Cause of mortality in insects under severe stress
Hitoshi Matsumoto, Kohjiro Tanaka, Hirofumi Noguchi and Yoichi Hayakawa
Institute of Low Temperature Science, Hokkaido University, Sapporo, Japan
Mortality in the host armyworm larvae Pseudaletia separata
parasitized by the parasitic wasp Cotesia kariyai was dra-
matically increased when they were simultaneously infected
by the entomopathogen Serratia marcescens. Previous
studies have shown that this strong insecticidal effect is due
to a metalloprotease-like insecticide (MPLI) released from
S. marcescens enterobacter. This study was conducted to
elucidate the exact cause of the mortality resulting from
MPLI.InjectionofMPLIcausedasharpincreasein
hemolymph dopamine concentration followed by elevated
levels of brain dopamine in armyworm larvae. [
3
H]Dop-
amine injected into the hemocoel, was incorporated into the
brains of MPLI-injected larvae to a level eight times greater
than in BSA-injected control larvae. Transmission electron
microscopy showed an obvious decrease in thickness and
density of the brain sheath in insects injected with MPLI.
This was probably due to the MPLI-induced elevation of
hemocyte metalloprotease activities. Further, electron
microscopic and TUNEL staining analyses showed a signi-
ficant increase in apoptotic cells in the brain 12 h after the
injection. Injection of 3-iodotyrosine (a tyrosine hydroxylase
inhibitor) before MPLI completely prevented the increase in
hemolymph dopamine in test larvae and their following
death. From these observations, we conclude that MPLI-
injected larvae may have suffered mortal damage through
increased apoptosis of brain cells caused by an influx of

microsporidia before or after parasitization, their premature
deaths have been observed before the wasp larvae have
completed their development [6]. In fact, most host Pseu-
daletia separata larvae die within 3 days of parasitization
by the wasp Cotesia kariyai when they are simultaneously
infected by the enterobacter Serratia marcescens. Previous
studies have shown that this mortality is mainly due to
metalloprotease-like insecticide (MPLI) released by S. mar-
cescens enterobacter [7]. Purified MPLI showed a strong
insecticidal effect with a median lethal dosage (LD
50
) of
13 pmol per larva. In preliminary experiments, we injected
purified MPLI into mice with the same dose per weight to
that used for the armyworm larvae, but we did not observe
any symptoms or disorders.
In this study, we tried to extended these experiments to
elucidate the mechanism by which MPLI kills armyworm
larvae within a few days of injection. Our results indicate
that the dopamine concentration in the hemolymph was
elevated by the injection of MPLI, resulting in influx of
dopamine into the brain through the externally damaged
sheath. At the same time, apoptosis of brain cells was
observed in the test larvae.
Materials and methods
Animals
P. separata larvae were reared on an artificial diet at
25 ± 1 °C with a photoperiod of 16-h light : 8-h dark.
Correspondence to Y. Hayakawa, Institute of Low Temperature
Science, Hokkaido University, Sapporo, Japan 060-0819.

i
(8 m
M
NaH
2
PO
4
,1.5m
M
KH
2
PO
4
, 137 m
M
NaCl, 2.7 m
M
KCl, pH 7.2) to 0.1 lgÆlL
)1
(final volume
10 lL) and injected into day-1 last-instar larvae of the
armyworm 8 h after lights on.
A 3-iodotyrosine-saturated solution was made by mixing
 5 mg 3-iodotyrosine with 1 mL NaCl/P
i
. Before injection
of MPLI, 10 lL of the 3-iodotyrosine solution was injected
twice into day-2 penultimate-instar and day-0 last-instar
larvae 6 h after lights on. MPLI was injected 2 h after the
second injection of 3-iodotyrosine.

, and injected into larvae 6 h after injection of
BSAorMPLI.After1h,thebrainwasdissected,washed
three times with NaCl/P
i
, and immediately homogenized
with 100 lL0.2
M
perchloric acid. Dopamine was separated
by paper chromatography, and its radioactivity counted
using a liquid-scintillation counter (Aloka LSC-5100).
Microscopic observation
Brains were dissected from test armyworm larvae and fixed
with 2.5% glutaraldehyde and 1% paraformaldehyde in
Pipes buffer (0.1
M
Pipes, 0.05
M
sucrose, pH 7.4) at 4 °C.
Post-fixation and staining was performed in 2% aqueous
OsO4 and 2% uranyl acetate, respectively. The tissue was
embedded in Epon 812 (TAAB Laboratories Equipment
Ltd, Aldermaston, Berkshire, UK) after dehydration. Thin
sections were cut on an Ultracut (Reichert-Jung, Wien,
Austria). For electron microscopy, thin sections were briefly
stained in 2% aqueous uranyl acetate and 0.1% lead citrate
[11]. Micrographs were taken with a JEM-1200EX (Jeol
Ltd) electron microscope.
Assay of metalloprotease activity
Hemolymph was collected into a chilled microtest tube
containing NaCl/P

i
andusedasbrainandfatbody
samples, respectively.
Three metalloprotease substrates (1,NH
2
-Mec-Ac-
Arg-Pro-Lys-Pro-Tyr-Ala-Nva-Trp-Met-Lys(Dnp)-NH
2
; 2,
NH
2
-Mec-Ac-Asp-Glu-Val-Asp-Ala-Pro-Lys(Dnp)-NH
2
; 3,
NH
2
-Mec-Ac-Pro-Leu-Gly-Leu-A2pr(Dnp)-Ala-Arg-NH
2
)
were solubilized in dimethyl sulfoxide (final concentration
10 m
M
) and used as stock solutions. The reaction mixture
(total volume 190 lL) consisted of 50 m
M
Tris/HCl buffer
(pH 7.5), 0.1
M
NaCl, 10 m
M

larval brains
Dopamine is the most abundant catecholamine in the insect
hemolymph and nervous tissues, where it may serve as a
hormone, neuromodulator and neurotransmitter [16,17].
Fig. 3. Morphological changes of the neural
sheath layer of MPLI-injected larval brains.
Brains were dissected from day-0 last-instar
larvae of the armyworm 12 h after injection of
17.4 pmol per larva of BSA (A,C) or MPLI
(B,D), and observed at a magnification
of ·5000 (A,B) and · 20 000 (C,D). Note that
the neurilemma of MPLI-injected larva is
thinner (indicated with bars shown in A and
B) and less dense (indicated with stars as
shown in C and D) than that of control BSA-
injected larva. Further, a gap between the
neurilemma and perineural cells is visible in
MPLI-injected larval brain (as indicated in D
with an arrow).
Fig. 2. [
3
H]Dopamine incorporation into MPLI-treated larval brains.
Brains were dissected from day-0 last-instar larvae of the armyworm
which had been injected with 17.4 pmol per larva of MPLI (closed
bar)orBSA(openbar)andtheninjectedwith10lCi [
3
H]dopamine
per larva. Radioactive dopamine was quantified as described in
Materials and methods. Each bar represents the mean ± SD from
four independent determinations.

The dopamine influx into the MPLI-injected larval brain
suggested that brains in the larvae may be damaged by the
MPLI treatment. To address this, brains of the MPLI-
injected larvae were analyzed by transmission electron
microscopy. Differences in the thickness and density of the
neurilemma are evident in Fig. 3: the neurilemma of the
MPLI-injected larvae is thinner and less dense than that of
the control larvae. Further, the perineural cells, which are
tightly attached to the neurilemma in the control brains, are
slightly separated from the neurilemma in the MPLI-treated
larval brains.
MPLI-induced enhancement of metalloprotease activity
in hemocytes
The structural changes in the brain sheath caused by the
MPLI treatment suggested that the neurilemma matrix
may be degraded by MPLI-induced proteolysis. To test this,
we first confirmed metalloprotease activity of MPLI using
three commercially available fluorogenic peptide substrates
[1,NH
2
-Mec-Ac-Arg-Pro-Lys-Pro-Tyr-Ala-Nva-Trp-Met-
Lys(Dnp)-NH
2
; 2,NH
2
-Mec-Ac-Asp-Glu-Val-Asp-Ala-
Pro-Lys(Dnp)-NH
2
; 3,NH
2

shown), indicating that MPLI did not stimulate the sole
metaloprotease in hemocytes after injection of MPLI.
However, there was no increase in the hydrolyzing activity
with synthetic substrates for serine proteases such as Var-
Pro-Arg-NH
2
-Mec, Ile-Glu-Gly-Arg- NH
2
-Mec, Phe-Ser-
Arg- NH
2
-Mec or Gln-Arg-Pro- NH
2
-Mec (data not
shown). Thus, as we speculated, injection of MPLI activated
hemocyte metalloproteases, which contributed to the deg-
radation of the neurilemmal matrix.
MPLI-induced apoptosis of brain cells
The final question is whether MPLI-induced elevation of
brain dopamine levels results in changes in the brain cells.
To examine this, brains dissected from the larvae 20 h after
injection of MPLI were studied by transmission electron
microscopy. Condensed chromatin was observed in the
brain cells of the MPLI-injected larvae (Fig. 5A), suggesting
that the brain cells were undergoing apoptosis as the result
of the MPLI injection. This was substantiated by evidence
that the number of TUNEL-positive cells was increased in
the brains of the larvae 12 h after injection of MPLI.
Prevention of the MPLI-induced dopamine elevation
decreased the insecticidal effect of MPLI

separate pools of dopamine: nervous system and integu-
ment [16,17,21,22]. Dopamine is the most abundant mono-
amine in the nervous system and may serve as a
neurotransmitter and neuromodulator [15,16]. Extremely
high concentrations of dopamine are present in the integu-
ment where it is used as an essential intermediate of cross-
linking agents in cuticle formation throughout insect
Fig. 6. Insecticidal effect of dopamine in hemolymph. (A) Hemolymph dopamine levels in armyworm larvae treated with MPLI and/or
3-iodotyrosine. Day-0 last-instar larvae of the armyworm were injected with 17.4 pmol per larva of MPLI (n ¼ 4) or BSA (n ¼ 4). 3-Iodotyrosine
was administered previously to the test larvae as described in Materials and methods. Hemolymph was collected for dopamine measurement 6 h
after injection with MPLI or BSA. Each column represents the mean ± SD for the number of determinations in parentheses. (B) Survival of larvae
treated with MPLI and/or 3-iodotyrosine. Day-0 last-instar larvae injected with 17.4 pmol per larva of MPLI (n ¼ 10) (m). Larvae pretreated with
3-iodotyrosine before injection with 17.4 pmol per larva of MPLI (n ¼ 10) (h). Larvae injected with 17.4 pmol per larva of BSA (n ¼ 10) (d).
This result was a typical case from four independent experiments, but the probabilities of significant survival difference between 3-iodotyrosine-
treated and nontreated animals were 100%. (C, D) TUNEL-staining of brains of the armyworm larvae pretreated with BSA (C) or 3-iodotyrosine
(D) 12 h after injection of MPLI. Note that 3-iodotyrosine pretreatment decreased the number of TUNEL-positive neural cells. Other explanations
as in Fig. 5.
3474 H. Matsumoto et al.(Eur. J. Biochem. 270) Ó FEBS 2003
development [21,22]. Previous studies indicate that the
dopamine concentration in the integument is about 50 times
that found in the hemolymph. Further, we found that
integument dopamine was secreted into incubation medium
in vitro [19]. Therefore, it is reasonable to expect that
integument dopamine is released into the hemolymph. If
this is true, the large increase in hemolymph dopamine
concentrationinMPLI-injectedlarvae(showninFig.1)
would also be due to its release from the integument.
Even though the dopamine concentration is significantly
increased in the hemolymph, dopamine cannot normally
penetrate the hemolymph/brain barrier because it is thought

such as parasitized insects simultaneously infected with the
entomopathogen, but also to dying insects under severe
stress. Further studies should improve our understanding of
the fundamental role of dopamine in insects as well as the
molecular mechanism of their death.
Acknowledgements
This work was supported by the Program for Promotion of Basic
Research Activities for Innovative Biosciences (Japan).
References
1. Heslop, J.P. & Ray, J.W. (1959) The reaction of cockroach Peri-
planeta americana L. to bodily stress and DDT. J. Insect Physiol. 3,
395–401.
2. Beament, J.W.L. (1958) A paralyzing agent in the blood of
cockroaches. J. Insect Physiol. 2, 199–214.
3. Cook, B.J. & Holman, G.M. (1985) Peptides and kinins. In Com-
prehensive Insect Physiology, Biochemistry and Pharmacology
(Kerkut,G.A.&Gilbert,L.I.,eds),Vol. 11, pp. 531–593. Perga-
mon Press, Oxford.
4. Ewing, L.S. (1967) Fighting and death from stress in a cockroach.
Science 155, 1035–1036.
5. Cook, B.J. & Holt, G.G. (1974) Neurophysiological changes
associated with paralysis arising from body stress in the cockroach
Periplaneta americana. J. Insect Physiol. 20, 21–40.
6. Brooks, W.M. (1993) Host–parasitoid–pathogen interaction. In
Parasites and Pathogens of Insects (Beckage, N.E., Thompson,
S.N. & Federici, B.A., eds), Vol. 2, pp. 231–272. Academic Press,
New York.
7. Matsumoto, H., Noguchi, H. & Hayakawa, Y. (1998) Primary
cause of mortality in the armyworm larvae simultaneously para-
sitizedbyparasiticwaspandinfectedwithbacteria.Eur. J. Bio-

Plenum Press, New York.
17. Evans, P.D. (1980) Biogenic amines in the insect nervous system.
Adv. Insect Physiol. 15, 317–473.
18. Noguchi, H., Hayakawa, Y. & Downer, R.G.H. (1995) Elevation
of dopamine levels in parasitized insect larvae. Insect Biochem.
Mol. Biol. 25, 197–201.
19. Noguchi, H. & Hayakawa, Y. (1996) Mechanism of parasitism-
induced elevation of dopamine levels in host insect larvae. Insect
Biochem. Mol. Biol. 26, 659–665.
20. Civelli, O., Bunzow, J.B. & Grandy, D.K. (1993) Molecular
diversity of the dopamine receptors. Annu. Rev. Pharmacol.
Toxicol. 33, 281–307.
21. Andersen, S.O., Rafn, K., Krogh, T.N., Hojrup, P. & Roepstorff,
P. (1995) Comparison of larval and pupal cuticular proteins in
Tenebrio molitor. Insect Biochem. Mol. Biol. 25, 177–187.
22. Hopkins, T.L., Morgan, T.D. & Kramer, K.J. (1984)
Catecholamines in hemolymph and cuticle during larval, pupal
and adult development of Manduca sexta (L.). Insect Biochem. 14,
533–540.
23. Scharrer, B.C.J. (1939) The differentiation between neuroglia and
connective tissue sheath in the cockroach (Periplaneta americana).
J. Comp. Neurol. 70, 77–88.
Ó FEBS 2003 Death from stress in insects (Eur. J. Biochem. 270) 3475
24. Wigglesworth, V.B. (1960) The nutrition of the central nervous
system in the cockroach, Periplaneta americana L.Theroleof
perineurium and glial cells in the mobilization of reserves. J. Exp.
Biol. 37, 500–512.
25. Skaer, H.L.E.B. & Lane, N.J. (1974) Junctional complexes, peri-
neurial and glia-axonal relationships and the ensheathing struc-
tures of the insect nervous system; a comparative study using