Journal of

www.elsevier.com/locate/yjipa

Journal of Invertebrate Pathology 96 (2007) 125–132

INVERTEBRATE

PATHOLOGY

Tolerance to Bacillus thuringiensis endotoxin inimmune-suppressed larvae of the flour moth Ephestia kuehniella

M. Mahbubur Rahman, Harry L.S. Roberts, Otto Schmidt *

Insect Molecular Biology Laboratory, School of Agriculture and Wine, University of Adelaide, Glen Osmond, SA 5064, Australia

Received 12 November 2006; accepted 19 March 2007Available online 2 April 2007

Abstract

Tolerance to Bacillus thuringiensis crystal endotoxins (Bt-toxins) is correlated with an elevated immune status in larvae of the flourmoth Ephestia kuehniella. To gain more specific information about the effector pathways involved in the protection against the toxin, westudied the effects of Bt-toxin formulations in susceptible (non-induced) and tolerant (immune-induced) larvae after natural (parasitism-mediated) and chemical (tropolone-mediated) suppression of defence reactions. Although melanization in hemolymph was significantlyreduced, there was no significant effect on susceptibility to the toxin in parasitised or tropolone-treated larvae. This suggests that mel-anization of hemolymph is correlated with an elevated immune status but not responsible for the observed tolerance to Bt-toxin. Toexamine whether hemolymph proteins exist in the gut lumen and function as pro-coagulants, we compared gut and plasma proteinsof immune-induced with those of non-induced larvae. Here we show that the lipid carrier lipophorin represents a major componentin the gut lumen and interacts with mature Bt-toxin to form a complex.� 2007 Elsevier Inc. All rights reserved.

Keywords: Ephestia kuehniella; Crystal toxin; Bacillus thuringiensis; Melanization; Coagulation; Tropolone; Parasitoid; Prophenoloxidase; Lipophorin

1. Introduction

Previous observations indicate that the immune systemof lepidopteran larvae is induced after feeding on a sub-lethal dose of Bt-formulations. The immune induction inthe hemolymph is probably caused by localised cellulardamage to the gut lining, exposing hemolymph and hemo-cytes to elicitors from the gut lumen (Ma et al., 2005; Rah-man et al., 2004a). The elevated immune status, which canbe correlated conveniently with the degree of plasma mel-anization, provides larvae with a limited but neverthelesssignificant protection against subsequent application oflethal doses of the Bt-formulation (Rahman et al.,2004a). The most likely explanation for the tolerance tothe toxin is the secretion of immune components into thegut of immune-induced larvae, where pathogens and toxinscan be inactivated by cell-free defence reactions (Rahman

0022-2011/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.jip.2007.03.018

* Corresponding author. Fax: +61 8303 7109.E-mail address: otto.schmidt@adelaide.edu.au (O. Schmidt).

et al., 2006). More specifically, the aggregation of a pro-coagulant by a lectin-like mature Bt-toxin can potentiallyinactivate the toxin in the gut lumen before it can reachthe brush border membrane (Sarjan, 2002).

Given that prophenoloxidase (PPO) and lipid particles,acting like pro-coagulants, are co-purified (Rahmanet al., 2006), which implies that melanization and coagula-tion reactions are closely interconnected (Li et al., 2002),the question is whether melanization is responsible forthe observed Bt-tolerance in immune-induced larvae? Ini-tial observations suggest that this is not the case: Firstly,while Bt-tolerant Helicoverpa armigera larvae show dark-ened peritrophic membranes (Ma et al., 2005), we did notobserve any signs of melanization in the gut tissue or thegut lumen of immune-induced Ephestia kuehniella larvae.Since PPO is involved in reactive oxygen production andprotein cross-linking apart from melanin synthesis, theobserved melanization in the gut of H. armigera may beone of several possible activities but not necessarily essen-tial. Thus in immune-induced E. kuehniella larvae the PO

126 M. Mahbubur Rahman et al. / Journal of Invertebrate Pathology 96 (2007) 125–132

activity in the gut may not be increased or become involvedin other than melanization reactions. Second, while an ele-vated immune status protects against some gut-derivedpathogens (Reeson et al., 1998), it does not cross-protectagainst insect parasitoids that oviposit inside the hemocoel(Rahman et al., 2004b).

This raises two questions: Firstly, how does immunesuppression by parasitoids affect melanization levels andBt-toxicity? Hymenopteran parasitoids that lay their eggsinside the hemocoel of another insect suppress the host’simmune system, precluding immune-mediated inactivationof the egg and emerging larvae (Schmidt et al., 2001). Thisincludes protection of the growing larvae against negativeeffects of melanization and coagulation of immune-activehost plasma inside the gut of the developing parasitoid.The suppression of the host immune system, which occursby maternal protein secretions of the female parasitoid,involving serpins (Beck et al., 2000; Nappi et al., 2005)and other immune suppressors (Asgari et al., 2003; Lavineand Beckage, 1995; Shelby et al., 2000) prevent activationof host melanization and coagulation reactions allowingthe development of the parasitoid inside the host larvae(Rahman et al., 2004b). The question is whether immunesuppression by parasitoids, which increases virulence ofviral pathogens (Washburn et al., 1996), can also increaseBt-toxicity in parasitised larvae.

The second question is whether melanization, which is aconvenient indicator of the immune status, is the causeof Bt-tolerance or represents an independent pathwaythat is less relevant to Bt-toxicity and Bt-tolerance. Weapproached these questions by in vivo suppression of melan-ization involving the inhibition of phenoloxidase by a metalion-chelating agent tropolone (Morita et al., 2003). Inhemolymph plasma tropolone interacts mainly with copper(Nomiya et al., 2004) present in active phenoloxidase(Chase et al., 2000; Fujimoto et al., 1995; Hall et al., 1995;Kawabata et al., 1995; Jiang et al., 1997) as an essentialredox-system of the enzyme (Decker and Terwilliger, 2000).

Here we show that, while plasma melanization is corre-lated with the immune status, it is not responsible for theobserved tolerance to the Bt-toxin in the gut lumen.Instead, the lipid carrier lipophorin, which is a pro-coagu-lant in hemolymph, is increased in the gut lumen of Bt-tol-erant larvae and able to bind to the toxin. This suggeststhat Bt-toxin can potentially be inactivated in the gutlumen by an aggregation of lipophorin particles sequester-ing the toxin into coagulation products.

2. Materials and methods

2.1. Insects

A Bt-susceptible colony of E. kuehniella was derivedfrom a long established laboratory colony that had beenmaintained without selection for over 100 generations.The Bt-tolerant E. kuehniella originated from the colonydescribed in Rahman et al. (2004a), and subsequently

maintained on diet containing 4000 or 8000 ppm of a Bt-formulation comprising a commercial formulation of Btendotoxins (DelfinWG, Sandoz (now Syngenta, NorthRyde, NSW), containing crystal toxins Cry1Aa, Cry1Ab,Cry1Ac, Cry2Aa, detergents, bacteria debris and spores.Both Bt-susceptible and a sub-population of Bt-tolerantlarvae used in the study were maintained on toxin-free dietof a 10:2:1 mixture of oat bran: wheat germ: brewers yeast.Before each experiment the Bt-tolerant larvae were kept ontoxin-free diet for a week, which removed any of the com-ponents of the Bt-formulation from the mid-gut.

The wasps were a clonal RP line of a thelytokousVenturia canescens laboratory culture previously estab-lished as reported by Beck et al. (1999). All experimentswere conducted and cultures maintained at 25 ± 1 �C,under a constant light-dark regime (L14:D10).

To obtain singly parasitised E. kuehniella larvae a singleV. canescens wasp was put together with 25–30 host larvae(3. and 4. instar) in a plastic container (7 cm in diameter,8 cm high). The parasitoids were observed during oviposi-tion and stinging attempts that resulted in a startleresponse from the larvae, combined with the characteristiccocking movements of the wasp’s ovipositor (Rogers, 1972)were considered as real oviposition events. Any stingingattempt that either did not evoke a startle response orwas not followed by a cocking movement was regardedas uncertain and the larva was discarded.

2.2. Melanization assays

Eight to ten larvae were chilled on ice for 5 min, washedwith ice-cold 70% ethanol and then ice-cold PBS. Haemol-ymph was extracted by cutting off a foreleg and bleedingeach larva directly into 1.5 ml ice-cold PBS. The solutionwas centrifuged for 5 min at 3000g and the cell-free super-natant transferred to a cuvette. The absorbance was firstmeasured at 280 nm to determine the relative protein con-centration, and then the absorbance at 490 nm wasrecorded every minute for 90 min on a Varian DMS100 s spectrophotometer.

Haemolymph melanization assays of parasitised larvaewere conducted 0, 2, 4, 6, 12, 24, 36, 48, 72 and 96 h post-par-asitism. To investigate the effects of physical insult on themelanization response, unparasitised larvae from the Bt-sus-ceptible line were each pierced once by a sterile micropin ofsimilar diameter to the ovipositor of V. canescens, and mel-anization assays conducted after 12 h. At least 5 replicateswere performed for each treatment. To investigate the effectsof ingested tropolone on the melanization response, 4thinstar larvae from the Bt-tolerant and susceptible lines weremaintained on standard diet containing 1 ppm tropolone for24 h and melanization assays conducted.

2.3. Bt-toxin bioassay of parasitised larvae

Three mating pairs of adult E. kuehniella from bothBt-susceptible and tolerant (4000 ppm F7 colony, LC50

M. Mahbubur Rahman et al. / Journal of Invertebrate Pathology 96 (2007) 125–132 127

2591 ppm) colonies were allowed to lay eggs on fresh foodfor two days at the control temperature. After 25 days halfthe resulting 4th instar larvae from each colony were singlyparasitised by V. canescens wasp (RP strain) as describedabove.

The parasitised and unparasitised larvae from both sus-ceptible and tolerant strains were then allotted into groupsof 20 larvae and transferred to diet containing, variously: 0,100, 1000, 2000, 4000 and 8000 ppm of the Bt-formulation.There were five replicates at each concentration. The Bt-susceptible and Bt-tolerant larvae were maintained for 10days, and larval survivorship was recorded.

2.4. Low-density gradient centrifugation of gut content

Fifty to 100 third to fifth instar larvae were washed inethanol and dried on filter paper. The midgut was removedby cutting off head and terminal segments and homoge-nized in an ice-cold anticoagulant solution (30 mM triso-dium citrate, 26 mM citric acid, 20 mM EDTA, 15 mMNaCl, pH 5.5) containing phenylthiourea (PTU). Thecrude extract was centrifuged at 10,000g for 10 min and1.5 ml of gut extract was added to 15 ml of a solution of44.3 g KBr in 100 ml and overlaid with 0.9% NaCl to afinal volume of 30 ml. The tube was inserted into a SWTi42rotor and spun at 24,000 rpm for 17 h at 10 �C. The gradi-ent was eluted in ca 1 ml fractions covering densities of1.15 g/ml (top fractions) to 1.45 g/ml (bottom fractions).

The separation of lipophorin particles and lipid-freeplasma components was visible after centrifugation bywhitish protein bands in the low-density area of thegradient.

2.5. Protein staining in Western blots and whole mounts

Aliquots of dialysed low-density gradient fractions weremixed with loading buffer, heated at 65 �C and analysed bySDS–PAGE and Western blots using antisera againstrecombinant vWD from Galleria mellonella, which is partof apolipoprotein I (Ma et al., 2006) and Cry1Ac (kind giftfrom Sarjeet Gill). After staining gels with Coomassie blueor Western blots with phosphatase-conjugated secondaryantibodies, the apolipoprotein I band at 230 kDa was usedas an indication for the presence of soluble lipid particles.

Whole mount staining was performed as describedrecently (Ma et al., 2005). Briefly, Bt-tolerant and suscepti-ble third to fifth instar larvae were washed in ethanol anddried on filter paper. The midgut was removed by cuttingoff head and terminal segments and incubates in fixationbuffer (4% paraformaldehyde containing 0.5% Tween 20).After extensive washing in PBS containing 0.1% Tween20, gut tissues were incubated with the first antiserum over-night and washed three times before adding the secondaryantibody. Rabbit antiserum against G. mellonella lipopho-rin and vWD domain of apolipophorin I (a 230 kDa pro-tein) was from G. Ma (Ma et al., 2006), antiserumagainst Lucillia cuprina (blow fly) lipophorin labelling apo-

lipoprotein I and II (ca. 70 kDa protein) was kindly pro-vided by S. Trowell, CSIRO, Canberra, antiserum againstG. mellonella arylphorin was kindly provided by K. Schal-ler (Wurzburg, Germany) and antiserum against PPO fromManduca sexta was kindly donated by M. Kanost (KansasUniversity, Manhattan, KS, USA). Antibodies werediluted 1:1000 before incubation with whole gut tissuesand visualised with alkaline phosphatase conjugated goatanti-rabbit antibodies.

2.6. Statistical analysis

The rate of the melanization reaction was estimated foreach replicate as the slope of the plot of absorbance againsttime (Rahman et al., 2004a). The effects of time interval fol-lowing parasitism on the rate of the melanization reactionwere analysed by one-way ANOVA, followed by Dunnet’stest with unparasitised larvae treated as the control, usingthe statistical software package JMP V4.0.4 (SAS 2001).

Median lethal concentration (LC50) values were esti-mated by probit analysis using POLO-PC software (LeOraSoftware, Berkeley, CA). Samples were considered signifi-cantly different if the 95% confidence intervals (95% CI)did not overlap.

3. Results

3.1. Plasma melanization in parasitised larvae

Melanization assays of cell-free haemolymph fromunparasitised larvae of the Bt-susceptible (non-induced)E. kuehniella line showed that the melanization reactionproceeded at a low rate (Rahman et al., 2004a), whichwas almost undetectable. In contrast, cell-free haemolymphfrom unparasitised E. kuehniella larvae from the Bt-toler-ant (immune-induced) line showed a high rate of melaniza-tion (F = 157.8, df = 1, 108, p < 0.0001). While the elevatedrate of melanization is significantly reduced in parasitisedlarvae it remains at a higher rate than in susceptible parasi-tised larvae (Fig. 1). A one-way ANOVA revealed the mel-anization rates following parasitism by V. canescens

showed a transient increase, which were significant(F = 16.62, df = 8, 54, p < 0.0001), similar to the transientincrease in Bt-susceptible larvae, except that the rate ofmelanization was higher in tolerant larvae.

To examine whether the observed transient inductionwas based on wounding or other stress factors, melaniza-tion assays were performed after stabbing larvae with asterile needle. The rate of plasma melanization from Bt-susceptible (non-induced) larvae conducted 12 h after thephysical insult revealed that the reaction proceeded at anundetectable rate, with no significant difference fromuntreated controls (not shown).

This suggests that parasitism causes significant reduc-tion in hemolymph melanization, which does not affect tox-icity in the gut. If phenoloxidase in the hemolymph and gutare regulated independently, this raises the question

02468

1012141618

UP 2 6 12 24 36 48 72 96Hours after parasitism

PPO

act

ivity

(Mel

aniz

atio

n)

Susceptible TolerantSusceptible Tolerant

Fig. 1. Mean rates of the melanization reaction [as the slope of the plot ofabsorbance against time (arbitrary units)] of cell-free haemolymph fromE. kuehniella larvae at different time intervals following parasitism byV. canescens. Rates of melanization from Bt-susceptible (grey bars) andthe Bt-tolerant (black bars) larvae are shown for comparison at each timeinterval. Bars represent SEM.

128 M. Mahbubur Rahman et al. / Journal of Invertebrate Pathology 96 (2007) 125–132

whether Bt-toxicity is affected in larvae with reduced phe-noloxidase activity in the gut lumen?

3.2. Plasma melanization in tropolone-treated larvae

Plasma from 4th instar larvae that were maintained ondiet containing 1 ppm tropolone for 24 h showed a signifi-cantly lower rate of melanization in the Bt-tolerant(immune-induced) larvae compared to untreated controls(F = 39.19, df = 1, 12, p < 0.0001, Fig. 2). There was nosignificant difference in the melanization rate between trop-olone-treated Bt-tolerant (immune-induced) larvae andboth tropolone-treated and untreated Bt-susceptible (non-induced) larvae.

3.3. Tropolone-treatment and parasitism success

Analysis of the wasp’s developmental parameters suchas head capsule width, development time and survivorship

0

0.5

1

1.5

2

2.5

3

3.5

4

Control Tropolone Control Tropolone

Bt-tolerant Bt-susceptible

Rat

e of

mel

anis

atio

n re

actio

n

Fig. 2. Mean rates of the melanization reaction [as the slope of the plot ofabsorbance against time (arbitrary units)] of cell-free haemolymph fromBt-tolerant and Bt- susceptible line E. kuehniella larvae maintained on dietcontaining 1 ppm tropolone for 24 h. Bars represent SEM.

revealed no significant differences between V. canescens

emerging from singly parasitised tropolone-treated(80.0% survival, n = 25) and non-treated (79.2% survival,n = 24) E. kuehniella larvae. This suggests that tropolone-treatment does not affect suitability of larvae as a host ofthe parasitoid. Alternatively any effect that tropolone mayhave on the physiology of non-parasitised larvae mayhave been abolished by the effects of epistatic immune-suppressive maternal secretions in parasitised larvae.

3.4. Bt-tolerance in parasitised larvae

There were no significant differences between the LC50

values for parasitised and unparasitised larvae of eitherthe Bt-tolerant strain (parasitised: LC50=5787, 95%CI = 1660–679,412; unparasitised: LC50 = 5371, 95%CI = 3547–239,170) or Bt-susceptible strain (parasitised:LC50 = 1147, 95% CI = 561–1867; unparasitised: LC50 =1859, 95% CI = 1197–2526).

This suggests that the observed tolerance to the Bt-toxinin the gut is not affected by plasma-mediated immune-sup-pression in the hemolymph but is regulated independentlyin the gut lumen.

3.5. Bt-tolerance in tropolone-treated larvae

Since there are no specific inhibitors of coagulation, wedecided to inhibit melanization in Bt-tolerant (immune-induced) larvae to determine whether tropolone-treatmentof larvae, which causes specific inhibition of melanization,can affect Bt-toxicity. If melanization is regulated indepen-dently from other humoral defence reactions, such as coag-ulation, reduction in melanization may or may not affectsusceptibility to the toxin.

Although tropolone-treatment reduced the rate of mel-anization in Bt-tolerant (immune-induced) larvae (Fig. 2),tolerance to the toxin in tropolone-treated larvae was sim-ilar to non-treated larvae (Table 1). This suggests thatreduction in melanization has no effect on toxicity, whichimplies that melanization is part of the elevated immunestatus in Bt-tolerant larvae, but not responsible for the pro-tection against the toxin.

3.6. Bt-toxin interacts with lipophorin particles in the gut

Western blots of Bt-tolerant (immune-induced) and sus-ceptible (non-induced) larvae were analysed for possibledifferences in the amounts of prophenoloxidase (PPO)using antibodies against PPO from Manduca sexta. In pro-tein extracts from hemolymph plasma the antibodies cross-reacted slightly with two or three bands around 70 kDa,which could represent isoforms and/or proteolyticallycleaved phenoloxidase (Fig. 3). Bt-tolerant larvae previ-ously exposed to Bt-formulations (as indicated in Fig. 3)have higher levels of PPO. The amount of phenoloxidaselabelling is not significantly reduced in Bt-tolerant

Table 1Exposure of Bt-tolerant strain with two different doses of the Bt-formulation after treatments with tropolonea

Tropolone (ppm) Bt-toxin formulation(ppm)

No. of larvaeexposed

Percentage larvalsurvival after 7 days

Percentage adultemergence

Mean develop-mental time (days)

0 5000 51 82 47 28.51000 5000 50 76 44 27.59

0 10,000 50 54 34 27.471000 10,000 51 75 33 28.53

a The experiment was performed with 4th instar larvae of an E. kuehniella strain, which has been kept for seven generations on food with 4000 ppm ofthe Bt-formulation under laboratory conditions. Exposure to Bt-formulation above 5000 ppm kills all non-induced (susceptible) larvae.

Res R100 R500 R2000 R4000 R8000 Sus M-B t

250

98

64

50

Res R100 M-B t

250

98

64

50

Fig. 3. Prophenoloxidase-detection of Western blots containing hemo-lymph plasma proteins from susceptible and Bt-tolerant larvae kept onvarious concentrations (110, 500, 2000, 4000 and 8000 ppm) of Bt-formulation (Dipel). Antibodies against PPO from M. sexta werevisualised with alkaline phosphatase-conjugated secondary antibodies.The level of phenoloxidase labelling in Bt-tolerant larvae is maintained forseveral generations in the absence of Bt (Res–Bt).

Fig. 4. Detection of lipophorin in the gut lumen of immune-induced(Bt-tolerant) and non-induced (susceptible) E. kueniella (Rahman et al.,

2004a) using antibodies against the vWD domain of apolipophorin I (Maet al., 2006). Note the homogeneous distribution of staining in the non-induced gut (upper panel), in contrast to the gut of immune-inducedlarvae, which contained non-digested food and dot-like stained coagula-tion products (lower panel). These dot-like coagulation products appearsto form in the absence of melanin synthesis. This is in contrast to otherlepidopteran species, such as Helicoverpa armigera, where melanizationreactions are observed inside the gut lumen (Ma et al., 2005).

M. Mahbubur Rahman et al. / Journal of Invertebrate Pathology 96 (2007) 125–132 129

(4000 ppm) larvae that have been kept in the absence of Btfor two subsequent generations (Res–Bt).

Since the toxin may be inactivated in the gut lumen by acoagulation reaction, we also analysed gut proteins for pos-sible changes in immune-induced and non-induced larvae.Protein analysis by electrophoresis from the gut showed asignificant amount of lipophorin, which was altered some-what in extracts from Bt-tolerant larvae (not shown).To examine the localization of lipophorin in the gut westained lipophorin particles on gut whole mounts usingG. mellonella antibodies against a conserved proteindomain in apolipophorin I (Ma et al., 2006). This showedan increased lipophorin staining in the gut of Bt-tolerantlarvae compared to susceptible larvae (Fig. 4). The stainingwas visible as bright spots inside the gut lumen and aroundnon-digested food particles that were much more abundantin tolerant larvae than in susceptible larvae.

3.7. Separation by low-density gradient centrifugation

Previous studies indicated that endotoxin interacts withlipophorin particles from hemolymph plasma from Galleria

mellonella (Sarjan, 2002) (Schmidt et al., 2005). To examinea possible interaction of gut lipophorin with endotoxin, wemixed gut extracts with mature Cry1Ac and separated the

lipophorin particles by low-density gradient centrifugation.Under these conditions some of the toxin was found inlow-density fractions (Fig. 5), which contained the majorlipophorin proteins. Some of the labelling is detected inlipophorin fractions at the top of the gel having formedlarge protein aggregates (not shown). This is a possibleindication that Cry1Ac binds to gut-derived lipophorinparticles.

Fig. 5. Low-density gradient centrifugation of gut extracts fromimmune-induced larvae mixed with trypsin-activated Cry1Ac. Aliquotsfrom the gradient were analysed on Western blots using antibodiesagainst Cry1Ac, van Willebrand Factor D domain of apolipophorin Ifrom G. mellonella (Ma et al., 2005). Trypsin-digested protoxincomprised a mixture of protein bands with a 69 kDa band as a majorprotein in addition to smaller bands at 60 kDa, and high molecularweight proteins, which could be toxin oligomers (Ma et al., 2005). Theidentity of plasma proteins was determined on separate Western blotsusing antisera against arylphorin, lipophorin, PPO (not shown) andCry1Ac. A Coomassie blue staining of hemolymph plasma fractions(below) separated in the same centrifugation run showing arylphorinfound only in high-density fractions (7–13), whereas apolipophorin IIfound only in low-density fractions (13–23) with a peak in fraction 17.Cry1Ac bands at 60 and 69 kDa are found across the low-density area ofthe gradient with a minor peak at fraction 17.

130 M. Mahbubur Rahman et al. / Journal of Invertebrate Pathology 96 (2007) 125–132

4. Discussion

Questions whether melanisation is causing coagulationreactions or constitute a visible but non-essential part ofpro-coagulant aggregations are at the core of invertebrateimmunity. The induction of the immune system is depen-dent on elicitors and other stress-related factors (Barnesand Siva-Jothy, 2000) and increased melanization inhemolymph and cuticle, determined mainly by phenoloxi-dase (PO) activity, has been implicated in increased resis-tance to pathogens (Reeson et al., 1998). Whilecoagulation and aggregation reactions are independentof prophenoloxidase (PPO) (Rizki et al., 1985) the twoare linked (Nagai and Kawabata, 2000). While theabsence of melanization caused a reduction in thedefence capacity of some insects (Charalambidis et al.,1994b) but not in others (Leclerc et al., 2006), the twoare regulated at the metabolic level by independentpathways (Charalambidis et al., 1994a), We have recentlydiscovered that PPO and immune-related proteins areassociated with lipophorin particles (Rahman et al.,2006). Thus PPO-containing lipid particles can beinduced to perform melanin synthesis or aggregate intocoagulation products.

The observation that immune suppression by theparasitoid Venturia does not affect susceptibility of hostEphestia larvae to Bt-formulations may have two reasonsthat are not mutually exclusive: Firstly, manipulation ofhost physiology by Venturia may only be partial and notaffect all aspects of defence reactions (Schmidt et al.,2001). Second, the suppression of host defence reactionsby parasitoids may be restricted to the hemocoel, where

the parasitoid develops and not affect defence reactions inthe gut lumen. The latter implies that defence componentslocated in the hemocoel and gut lumen are regulated inde-pendently with implications for the regulation of PO,which in lepidopteran insects is derived exclusively fromhemocytes and transported to other tissues of the body,where it is involved in processes, such as cuticle hardening(Sugumaran, 1996) and defence (Brey et al., 1993). Thisstudy indicates that while melanization is a good indicatorfor immune induction and suppression, phenoloxidase isnot directly involved in the response to the toxin, as hasbeen shown in other systems (Leclerc et al., 2006). Thisraises important questions: If melanization is not relevantto the protection against the toxin, why is prophenoloxi-dase activated in Bt-tolerant larvae? More importantly,what are the alternative immune pathways that protectagainst the toxin?

One possible answer to the first question is that themechanism of up- or down-regulation of immune reactionsmay include melanization in conjunction with other reac-tions, such as coagulation, even though these can be sepa-rated at the genetic (Rizki et al., 1985) or metabolic level(Charalambidis et al., 1994a).

If melanization is not relevant for Bt-tolerance, whichdefence pathways protect the larva against the toxin?Immune suppression by parasitoids has been used toexamine baculovirus virulence (Washburn et al., 1996).Co-infection of M. sexta larvae with polydnavirus fromCotesia congregata increases susceptibility to fatal infec-tion by Autographa californica M Nucleopolyhedrovirus(Washburn et al., 2000). Conversely, preliminary experi-ments in H. armigera indicate that the elevated immunestatus, which provides protection against Bt-toxin (Maet al., 2005), also protect against baculovirus infection(unpubl.). While baculoviruses and Bt-toxin share simi-larities, there are differences, both in the path of infec-tion and in host defence. Firstly, while a protective lineof defence inside the gut lumen is not ruled out, mostof the virus-related defence reactions appear to takeplace after the virions are released from infected gut cells(Federici, 1993). According to one model, virions mayspread through the larval body in extracellular matrixspace provided by trachea and other tissues attached tothe gut (Kirkpatrick et al., 1994). In immune-competentlarvae virions are trapped by the immune system andprevented from spreading (Engelhard et al., 1994; Wash-burn et al., 1999, 2000). Since there is no evidence thatvirions are phagocytized by tracheal cells and since insecthemocytes are unable to cross the basement membranelining of the hemolymph, the only inducible defencepathway that can prevent the spreading of virions in thistissue environment is the inactivation by aggregation andcoagulation of defence molecules, where the virion-con-taining coagulation products inside the trachea attracthemocytes to form nodules (Washburn et al., 2000).The relevance to the protection against Bt-toxin is thefact that recognition of the pathogen and inactivation

M. Mahbubur Rahman et al. / Journal of Invertebrate Pathology 96 (2007) 125–132 131

(at least at an initial stage) occurs outside the hemocoellining. This implies that the factors involved in recogni-tion and inactivation of pathogens and toxins are ableto operate away from hemocytes, if not in theirabsence.

What are the candidates for a cell-free recognition andinactivation system? While a number of possible pro-coagulants have been identified in arthropods, the possiblefunction and mode of action seems to differ among differ-ent species (Duvic and Brehelin, 1998; Hall et al., 1999;Iwanaga et al., 1998; Korayem et al., 2004; Li et al.,2002; Muta and Iwanaga, 1996; Theopold et al., 2002,2004). The main feature of pro-coagulants is their uniqueability to recognize and form specific aggregates aroundpathogens and toxins, which effectively separate the dam-aging effects from the surrounding physiological environ-ment of the host. Many of the pro-coagulants, such aslipophorin (Li et al., 2002), vitellogenins (Hall et al.,1999) and hexamerin (Ma et al., 2005), serve metabolicand storage functions and in performing those functionsare transferred across the hemocoel lining to accumulatein other tissues and body cavities, such as oocytes, epider-mis and the gut lumen. Given their ubiquitous presenceand their potential function as pro-coagulants, these struc-tures could form a first line of defence against intrudingpathogens. For example, lipophorin particles, which aresecreted into the gut lumen for lipid transport betweenthe gut nutrients and the gut cell lining may attractBt-toxin molecules, which may exist in oligomeric formoutside the membrane (Bravo et al., 2004), causing seques-tration of the mature toxin before it can reach the brushborder membrane.

The observation that mature Cry1Ac is found with lipo-phorin after low-density gradient centrifugation, suggestthat Cry1Ac binds to lipophorin particles and can be trans-ported into low-density areas of the gradient, where someof the proteins aggregate into coagulation products thatare not resolved by SDS–PAGE.

In summary, the observations that lipophorin particlesexist in the gut lumen that can act as a pro-coagulant arecompatible with the existence of an immune-inducibleBt-tolerance mechanism based on the sequestration of thetoxin by a coagulation reaction inside the gut lumen. Theemergence of insect pest populations that become tolerantto Bt-toxin by transient induction mechanisms that can betransmitted to subsequent generations by a maternal effect(Ma et al., 2005; Rahman et al., 2004a) has possible impli-cations for resistance management strategies in the field.While the level of tolerance is well below the toxin levelsexpressed in most transgenic plants, inducible tolerancemechanisms may nevertheless become relevant in situationswhere insects are exposed to low levels of the toxin orwhere semi-dominant heterozygotes of genetic mutantinsects emerge with increased levels of resistance. If thetwo mechanisms are additive, the combined levels of resis-tance may become a threat to transgenic resistancestrategies.

Acknowledgments

This work was supported by a BioInnovation SA grantto OS. We acknowledge the receipt of Ph.D scholarshipsfor MMR and HSLR from the University of Adelaide.

References

Asgari, S., Zhang, G., Zareie, R., Schmidt, O., 2003. A serine proteinasehomolog venom protein from an endoparasitoid wasp inhibitsmelanization of the host hemolymph. Insect Biochem. Mol. Biol. 33,1017–1024.

Barnes, A.I., Siva-Jothy, M.T., 2000. Density-dependent prophylaxis inthe mealworm beetle Tenebrio molitor L-(Coleoptera: Tenebrionidae):cuticular melanization is an indicator of investment in immunity. Proc.Roy. Soc. (Lond.) B Biol. Sci. 267, 177–182.

Beck, M., Siekmann, G., Li, D.M., Theopold, U., Schmidt, O., 1999. Amaternal gene mutation correlates with an ovary phenotype in aparthenogenetic wasp population. Insect Biochem. Mol. Biol. 29,453–460.

Beck, M., Theopold, U., Schmidt, O., 2000. Evidence for serine proteaseinhibitor activity in the ovarian calyx fluid of the endoparasitoidVenturia canescens. J. Insect Physiol. 46, 1275–1283.

Bravo, A., Gomez, I., Miranda-CasoLuengo, R., Conde, J., Munoz-Garay, C., Sanchez, J., Zhuang, M., Gill, S.S., Soberon, M., 2004.Manduca sexta aminopeptidase N drives the oligomeric form ofBacillus thuringiensis Cry1Ab toxin and the cadherin receptor toplasma membrane microdomains. Biochim. Biophys. Acta 1667,38–46.

Brey, P.T., Lee, W.J., Yamakawa, M., Koizumi, Y., Perrot, S., Francois,M., Ashida, M., 1993. Role of the integument in insect immunity—epicuticular abrasion and induction of cecropin synthesis in cuticularepithelial cells. Proc. Natl. Acad. Sci. (USA) 90, 6275–6279.

Charalambidis, N.D., Bournazos, S.N., Lambropoulou, M., Marmaras,V.J., 1994a. Defense and melanization depend on the eumelaninpathway, occur independently and are controlled differentially indeveloping Ceratitis capitata. Insect Biochem. Mol. Biol. 24,655–662.

Charalambidis, N.D., Lambropoulou, M., Marmaras, V.J., 1994b.Mutation white pupae in the integument of Ceratitis capitata affectsboth defense and melanogenesis. Dev. Comp. Immunol. 18, 193–200.

Chase, M.R., Raina, K., Bruno, J., Sugumaran, M., 2000. Purification,characterization and molecular cloning of prophenoloxidases fromSarcophaga bullata. Insect Biochem. Mol. Biol. 30, 953–967.

Decker, H., Terwilliger, N., 2000. Cops and robbers: putative evolution ofcopper oxygen-binding proteins. J. Exp. Biol. 203, 1777–1782.

Duvic, B., Brehelin, M., 1998. Two major proteins from locust plasma areinvolved in coagulation and are specifically precipitated by laminarin,a beta-1,3-glucan. Insect Biochem. Mol. Biol. 28, 959–967.

Engelhard, E.K., Kammorgan, L., Washburn, J.O., Volkman, L.E., 1994.The insect tracheal system—a conduit for the systemic spread ofAutographa-californica—M nuclear polyhedrosis virus. Proc. Natl.Acad. Sci. (USA) 91, 3224–3227.

Federici, B.A., 1993. Viral pathobiology in relation to insect control. In:Beckage, N.E., Thompson, S.N., Frederici, B.A. (Eds.), Parasites andPathogens of Insects. Academic Press Inc., San Diego, pp. 81–101.

Fujimoto, K., Okino, N., Kawabata, S., Iwanaga, S., Ohnishi, E., 1995.Nucleotide sequence of the cDNA encoding the proenzyme of phenoloxidase a(1) of Drosophila melanogaster. Proc. Natl. Acad. Sci. (USA)92, 7769–7773.

Hall, M., Scott, T., Sugumaran, M., Soderhall, K., Law, J.H., 1995.Proenzyme of Manduca sexta phenol oxidase—purification, activation,substrate specificity of the active enzyme, and molecular cloning. Proc.Natl. Acad. Sci. (USA) 92, 7764–7768.

Hall, M., Wang, R., van Antwerpen, R., Sottrup-Jensen, L., Soderhall, K.,1999. The crayfish plasma clotting protein: a vitellogenin-related

132 M. Mahbubur Rahman et al. / Journal of Invertebrate Pathology 96 (2007) 125–132

protein responsible for clot formation in crustacean blood. Proc. Natl.Acad. Sci. (USA) 96, 1965–1970.

Iwanaga, S., Kawabata, S., Muta, T., 1998. New types of clotting factorsand defense molecules found in horseshoe crab hemolymph—theirstructures and functions. J. Biochem. 123, 1–15.

Jiang, H.B., Wang, Y., Ma, C.C., Kanost, M.R., 1997. Subunit compo-sition of pro-phenol oxidase from Manduca sexta: Molecular cloningof subunit ProPO-P1. Insect Biochem. Mol. Biol. 27, 835–850.

Kawabata, T., Yasuhara, Y., Ochiai, M., Matsuura, S., Ashida, M., 1995.Molecular cloning of insect pro-phenol oxidase—a copper-containingprotein homologous to arthropod hemocyanin. Proc. Natl. Acad. Sci.(USA) 92, 7774–7778.

Kirkpatrick, B.A., Washburn, J.O., Engelhard, E.K., Volkman, L.E.,1994. Primary infection of insect tracheae by Autographa californica Mnuclear polyhedrosis virus. Virology 203, 184–186.

Korayem, A.M., Fabbri, M., Takahashi, K., Scherfer, C., Lindgren, M.,Schmidt, O., Ueda, R., Dushay, M.S., Theopold, U., 2004. ADrosophila salivary gland mucin is also expressed in immune tissues:evidence for a function in coagulation and the entrapment of bacteria.Insect Biochem. Mol. Biol. 34, 1297–1304.

Lavine, M.D., Beckage, N.E., 1995. Polydnaviruses—potent mediators ofhost insect immune dysfunction. Parasitol. Today 11, 368–378.

Leclerc, V., Pelte, N., Chamy, L.E., Martinelli, C., Ligoxygakis, P.,Hoffmann, J.A., Reichhart, J., 2006. Prophenoloxidase activation isnot required for survival to microbial infections in Drosophila. EMBOReports 7, 231–235.

Li, D., Scherfer, C., Korayem, A.M., Zhao, Z., Schmidt, O., Theopold,U., 2002. Insect hemolymph clotting: evidence for interaction betweenthe coagulation system and the prophenoloxidase activating cascade.Insect Biochem. Mol. Biol. 32, 919–928.

Ma, G., Hay, D., Li, D., Asgari, S., Schmidt, O., 2006. Recognition andinactivation of LPS by lipophorin particles. Dev. Comp. Immunol. 30,619–626.

Ma, G., Roberts, H., Sarjan, M., Featherstone, N., Lahnstein, J., Akhurst,R., Schmidt, O., 2005. Is the mature endotoxin Cry1Ac from Bacillus

thuringiensis inactivated by a coagulation reaction in the gut lumen ofresistant Helicoverpa armigera larvae? Insect Biochem. Mol. Biol. 35,729–739.

Morita, Y., Matsumura, E., Okabe, T., Shibata, M., Sugiura, M., Ohe, T.,Tsujibo, H., Ishida, N., Inamori, Y., 2003. Biological activity oftropolone. Biol. Pharm. Bull. 26, 1487–1490.

Muta, T., Iwanaga, S., 1996. The role of hemolymph coagulation in innateimmunity. Curr. Opinion Immunol. 8, 41–47.

Nagai, T., Kawabata, S., 2000. A link between blood coagulation andprophenol oxidase activation in arthropod host defense. J. Biol. Chem.275, 29264–29267.

Nappi, A.J., Frey, F., Carton, Y., 2005. Drosophila serpin 27A is a likelytarget for immune suppression of the blood cell-mediated melanoticencapsulation response. J. Insect Physiol. 51, 197–205.

Nomiya, K., Yoshizawa, A., Kasuga, N.C., Yokoyama, H., Hirakawa, S.,2004. Synthesis, solid-state characterization and antimicrobialactivities of three different polymorphs of a copper(II) complex with

4-isopropyltropolone (hinokitiol). Inorganica Chimica Acta 357,1168–1176.

Rahman, M.M., Ma, G., Roberts, H.L.S., Schmidt, O., 2006. Cell-freeimmune reactions in insects. J. Insect Physiol. 52, 754–762.

Rahman, M.M., Roberts, H.L.S., Sarjan, M., Asgari, S., Schmidt, O.,2004a. Induction and transmission of Bacillus thuringiensis tolerance inthe flour moth Ephestia kuehniella. Proc. Natl. Acad. Sci. (USA) 101,2696–2699.

Rahman, M.M., Roberts, H.L.S., Schmidt, O., 2004b. The developmentof the endoparasitoid Venturia canescens in Bt-tolerant, immuneinduced larvae of the flour moth Ephestia kuehniella. J. Invertebr.Pathol. 87, 129–131.

Reeson, A.F., Wilson, K., Gunn, A., Hails, R.S., Goulson, D., 1998.Baculovirus resistance in the Noctuid Spodoptera exempta is pheno-typically plastic and responds to population density. Proc. Roy. Soc.Lond. B Biol. Sci. 265, 1787–1791.

Rizki, T.M., Rizki, R.M., Bellotti, R.A., 1985. Genetics of a Drosophila

phenoloxidase. Mol. Genet. Genomics 201, 7–13.Rogers, D., 1972. The ichneumonid wasp Venturia canescens: oviposition

and avoidance of superparasititism. Ent. Exp. Appl. 15, 190–194.Sarjan, M., 2002. Resistance against Endotoxin from Bacillus thuringiensis

in Lipidopteran Insects. University of Adelaide, Adelaide.Schmidt, O., Theopold, U., Strand, M., 2001. Innate immunity and its

evasion and suppression by hymenopteran endoparasitoids. Bioessays23, 344–351.

Schmidt, O., Rahman, M.M., Ma, G., Theopold, U., Sun, Y., Sarjan, M.,Fabbri, M., Roberts, H.S.L., 2005. Mode of action of antimicrobialproteins, pore-forming toxins and biologically active peptides(Hypothesis). Invert. Surviv. J. 2, 82–90.

Shelby, K.S., Adeyeye, O.A., Okot-Kotber, B.M., Webb, B.A., 2000.Parasitism-linked block of host plasma melanization. J. Invert. Pathol.75, 218–225.

Sugumaran, M., 1996. Role of insect cuticle in insect immunity. In:Soderhall, K., Iwanaga, S., Vasta, G. (Eds.), New Directions inInvertebrate Immunology. SOS Publication, New Haven, pp. 355–374.

Theopold, U., Li, D., Fabbri, M., Scherfer, C., Schmidt, O., 2002. Thecoagulation of insect hemolymph. Cell. Mol. Life Sci. 59, 363–372.

Theopold, U., Schmidt, O., Soderhall, K., Dushay, M.S., 2004. Coagu-lation in arthropods: defence, wound closure and healing. TrendsImmunol. 25, 289–294.

Washburn, J.O., Haas-Stapleton, E.J., Tan, F.F., Beckage, N.E., Volk-man, L.E., 2000. Co-infection of Manduca sexta larvae with polydna-virus from Cotesia congregata increases susceptibility to fatal infectionby Autographa californica M Nucleopolyhedrovirus. J. Insect Physiol.46, 179–190.

Washburn, J.O., Kirkpatrick, B.A., Volkman, L.E., 1995. Comparativepathogenesis of Autographa californica M nuclear polyhedrosis virusin larvae of Trichoplusia ni and Heliothis virescens. Virology 209,561–568.

Washburn, J.O., Kirkpatrick, B.A., Volkman, L.E., 1996. Insect protec-tion against viruses. Nature 383, 767.