Herbivore exploits orally secreted bacteria
to suppress plant defenses
Seung Ho Chunga, Cristina Rosaa, Erin D. Scullyb, Michelle Peiffera, John F. Tookera, Kelli Hoovera, Dawn S. Luthec,
and Gary W. Feltona,1
Departments ofaEntomology andcPlant Science, Center for Chemical Ecology, andbIntercollege Program in Genetics, Huck Institutes of the Life Sciences,
Pennsylvania State University, University Park, PA 16802
Edited by James H. Tumlinson, Pennsylvania State University, University Park, PA, and approved July 11, 2013 (received for review May 15, 2013)
Induced plant defenses in response to herbivore attack are mod-
ulated by cross-talk between jasmonic acid (JA)- and salicylic acid
(SA)-signaling pathways. Oral secretions from some insect herbi-
vores contain effectors that overcome these antiherbivore de-
fenses. Herbivores possess diverse microbes in their digestive
systems and these microbial symbionts can modify plant–insect
interactions; however, the specific role of herbivore-associated
microbes in manipulating plant defenses remains unclear. Here,
we demonstrate that Colorado potato beetle (Leptinotarsa decem-
lineata) larvae exploit bacteria in their oral secretions to suppress
antiherbivore defenses in tomato (Solanum lycopersicum). We
found that antibiotic-untreated larvae decreased production of
JA and JA-responsive antiherbivore defenses, but increased SA
accumulation and SA-responsive gene expression. Beetles benefit
from down-regulating plant defenses by exhibiting enhanced lar-
val growth. In SA-deficient plants, suppression was not observed,
indicating that suppression of JA-regulated defenses depends on
the SA-signaling pathway. Applying bacteria isolated from larval
oral secretions to wounded plants confirmed that three microbial
symbionts belonging to the genera Stenotrophomonas, Pseudo-
monas, and Enterobacter are responsible for defense suppression.
Additionally, reinoculation of these bacteria to antibiotic-treated
larvae restored their ability to suppress defenses. Flagellin isola-
ted from Pseudomonas sp. was associated with defense suppression.
Our findings show that the herbivore exploits symbiotic bacteria as
a decoy to deceive plants into incorrectly perceiving the threat as
microbial. By interfering with the normal perception of herbivory,
beetles can evade antiherbivore defenses of its host.
plants recognize insect-derived cues such as touch, tissue dis-
ruption, oviposition, and oral secretions (OSs; saliva and/or regur-
gitant) (1, 2). Phytohormones mediate specific defense responses
depending on biotic attackers (3). Microbial pathogens that feed on
dead tissues or cells and most herbivorous insects are susceptible to
jasmonic acid (JA)/ethylene (ET)-regulated defenses, whereas mi-
crobial pathogens that require living hosts are susceptible to sali-
cylic acid (SA)-regulated defense (3). It is well documented that
cross-talk between JA/ET and SA plays an important role in fine
tuning induced defenses to enhance plant fitness (3, 4).
As plants have evolved defense strategies in response to biotic
attack, some herbivores produce effector molecules that sup-
press induced defenses of plants (5). Herbivore effectors are
found in OSs and eggs but only a few of the effectors have been
identified (7–9). Herbivore suppression of JA-regulated defenses
can be mediated by cross-talk with the SA-signaling pathway or
through other SA-independent means (6, 8, 10).
Many herbivorous insects harbor microbial symbionts that
provide essential nutrients or vitamins and/or are associated with
insect defenses against predators or parasites (11). In addition,
symbionts can influence plant–insect interactions (12, 13). Sym-
biotic bacteria may allow some herbivores to expand the range of
available host plants. When endosymbiotic bacteria of stinkbugs
(Megacopta punctatissima), which feed on both leguminous crops
and wild legumes, were transferred to Megacopta cribaria, whose
lants defend themselves against herbivores via the induction
of defense metabolites or proteins that are triggered when
host plants are restricted to wild legumes, M. cribaria performed
better on soybean crops (14). Moreover, insect-associated
microbes can change host plant physiology to benefit their insect
host. Wolbachia-infected leaf miners (Phyllonorycter blancardella)
elicit a green-island phenotype in apple, preserving photosyn-
thetically active tissues in senescent leaves, and in turn, increase
leaf miner performance (12). In maize roots, Wolbachia in
western corn rootworms (Diabrotica virgifera virgifera) may sup-
press defense gene expression (13), although this is disputed
(15). Despite a few examples, the role of insect symbionts as
elicitors or effectors in modifying plant–insect interactions has
received scant attention.
Previous research has shown that application of OSs from
Colorado potato beetle (CPB; Leptinotarsa decemlineata) larvae
to mechanically wounded plants suppressed induced defenses in
tomato and potato compared with wounded and water-treated
plants (16–18). To investigate potential mechanisms of defense
suppression by CPB larvae, we asked if microbes in OSs from
CPBs manipulate induced defenses. We found that larvae that
fed on antibiotic-treated leaves failed to suppress JA-regulated
defenses, whereas larvae not treated with antibiotics were able
to suppress these defenses. The suppression of antiherbivore
defenses was dependent on a functional SA-signaling pathway.
Moreover, we identified symbiotic bacteria in OSs that sup-
pressed plant defenses and this suppression increased larval
performance. Thus, these data show that CPB larvae use orally
The role of herbivore-associated microbes in modifying plant
defenses has received scant attention. The Colorado potato
beetle secretes symbiotic bacteria to wounds to manipulate
plant defenses. The bacteria elicit salicylic acid (SA)-regulated
defenses, and because SA signaling often negatively cross-talks
with jasmonate signaling, plants are unable to fully activate
their jasmonate-mediated resistance against the herbivore.
From the plants’ perspective, they recognize herbivores not as
such, but as microbial threats. We identified the specific bac-
teria from the beetle secretions and also characterized one of
the bacterial effectors responsible for defense suppression.
This clever, deceptive strategy for suppressing defenses has
not been previously documented. Our results add a significant,
unique concept to plant–insect interactions and how herbi-
vores hijack plant defense signaling.
Author contributions: S.H.C., C.R., E.D.S., K.H., and G.W.F. designed research; S.H.C. and
M.P. performed research; J.F.T. and D.S.L. contributed new reagents/analytic tools; S.H.C.
and G.W.F. analyzed data; and S.H.C., C.R., E.D.S., J.F.T., K.H., D.S.L., and G.W.F. wrote the
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
Freely available online through the PNAS open access option.
Data deposition: The sequences reported in this paper have been deposited in the Gen-
Bank database (accession nos. JX296529, JX296530, JX296531, and KC977253–KC977257).
1To whom correspondence should be addressed. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
www.pnas.org/cgi/doi/10.1073/pnas.1308867110 PNAS Early Edition
| 1 of 6
secreted bacteria to manipulate plant defenses by triggering an
ineffective defense pathway.
Symbiotic Microbes in OSs Are Deposited During Larval Feeding. CPB
larvae produce oral secretions when they feed on tomato leaves
(18) (Movie S1). To determine if microbes in OSs are deposited
on leaves during herbivory, we performed scanning electron
microscopy (SEM) of leaves after 1 h of larval feeding, during
which frass was not excreted. We found abundant bacteria on the
wounded area damaged by larvae that fed on artificial diet
without the addition of antibiotic (AB) compared with leaves
damaged by larvae fed with AB (Fig. 1A). We also confirmed
that larvae that fed on leaves treated or untreated with AB,
hereafter referred to as AB-treated or untreated larvae, secreted
similar amounts of OSs on leaves (SI Appendix, Figs. S1 and S2).
These results indicate that CPB larvae deposited symbiotic
microbes in their OSs.
Presence of Symbiotic Microbes Suppresses Plant Defense and
Enhances Larval Growth. To investigate whether the presence of
symbiotic microbes benefits CPB larvae, we measured growth of
neonate larvae on detached leaves damaged by AB-treated or
untreated larvae and polyphenol oxidase (PPO) activity in
plants damaged by these larvae. Measurement of PPO activity
serves as a rapid, sensitive, and quantitative assay to study JA-
responsive defenses. The neonate larvae that fed on leaves
damaged by untreated larvae gained more weight than those
that fed on leaves damaged by AB-treated larvae (Fig. 1B).
Plants damaged by untreated larvae showed lower PPO activi-
ties than those damaged by AB-treated larvae (Fig. 1C). In-
terestingly, the neonate larvae that fed on leaves damaged by
untreated larvae grew as well as those on undamaged leaves
(Fig. 1B), although PPO activities in undamaged leaves were
lower than leaves damaged by untreated larvae (Fig. 1C). Three
of four independent experiments showed similar results (SI
Appendix, Fig. S3). These results indicate that the suppression
of induced defenses by symbionts was sufficient to enhance
Plant Defenses Are Suppressed by Symbiotic Microbes in OSs. To
investigate if microbial symbionts present in OSs affect induced
defenses, we measured the expression of selected defense genes
and enzymatic activities in plants damaged by AB-treated or
untreated larvae. The cysteine proteinase inhibitor (CysPI) and
polyphenol oxidase F/B (PPOF/B) were selected as JA-marker
genes because they are well-established JA-inducible proteins
that function against CPBs (19–21). For an SA-marker gene,
pathogenesis-related 1 (PR-1(P4)) was selected (22). Feeding
by untreated larvae decreased expression levels of the JA-
responsive CysPI and PPOF/B compared with feeding by AB-
treated larvae, whereas feeding by untreated larvae increased
SA-responsive PR-1(P4) expression (Fig. 2). PPO activities in
plants damaged by AB-treated larvae were higher than plants
damaged by untreated larvae (Fig. 2). To further support our
hypothesis, we applied OSs from AB-treated or untreated larvae
to mechanically wounded plants. Plants wounded and treated
with OSs from untreated larvae had lower PPO activity than
those wounded and treated with water or OSs from AB-treated
larvae (SI Appendix, Fig. S4). However, PPO activities in plants
treated with OSs from AB-treated larvae were similar to plants
treated with wounding and water (SI Appendix, Fig. S4). Appli-
cation of AB to wounded plants had no effect on gene expression
or PPO activity (SI Appendix, Fig. S5). These results demonstrate
that the microbes in OSs from CPB larvae suppressed antiher-
Defense Suppression by Symbiotic Microbes in OSs Is Dependent on
SA-Signaling Pathway. Negative cross-talk between the JA- and
SA-signaling pathways plays a major role in suppression of in-
duced defenses mediated by herbivore effectors (10). To de-
termine if an antagonistic interaction between the JA- and SA-
signaling pathways is mediated by symbionts, we measured cis-JA
and SA levels in plants damaged by AB-treated or untreated
larvae. Compared with feeding by AB-treated larvae, feeding by
untreated larvae decreased cis-JA accumulation, but increased
SA production at 2 h and 4 h after insect infestation (SI Ap-
pendix, Fig. S6). We also measured defense gene expression in
Moneymaker (wild type) and SA-deficient NahG tomato plants
damaged by AB-treated or untreated larvae. In Moneymaker,
feeding by untreated larvae decreased expression of JA-responsive
genes but increased SA-responsive gene expression (Fig. 3). In
contrast, suppression of defense gene expression did not occur in
NahG plants attacked by AB-untreated larvae (Fig. 3). Moreover,
feeding by AB-untreated larvae decreased PPO activities in Mon-
eymaker, but not in NahG plants (SI Appendix, Fig. S7). These
results indicate that negative cross-talk between the JA- and SA-
dependent pathways was involved in defense suppression.
Symbiotic Bacteria Isolated in OSs Suppress Plant Defenses. To in-
vestigate whether fungi and/or bacteria in OSs from CPB larvae
suppress plant defenses, we fed larvae with leaves containing
antibacterial (anti-B) agents, antifungal agents (anti-F), or both
(anti-B/F). Feeding by untreated larvae or anti-F–treated larvae
decreased PPO activities compared with feeding by anti-B or
anti-B/F–treated larvae (SI Appendix, Fig. S8), indicating that
bacteria, but not fungi, are involved in suppressing this JA-
responsive defense. To characterize the symbiotic bacteria re-
sponsible for defense suppression, we isolated 22 bacterial col-
onies in OSs from untreated larvae. We cultured each isolate in
liquid media (2xYT, 2x yeast extract and trypton) and then ap-
plied a specific amount of each isolate to mechanically wounded
plants. Application of culture media to wounds (W + 2xYT)
did not affect PPO activities compared with wounding followed
by application of water (SI Appendix, Fig. S9A). Of the 22
bacterial isolates, three isolates significantly decreased PPO
activities compared with W + 2xYT treatment (SI Appendix,
Fig. S9). Based on the 16S rRNA gene sequence, we classified
leaves during feeding by CPB larvae after first feeding on an artificial diet
with antibiotic (Right) or without antibiotic (Left). Damaged leaf tissues
were prepared for SEM images 1 h after larval feeding. (Scale bar, 1 μm.) (B)
Larval growth and (C) polyphenol oxidase (PPO) activities in plants damaged
by larvae that fed on AB-treated or untreated leaves. Neonates were
allowed to feed on excised leaflets from each treatment for 5 d and larval
mass was determined. Values are means ± SEM. Different letters represent
significant differences [ANOVA, P < 0.05; followed by LSD test; F(2,71)= 6.63,
P = 0.0023, n = 23–26]. PPO activities were measured on subsamples from
each treatment 48 h after insect feeding [F(2,9)= 27.13, P = 0.0002, n = 4]. To
collect subsamples, two leaf discs from each of two leaves were pooled as
one replicate. AB(−), plants damaged by untreated larvae; AB(+), plants
damaged by AB-treated larvae; Con, undamaged plants.
(A) Scanning electron microscopy images of bacteria secreted onto
2 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1308867110Chung et al.
the isolates that suppressed PPO activity as members of the
genera Stenotrophomonas, Pseudomonas, and Enterobacter. Of
the remaining isolates that did not suppress PPO activities, we
sequenced some of the nonsuppressing bacteria and classified
them as Raoultella sp., Pseudomonas sp., and Enterobacteriaceae.
We also determined if the concentration of symbiotic bacteria
applied is critical for defense suppression. We applied serial
dilutions of the suppressing Stenotrophomonas sp., Pseudomonas
sp., and Enterobacter sp. and the nonsuppressing bacterial iso-
lates Raoultella sp., Pseudomonas sp., and Enterobacteriaceae to
wounded plants. Defense suppression by the three suppressing
bacteria was dose dependent, with a threshold of 105CFU/mL
(SI Appendix, Fig. S10). When we applied the same concentra-
tion of each isolate (109CFU/mL) to wounded plants, the sup-
pressing Pseudomonas sp. decreased PPO activities compared
with W + 2xYT treatment. Application of other nonsuppressing
bacteria did not decrease PPO activity (SI Appendix, Fig. S11).
In addition to suppression of PPO activity, we measured ex-
pression levels of JA/SA-regulated genes in plants that were
wounded and treated with the suppressing Pseudomonas sp. cul-
ture. We found that this isolate decreased JA-responsive CysPI
and PPOF expression, but increased SA-regulated PR-1(P4) ex-
pression compared with W + 2xYT (SI Appendix, Fig. S12).
To investigate the effects of a mixture of the suppressing
bacterial isolates on plant defenses, we applied all possible com-
binations by applying two or three isolates at a time to wounded
plants. We did not find synergistic, antagonistic, or additive effects
of the mixture on suppression of PPO activities, but all combi-
nations decreased PPO activities compared with W + 2xYT
treatment (SI Appendix, Fig. S13). These results indicate that the
presence of at least one of the three bacteria, Stenotrophomonas
sp., Pseudomonas sp., or Enterobacter sp. in OSs was sufficient to
suppress induced defenses.
Reintroduction of the Suppressing Bacteria to AB-Treated Larvae
Suppresses Plant Defenses. To confirm that the AB treatment
used in our studies was indeed active on the bacteria of interest,
we performed zone of inhibition assays with the three isolates of
cultured bacteria described above. We found that the AB mix-
ture inhibited growth of the three bacterial spp. that suppress
plant defenses (SI Appendix, Table S1). In addition, we quanti-
fied the amount of Pseudomonas sp. on leaves after larval
feeding by measuring rpoD (sigma factor subunit of RNA poly-
merase) abundance. Relative abundance of rpoD on plants
damaged by untreated larvae was much higher than plants fed on
by AB-treated larvae (SI Appendix, Fig. S14). We could not de-
tect rpoD on undamaged plants. These data indicated that
Pseudomonas sp. was deposited on leaves by untreated larvae
and that AB treatment effectively inhibits Pseudomonas sp. in
We further tested whether bacteria in OSs from untreated
larvae were responsible for manipulation of induced defenses by
reinoculation of the suppressing bacteria individually into AB-
treated larvae. AB-treated or untreated larvae fed on leaves that
received cultured bacteria or control buffer for 2 d. Addition of
cultured-suppressing bacteria to untreated larvae did not affect
PPO activities compared with the addition of control buffer to
untreated larvae. Feeding by untreated larvae decreased PPO
activities compared with feeding by AB-treated larvae that fed
on leaves with control buffer. In contrast, reinoculating sup-
pressing bacteria into AB-treated larvae restored defense sup-
pression so that PPO activities in plants damaged by AB-treated
larvae that received Stenotrophomonas sp., Pseudomonas sp., or
Enterobacter sp. were similar to plants damaged by untreated
larvae with control buffer or the suppressing bacteria (Fig. 4).
These data demonstrate that the suppressing bacteria were de-
livered to wounded plants through OSs and that they significantly
down-regulate induced defenses.
Flagellin Suppresses Plant Defenses. Because complete genome se-
quences and proteome databases are available for several Pseudo-
monas sp., we focused on Pseudomonas sp. among the suppressing
bacteria to investigate molecular mechanisms underlying suppres-
sion of plant defenses. To identify potential effectors in the sup-
pressing bacteria, we purified flagellin from Pseudomonas sp. and
applied it to wounded plants. Purified flagellin was analyzed by
SDS/PAGE and the band with ∼67 kDa was excised and subjected
to MALDI-TOF/TOF mass spectrum analysis (SI Appendix, Fig.
S15A). The mass spectra of the purified flagellin were identified as
Pseudomonas flagellin (SI Appendix, Table S2). Application of
varying flagellin concentrations decreased PPO activities com-
pared with wounding and buffer treatment (SI Appendix, Fig.
S15B), showing that flagellin was an effector protein involved in
suppressing plant defenses.
Microbial symbionts provide important roles in the survival of
their hosts including providing nutrition, detoxifying toxins, and
priming of host immunity, among others. In the case of certain
parasitic insects, symbiotic viruses suppress the defensive
responses of their hosts (23). However, in the case of herbivores,
very little is known about how their symbionts may mediate host
plant defenses. Here, we demonstrate that CPB larvae secrete
bacteria in their OSs to evade antiherbivore defenses in tomato.
These responses did not occur with plants, which have a de-
fective SA-signaling pathway (Fig. 3 and SI Appendix, Fig. S7);
the plant responded as if it was being attacked by a pathogen
rather than an herbivore. Thus, the beetle, by presenting bacteria
in its oral secretions, is hijacking the defensive machinery of the
host plant for its own benefit. In essence, the beetle is diverting
resources away from the appropriate immune response of its
plants damaged by larvae that fed on AB-treated or untreated
leaves. Gene expression was measured 24 h after initiation of
insect feeding. Values are untransformed means ± SEM (n = 4–
5). Different letters represent significant differences [ANOVA,
P < 0.05; followed by LSD test; CysPI, F(2,11)= 214.7, P < 0.0001;
PPOF, F(2,11)= 185.5, P < 0.0001; PPOB, F(2,12)= 36.7, P < 0.0001;
PR-1(P4), F(2,11)= 19.7, P = 0.0002]. AB(−), plants damaged by
untreated larvae; AB(+), plants damaged by AB-treated larvae;
Con, undamaged plants. CysPI, cysteine proteinase inhibitor;
PPOF/B, polyphenol oxidase F/B; PR-1(P4), pathogenesis-related
protein 1 (P4).
Expression levels of JA- and SA-regulated genes in
Chung et al. PNAS Early Edition
| 3 of 6
host plant. Conversely, a common strategy used by plant patho-
gens (e.g., Pseudomonas syringae pathovar tomato) is to hijack
the JA pathway at the expense of resources to the SA immune
response (24). Our findings indicate that the beetles, by har-
boring bacteria in their oral secretions, are recognized, in part, by
plants as a microbial threat. We do not know the extent of this
exploitive strategy, but in a few other examples that we have
tested, it appears that beetles routinely release copious amounts
of oral secretions during feeding. The role of microbial sym-
bionts in the coevolution of plants and herbivores has no doubt
received inadequate attention.
When plants are attacked by herbivores, plants perceive elic-
itors or herbivore-associated molecular patterns in OSs and
trigger biosynthesis of phytohormones and antiherbivore defense
responses (2). As microbial pathogens secrete effector proteins
into their host to suppress immune defenses that are triggered by
microbial-associated molecular patterns (MAMPs) (25), insect
herbivores often evolve strategies to circumvent plant defenses.
Although there is an emerging literature indicating that some
chewing/piercing-sucking herbivores use effectors to manipulate
plant defenses, there is scant evidence that symbiotic microbes of
herbivores manipulate plant defenses. Among herbivore effec-
tors identified so far, glucose oxidase in saliva of Helicoverpa zea
caterpillar was the first effector identified reducing JA-regulated
nicotine production in Nicotiana tabacum (6). In aphids (Myzus
persicae), the salivary protein MpCOO2 has been shown to be an
effector and its overexpression in Nicotiana benthamiana pro-
moted aphid fecundity (7). It also has been reported that non-
proteinaceous components of the OSs of Pieris brassicae and
Spodoptera littoralis caterpillars reduced induction of anti-
herbivore defense genes in Arabidopsis in a manner independent
of JA/SA-signaling pathways (8). Saliva of Spodoptera exigua
suppressed JA-regulated defense responses in Arabidopsis (10).
Spider mites (Tetranychus evansi) suppressed direct defenses so
that mites had higher oviposition and survival rates on plants that
they attacked (26). Because spider mites inject saliva into host
plants (27), it is possible that unknown effectors present in T.
evansi saliva manipulate plant defenses. Silverleaf whiteflies
(Bemisia tabaci) decreased JA-responsive gene transcripts and
increased SA-regulated gene transcripts, which allowed faster
nymph development (28). It is presumed that silverleaf whiteflies
secrete saliva to suppress plant defenses (28). However, none of
the above studies indicated an involvement of microbial sym-
bionts in modifying plant defenses.
To our knowledge, there has been only one previous report of
defense suppression by herbivore-harbored microbes. In maize
roots, larvae of the western corn rootworm (D. virgifera virgifera)
inhibited the induction of defense gene expression, which was
not observed when insects were treated with the antibiotic tet-
racycline (13). The authors reported that larvae with Wolbachia
infection suppressed plant-defense–related gene transcripts, al-
though the effect of suppression on insect performance was not
examined. Additionally, it is unclear whether Wolbachia or
Wolbachia-derived compounds are secreted into plants when
western corn rootworm larvae feed on root tissues. More re-
cently, it was shown that Wolbachia of the rootworm did not
suppress maize defense responses (15). In contrast to herbivore
symbionts, which are not pathogenic to plants, phytopathogens
associated with insects can modify plant responses (29). The
plant pathogenic bacterium, Candidatus Liberibacter psyllaurous
(Lps), vectored by tomato psyllids (Bactericerca cockerelli), sup-
pressed JA/SA-regulated defense transcripts (29), although it
was not determined whether the modified responses were ben-
eficial to psyllids or Lps and how effectors from psyllids or Lps
change defense responses. In this study, we clearly demonstrated
that CPB larvae released symbiotic bacteria, which was con-
firmed by confocal microscopy (SI Appendix, Fig. S1 and Movie
S1), electron microscopy (Fig. 1A), and quantitative PCR
(qPCR) (SI Appendix, Fig. S14). Further we demonstrated that
specific bacteria in OSs suppressed the induced defenses in
type Moneymaker and SA-deficient NahG plants damaged by
larvae that fed on AB-treated or untreated leaves. Values are
untransformed means ± SEM (n = 4–5). Gene expression was
measured 24 h after initiation of insect feeding. Different let-
ters represent significant differences [ANOVA, P < 0.05; fol-
lowed by LSD test; for Moneymaker, CysPI, F(2,12)= 176, P <
0.0001; PPOF, F(2,12)= 49.7, P < 0.0001; PPOB, F(2,1)= 14.7, P =
0.0006; PR-1(P4), F(2,11)= 17.5, P = 0.0004; for NahG, CysPI, F(2,12)=
Con, undamaged plants. CysPI, cysteine proteinase inhibitor;
PPOF/B, polyphenol oxidase F/B; PR-1(P4), pathogenesis-related
protein 1 (P4).
Expression levels of JA- and SA-regulated genes in wild-
treated or untreated larvae following reinoculation
of the larvae with three bacterial isolates (Steno-
trophomonas sp., Pseudomonas sp., and Enter-
obacter sp.) cultured from CPB larvae and found to
suppress JA-mediated plant defenses. Larvae were
allowed to feed on AB-treated or untreated leaves
for 2 d and then fed on leaves that were reinocu-
lated with suspension buffer (10 mM MgCl2) or the
bacterial isolates in suspension buffer for 2 d. PPO
activities were measured 48 h after initiation of insect feeding. Values are means ± SEM (n = 6). Different letters represent significant differences [ANOVA, P <
0.05; followed by LSD test; (A) Stenotrophomonas sp., F(4,25)= 33.8, P < 0.0001; (B) Pseudomonas sp., F(4,25)= 22.6, P < 0.0001; (C) Enterobacter sp., F(4,25)= 128,
P < 0.0001]. Buffer, 10 mM MgCl2; En, Enterobacter sp.; Ps, Pseudomonas sp.; St, Stenotrophomonas sp. AB(−), plants damaged by untreated larvae; AB(+),
plants damaged by AB-treated larvae; Con, undamaged plants.
PPO activities in plants damaged by AB-
4 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1308867110Chung et al.
tomato (SI Appendix, Figs. S9–S13), which promotes CPB larval
growth (Fig. 2).
It is important to note that neonate larvae gained more weight
on leaves that were damaged by untreated larvae than on those
that were damaged by AB-treated larvae (Fig. 1 and SI Appendix,
Fig. S3). Also, increased larval growth corresponded with de-
creased expression of JA-responsive antiherbivore genes (CysPI
and PPOF/B) and reduced PPO activity, when plants were
damaged by AB-untreated larvae (Fig. 2). Thus, suppression of
antiherbivore defenses is highly beneficial to the performance of
CPB larvae. Earlier reports have shown that defense suppression
mediated by herbivore effectors enhance herbivore performance
(6, 7, 26, 28). For example, S. littoralis larvae grew better on
Arabidopsis treated with OSs of S. littoralis than on wounded
plants (8). The growth of neonate larvae on plants damaged by
untreated larvae was similar to that on undamaged plants. It may
be attributed to the fact that untreated larvae induced in-
sufficient production of defensive proteins such as CysPI and
PPO to retard neonatal growth. Alternatively, it is possible that
CPB larvae adapted to lower levels of defenses induced by un-
treated larvae. It was shown that CPB larvae can overcome
induced defenses by changing the profile of their digestive pro-
Among the bacterial colonies isolated from OSs of untreated
larvae, three isolates, genera Stenotrophomonas, Pseudomonas,
and Enterobacter, suppressed PPO activities when applied to
wounded plants (SI Appendix, Figs. S9 and S10). Similar bacteria
have been found in other geographic locations. CPB larvae from
potato fields in Maryland and Virginia contained these genera
(31) and CPB populations in Turkey had Pseudmononas sp. (32),
but the role(s) these bacteria may play in association with the
insect was not investigated in these cases. These bacteria from
field populations showed high sequence similarities (>97%) to
the 16S rRNA gene sequences from the isolates we cultured
from CPB in our laboratory colony. Furthermore, CPB larvae
collected from potato fields in Pennsylvania have symbiotic
bacteria that suppress plant defenses. Plants damaged by un-
treated larvae from this field population showed lower PPO ac-
tivities than those damaged by AB-treated larvae (SI Appendix,
Fig. S16). We also detected the same Pseudomonas sp. from the
field population (12 larvae and four adults; n = 16) as the sup-
pressing Pseudomonas sp. in our colony (9 larvae and two adults;
n = 11) (SI Appendix, Fig. S17). The sequence similarity of the
rpoD gene between these samples was 100%. Thus, it is note-
worthy to investigate whether symbiotic bacteria isolated from
other field populations of CPB suppress plant defenses.
We found that antiherbivore defenses were suppressed by
symbiotic microbes in OSs in an SA-dependent manner (Figs. 2
and 3 and SI Appendix, Fig. S6). When NahG plants were
damaged by AB-treated or untreated larvae, suppression of JA-
responsive genes and PPO activity was not observed (Fig. 3 and
SI Appendix, Fig. S7). Application of suppressing bacteria iso-
lated from OSs to wounded plants reduced both JA-responsive
defense gene expression and PPO activity compared with plants
wounded and treated with culture medium (SI Appendix, Figs.
S9–S12). Flagellin isolated from the culture medium of Pseu-
domonas sp., one of the suppressing bacteria, decreased PPO
activities (SI Appendix, Fig. S15). In Arabidopsis, MAMPs, such
as flagellin, induced SA biosynthesis and triggered SA-regulated
responses (33). Tomato can recognize flagellin and its conserved
domain (flg15) through FLS2 (Flagellin Sensing 2), a flagellin
receptor (34). Flagellin from Pseudomonas sp. in OSs thus acts as
an effector of CPB larvae by eliciting a SA-signaling pathway to
down-regulate JA-regulated defenses triggered by feeding.
In addition to flagellin, other components may be produced by
CPB symbiotic bacteria, which suppress JA-regulated defense
Other studies have shown that phytopathogens suppress host
plant defenses to enhance the performance of their insect vectors
(35). For example, Tomato spotted wilt virus vectored by western
flower thrips (Frankliniella occidentalis) decreased JA-regulated
defenses through induction of the SA-signaling pathway, bene-
fitting the insect vector (36). Phytoplasmas secreted the protein
SAP11 to down-regulate lipoxygenase and inhibit JA biosyn-
thesis, which increases survival and fecundity of the leafhopper
vector and phytoplasma transmission (37). It remains to be de-
termined if other metabolites produced by CPB symbiotic bac-
teria suppress plant defenses or if other MAMPs from these
bacteria activate the SA-signaling pathway to interfere with JA-
regulated defenses. We cannot rule out the possibility that uncul-
turable, AB-sensitive microbes in OSs also could be involved in
In conclusion, we provide direct evidence that orally secreted
microbial symbionts function as effectors and mediate plant–
insect interactions. In other words, microbial symbionts in OSs
from CPB larvae are secreted onto plant tissues during herbivory
and manipulate induced defenses through negative cross-talk
between JA- and SA-regulated signaling pathways. Thus, CPB
larvae hijack these signaling pathways, which benefits the larvae
by improving their growth. Our results provide important clues
to how these unrecognized and underappreciated players (i.e.,
insect-associated microbes) shape the complex interaction be-
tween plants and insect herbivores. This adds an exciting unique
dimension to our understanding of how herbivores deploy
counterdefense strategies in response to plant defenses.
Materials and Methods
Plant and Insect Materials. Tomato (Solanum lycopersicum cv. Betterboy) and
transgenic NahG and its wild-type cv. Moneymaker plants were grown in
Promix potting soil (Premier Horticulture) in a greenhouse. The laboratory
and field colonies of CPBs (L. decemlineata) were maintained separately.
Details are provided in SI Appendix, Materials and Methods.
Fluorescent Pictures and Movie of Regurgitation by CPB Larvae. To demon-
strate that CBPlarvaedepositOSs duringherbivory, thefluorescent dyeAlexa
Fluor 488 (Invitrogen) (38) was fed to the larvae, which were then trans-
ferred to fresh leaf tissues. Details are provided in SI Appendix, Materials
Antibiotics Treatment. To reduce microbes present in OSs from larvae, larvae
were fed antibacterial and/or antifungal agents (SI Appendix, Materials
SEM Images. Leaves were damaged by larvae that fed on AB-treated or
untreated artificial diets for 2–3 d. The larvae were then transferred to fresh
leaves. After 1 h of herbivory, the damaged sections of leaves were prepared
for SEM (details are provided in SI Appendix, Materials and Methods).
Herbivore Treatment and Application of OSs to Wounded Plants. To evaluate
the impact of larval feeding on induced plant responses, one AB-treated or
untreated larva was placed on the terminal leaflet of each four-leaf stage
plant using clip cages. Undamaged control plants received the clip cage
without insects. After 100% of the area confined by the cage was consumed,
leaf tissues were harvested for gene expression and defense enzymatic ac-
tivity (polyphenol oxidase activity). To verify that observed plant responses
required the presence of microbes in the OSs from the larvae and not just
wounding, OSs from AB-treated or untreated larvae were applied to
mechanically wounded plants. Details are provided in SI Appendix, Materials
Impact of JA-Mediated Defense Suppression on CPB Performance. To in-
vestigate whether the suppression of JA-dependent plant defenses enhances
larval performance, we measured larval growth. Plants were damaged by AB-
treated or untreated larvae that were confined toa single leaflet as described
above and the damaged leaflets were detached after 48 h. One neonate was
then placed on an excised leaflet in a 1-oz cup containing 1% agar to
maintain leaf moisture. Control larvae received the leaflets from undamaged
treatment. To collect subsamples for PPO activity, one leaf disc (diameter, 1.8
cm) was removed surrounding the leaf tissue that had been consumed by
the larvae and two leaf discs from each of two leaves were pooled as one
replicate. This feeding experiment was repeated four times.
Chung et al.PNAS Early Edition
| 5 of 6
PPO Activity. PPO activities for all experiments were measured 48 h after Download full-text
treatment using caffeic acid as the substrate as described previously (39).
Quantification of JA and SA. Plants were damaged by AB-treated or untreated
larvae that were confined to a single leaflet as described above. Leaf tissues
were harvested 2, 4, 24, and 48 h after placing insects. JA and SA were ex-
tracted and measured using GC/MS as described previously (40).
DNA and RNA Extractions and Quantitative Real-Time PCR. Leaf tissues were
ground in liquid nitrogen. RNA was extracted with an RNeasy Plus Mini kit
(Qiagen) following the manufacturer’s protocol and quantified using
NanoDrop (Thermo Scientific). Total genomic DNA from plants was extrac-
ted using DNeasy Plant Mini kit (Qiagen) following the manufacturer’s
protocol and quantified using NanoDrop (Thermo Scientific). Details are
provided in SI Appendix, Materials and Methods.
Isolation of Bacteria in OSs and Application of the Bacteria to Wounded Plants.
To isolate bacteria in OSs from CPB larvae, OSs was collected from AB-un-
treated larvae and cultured on 2xYT agar plates. Cultured bacterial isolates
were applied to wounded plants. Details are provided in SI Appendix,
Materials and Methods.
Reinoculation of Symbiotic Bacteria to Larvae and Zone of Inhibition Assays. To
verify that the cultured bacterial isolates found to suppress JA-dependent
responses are indeed secreted by larvae onto leaves, we reintroduced the
bacterial isolates to AB-treated larvae and again evaluated plant responses.
Details are provided in SI Appendix, Materials and Methods.
DNA Extraction, PCR, 16S rRNA, and rpoD Gene Sequencing to Taxonomically
Classify Bacterial Isolates. Direct colony PCR followed by Sanger sequencing
was performed toidentify bacterial isolates using universal 16S rRNA primers.
Details are provided in SI Appendix, Materials and Methods.
Flagellin Purification and Identification. Flagellin was purified as described
previously (41). Details are provided in SI Appendix, Materials and Methods.
Statistical Analysis. All data were analyzed using analysis of variance or
Student’s t test. Details are provided in SI Appendix, Materials and Methods.
ACKNOWLEDGMENTS. We thank J. Kim for helpful discussion, J. H. Tumlinson
and N. McCartney for phytohormone extraction, and P. Schmitt for use of
ultracentrifuge. This work was supported by US Department of Agriculture–
National Institute of Food and Agriculture Grant 2011-67013-30352 (to G.W.F.
and D.S.L.) and National Science Foundation Grant IOS-1256326 (to G.W.F., C.R.,
1. Howe GA, Jander G (2008) Plant immunity to insect herbivores. Annu Rev Plant Biol
2. Felton GW, Tumlinson JH (2008) Plant-insect dialogs: Complex interactions at the
plant-insect interface. Curr Opin Plant Biol 11(4):457–463.
3. Pieterse CMJ, Leon-Reyes A, Van der Ent S, Van Wees SCM (2009) Networking by
small-molecule hormones in plant immunity. Nat Chem Biol 5(5):308–316.
4. Thaler JS, Humphrey PT, Whiteman NK (2012) Evolution of jasmonate and salicylate
signal crosstalk. Trends Plant Sci 17(5):260–270.
5. Walling LL (2009) Advances in Botanical Research, ed Van Loon LC (Academic, Lon-
don), pp 551–612.
6. Musser RO, et al. (2002) Herbivory: Caterpillar saliva beats plant defences. Nature
7. Bos JIB, et al. (2010) A functional genomics approach identifies candidate effectors
from the aphid species Myzus persicae (green peach aphid). PLoS Genet 6(11):
8. Consales F, et al. (2012) Insect oral secretions suppress wound-induced responses in
Arabidopsis. J Exp Bot 63(2):727–737.
9. Atamian HS, et al. (2013) In planta expression or delivery of potato aphid Macro-
siphum euphorbiae effectors Me10 and Me23 enhances aphid fecundity. Mol Plant
Microbe Interact 26(1):67–74.
10. Weech M-H, Chapleau M, Pan L, Ide C, Bede JC (2008) Caterpillar saliva interferes with
induced Arabidopsis thaliana defence responses via the systemic acquired resistance
pathway. J Exp Bot 59(9):2437–2448.
11. Oliver KM, Degnan PH, Burke GR, Moran NA (2010) Facultative symbionts in aphids
and the horizontal transfer of ecologically important traits. Annu Rev Entomol
12. Kaiser W, Huguet E, Casas J, Commin C, Giron D (2010) Plant green-island phenotype
induced by leaf-miners is mediated by bacterial symbionts. Proc R Soc B Biol Sci
13. Barr KL, Hearne LB, Briesacher S, Clark TL, Davis GE (2010) Microbial symbionts in
insects influence down-regulation of defense genes in maize. PLoS ONE 5(6):e11339.
14. Hosokawa T, Kikuchi Y, Shimada M, Fukatsu T (2007) Obligate symbiont involved in
pest status of host insect. Proc R Soc B Biol Sci 274(1621):1979–1984.
15. Robert CAM, et al. (2013) Direct and indirect plant defenses are not suppressed by
endosymbionts of a specialist root herbivore. J Chem Ecol 39(4):507–515.
16. Lawrence SD, Novak NG, Blackburn MB (2007) Inhibition of proteinase inhibitor
transcripts by Leptinotarsa decemlineata regurgitant in Solanum lycopersicum.
J Chem Ecol 33(5):1041–1048.
17. Lawrence SD, Novak NG, Ju CJ, Cooke JE (2008) Potato, Solanum tuberosum, defense
against Colorado potato beetle, Leptinotarsa decemlineata (Say): Microarray gene
expression profiling of potato by Colorado potato beetle regurgitant treatment of
wounded leaves. J Chem Ecol 34(8):1013–1025.
18. Chung SH, Felton GW (2011) Specificity of induced resistance in tomato against
specialist lepidopteran and coleopteran species. J Chem Ecol 37(4):378–386.
19. Brunelle F, Girard C, Cloutier C, Michaud D (2005) A hybrid, broad-spectrum inhibitor
of Colorado potato beetle aspartate and cysteine digestive proteinases. Arch Insect
Biochem Physiol 60(1):20–31.
20. Girard C, et al. (2007) A multicomponent, elicitor-inducible cystatin complex in to-
mato, Solanum lycopersicum. New Phytol 173(4):841–851.
21. Thipyapong P, Stout MJ, Attajarusit J (2007) Functional analysis of polyphenol oxi-
dases by antisense/sense technology. Molecules 12(8):1569–1595.
22. van Kan JA, Joosten MH, Wagemakers CA, van den Berg-Velthuis GC, de Wit PJ (1992)
Differential accumulation of mRNAs encoding extracellular and intracellular PR
proteins in tomato induced by virulent and avirulent races of Cladosporium fulvum.
Plant Mol Biol 20(3):513–527.
23. Strand MR (2010) Insect Virology, eds Asgari S, Johnson KN (Caister Academic Press,
Wymondham, UK), pp 171–197.
24. Zhao Y, et al. (2003) Virulence systems of Pseudomonas syringae pv. tomato promote
bacterial speck disease in tomato by targeting the jasmonate signaling pathway.
Plant J 36(4):485–499.
25. Jones JDG, Dangl JL (2006) The plant immune system. Nature 444(7117):323–329.
26. Sarmento RA, et al. (2011) A herbivore that manipulates plant defence. Ecol Lett
27. Takabayashi J, Shimoda T, Dicke M, Ashihara W, Takafuji A (2000) Induced response
of tomato plants to injury by green and red strains of Tetranychus urticae. Exp Appl
28. Zarate SI, Kempema LA, Walling LL (2007) Silverleaf whitefly induces salicylic acid
defenses and suppresses effectual jasmonic acid defenses. Plant Physiol 143(2):
29. Casteel CL, Hansen AK, Walling LL, Paine TD (2012) Manipulation of plant defense
responses by the tomato psyllid (Bactericerca cockerelli) and its associated endo-
symbiont Candidatus Liberibacter psyllaurous. PLoS ONE 7(4):e35191.
30. Gruden K, et al. (2004) Molecular basis of Colorado potato beetle adaptation to
potato plant defence at the level of digestive cysteine proteinases. Insect Biochem
Mol Biol 34(4):365–375.
31. Blackburn MB, Gundersen-Rindal DE, Weber DC, Martin PAW, Farrar RR, Jr. (2008)
Enteric bacteria of field-collected Colorado potato beetle larvae inhibit growth of the
entomopathogens Photorhabdus temperata and Beauveria bassiana. Biol Control
32. Muratoglu H, Demirbag Z, Sezen K (2011) The first investigation of the diversity of
bacteria associated with Leptinotarsa decemlineata (Coleoptera: Chrysomelidae). Bi-
ologia (Bratisl) 66(2):288–293.
33. Tsuda K, Sato M, Glazebrook J, Cohen JD, Katagiri F (2008) Interplay between MAMP-
triggered and SA-mediated defense responses. Plant J 53(5):763–775.
34. Robatzek S, et al. (2007) Molecular identification and characterization of the tomato
flagellin receptor LeFLS2, an orthologue of Arabidopsis FLS2 exhibiting characteris-
tically different perception specificities. Plant Mol Biol 64(5):539–547.
35. Belliure B, Janssen A, Maris PC, Peters D, Sabelis MW (2005) Herbivore arthropods
benefit from vectoring plant viruses. Ecol Lett 8(1):70–79.
36. Abe H, et al. (2012) Antagonistic plant defense system regulated by phytohormones
assists interactions among vector insect, thrips and a tospovirus. Plant Cell Physiol
37. Sugio A, Kingdom HN, MacLean AM, Grieve VM, Hogenhout SA (2011) Phytoplasma
protein effector SAP11 enhances insect vector reproduction by manipulating plant
development and defense hormone biosynthesis. Proc Natl Acad Sci USA 108(48):
38. Peiffer M, Felton GW (2009) Do caterpillars secrete “oral secretions”? J Chem Ecol
39. Felton GW, Donato K, Vecchio RJ, Duffey SS (1989) Activation of plant foliar oxidases
by insect feeding reduces nutritive quality of foliage for noctuid herbivores. J Chem
40. Tooker JF, Rohr JR, Abrahamson WG, De Moraes CM (2008) Gall insects can avoid and
alter indirect plant defenses. New Phytol 178(3):657–671.
41. Felix G, Duran JD, Volko S, Boller T (1999) Plants have a sensitive perception system
for the most conserved domain of bacterial flagellin. Plant J 18(3):265–276.
6 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1308867110Chung et al.