility of this technology in
Fig. 5. MALDI-TOF imaging to map the distribution of natural products in biological tissues. (A) two preliminary studies (E.E., R.C.C.,
Schematic for procedure of MALDI-TOF imaging. (B) From left to right, MALDI-TOF imaging of protonated T.L.S., D.G., W.H.G., and P.C.D., un-
curacin A (12) at m/z 374 (red), sodiated jamaicamide B (19) at m/z 511 (green), the simultaneous published work). First, two L. majuscula
superposition of both m/z 374 and 511, superposition of both masses plus photograph of filaments, strains (JHB from Jamaica and 3L from
photograph of both filaments identified as strains ‘‘3L’’ (⫽ curacin A producer) and JHB[ ⫽ jamaicamide
Curaçao) were mixed and analyzed by
A (18) and B (19) producer], and partial mass spectrum of JHB strain showing the molecular ion cluster of
jamaicamide B with its distinctive chlorine isotope pattern.
MALDI-TOF imaging. It is known that
the JHB strain produces jamaicamide B
(16) and the 3L strain produces curacin
cause of its lack of basic nitrogen atoms. location of two brominated metabolites, A (10) (54, 62, 81). Sodiated jamaicam-
This example demonstrates that MALDI aerothionin (12) and homoaerothionin, ide B has an m/z of 511 and is colored
and ESI are complementary approaches within a sponge–microorganism com- green on the image, whereas protonated
and that both should be used in natural plex. Although these experiments re- curacin A has an m/z of 374 and is
product dereplication and discovery corded increased bromine in a particular shown in red (Fig. 5). The MALDI-
programs. cell type (spherulous sponge cells), it TOF image clearly distinguishes the two
Other important MS advances stem was not certain that the detected bro- producing strains by the secondary me-
from the development of novel ionization mine was actually part of the target tabolites they contain. Moreover, it is
methods and new mass analyzers. For ex- compounds. Subsequently, the location clear that jamaicamide B is chlorinated
ample, desorption electrospray ionization of the polyaromatic alkaloid dercitamide from the observed 35Cl/37Cl isotopic
(13) was directly imaged by using epiflu- ratio.
(DESI), electrospray-assisted laser desorp-
orescence and laser-scanning confocal In a second example, we analyzed the
tion ionization (ELDI), and direct analysis
microscopy in conjunction with TEM. marine sponge Dysidea herbaceae. Be-
in real time (DART) are newly developed
Interestingly, this study localized the cause of its relative availability, abun-
ionization methods that analyze samples dance of described natural products, and
natural product exclusively to the
directly and with minimal preparation associated cyanobacterium (Oscillatoria
sponge cells (Oceanapia sagittaria) and
(64–68). The new orbitrap Fourier trans- spongeliae), D. herbaceae has developed
not to the associated microbes as was
form (FT) mass analyzer can provide high as a useful model for secondary metabo-
expected (73). In another early study,
resolution without the required expertise the imaging of latrunculin B (14) in the lite and biosynthetic gene cluster local-
that an FT-ICR instrument requires (69, sponge Negombata magnifica used im- ization studies (see below, Gene-based
70), a characteristic that has made this the munogold staining with light microscopy imaging methods). Here, we prepared
instrument of choice in metabolomic stud- and TEM. These studies also revealed thin sections of sponge-cyanobacterium
ies (71). Although none of these more that sponge cells were the site of storage tissue using a cryogenic microtome for
recent methods have yet been applied to and presumed biosynthesis (74). In light analysis by epifluorescence microscopy
the study of cyanobacterial natural prod- of subsequent natural product and bio- and MALDI imaging. Fig. 6 shows the
ucts or complexes of coexisting species, it synthetic investigations, these results portion of a thin section of sponge-
is likely that some will offer advantages with latrunculin B are particularly in- cyanobacterial tissue used in the imag-
over current methods. triguing because related compounds ing analysis by autofluorescence at two
have been shown to be produced by mi- wavelengths and then by MALDI-TOF.
Imaging Techniques. Background. In 1983, croorganisms using assembly-line type The mass at m/z 530 (Fig. 6 D, E, and
Faulkner and colleagues (72) used x-ray synthases typical of bacteria (2, 11). G) and molecular ion isotope cluster
analysis coupled with transmission elec- MALDI imaging. It is clear that single- (data not shown) are consistent with a
tron microscopy (TEM) to examine the filament MALDI-TOF analysis is a hexachlorinated peptide such as 13-

Simmons et al. PNAS 兩 March 25, 2008 兩 vol. 105 兩 no. 12 兩 4591
 horseradish peroxidase detection system.
Larvae treated with this probe were la-
beled only in a thin band, the pallial
sinus, which was shown to contain a new
␥-proteobacterial strain, Endobugula ser-
tula. Although it was not conclusively
demonstrated at this point whether the
KS was specifically involved in bryosta-
tin biosynthesis, subsequent gene clon-
ing work indicated that it very likely was
(11). Our analysis shows a 98% se-
quence identity at the DNA base level
between the original KS and the BryB
KS present in the subsequently cloned
and sequenced bryostatin gene cluster.
In 1996, we reported the structure of
the chlorinated peptide, barbamide (18),
from a Curaçao collection of the marine
cyanobacterium L. majuscula (84). Barb-
Fig. 6. MALDI-TOF imaging of the marine sponge Dysidea herbaceae. (A) MALDI target plate with thin amide is strongly molluscicidal to the
section in place. (B) Autofluorescence of D. herbaceae thin section at 590 nm⫹. (C) Autofluorescence of intermediate snail host involved in schis-
D. herbaceae thin section at 420 nm⫹. (D) MALDI image of m/z 530. This mass and the complex molecular tosomiasis, Biomphaleria glabrata. The
ion isotope cluster suggests its identity as the hexachlorinated peptide 13-demethylisodysidenin (20). (E)
halogen atoms comprising the unique
Autofluorescence (420 nm) image overlaid by MALDI image at m/z 530 indicating localization of com-
pound with this mass. (F) MALDI image of unknown compound m/z 1,028. (G) Autofluorescence (420 nm)
trichloromethyl group are located at a
image overlaid with m/z 530 (red) and m/z 1,028 (green) showing the differential localization of these two position without an adjacent activating
molecules. functionality, and this suggested a novel
route for their biochemical incorpora-
tion (85, 86). These barbamide haloge-
demethylisodysidenin (17). Another terization in conjunction with imaging nases and related enzymes have since
mass, m/z 1,028 (Fig. 6 F and G), is for are MALDI imaging-FT-ICR and the been characterized as a new class of
an unknown metabolite with a very dif- LTQ imaging system with tandem MS ‘‘radical halogenases’’ capable of direct
ferent distribution in the sponge matrix capabilities (77, 83). halogenation of unreactive carbon atoms
and clearly illustrates the power of Gene-based imaging methods. The Haygood in natural products, such as methyl
MALDI-TOF imaging in differentially laboratory has examined the biosyn- groups (87–89). Early on, we recognized
localizing secondary metabolites from thetic source of the anticancer bryo- the striking structural similarity of barb-
complex environments. statins (e.g., 2). Originally isolated from amide to a number of Dysidea‘‘sponge’’
At present, MALDI imaging is lim- the bryozoan B. neritina, the bryostatin chlorinated peptides. We and others
ited by its spatial resolution, due in part structures are consistent with a micro- proposed that the cyanobacterial symbi-
to the requirement for a uniform crys- bial polyketide synthase origin. To inves- ont (Oscillatoria spongeliae) found grow-
talline matrix coating. The matrix coat- tigate the metabolic origin of 2, this ing among the sponge cells was likely
ing allows for molecular diffusion of the group (28) used degenerate primers to responsible for chlorinated peptide pro-
analyte in the crystalline lattice, limiting clone a component of the polyketide duction (90, 91). Cell sorting experi-
ments, as noted earlier, supported this
the spatial resolution to the size of the synthase, the ketosynthase (KS) domain,
hypothesis as the peptides were recov-
crystals or the width of a typical N2 la- from both larval and adult forms of the
ered from the purified cyanobacterial
ser (⬇100 ␮m). With improvements in bryozoan. A total of 9 unique KS
fraction (90, 91).
both matrix deposition and laser tech- sequences were recovered, and one pre- The DNA sequences of the barbamide
nologies, it is anticipated that resolution dominated in the larvae. This latter halogenases BarB1 and BarB2 allowed a
in the range of 7–10 ␮m will become sequence was used to produce a biotin- gene-based approach to investigate and
available (82). Finally, two recent ad- labeled RNA gene probe that was confirm the identity of the producing
vances that will allow structural charac- subsequently visualized with an avidin- organism in this symbiosis (92). First,
homology-based cloning from the mixed
sponge/cyanobacterial tissue (Fig. 2)
yielded a barB1 homolog, dysB1, encod-
ing a protein sequence with 93% amino
acid identity to BarB1. The dysB1 se-
quence in combination with cyanobacte-
rial-specific 16S rDNA sequences was
used in FISH experiments to probe thin
sections of the sponge containing the
associated cyanobacterial symbiont. The
cyanobacteria were identified in the thin
sections by their chlorophyll fluores-
cence and by localization of the fluores-
Fig. 7. CARD-FISH analysis of Dysidea herbacea showing autofluorescence (522 nm) of cyanobacterial cently labeled cyanobacterial-specific
cells due to chlorophyll (A) and probe-specific hybridization with the complimentary oligonucleotide to 16S rDNA gene probe. The crucial ex-
dysB1, a homolog of the barbamide barB1 biosynthetic gene (B). [Reprinted with permission from ref. 92 periment showed colocalization of the
(Copyright 2005, Springer).] fluorescently labeled dysB1 probe and

4592 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0709851105 Simmons et al.

16S rDNA probe in cyanobacterial cells spongeliae provides an example of RNA and localized by subsequent MALDI
but not in sponge cells (Fig. 7). Cya- detection (92); uncovering the gene for imaging.
nobacterial cells were thereby unequivo- patellamide biosynthesis in isolated cells Exciting hybrid technologies are be-
cally shown to possess the halogenase of Prochloron sp. from the tunicate Lis- coming available that will increase the
mRNA necessary for chlorinated pep- soclinum patella is an example of detec- rigor and ease of demonstrating the site
tide biosynthesis. tion at the DNA level (18, 19). A con- of biosynthesis in complex assemblages.
siderable hurdle in applying most of Cell isolations and genome determina-
Conclusions these gene-based technologies is that the tions are now possible on a single cell
The presence of a natural product biosynthesis needs to be sufficiently un- [e.g., using multiple displacement poly-
within a particular cell type can provide derstood such that the genes can be rec- merase chain reaction technology (96)],
only circumstantial evidence as to its ognized with certainty. The standard from which biosynthetic clusters can be
biosynthetic origin. Above, various older methods for demonstrating that a given deduced and candidate structures gener-
techniques for showing the presence of gene is involved in the biosynthesis of a ated by using established biosynthetic
a natural product have been reviewed compound involve either gene knockout logic. Structures of sufficient structural
and newer experimental approaches dis- experiments or heterologous expression or anticipated biological interest can be
located in cultured or collected biomass
cussed. Some of these are appropriate of the biosynthetic pathway. Both of
with increasing facility (97). Alterna-
for analysis of isolated cells, including these methods require genetic manipula-
tively, synthetic genes of substantial size
HPLC, NMR, and, most powerfully, tions that can be extremely challenging (e.g., 50,000 bp) can be produced with
MS. Other techniques are well suited in a little studied marine organism. To optimized codon usage and promoter
for sample imaging, such as x-ray or flu- date, the function of most marine natu- sequences for effective heterologous ex-
orescence microscopy, and MALDI ral product biosynthetic genes have been pression (98–100). Indeed, the coming
approaches. Nevertheless, all of these deduced through bioinformatic analysis. years will witness amazing technological
approaches fall short of providing un- At the protein level, little has been advances and allow us to ask and answer
equivocal proof for the site of biosyn- achieved because there is so little long standing, profound, and currently
thesis of a given compound. knowledge of biosynthetic proteins in unimagined questions in chemical
Conceptually, demonstrating the bio- general and even less so from members biology.
synthesis of a natural product in a given of complex marine environments. How- Details of the experiments described in
cell type can be achieved on chemical, ever, knowledge and techniques are this article can be found in SI Methods
protein, or genetic levels. At the chemi- emerging that will allow powerful inter- and Data.
cal level, demonstration of new com- rogation of the site of biosynthesis using
pound synthesis in an isolated cell type biosynthetic proteins. An increasing ACKNOWLEDGMENTS. We thank P. Crews and K.
is compelling and may involve any of a number of marine biosynthetic pathways Tenney (University of California, Santa Cruz) for
samples of the sponge Dysidea herbaceae, B.
number of analytical chemical ap- are being cloned and characterized. Palenik and B. Brahamsha for use of their micro-
proaches, possibly involving isotope- Complementing this is a dramatic in- scopes, L. Gerwick for bioinformatics analysis of
labeled precursors. At the genetic level, crease in the number and diversity of the curacin A gene cluster, R. Grindberg for anal-
ysis of BryB in the bryostatin gene cluster, T. Ma-
DNA or RNA sequences encoding the marine genome and metagenome se- tainaho and the captains/crews of the Golden
biosynthetic enzymes can be localized, quences (93–95). In principle, this ge- Dawn and Teleta Dive Boats for their assistance
either in isolated cells or through imag- netic sequence information can be used with cyanobacterial collections, and the govern-
ing techniques. The CARD-FISH to identify diagnostic peptide fragments ments of Papua New Guinea and Curaçao for
permission to make these collections. This work
analysis of the sponge/cyanobacterium from putative biosynthetic enzymes that was supported by National Institutes of Health
complex Dysidea herbacea/Oscillatoria can be liberated by protease digestion Grants CA52955 and CA100851.

1. Newman DJ, Cragg GM (2007) J Nat Prod 70:461– 477. 15. Eyberger AL, Dondapati R, Porter JR (2006) J Nat Prod 30. Kerr RG, Brueck T (2007) PCT Int Appl WO 20077037791.
2. Ramaswamy AV, Flatt PM, Edwards DJ, Simmons TL, 69:1121–1124. 31. Adams DG (2001) Microbiol Today 28:131–133.
Han B, Gerwick WH (2006) in Frontiers in Marine Bio- 16. Puri SC, Verma V, Amna T, Qazi GN, Spiteller M (2005) 32. Ferris MJ, Hirsch CF (1991) Appl Environ Microbiol
technology, eds Proksch P, Müller WEG (Horizon Bio- J Nat Prod 68:1717–1719. 57:1448 –1452.
science, Norfolk, England), pp 175–224. 17. Salomon CE, Faulkner DJ (2002) J Nat Prod 65:689 – 692. 33. Gerwick WH, Roberts MA, Proteau PJ, Chen J (1994)
3. Simmons TL, Andrianasolo E, McPhail K, Flatt P, Ger- 18. Schmidt EW, Nelson JT, Rasko DA, Sudek S, Eisen JA, J Appl Phycol 6:143–149.
wick WH (2005) Mol Cancer Ther 4:333–342. Haygood MG, Ravel J (2005) Proc Natl Acad Sci USA 34. Paerl HW (1982) in The Biology of Cyanobacteria, eds
4. Piel J (2006) Curr Med Chem 13:39 –50. 102:7315–7320. Carr NG, Whitton BA (Univ of California Press, Berke-
5. Simmons TL, Gerwick WH (2008) in Oceans and Human 19. Milne BF, Long PF, Starcevic A, Hranueli D, Jaspers M ley), pp 441– 461.
Health, eds Walsh P, Solo-Gabriele H, Fleming LE, Smith (2006) Org Biomol Chem 4:631– 638. 35. Zehr JP, Mellon M, Braun S, Litaker W, Steppe T, Paerl
SL, Gerwick WH (Elsevier, New York), in press. 20. Kobayashi J, Ishibashi M (1993) Chem Rev 93:1753– HW (1995) Appl Environ Microbiol 61:2527–2532.
6. Hildebrand M, Waggoner LE, Liu H, Sharp KH, Ridley 1769. 36. Bewley CA, Faulkner DJ (1998) Angew Chem Int Ed
CP, Haygood MG (2004) Nat Prod Rep 21:122–142. 21. Friedrich AB, Merkert H, Fendert T, Hacker J (1999) Mar
37:2162–2178.
7. Pettit GR, Kamano Y, Herald CL, Tuinman AA, Boettner Biol 134:461– 470.
37. Perez-Matos AE, Rosado W, Govind NS (2007) Antonie
FE, Kizu H, Schmidt JM, Baczynski L, Tomer K, Bontems 22. Ridley CP, Bergquist PR, Harper MK, Faulkner DJ,
van Leeuwenhoek 92:155–164.
RJ (1987) J Am Chem Soc 109:6883– 6885. Hooper JN, Haygood MG (2005) Chem Biol 12:397– 406.
38. Kraus MP (1966) Nature 211:310 –311.
8. Pettit GR, Kamano Y, Dufresne C, Cerny RL, Herald CL, 23. Bewley CA, Holland ND, Faulkner DJ (1996) Experientia
39. Giovannoni SJ, Stingl U (2005) Nature 437:343–348.
Schmidt JM (1989) J Org Chem 54:6005– 6006. 52:716 –722.
40. Tyers M, Mann M (2003) Nature 422:193–197.
9. Leusch H, Moore RE, Paul VJ, Mooberry SL, Corbett TH 24. Schmidt EW, Bewley CA, Faulkner DJ, Haygood MG
41. Marshall AG, Hendrickson CL, Emmett MR, Rodgers RP,
(2001) J Nat Prod 64:907–910. (1998) J Org Chem 63:1254 –1258.
Blakney GT, Nilsson CL (2007) Eur J Mass Spectrom
10. Williamson RT, Chapin EL, Carr AW, Gilbert JR, 25. Andrainansolo EH, Gross H, Goeger D, Musafija-Girt M,
Graupner PR, Lewer P, McKamey P, Carney JR, Gerwick McPhail K, Leal RM, Mooberry SL, Gerwick WH (2005) 13:57–59.
WH (2000) Org Lett 2:289 –292. Org Lett 7:1375–1378. 42. Liu E, Gerszten RE (2006) Cardiovasc Biomarkers 575–
11. Sudek S, Lopanik NB, Waggoner LE, Hildebrand M, 26. Enticknap JJ, Kelly M, Peraud O, Hill RT (2006) Appl 586.
Anderson C, Liu H, Patel A, Sherman DH, Haygood MG Environ Microbiol 72:3724 –3732. 43. Coon JJ, Syka JEP, Shabanowitz J, Hunt DF (2005) Bio-
(2007) J Nat Prod 70:67–74. 27. Thiel V, Neulinger SC, Staufenberger T, Schmaljohann Techniques 38:519 –523.
12. Paul VJ, Puglisi MP, Ritson-Williams R (2006) Nat Prod R, Imhoff JF (2007) FEMS Microbiol Ecol 59:47– 63. 44. Zhang J (2007) Proteomics Clin Appl 1:805– 819.
Rep 23:153–180. 28. Davidson SK, Allen SW, Lim GE, Anderson CM, Haygood 45. Boersema PJ, Divecha N, Heck AJR, Mohammed S (2007)
13. Pawlik JR (1993) Chem Rev 93:1911–1922. MG (2001) Appl Environ Microbiol 67:4531– 4537. J Proteome Res 6:937–946.
14. Faulkner DJ, Ghiselin MT (1983) Mar Ecol Prog Ser 29. Coleman AC, Kerr RG (2000) Tetrahedron 56:9569 – 46. Motoyama A, Venable JD, Ruse CI, Yates JR, III (2006)
13:295–301. 9574. Anal Chem 78:5109 –5118.

Simmons et al. PNAS 兩 March 25, 2008 兩 vol. 105 兩 no. 12 兩 4593
47. Nozawa K, Tsuda M, Ishiyama H, Sasaki T, Tsuruo T, 65. Cooks RG, Ouyang Z, Takats Z, Wiseman JM (2006) 83. Verhaert PD, Conaway MCP, Pekar TM, Miller K (2007)
Kobayashi J (2006) Bioorg Med Chem 14:1063–1067. Science 311:1566 –1570. Int J Mass Spectrom 260:177–184.
48. Kobayashi J, Cheng J, Ohta T, Nakamura H, Nozoe S, 66. Shiea J, Huang M-Z, HSu H-J, Lee C-Y, Yuan C-H, Beech 84. Orjala J, Gerwick WH (1996) J Nat Prod 59:427– 430.
Hirata Y, Ohizumi Y, Sasaki T (1988) J Org Chem I, Sunner J (2005) Rapid Commun Mass Spectrom 85. Sitachitta N, Rossi J, Roberts MA, Gerwick WH, Fletcher
53:6147– 6150. 19:3701–3704. M, Willis CL (1998) J Am Chem Soc 120:7131–7132.
49. Feng X, Siegel MM (2007) Anal Bioanal Chem 67. Haefliger OP, Jeckelmann N (2007) Rapid Commun 86. Sitachitta N, Marquez BM, Williamson RT, Rossi J, Rob-
389:1341–1363. Mass Spectrom 21:1361–1366. erts MA, Gerwick WH, Nguyen V-A (2000) Tetrahedron
50. Welker M, Brunke M, Preussel K, Lippert I, von Döhren 68. Sampson JS, Hawkridge AM, Muddiman DC (2006) J Am 56:9103–9113.
H (2004) Microbiology 150:1785–1796. Soc Mass Spectrom 17:1712–1716. 87. Chang Z, Flatt P, Gerwick WH, Nguyen V-A, Willis CL,
51. Dieckmann R, Graeber I, Kaesler I, Szewzyk U, von 69. Lim H-K, Chen J, Sensenhauser C, Cook K, Subrah- Sherman DL (2002) Gene 296:235–247.
Döhren H (2005) Appl Microbiol Biotechnol 67:539 – manyam V (2007) Rapid Commun Mass Spectrom 88. Flatt PM, O’Connell SJ, McPhail KL, Zeller G, Willis CL,
548. 21:1821–1832. Sherman DL, Gerwick WH (2006) J Nat Prod 69:938 –944.
52. Luesch H, Yoshida WY, Moore RE, Paul VJ, Corbett TH 70. Hu Q, Noll RJ, Li H, Makarov A, Hardman M, Cooks RG 89. Galonic DP, Vaillancourt FH, Walsh CT (2006) J Am
(2001) J Am Chem Soc 123:5418 –5423. (2005) J Mass Spectrom 40:430 – 443. Chem Soc 128:3900 –3901.
53. Luesch H, Chandra SK, Raya RM, DeJesus PD, Orth AP, 71. Breitling R, Pitt AR, Barrett MP (2006) Trends Biotech- 90. Faulkner DJ, Unson MD, Bewley CA (1994) Pure Appl
Walker JR, Belmonte JCI, Schultz PG (2006) Nat Chem nol 24:543–548. Chem 66:1983–1990.
Biol 2:158 –167. 72. Thompson JE, Barrow KD, Faulkner DJ (1983) Acta Zool 91. Flowers AE, Garson MJ, Webb RI, Dumdei EJ, Charan RD
54. Gerwick WH, Proteau PJ, Nagle DG, Hamel E, Blohkin A, 64:199 –210. (1998) Cell Tissue Res 292:597– 607.
Slate DL (1994) J Org Chem 59:1243–1245. 73. Salomon CE, Deerinck T, Ellisman MH, Faulkner DJ 92. Flatt PM, Gautschi JT, Thacker RW, Musafija-Girt M,
55. Verdier-Pinard P, Sitachitta N, Rossi JV, Sackett DL, (2001) Mar Biol 139:313–319. Crews P, Gerwick WH (2005) Mar Biol 147:761–774.
Gerwick WH, Hamel E (1999) Arch Biochem Biophys 74. Gillor O, Carmeli S, Rahamim Y, Fishelson Z, Ilan M 93. Cheng Q, Xiang L, Izumikawa M, Meluzzi D, Moore BS
370:51–58. (2000) Mar Biotechnol 2:213–223. (2007) Nat Chem Biol 3:557–558.
56. Wipf P, Reeves JT, Day BW (2004) Curr Pharm Des 75. McDonnell LA, Heeren RMA (2007) Mass Spectrom Rev 94. Dunlap WC, Battershill CN, Liptrot CH, Cobb RE, Bourne
10:1417–1437. 26:606 – 643. DG, Jaspers M, Long PF, Newman DJ (2007) Methods
57. Yoo H-D, Gerwick WH (1995) J Nat Prod 58:1961–1965. 76. Signor L, Varesio E, Staack RF, Starke V, Richter WF, 42:358 –376.
58. Marquez B, Verdier-Pinard P, Hamel E, Gerwick WH Hopfgartner G (2007) J Mass Spectrom 42:900 –909. 95. Rusch DB, Halpern AL, Sutton G, Heidelberg KB, Wil-
(1998) Phytochemistry 49:2387–2389. 77. Taban IM, Altelaar AFM, van der Burgt YEM, McDon- liamson S, Yooseph S, Wu D, Eisen JA, Hoffman JM,
59. Vaillancourt FH, Yeh E, Vosburg DA, O’Connor SE, nell LA, Heeren RMA, Fuchser J, Baykut G (2007) J Am Remington K, et al. (2007) PLoS Biol 5:e77.
Walsh CT (2005) Nature 436:1191–1194. Soc Mass Spectrom 18:145–151. 96. Lasken RS (2007) Curr Opin Microbiol 10:1–7.
60. Chang Z, Sitachitta N, Rossi JV, Roberts MA, Flatt PM, Jia 78. McLean JA, Ridenour WB, Caprioli RM (2007) J Mass 97. McAlpine JB, Bachmann BO, Piraee M, Tremblay S,
J, Sherman DH, Gerwick WH (2004) J Nat Prod 67:1356 – Spectrom 42:1099 –1105. Alarco A-M, Zazopoulos E, Farnet CM (2005) J Nat Prod
1367. 79. Cornett DS, Reyzer ML, Chaurand P, Caprioli RM (2007) 68:493– 496.
61. Gu L, Jia J, Liu H, Håkansson K, Gerwick WH, Sherman Nat Methods 4:828 – 833. 98. Menzella HG, Reisinger SJ, Welch M, Kealey JT,
DH (2006) J Am Chem Soc 128:9014 –9015. 80. Hsieh Y, Chen J, Korfmacher WA (2007) J Pharmacol Kennedy J, Reid R, Tran CQ, Santi DV (2006) J Ind
62. Verdier-Pinard P, Lai J-Y, Yoo H-D, Yu J, Marquez B, Toxicol Methods 55:193–200. Microbiol Biotechnol 33:22–28.
Nagle DG, Nambu M, White JD, Falck JR, Gerwick WH, 81. Edwards DJ, Marquez BL, Nogle LM, McPhail K, Goeger 99. Kodumal SJ, Patel KG, Reid R, Menzella HG, Welch M,
et al. (1998) Mol Pharmacol 53:62–76. DE, Roberts MA, Gerwick WH (2004) Chem Biol 11:817– Santi DV (2004) Proc Natl Acad Sci USA 101:15573–
63. Ming L-J, Epperson JD (2002) J Inorg Biochem 91:46 –58. 833. 15578.
64. Keil A, Talaty N, Janfelt C, Noll RJ, Gao L, Ouyang Z, 82. Chaurand P, Schriver KE, Caprioli RM (2007) J Mass 100.Smith HO, Hutchison CA, III, Pfannkoch C, Venter JC
Cooks RG (2007) Anal Chem 79:7734 –7739. Spectrom 42:476 – 489. (2003) Proc Natl Acad Sci USA 100:15440 –15445.

4594 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0709851105 Simmons et al.