Chemistry & Biology

Article

A Bacterial Source for Mollusk Pyrone PolyketidesZhenjian Lin,1 Joshua P. Torres,6 Mary Anne Ammon,6 Lenny Marett,1,2 Russell W. Teichert,3 Christopher A. Reilly,4

Jason C. Kwan,1 RonaldW. Hughen,2 Malem Flores,6 Ma. Diarey Tianero,1 Olivier Peraud,1 James E. Cox,5 Alan R. Light,2

Aaron Joseph L. Villaraza,7 Margo G. Haygood,8 Gisela P. Concepcion,6 Baldomero M. Olivera,3 and Eric W. Schmidt1,3,*1Department of Medicinal Chemistry2Department of Anesthesiology3Department of Biology4Department of Pharmacology and Toxicology

University of Utah, Salt Lake City, UT 84112, USA5Metabolomics Core Research Facility and Department of Biochemistry, University of Utah Health Sciences, Salt Lake City, UT 84112, USA6Marine Science Institute7Institute of Chemistry

University of the Philippines, Diliman, Quezon City 1101, Philippines8Department of Environmental and Biomolecular Systems, OGI School of Science and Engineering, Oregon Health & Science University,

Beaverton, OR 97006, USA

*Correspondence: ews1@utah.eduhttp://dx.doi.org/10.1016/j.chembiol.2012.10.019

SUMMARY

In the oceans, secondary metabolites often protectotherwise poorly defended invertebrates, such asshell-less mollusks, from predation. The origins ofthese metabolites are largely unknown, but many ofthem are thought to be made by symbiotic bacteria.In contrast, mollusks with thick shells and toxicvenoms are thought to lack these secondary metab-olites because of reduced defensive needs. Here, weshow that heavily defended cone snails also occa-sionally contain abundant secondary metabolites,g-pyrones known as nocapyrones, which aresynthesized by symbiotic bacteria. The bacteria,Nocardiopsis alba CR167, are related to widespreadactinomycetes that we propose to be casual symbi-onts of invertebrates on land and in the sea. Thenatural roles of nocapyrones are unknown, but theyare active in neurological assays, revealing thatmollusks with external shells are an overlookedsource of secondary metabolite diversity.

INTRODUCTION

Many marine animals are protected by arsenals of defensive

compounds, which have been exploited in the discovery of phar-

maceuticals (Putz and Proksch, 2010). These chemicals can be

divided into two groups: proteins and peptides (large molecules)

and secondary metabolites (small molecules). Protein-based

toxins include venoms, which are synthesized by the animals

themselves and are encoded in the animal genomes (Terlau

and Olivera, 2004). Small molecules are highly diverse in terms

of their chemical structures, and they are more commonly asso-

ciated with soft-bodied, ‘‘defenseless’’ animals, such as ascid-

ians, sponges, and nudibranch mollusks (Blunt et al., 2011). In

contrast to the casewith protein toxins, increasing evidence indi-

cates that symbiotic bacteria synthesize at least some of these

Chemistry & Biology 20,

small molecules, although only a few systems have been studied

in any detail (Piel, 2009; Kubanek et al., 1997).

Mollusks lacking shells are sometimes defended instead by

secondary metabolites (Benkendorff, 2010). Often, the mole-

cules seem to have a dietary origin (Fontana, 2006; Gavagnin

et al., 1994). However, a widely occurring group of mollusk

secondary metabolites, pyrones and structurally related polyke-

tides, are generated in the animals de novo, possibly even by the

mollusks themselves (Davies-Coleman and Garson, 1998; Marin

et al., 1999; Cimino et al., 1987; Ireland and Scheuer, 1979). The

close structural resemblance of these polyketides to bacterial

metabolites has led to the hypothesis that symbiotic bacteria,

and not the animals, produce the compounds in question

(Cutignano et al., 2009). This idea is complicated by the fact

that there are many convergent biosynthetic routes to pyrones

(Busch and Hertweck, 2009), so the ultimate source is a matter

of debate. Molecular or genetic evidence is still lacking to defin-

itively tie production of any gastropod metabolite to symbiotic

bacteria or to any other source.

Mollusks with shells, such as cone snails and turrids (Terlau

and Olivera, 2004; Lopez-Vera et al., 2004), synthesize animal

genome-derived peptide toxins, both for predation and for

defense (Olivera, 2002). In contrast to their shell-less relatives,

mollusks with external shells are usually believed to be deficient

in secondary metabolites (Benkendorff, 2010), leading re-

searchers to avoid the animals as sources of new natural prod-

ucts. For venom-containing mollusks, like cone snails, peptide

toxins should provide ample defense against predators.

Recently, we began to question whether cone snails might

also harbor symbiont-derived metabolites. If so, this might chal-

lenge current ideas about distribution and roles of marine natural

products and suggest that otherwise protected animals are

good sources for discovery. Because by far most mollusks are

defended by shells, this would open a new area for compound

discovery. We reported that at least three species of cone snails

contain abundant associated actinomycete bacteria, which are

well known to produce secondary metabolites in culture (Peraud

et al., 2009). There are now many reports of actinomycetes iso-

lated from marine animals. However, these actinomycetes have

73–81, January 24, 2013 ª2013 Elsevier Ltd All rights reserved 73

Figure 1. Actinobacteria in C. rolani

(A) Live sample of C. rolani used in this study.

(B) Cultivated N. alba CR167 from C. rolani.

See also Tables S1 and S2.

Figure 2. Structures of Nocapyrones H-Q (1–10) and A-C (11–13)See also Figures S1–S3.

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

not been tied to animal biology, and one hypothesis is that these

might usually be derived from very casually occurring bacteria or

even from spores. On land, actinomycetes have been shown to

form casual symbioses with a variety of insects, where they

secrete compounds implicated in defense against infection

(Schoenian et al., 2011; Kroiss et al., 2010). Because of the abun-

dance of actinomycetes in certain cone snails, we thought it

possible that actinomycetes might also play central roles in

marine organisms.

Here, we report a case in which actinomycete bacteria

synthesize polyketide pyrones in the cone snails, Conus rolani

and Conus tribblei. These results demonstrate that bacteria

produce gastropod compounds within their hosts, answering

a long-standing question about origins of mollusk metabolites.

The results also suggest that shelled mollusks are overlooked

sources of secondary metabolic diversity.

RESULTS

Neuroactive Nocapyrones from Cone-Snail-AssociatedBacteriaStrain N. alba CR167 was isolated from the venom duct of

C. rolani, collected in Mactan Island, Philippines, in 2007 with

informed consent and permission from local and national

governments, as previously reported (Figure 1) (Peraud et al.,

2009). The extract from N. alba CR167 was neuroactive, as as-

sayed by calcium imaging of dissociated dorsal root ganglion

(DRG) neurons frommice. Activity was followed through a series

of chromatographic steps, leading to purification of 13 com-

pounds (Figure 2). We identified ten compounds, nocapyrones

H-Q (1–10), as well as the three previously described com-

pounds, nocapyrones A-C (11–13) (Schneemann et al., 2010).

Their structures were elucidated on the basis of one- and two-

dimensional nuclear magnetic resonance (NMR) and by mass

spectrometry (Figure S1 available online; Table S1). Briefly,

high-resolution mass spectrometry was consistent with the

molecular formulae of each molecule, whereas tandem mass

spectrometry (MS/MS) showed that the nocapyrones followed

a very consistent fragmentation pattern (Table S2) for g-pyrones

(Porter and Baldas, 1971). 13C and 1H resonanceswere assigned

on the basis of 1H, 13C, COSY, HSQC, and HMBC spectra (Fig-

ure S1; Table S1).

Because configurational information is lacking for this

compound class, we determined the stereochemistry of 1-13.

74 Chemistry & Biology 20, 73–81, January 24, 2013 ª2013 Elsevier

First, we examined the configuration of the C-6 OH, present in

compounds 3, 6, and 8. The Mosher’s esters (Dale and Mosher,

1973) were synthesized and analyzed by 1H NMR, revealing that

C-6 in all compounds was in the S configuration (Figure S2). We

used the same method to examine the C-11 OH, present in

compounds 2, 6, and 11. In 2, C-11 was in the R configuration,

whereas for 6 and 11, it was in the S configuration (Figure S2).

We next sought to determine the configuration of the remote

methyl groups at C-10. Compounds 6 and 11 were isolated as

inseparable mixtures of diasteromers in the side chain. Careful

examination showed that11wasalsopreviously isolatedasadia-

stereomeric mixture in the initial report of its structure (Schnee-

mann et al., 2010). The diastereomeric ratio of 6was determined

as 1.3:1 by integration of O-methine protons at d 3.62 and 3.55

(Figure S1), with a similar ratio observed for 11. Comparison of

the chemical shifts and coupling constants ofO-methine protons

in 6 and 11 with literature values for (2R*,3S*)- and (2R*,3R*)-3-

methyltridecan-2-ol (Larsson et al., 2004) showed that the major

isomer in both 6 and 11 had a 10R*,11S* relative configuration.

Initially, upon isolation we thought that this was likely due to

partial racemization at the OH-bearing C-11 position, but the

Mosher’s analysis described above revealed that only single

C-11 isomers were present (Figure S2). Thus, compounds 6a

Ltd All rights reserved

Table 1. Nocapyrones Detected in Methanolic Extracts of Conus rolani

Tissue Date Collected

Pyrones Detected

1 3 12

Mean ± SD

(mg/g)a d (ppm)bMean ± SD

(mg/g) d (ppm)

Mean ± SD

(mg/g) d (ppm)

1 Whole snail (n = 1)c Sept. 2009 12.4 ± 2.7 1.67 1,250 ± 154 �0.87 1.57 ± 0.26 1.22

2 Whole snail (n = 3)c Sept. 2010 7.84 ± 0.34 0.97 507 ± 31 �0.34 10.2 ± 1.5 0.87

3 Whole snail (n = 4)c Nov. 2011 nd 129 ± 37 �0.68 nd

4 Venom duct (n = 20)c July 2011 0.404 ± 0.012d 1.30 11.0 ± 2.2d 2.05 0.366 ± 0.095d 1.88

4a Other anatomical

regions (n = 2)

July 2011 nd N.D. nd

5 Mucus (n = 4)c Oct. 2011 44.7 ± 5.2e 0.86 105 ± 7.2e �1.77 31.1 ± 2.48e 1.44

6 Whole snail and

mucus (n = 20)cMar. 2012 nd nd nd

Mean over three independent measurements. Data analysis was done using Analyst QS software 2.0. nd, below detection limit. Statistical limit of

quantification: 0.29 ng/ml; practical detection limit: 0.10 ng/ml. See also Figures S4 and S5.aMicrogram compound per gram of snail tissue dry weight (each snail weighs �1 g).bParts per million (ppm) error of high-resolution mass spectrometric measurement.cn, number of samples averaged to determine quantity but with each individual sample measured separately. In these cases, each snail contained the

compounds, and averaging was done for ease of presentation.dNumbers reflect microgram per venom duct rather than per gram dry weight.eNumbers reflect microgram per snail rather than per gram dry weight.

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

and 11a have the 10R, 11S configuration, whereas 6b and 11b

have the 10S, 11S configuration.

Using Mosher’s method, compound 7 was shown to be

present as an enatiomeric mixture, with a S:R ratio of 10:1 (Fig-

ure S3). Although in this case a primary alcohol was esterified,

whereas the Mosher’s method is used mainly for secondary

alcohols, an excellent literature precedent was available for

compounds with related functional groups: (2S)-2-methylhexyl

(2S)-3,3,3-trifluoro-2-methoxy-2-phenylpropanoate and (2R)-2-

methylhexyl (2S)-3,3,3-trifluoro-2-methoxy-2-phenylpropanoate

(Guintchin and Bienz, 2003). The Mosher’s ester of the major

isomer 7a exhibited chemical shifts of 4.24, 4.08 (AB of ABX,

JAB = 10.7 Hz, JAX = 6.6 Hz, JBX = 5.7 Hz, OCH2), which were

very similar to those reported in the literature precedent for the

S isomer (Figure S3). Likewise, 7b exhibited chemical shifts

consistent with the R isomer. Thus, the major 7a was present

as the S isomer.

Wewished to determine the remotemethyl center of 5 directly,

but no functional handle or standard compound was available. A

closely related fatty acid in the S configuration has a reported

optical rotation of [a]D = + 8.0 (Sonnet et al., 1990), whereas 5

has a measured rotation of [a]D = + 9.0. In addition, the major

isomers of related compounds 6 and 11 are in the S configura-

tion. Thus, we tentatively propose that 5 is likely also in the S

configuration.

Nocapyrones from Methanolic Extracts of C. rolaniWe noticed that nocapyrones were structurally related to previ-

ously reported polyketides from other gastropod mollusks.

Several pyrones are known to be made within mollusks based

upon labeling studies (Ireland and Scheuer, 1979; Di Marzo

et al., 1991), but whether by the animal, symbiotic bacteria, or

other organism was unknown. We hypothesized that N. alba

CR167 synthesizes pyroneswithin cone snail tissues.We already

Chemistry & Biology 20,

had demonstrated that actinobacteria live in the foot of snails,

indicating that cultivated strains, such asNocardiopsis sp.,might

be alive within the snails, rather than existing as inactive spores

(Peraud et al., 2009). However, this hypothesis still seemed

highly speculative because polyketides had never before been

reported from thickly shelled mollusks, such as cone snails.

Therefore, finding the Nocardiopsis pyrones within whole snail

tissue would provide strong support for the hypothesis.

To test this hypothesis, we extracted the whole animals with

ethanol and analyzed their organic content. Surprisingly, 1, 3,

and 12 were among the major organic components that could

be isolated from an individual snail collected in 2009. This

prompted us to perform a more detailed chemical analysis (Fig-

ure S4), collecting and analyzing snails from the same site near

Sogod, Cebu Island, during the period of 2009–2012 and deter-

mining the tissue localization and relative abundance of com-

pounds (Table 1). We synthesized a 13C-labeled pyrone as an

internal control to accurately calibrate the quantity of com-

pounds and performed control experiments to accurately deter-

mine concentrations in the snails (Figure S5). To ensure that

results would not be compromised by laboratory contamination,

work was divided between two geographical locations.

Compounds 1, 3, and 12 were variable in abundance in

C. rolani and C. tribblei samples and were sometimes not

present (Table 1; Figure S5). When present, the compounds

were relatively abundant, totaling between 0.01%–0.1% of snail

dry weight (not including the shell). Linearity was demonstrated

from calibration curves in the range 0.010 to 10.0 mg/ml with

correlation coefficients of at least 0.9996. Percent recovery of in-

jected standard (1 mg) from live snails was found to be 87.78% ±

3.71% (n = 3). This abundance is on par for what is commonly

found for other marine animal metabolites in better-studied

systems, such as sponges, ascidians, and soft-bodiedmollusks,

indicating that the compounds are present in a physiologically

73–81, January 24, 2013 ª2013 Elsevier Ltd All rights reserved 75

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

relevant concentration range. Variability of secondary metabo-

lites between samples is frequently encountered in marine

natural products (Donia et al., 2011).

We investigated the tissue specificity of nocapyrones. Com-

pounds were only found in two locations and were absent else-

where (Table 1). Nocapyrones appeared to be largely localized

to the mucus, with a small amount (�1% of the compounds)

found in the venom duct. Pyrones and other polyketides have

previously been isolated from themucusof soft-bodiedmollusks,

where they are proposed to function in defense or communica-

tion (Di Marzo et al., 1991; Sleeper and Fenical, 1977).

Origin of Nocapyrones in Cone SnailThe data described above strongly supported the hypothesis

that bacteria produce cone snail pyrones, but there are many

potential caveats that necessitated further study (Schmidt,

2008). For example, convergent evolution or horizontal gene

transfer could lead to more complex scenarios (Schmidt,

2008). To tie the bacteria to pyrone production within the

animals, we identified the probable pyrone biosynthetic gene

cluster and showed that both this cluster and N. alba bacteria

were present in the snails themselves.

At least three convergent routes lead to the biosynthesis of py-

rones in nature (Busch and Hertweck, 2009). Therefore, to iden-

tify the pyrone biosynthetic gene cluster, we sequenced the

genome of N. alba CR167 to an average of 250 3 coverage.

The resulting genome was assembled into 353 contigs (N50 =

44.1 kbp); because of this fragmentation we analyzed both

assembled contigs and raw reads to find the pyrones gene

cluster. The central metabolic genes in the strain were similar to

those from the previously sequenced Nocardiopsis dassonvillei

(GenBank CP002040) (Sun et al., 2010), although both synteny

and sequence identity were relatively low between the genomes.

Very recently, the genome sequence of Nocardiopsis alba ATCC

BAA-2165, cultivated from the digestive tract of honeybees in

Ohio, was published in GenBank (CP003788). This genome

was >95% DNA sequence identical to that of N. alba CR167

and shared 16 out of 18 identified biosynthetic gene clusters,

usually at >99% DNA sequence identity.

Pyrones are known to be made by various types of polyketide

synthase (PKS) proteins. Using BLAST and antiSMASH (Me-

dema et al., 2011) analysis, only three PKS gene clusters were

present in N. alba CR167, including a type II PKS and two type

I PKS pathways. No type III PKSs were present. Nearly identical

pathways were also found in the well-assembled N. alba ATCC

BAA-2165 genome. Of the three PKS clusters, only one con-

tained a methyltransferase. That PKS cluster was also the only

one that appeared to have the correct domain specificity and

module architecture to produce pyrones. In addition, there is

a known pyrone g-O-methyltransferase from the jerangolid

gene cluster (Julien et al., 2006). BLAST searching for homologs

in the genome revealed only a single significant hit, also corre-

sponding to this PKS-clustered methyltransferase. Thus,

genomic data strongly implicated the identified ncp cluster as

being responsible for nocapyrone biosynthesis (Figure 3).

To provide initial chemical evidence in support of the bio-

informatics results, the jerangolid-like methyltransferase NcpB

was produced in Escherichia coli, purified (Figure S6), and

used in enzyme assays with various substrates (Nelson et al.,

76 Chemistry & Biology 20, 73–81, January 24, 2013 ª2013 Elsevier

2007). An authentic substrate was generated by chemical deme-

thylation of the natural products. Close structural homologs

were also available for assay. Although the natural substrates

were efficiently methylated with the anticipated regiochemistry,

chemically similar compounds (4-hydroxy-3,6-dimethyl-2-

pyrone, 4-hydroxy-6-methyl-2-pyrone, and germicidin A) were

not substrates (Figure 4). The narrow substrate selectivity of

this enzyme strongly supports the assignment of ncp as the

nocapyrones biosynthesis cluster. Genetic knockout data would

be desirable to reinforce this finding, but in initial experiments the

strain was resistant to transformation by standard methods

(Kieser et al., 2000).

On this basis, we proposed that ncp (deposited in GenBank,

JN792621) was responsible for pyrone synthesis (Figure 3).

ncp encodes a cluster of four genes, ncpA (oxidoreductase),

ncpB (methyltransferase), ncpC (PKS), and ncpD (putative free-

standing acyltransferase). The correct assembly of this gene

cluster was confirmed by PCR, as well as by a nearly identical

and syntenic cluster encoded in the chromosome of the recently

deposited N. alba ATCC BAA-2165. Based upon the domain

architecture of NcpC, and in analogy to other pyrone biosyn-

thetic pathways, a hypothetical biogenetic scheme can be

proposed (Figure 3). First, diverse fatty acyl CoA esters are

loaded onto NcpC. Subsequently, the PKS domains act itera-

tively, extending the fatty acid with two units of methylmalonyl

CoA. As is found with some other PKS proteins, in the absence

of a thioesterase cleavage of a diketoester proceeds in tandem

with pyrone formation (Frank et al., 2007). Finally, the resulting

products are substrates forO-methylation by NcpB. NcpAmight

catalyze C-oxidation. Existing data do not strongly define the

role for NcpD, although it may be involved in starter unit loading

or product offloading. It should be noted that other biosynthetic

routes are possible. We favor this scheme because the NcpC

acyltransferase is likely methylmalonate specific, and the side

chain variation in the pyrone series is strongly reminiscent of

the normal mixture found in actinomycete fatty acids. Finally,

a subtle point is that only the linear side-chain pyrones were iso-

lated from cone snails, which is expected in animals if the source

fatty acids originate from primary metabolism.

With the biosynthetic cluster in hand, PCR primers (Figure S7)

were designed to specifically detect the presence of the same

genes in whole snail tissues. Indeed, both ncpB and ncpC could

be amplified from snail tissue, as could the N. alba CR167 16S

rRNA gene sequence (Figure 5). The variable nature of

compound production and the relative difficulty of obtaining

samples (the snails live at �70 m) precluded more detailed

colocalization studies. Indeed, the characteristics of N. alba

CR167 are consistent with a lifestyle that is not host restricted.

N. alba CR167 was readily cultured and maintained in the labo-

ratory. Therefore, these bacteria are casual associates of cone

snails, rather than obligate symbionts. Finally, previously we

showed that actinobacteria specifically inhabit mucus-gener-

ating cells within C. rolani (Peraud et al., 2009). This location is

consistent with the chemical results found here.

ConeSnails as aSource of SmallMolecules for BioactiveCompound DiscoveryNocapyrones modulated nerve cell depolarization, with the most

active compounds achieving an IC50 of 2 mM. Nocapyrones

Ltd All rights reserved

Figure 3. Organization of the Nocapyrone Biosynthetic Gene Clusters and Model for Nocapyrone Biosynthesis

KS, b-keto acyl synthase; AT, acyl transferase; ACP, acyl carrier protein; mMCoA, methyl malonyl CoA;MT, methyltransferase; Ox, oxidoreductase. This scheme

shows the hypothetical biogenesis (see text). See also Figure S6.

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

B (12) and H (1) were active against nearly all DRG neuronal cell

types at 50 mM in the calcium-imaging assay, whereas 3 was

inactive in the assay. We use the DRG assay because it provides

a rapid means to test response of diverse receptors in diverse

neuronal subtypes (Teichert et al., 2012a, 2012b; Lin et al.,

2011). DRG neurons are known to be heterogeneous, with

a variety of different cell types that respond to different stimuli.

Using the calcium-imaging assay, cell types can be identified

by their differential responses to discrete reagents (Teichert

et al., 2012a).

Depending on the DRG neuronal cell type, nocapyrones either

inhibited or amplified responses that were elicited by depolariz-

ing the cells with a brief application of high extracellular potas-

sium (KCl pulse). For example, both compounds 1 and 12

partially blocked depolarization-elicited (i.e., KCl-elicited)

increases in cytoplasmic calcium in large-diameter cells that

were capsaicin resistant, whereas they amplified the KCl-elicited

increases in cytoplasmic calcium of small-diameter, capsaicin-

sensitive cells (Figure 6). In additional tests, the effects of

compound 1 did not appear to correlate with acetylcholine-

sensitivity, whereas an amplification was observed consistently

in small, capsaicin- and histamine-sensitive cells. A near-

complete block was observed in a subset of menthol-sensitive

cells (Figure S8). The compounds did not affect sodium channels

Chemistry & Biology 20,

or a panel of �60 human channels and receptors (http://pdsp.

med.unc.edu/). We also tested compounds 1, 5, and 12 against

a panel of human cells overexpressing various transient receptor

potential (TRP) channels, and they were able to activate or inhibit

Ca2+ flux into those cells with IC50s between 2 and 70 mM, de-

pending upon the agent and channel subtype (Table S3).

Notably, the menthol-sensitive TRPM8 channel was inhibited,

whereas TRPA1 was activated. This is consistent with findings

using the DRG assay, where menthol-sensitive cells were

inhibited, whereas TRPA1 is abundantly expressed in capsa-

icin-sensitive neurons. However, the broad effects of the noca-

pyrones on TRP channels suggests that they may affect an

underlying channel-regulating element that leads to cell-type

specific effects, rather than acting directly on the channels or

cell-surface receptors tested.

Compounds 1 and 12 were also tested against the human

breast adenocarcinoma (MCF-7) and Chinese hamster ovary

(AA8) cell lines. Against MCF-7, 1 and 12 were cytotoxic with

IC50 values of 8.7 and 22.2 mM, respectively, whereas 3was inac-

tive against MCF-7 (>100 mM), and against AA8 only 1was cyto-

toxic with an IC50 of 10.2 mM. By contrast, the compounds were

not bacteriostatic or bactericidal. Schneemannet al. (2010) previ-

ously showed that nocapyrones A-D were inactive against an

even broader panel of bacteria and the yeast Candida glabrata.

73–81, January 24, 2013 ª2013 Elsevier Ltd All rights reserved 77

Figure 4. NcpB Is the Nocapyrone Ο-methyltransferase

High-performance liquid chromatography (HPLC) traces are shown with detection at 228 nm.

(A) Standards of pure 9 and 9a.

(B) Enzyme assay containing 9a, NcpB, and SAM.

(C) Control enzyme assay lacking SAM.

(D) Control enzyme assay with boiled NcpB.

(E–G) Other pyrone substrates were not accepted by the enzyme, using the same reaction conditions as shown in (B).

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

DISCUSSION

Here, we show that bacteria produce pyrone polyketides in cone

snails. Pyrones and related polyketides have been isolated from

many different mollusks from many habitats, including diverse

sacoglossan, siphonarid, and cephalispidean mollusks (Gavag-

nin et al., 1994; Davies-Coleman and Garson, 1998; Ireland

and Scheuer, 1979; Di Marzo et al., 1991, 1993; Sleeper and

Fenical, 1977; Manker et al., 1988; Darias et al., 2006), where

they are produced de novo (Ireland and Scheuer, 1979; Sleeper

and Fenical, 1977; Manker et al., 1988). However, this informa-

tion does not prove the source. For example, sacoglossans eat

eukaryotic green algae and employ both the chloroplasts and

algal mRNA to survive on sunlight for extended periods (Rumpho

et al., 2000; Mujer et al., 1996). Feeding studies demonstrate that

sacoglossan pyrones are labeled by carbonate so that they obvi-

ously arise from intermediates of photosynthesis mediated by

the hosted chloroplasts (Ireland and Scheuer, 1979). The pyrone

C-methyl groups arise from propionate, which might indicate

a bacterial source (Di Marzo et al., 1991; Cutignano et al.,

2012). This is due to the fact that a-methylated polyketides often

arise from incorporation of the polyketide precursor methylmal-

onate (propionate) in bacteria, whereas in animals and other

78 Chemistry & Biology 20, 73–81, January 24, 2013 ª2013 Elsevier

eukaryotes a-methylated polyketides usually arise from later

methylation of polyketide intermediates.

Molecular genetics can provide the definitive answer to the

source question for marine natural products (Piel, 2009). Here,

we used multiple types of chemical and molecular biological

evidence to show that N. alba CR167 bacteria make in the

cone snail species, C. rolani and C. tribblei. These multiple

methods were required to demonstrate that the bacteria are

important to their hosts. This is not a given; indeed, bacteria culti-

vated from animals are most likely to be irrelevant to the core

biology of the animals (Donia et al., 2011; Klepzig et al., 2009).

As an example, actinomycetes have been previously isolated

from many different marine animals, where they have been

shown to be an excellent source of promising drug leads (Blunt

et al., 2011). However, in no prior case has an actinomycete

been directly tied to actual function in marine animals, to the

best of our knowledge. Prior knowledge (Donia et al., 2011)

suggests that the majority of such actinomycete isolates may

not actively grow with their hosts and so do not contribute abun-

dant chemicals to the whole animals. Tying individual cultivated

bacterial strains to function in vivo requires careful study.

Actinomycete symbionts are well known to produce

secondary metabolites used by terrestrial insects (Barke et al.,

Ltd All rights reserved

Figure 5. Genes for Nocapyrone Synthesis Are Found in Cone Snail

Tissue

TheN. albaCR167 ncpB, ncpC, and 16S rRNA genes were amplified from both

C. rolani and C. tribblei metagenome DNA by nested PCR experiments. The

PCR products were confirmed to be identical to the positive controls by

Sanger sequencing the gel-extracted PCR products.

(A) Primers noc16s_Fwd and noc16s_Rev.

(B) Primers MT_in_Fwd and MT_in_Rev.

(C) Primers PKS_in_Fwd and PKS_in_Rev. Template DNA: (1) C. tribblei; (2)

C. rolani; (3) N. alba CR167; (4) no DNA control; 0) DNA ladder.

See also Figure S7.

Figure 6. Activity of Compound 12 Observed by Calcium Imaging of

Dissociated DRG Neurons in Culture

Each trace is the response of a single neuron. Responses from �100 neurons

were monitored individually and simultaneously in a given experimental trial.

Selected traces are shown from a single experimental trial. The y axis is

a measure of relative intracellular (cytoplasmic) calcium concentration, [Ca2+]i,

obtained by standard ratiometric calcium-imaging techniques (i.e., ratio of

340 nm/380 nm excitation while monitoring fluorescence emission at 510 nm).

The x axis is time in minutes. Increases in [Ca2+]i were elicited by briefly de-

polarizing the neurons with 25 mM potassium (KCl pulse) at regular 7 min

intervals, as indicated by vertical arrows. Each KCl pulse elicited an increase in

[Ca2+]i (by activating voltage-gated calcium channels) that is observed as

a peak in each trace. Capsaicin (300 nM) was applied to the neurons for 1 min

at the end of the experiment, as indicated by the black circle. After the fourth

KCl pulse, the compound was applied to the cells for 6 min (horizontal bar). In

many small-diameter, capsaicin-sensitive DRG neurons, the compound

caused an amplification of the calcium-transient elicited by a KCl pulse

(bottom three traces) or directly elicited a calcium-transient (fourth trace from

the bottom). In contrast, in many large-diameter, capsaicin-resistant neurons,

the compound partially blocked the calcium-transient elicited by a KCl pulse

(top four traces). All effects of the compound were reversible, as shown.

See also Figure S8 and Table S3.

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

2010), whereas in most marine symbioses so far studied

secondary metabolites are produced by proteobacteria or cya-

nobacteria. In ants, the actinomycete symbionts and their

produced chemicals vary over time and space, even in single

types of animals (Barke et al., 2010). Whether the symbionts

are variable or stable, the resulting compounds play important

ecological roles (Currie et al., 2003; Seipke et al., 2011). The

symbiosis reported here has many similarities to the casual

actinomycete symbioses reported on land but extends these

observations to the marine environment. It is noteworthy that

our data were acquired from repeated sampling of wild popula-

tions so that the results are relevant to the chemical ecology of

the organisms.

In 2010, Patil and colleagues isolatedN. alba ATCC BAA-2165

from the digestive tract of honeybees in Ohio (GenBank:

CP003788.1) (Patil et al., 2010). The bacteria could be reliably

cultivated as one of the major putative symbionts over all

seasons in a year. The strain showed modest antibiotic activity,

although for unknown reasons. While this manuscript was under

review, the genome of this strain was released, showing that

overall it is nearly identical to that of N. alba CR167. This

extremely close relationship shows that N. alba is a common

but casual symbiont of both marine and terrestrial invertebrates.

Moreover, the more distantly related N. dassonvillei is also an

animal associate, including very rarely acting as a human

pathogen (Penn and Jensen, 2012; Lejbkowicz et al., 2005).

Finally, nocapyrones 11–13 were previously reported from a

Nocardiopsis strain isolated from sponges. We propose that

N. alba, and possibly related Nocardiopsis strains, are widely

occurring, casual symbionts of diverse invertebrates and that

pyrones play a role in these symbioses. As the sponge and

honeybee cases have not been functionally studied, it is

unknown whether pyrones are produced in other invertebrates

or what the roles of these symbionts might be.

Our results show that mollusks that are heavily defended both

with a heavy shell and toxic venom can also harbor bioactive

small molecules in their mucus. These compounds were present

at concentrations relevant to chemical defense and that are

below those needed to directly affect mammalian cells and

neurons. Our system is challenging for ecology studies in that

Chemistry & Biology 20,

the snails live in relatively deep water (�70 m) and do not survive

well in aquaria. However, related molecules from shell-less

mollusks have been experimentally determined to serve roles

73–81, January 24, 2013 ª2013 Elsevier Ltd All rights reserved 79

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

such as defense, regeneration, and communication (Gavagnin

et al., 1994; Davies-Coleman and Garson, 1998; Ireland and

Scheuer, 1979; Di Marzo et al., 1991, 1993; Sleeper and Fenical,

1977; Manker et al., 1988; Darias et al., 2006). Further study is

required to determine whether the pyrones play these or other

roles within cone snail mucus. It should be clear that, in this con-

text, communication hasmany possible meanings. For example,

the compounds speculatively could play a role in communication

between snails, between snails and bacteria, or even between

bacteria in roles such as quorum sensing.

SIGNIFICANCE

Otherwise defenseless, mollusks are known to be good

sources of small molecules for drug discovery. We show

that even thickly shelled gastropods are a good source of

secondary metabolites with medicinal potential. Dogma

suggested that these mollusks may be poor sources of

secondary metabolites, but mollusks with shells have not

been widely explored recently. Given the improvement in

analytical methods since the dawn of marine natural prod-

ucts research, gastropods with shells warrant a second

look as a source of new compounds.

Bacteria are often thought to be the ultimate source of

small molecules in animals, but this has only been closely

investigated in a relatively small number of studies. Although

many bacteria have been cultivated from marine animals, it

is hard to determine whether those bacteria play a role in

the biology of the animals and whether they produce

compounds within their hosts. Here, we provide evidence

that actinomycetes make secondary metabolites in marine

animals, using a series of methods to rigorously tie bacterial

biosynthesis to phenotype within animals. These relation-

ships are, surprisingly, closely related to interactions found

in very distant habitats on land, showing awidespread pene-

tration of casual actinomycete symbionts in invertebrates.

Finally, pyrones and related polyketides are a widespread

family of mollusk compounds, for which the ultimate biosyn-

thetic source was not known. Here, we show that these

compounds are produced by symbiotic bacteria within

cone snails. Although further study is required to determine

whether other mollusk polyketides are bacterially synthe-

sized, the methods we use are widely applicable.

EXPERIMENTAL PROCEDURES

See the Supplemental Experimental Procedures.

SUPPLEMENTAL INFORMATION

Supplemental Information includes eight figures, three tables, and Supple-

mental Experimental Procedures and can be found with this article online at

http://dx.doi.org/10.1016/j.chembiol.2012.10.019.

ACKNOWLEDGMENTS

This work was funded by an ICBG grant (U01TW008163) from Fogarty

(National Institutes of Health [NIH]). We thank the government of the

Philippines and the community of Mactan Island for permission to conduct

this study. We thank Brian Dalley and Brett Milash (University of Utah Geno-

mics Core Facility) for sequencing the genome of CR167 and aiding with

80 Chemistry & Biology 20, 73–81, January 24, 2013 ª2013 Elsevier

data transfer and Wes Tolman (University of Utah), Joshua Orvis, Kevin Ga-

lens, and Chris Hemmerich (all of the The Joint Genome Institute) for helping

us to install the Ergatis software and bioinformatics database in-house.

Received: July 18, 2012

Revised: October 11, 2012

Accepted: October 24, 2012

Published: January 24, 2013

REFERENCES

Barke, J., Seipke, R.F., Gruschow, S., Heavens, D., Drou, N., Bibb, M.J., Goss,

R.J.M., Yu, D.W., and Hutchings, M.I. (2010). Amixed community of actinomy-

cetes produce multiple antibiotics for the fungus farming ant Acromyrmex

octospinosus. BMC Biol. 8, 109.

Benkendorff, K. (2010). Molluscan biological and chemical diversity:

secondary metabolites and medicinal resources produced by marine

molluscs. Biol. Rev. Camb. Philos. Soc. 85, 757–775.

Blunt, J.W., Copp, B.R., Munro, M.H.G., Northcote, P.T., and Prinsep, M.R.

(2011). Marine natural products. Nat. Prod. Rep. 28, 196–268.

Busch, B., and Hertweck, C. (2009). Evolution of metabolic diversity in polyke-

tide-derived pyrones: using the non-colinear aureothin assembly line as

a model system. Phytochemistry 70, 1833–1840.

Cimino, G., Sodano, G., and Spinella, A. (1987). New propionate-derived

metabolites from Aglaja depicta and from its prey Bulla striata (opisthobranch

mollusks). J. Org. Chem. 52, 5326–5331.

Cutignano, A., Cimino, G., Villani, G., and Fontana, A. (2009). Shaping the

polypropionate biosynthesis in the solar-powered mollusc Elysia viridis.

ChemBioChem 10, 315–322.

Cutignano, A., Villani, G., and Fontana, A. (2012). One metabolite, two path-

ways: convergence of polypropionate biosynthesis in fungi and marine

molluscs. Org. Lett. 14, 992–995.

Currie, C.R., Wong, B., Stuart, A.E., Schultz, T.R., Rehner, S.A., Mueller, U.G.,

Sung, G.H., Spatafora, J.W., and Straus, N.A. (2003). Ancient tripartite coevo-

lution in the attine ant-microbe symbiosis. Science 299, 386–388.

Dale, J.A., and Mosher, H.S. (1973). Nuclear magnetic resonance enantiomer

regents. Configurational correlations via nuclear magnetic resonance

chemical shifts of diastereomeric mandelate, O-methylmandelate, and

a-methoxy-a-trifluoromethylphenylacetate (MTPA) esters. J. Am. Chem.

Soc. 95, 512–519.

Darias, J., Cueto, M., and Dıaz-Marrero, A.R. (2006). The chemistry of marine

pulmonate gastropods. Prog. Mol. Subcell. Biol. 43, 105–131.

Davies-Coleman, M.T., and Garson, M.J. (1998). Marine polypropionates. Nat.

Prod. Rep. 15, 477–493.

Di Marzo, V., Vardaro, R.R., De Petrocellis, L., Villani, G., Minei, R., and Cimino,

G. (1991). Cyercenes, novel pyrones from the ascoglossan mollusk Cyerce

cristallina. Tissue distribution, biosynthesis and possible involvement in

defense and regenerative processes. Cell. Mol. Life Sci. 47, 1221–1227.

DiMarzo, V., Marin, A., Vardaro, R.R., De Petrocellis, L., Villani, G., andCimino,

G. (1993). Histological and biochemical bases of defense mechanisms in four

species of Polybranchioidea ascoglossan mollusks. Mar. Biol. 117, 367–380.

Donia, M.S., Fricke, W.F., Ravel, J., and Schmidt, E.W. (2011). Variation in

tropical reef symbiont metagenomes defined by secondary metabolism.

PLoS ONE 6, e17897.

Fontana, A. (2006). Biogenetic proposals and biosynthetic studies on

secondary metabolites of opisthobranch molluscs. Prog. Mol. Subcell. Biol.

43, 303–332.

Frank, B., Wenzel, S.C., Bode, H.B., Scharfe, M., Blocker, H., and Muller, R.

(2007). From genetic diversity to metabolic unity: studies on the biosynthesis

of aurafurones and aurafuron-like structures in myxobacteria and streptomy-

cetes. J. Mol. Biol. 374, 24–38.

Gavagnin, M., Marin, A., Mollo, E., Crispino, A., Villani, G., and Cimino, G.

(1994). Secondary metabolites from the Mediterranean Elysioidea: origin and

biological role. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 108, 107–115.

Ltd All rights reserved

Chemistry & Biology

A Bacterial Source for Mollusk Pyrone Polyketides

Guintchin, B., and Bienz, S. (2003). Stereoconvergent preparation of chiral

vinylsilanes by cuprate substitution of a-acetoxyallylsilanes. Application to

the synthesis of (S)-(+)-bishomomanicone. Tetrahedron 59, 7527–7533.

Ireland, C., and Scheuer, P.J. (1979). Photosynthetic marine mollusks: in vivo

14c incorporation into metabolites of the sacoglossan Placobranchus ocella-

tus. Science 205, 922–923.

Julien, B., Tian, Z.Q., Reid, R., and Reeves, C.D. (2006). Analysis of the ambru-

ticin and jerangolid gene clusters of Sorangium cellulosum reveals unusual

mechanisms of polyketide biosynthesis. Chem. Biol. 13, 1277–1286.

Kieser, T., Bibb, M.J., Buttner, M.J., Chater, K.F., and Hopwood, D.A. (2000).

Practical Streptomyces Genetics (Norwich: The John Innes Foundation).

Klepzig, K.D., Adams, A.S., Handelsman, J., and Raffa, K.F. (2009).

Symbioses: a key driver of insect physiological processes, ecological interac-

tions, evolutionary diversification, and impacts on humans. Environ. Entomol.

38, 67–77.

Kroiss, J., Kaltenpoth, M., Schneider, B., Schwinger, M.G., Hertweck, C.,

Maddula, R.K., Strohm, E., and Svatos, A. (2010). Symbiotic Streptomycetes

provide antibiotic combination prophylaxis for wasp offspring. Nat. Chem.

Biol. 6, 261–263.

Kubanek, J., Graziani, E.I., and Andersen, R.J. (1997). Investigations of terpe-

noid biosynthesis by the dorid nudibranch Cadlina luteomarginata. J. Org.

Chem. 62, 7239–7246.

Larsson, M., Galandrin, E., and Hogberg, H.-E. (2004). Diastereoselective

addition of organozinc reagents to 2-alkyl-3-(arylsulfanyl)propanals.

Tetrahedron 60, 10659–10669.

Lejbkowicz, F., Kudinsky, R., Samet, L., Belavsky, L., Barzilai, M., and

Predescu, S. (2005). Identification of Nocardiopsis dassonvillei in a blood

sample from a child. Am. J. Infect. Dis. 1, 1–4.

Lin, Z., Reilly, C.A., Antemano, R., Hughen, R.W.,Marett, L., Concepcion, G.P.,

Haygood, M.G., Olivera, B.M., Light, A., and Schmidt, E.W. (2011).

Nobilamides A-H, long-acting transient receptor potential vanilloid-1

(TRPV1) antagonists from mollusk-associated bacteria. J. Med. Chem. 54,

3746–3755.

Lopez-Vera, E., Heimer de la Cotera, E.P., Maillo, M., Riesgo-Escovar, J.R.,

Olivera, B.M., and Aguilar, M.B. (2004). A novel structural class of toxins: the

methionine-rich peptides from the venoms of turrid marine snails (Mollusca,

Conoidea). Toxicon 43, 365–374.

Marin, A., Alvarez, L.A., Cimino, G., and Spinella, A. (1999). Chemical defence

in cephalaspidean gastropods: origin, anatomical location and ecological role.

J. Molluscan Stud. 65, 121–131.

Manker, D.C., Garson,M.J., and Faulkner, D.J. (1988). De novo biosynthesis of

polypropionate metabolites in the marine pulmonate Siphonaria denticulata.

J. Chem. Soc. Chem. Commun. 16, 1061–1062.

Medema, M.H., Blin, K., Cimermancic, P., de Jager, V., Zakrzewski, P.,

Fischbach, M.A., Weber, T., Takano, E., and Breitling, R. (2011). antiSMASH:

rapid identification, annotation and analysis of secondary metabolite biosyn-

thesis gene clusters in bacterial and fungal genome sequences. Nucleic

Acids Res. 39 (Web Server issue), W339–W346.

Mujer, C.V., Andrews, D.L., Manhart, J.R., Pierce, S.K., and Rumpho, M.E.

(1996). Chloroplast genes are expressed during intracellular symbiotic associ-

ation of Vaucheria litorea plastids with the sea slug Elysia chlorotica. Proc. Natl.

Acad. Sci. USA 93, 12333–12338.

Chemistry & Biology 20,

Nelson, J.T., Lee, J., Sims, J.W., and Schmidt, E.W. (2007). Characterization of

SafC, a catechol 4-O-methyltransferase involved in saframycin biosynthesis.

Appl. Environ. Microbiol. 73, 3575–3580.

Olivera, B.M. (2002). Conus venom peptides: reflections from the biology of

clades and species. Annu. Rev. Ecol. Syst. 33, 25–47.

Patil, P.B., Zeng, Y., Coursey, T., Houston, P., Miller, I., and Chen, S. (2010).

Isolation and characterization of a Nocardiopsis sp. from honeybee guts.

FEMS Microbiol. Lett. 312, 110–118.

Penn, K., and Jensen, P.R. (2012). Comparative genomics reveals evidence of

marine adaptation in Salinispora species. BMC Genomics 13, 86.

Peraud, O., Biggs, J.S., Hughen, R.W., Light, A.R., Concepcion, G.P., Olivera,

B.M., and Schmidt, E.W. (2009). Microhabitats within venomous cone snails

contain diverse actinobacteria. Appl. Environ. Microbiol. 75, 6820–6826.

Putz, A., and Proksch, P. (2010). Chemical defence in marine ecosystems.

Annual Plant Rev. 39, 162–213.

Piel, J. (2009). Metabolites from symbiotic bacteria. Nat. Prod. Rep. 26,

338–362.

Porter, Q.N., and Baldas, J. (1971). Mass Spectrometry of Heterocyclic

Compounds (New York: Wiley-Interscience), 139–147.

Rumpho, M.E., Summer, E.J., and Manhart, J.R. (2000). Solar-powered sea

slugs. Mollusc/algal chloroplast symbiosis. Plant Physiol. 123, 29–38.

Schneemann, I., Ohlendorf, B., Zinecker, H., Nagel, K., Wiese, J., and Imhoff,

J.F. (2010). Nocapyrones A-D, g-pyrones from a Nocardiopsis strain isolated

from the marine sponge Halichondria panicea. J. Nat. Prod. 73, 1444–1447.

Schmidt, E.W. (2008). Trading molecules and tracking targets in symbiotic

interactions. Nat. Chem. Biol. 4, 466–473.

Schoenian, I., Spiteller, M., Ghaste, M., Wirth, R., Herz, H., and Spiteller, D.

(2011). Chemical basis of the synergism and antagonism in microbial

communities in the nests of leaf-cutting ants. Proc. Natl. Acad. Sci. USA

108, 1955–1960.

Seipke, R.F., Barke, J., Brearley, C., Hill, L., Yu, D.W., Goss, R.J.M., and

Hutchings, M.I. (2011). A single Streptomyces symbiont makes multiple

antifungals to support the fungus farming ant Acromyrmex octospinosus.

PLoS ONE 6, e22028.

Sleeper, H.L., and Fenical, W. (1977). Navenones A-C: trail-breaking alarm

pheromones from the marine opisthobranch Navanax inermis. J. Am. Chem.

Soc. 99, 2367–2368.

Sonnet, P.E., Gazzillo, J.A., Dudley, R.L., and Boswell, R.T. (1990). Synthesis

and characterization of enantiomers of 5- and 6-methyloctanoic acids. Chem.

Phys. Lipids 54, 205–214.

Sun, H., Lapidus, A., Nolan, M., Lucas, S., Del Rio, T.G., Tice, H., Cheng, J.F.,

Tapia, R., Han, C., Goodwin, L., et al. (2010). Complete genome sequence of

Nocardiopsis dassonvillei type strain (IMRU 509). Stand. Genomic Sci. 3,

325–336.

Teichert, R.W., Smith, N.J., Raghuraman, S., Yoshikami, D., Light, A.R., and

Olivera, B.M. (2012a). Functional profiling of neurons through cellular neuro-

pharmacology. Proc. Natl. Acad. Sci. USA 109, 1388–1395.

Teichert, R.W., Raghuraman, S., Memon, T., Cox, J.L., Foulkes, T., Rivier, J.E.,

and Olivera, B.M. (2012b). Characterization of two neuronal subclasses

through constellation pharmacology. Proc. Natl. Acad. Sci. USA 109,

12758–12763.

Terlau, H., and Olivera, B.M. (2004). Conus venoms: a rich source of novel ion

channel-targeted peptides. Physiol. Rev. 84, 41–68.

73–81, January 24, 2013 ª2013 Elsevier Ltd All rights reserved 81