ARTICLE IN PRESS

Journal of Insect Physiology 52 (2006) 754–762

0022-1910/$ - se

doi:10.1016/j.jin

�CorrespondE-mail addr

www.elsevier.com/locate/jinsphys

Cell-free immune reactions in insects

M. Mahbubur Rahman, Gang Ma, Harry L.S. Roberts, Otto Schmidt�

Insect Molecular Biology, University of Adelaide, Glen Osmond, SA 5064, Australia

Received 3 April 2006; received in revised form 4 April 2006; accepted 4 April 2006

Abstract

Insects, like many other multicellular organisms, are able to recognise and inactivate potential pathogens and toxins in the absence of

cells. Here we show that the recognition and inactivation of lipopolysaccharides (LPS) and bacteria is mediated by lipophorin particles,

which are the lipid carrier in insects. In immune-induced insects sub-populations of lipophorin particles are associated with pattern

recognition proteins and regulatory proteins that activate prophenoloxidase. Moreover, interactions with lectins result in the assembly of

lipophorin particles into cage-like coagulation products, effectively protecting the surrounding tissues and cells from the potentially

damaging effects of pathogens and phenoloxidase products. The existence of cell-free defence reactions implies that immune signals exist

upstream of cell-bound receptors.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Lipopolysaccharide; Prophenoloxidase; Lipophorin; Coagulation; Pattern recognition proteins

1. Introduction

A fundamental question in innate immunity is howmicrobes and other potentially damaging objects arerecognised and inactivated in the absence of cells (Hall etal., 1999; Kanost et al., 2004; Karlsson et al., 2004).Invertebrates use proteolytic coagulation cascades andantimicrobial peptides (Boman and Hultmark, 1987) toinactivate potentially damaging organisms and toxins(Hoffmann et al., 1999). Previous observations involvingcell-free plasma proteins, such as hemolin interacting withbacteria, suggested a Lipid A-mediated binding of patternand other recognition proteins (Daffre and Faye, 1997),followed by a Ca-dependent coagulation reaction, which isdependent on LPS-sugar moieties (Schmidt et al., 1993;Sun et al., 1990). Although recent experiments indicatedthat lipophorin and phenoloxidase are involved in coagula-tion (Li et al., 2002) and that the addition of LPS topurified lipophorin particles from insect larvae causesaggregation and inactivation of the toxin (Ma et al., 2006),the question remains: How is the binding of LPS to patternrecognition proteins translated into the inactivation of

e front matter r 2006 Elsevier Ltd. All rights reserved.

sphys.2006.04.003

ing author. Tel.: +618 8303 7252; fax: +618 8303 7109.

ess: otto.schmidt@adelaide.edu.au (O. Schmidt).

pathogens and toxins? Indirect evidence suggests thatinvertebrates with an open circulatory system sequesterdamaging microorganisms and toxins by a combination ofcoagulation (Nagai and Kawabata, 2000) and melanisationreactions (Kanost et al., 2004), involving adhesive (Lee etal., 1998) and covalent cross-linking of plasma components(Jiang et al., 2003), melanin synthesis and reactive oxygenproduction (Nappi and Ottaviani, 2000). Since themechanisms underlying these processes are not known,we asked whether lipophorin particles are the regulatoryand effector components for cell-free immune reactions.

2. Material and methods

2.1. Low-density gradient centrifugation

50–100 fourth to fifth instar larvae from Ephestia

kuehniella were washed in ethanol and dried on filter paperbefore extracting hemolymph by cutting a for-leg andbleeding into a ice-cold anticoagulant solution (30mMtrisodium citrate, 26mM citric acid, 20mM EDTA, 15mMNaCl, pH 5.5) containing phenylthiourea (PTU). Hemo-cytes were removed by centrifugation at 5000 g for 3min.Cell-free plasma (ca 1.5ml) was added to 15ml of asolution of 44.3 g KBr in 100ml and overlaid with 0.9%

ARTICLE IN PRESSM. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762 755

NaCl to a final volume of 30ml. The tube was spun at24,000 rpm (SW42Ti rotor) for 17 h at 10 1C. The gradientwas eluted in ca 1.2ml 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 bygreen pigment-containing proteins in the plasma fractionand melanised lipophorin sub-fractions in the low-densityarea of the gradient. To reduce melanisation and lipidoxidation during centrifugation b-mercaptoethanol (b-ME,final concentration 2mM) and a reactive oxygen scavengernitro blue trizolium (NBT, final concentration 10 mg/ml)were added to protein extracts. Under these conditions thebrown-coloured fraction was replaced by pink-colouredoxidised NBT (not shown) an indication for the productionof reactive electrons in lipophorin-containing fractions.Although attempts were made to preclude oxidisationresidual pink-coloured NBT was still detectable in lipo-phorin-containing gradient fractions. To isolate lessreactive particles, we used hemolymph from younger (thirdinstar) larvae of a larger moth, Galleria mellonella.

2.2. LPS bioassay

We used LPS-induced melanin synthesis (Rahman et al.,2004) and lipophorin particle aggregation (Ma et al., 2006)as an indication for elicitor-mediated cell-free immunefunctions of plasma fractions obtained by low-densitygradient centrifugation. For melanin measurements plasmafractions were diluted with PBS to protein concentrationsbetween 2.4 and 2.8OD280 and after addition of DOPA(final concentration of 10mM) melanisation was measuredat OD490 in the presence of calcium (final concentration1mM) and LPS (final concentration of 10 mg/ml). Inlipophorin-free plasma fractions measurements were per-formed with the amount of proteins ranging from 2.4 up to4.3OD280. Comparative measurements were always per-formed on the same day using fraction material from thesame gradient centrifugation.

For the measurement of LPS-induced aggregationreactions aliquots of gradient fractions were mixed withLPS at various concentrations in the presence and absenceof calcium (final concentration 1mM) and incubated for1 h at RT. Proteins were dissolved in loading buffer, heatedfor 10min at 65 1C and separated by SDS–PAGE. Theamount of LPS-mediated covalent cross-linking in eachsample was estimated by the relative amount of Coomassieblue or antibody-stained apolipophorin I, using a pixelestimation program. To estimate LPS-induced aggregationand distinguish adhesive linkages from covalent cross-linking, aliquots were centrifuged (10,000 g for 10min)after incubation and supernatants (and pellets) dissolved inloading buffer and after heating for 10min at 65 1Canalysed separately by SDS-PAGE. While amounts ofproteins in supernatant and pellet added up in adhesiveaggregation reactions, less protein was detected in thepellets under conditions enhancing covalent cross-linking,

which presumably prevented complete dissolution ofaggregated proteins in the pellets.

2.3. Lectin bioassays

For the measurement of lectin-mediated aggregation,aliquots of gradient fractions were mixed with lectins atvarious concentrations in the presence and absence ofcalcium (final concentration 1mM) and incubated for 1 hat RT. Proteins were dissolved in loading buffer, heated for10min at 65 1C and separated by SDS-PAGE. To examinesugar-binding properties, incubations were performed inthe presence of various sugars to final concentrations of10mM.

3. Results

3.1. Lipophorin particles contain immune proteins

When lipoprotein particles from the flour mothE. kuehniella were isolated by low-density gradientcentrifugation a brown-coloured band was clearly visiblein the low-density area separated from a blue-pigmentedplasma protein complex (not shown), which could indicatemelanisation reactions in association with lipoproteinparticles. Examination of proteins from gradient-separatedparticles revealed two major apolipoproteins (apo I/II),representing several sub-populations of lipophorin parti-cles of different relative lipid content and protein associa-tions, such as pattern recognition proteins (Fig. 1a, bgbp),imaginal disc growth factors (Ma et al., 2006) andmorphogens (Panakova et al., 2005). In addition to apoI/II, other protein bands were identified in low-densityfractions of the gradient as two phenoloxidase proteinvariants (Fig. 1a, PPO). This indicates that lipophorinparticles are associated with immune proteins and possiblyrecognise microbes via pattern recognition proteins.We therefore incubated gradient fractions with (S15) and

without (S9) lipophorin particles with bacteria andmonitored aggregation. Under these conditions bacteriawere found in large aggregates in the presence oflipophorin particles, whereas plasma fractions from high-density areas of the gradient were not active (Fig. 2a). Thissuggests that lipophorin particles containing immune-related proteins are able to recognise elicitors andinactivate potential pathogens by forming large aggregates.

3.2. Lipophorin particles in immune-induced insects

To examine the distribution of pattern recognitionproteins and phenoloxidases in immune-induced larvae,we compared an E. kuehniella strain with an elevatedimmune status (Rahman et al., 2004) with the non-inducedstrain. In plasma from immune-induced larvae, the relativeamounts of b-glucan-binding proteins in lipophorinparticle fractions is higher and extends to very low-densityfractions compared to plasma from non-induced larvae

ARTICLE IN PRESS

Fig. 1. Lipophorin particles are the regulatory units of cell-free immunity (a) Sub-populations of lipophorin particles (low-density and very low-density)

are associated with proteins involved in pattern recognition and proteolytic activation of prophenoloxidase. Western blots of low-density gradient

fractions from immune-induced (R) and non-induced (S) larvae (Rahman et al., 2004) probed with antibodies against b-glucan binding proteins (bgbp)from Manduca sexta (Ma and Kanost, 2000), van Willebrand factor domain (vWD) from Galleria mellonella, which binds to apolipophorin I (Ma et al.,

2006), and prophenoloxidase from M. sexta (Jiang et al., 1997), which binds to two protein variants (Jiang et al., 1997). Note that the b-glucan binding

protein is a dimer, which is stable in SDS solutions (Ma et al., 2006). (b) Relative amounts of b-glucan binding protein (bgbp), phenoloxidase and

phenoloxidase-activating proteins (PAP2 and PAP3) from M. sexta (Jiang et al., 2003) in low-density (fraction 17) and very low-density (fraction 23)

gradient fractions containing lipophorin particles derived from immune-induced (R) and non-induced (S) larvae (Rahman et al., 2004). M, molecular

weight markers.

M. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762756

(Fig. 1a). In contrast to pattern recognition proteins, thedistribution of phenoloxidase did not show significantdifferences in gradient fractions from induced and non-induced insects (Fig. 1b). Likewise, phenoloxidase-activat-ing proteases (PAP2 and PAP3) did not show significantdifferences in immune-induced and non-induced larvae andwere not visible in lipid-rich (very low-density) lipophorinparticles (Fig. 1b). The presence of phenoloxidase andabsence of prophenoloxidase activating proteases in lipid-rich particles suggests that phenoloxidase in these particlesmay not become readily activated and therefore mayrepresent a sub-population of particles with other func-tions, such as the transport and storage of oxygen.

3.3. LPS-induced melanisation is associated with lipophorin

particles

Given that prophenoloxidase co-migrates with lipophor-in particles, but are also detected in lipophorin-free plasmafractions, we asked whether both fractions respond to LPSwith melanin synthesis reactions. When lipophorin frac-tions were mixed with LPS in the presence of DOPA andcalcium, a phenoloxidase-mediated melanisation reactionwas observed (Fig. 2b). No melanin was detectable in theabsence of calcium or in lipophorin-free plasma fractions(Fig. 2), despite the presence of pattern recognition andphenoloxidase proteins in these fractions (Fig. 1a). Thissuggests that LPS-mediated phenoxidase activation re-quires an association with lipophorin particles. Given thatlipid transfer proteins are similar among vertebrates andinvertebrates (Babin et al., 1999) and some LPS-bindingproteins are evolutionarily related to phospholipid transferproteins (Yamashita et al., 2001), this could imply that LPS

molecules are taken up into the lipid moiety of the lipidcarrier, followed by a Ca-dependent activation of pheno-loxidase. This is consistent with a small but significantdelay in melanisation reactions observed after addition ofLPS to lipophorin particles (Fig. 2b), which could be anindication that LPS-uptake into lipophorin particles is arate-limiting step.

3.4. Lectin-induced aggregation of lipophorin particles

While melanin synthesis usually occurs together withcoagulation the two processes may be regulated separately.We therefore examined the LPS-triggered aggregation andcovalent cross-linking of lipophorin particles, using thedisappearance of apolipoproteins from detergent-solubleextracts as an indicator for coagulation-related aggregationof lipophorin proteins (Fig. 3). Time course experimentsallowed us to differentiate between adhesive aggregationand covalent cross-linking reactions when we noticed thatcovalent cross-linking is only inducible within several daysafter gradient centrifugation and in the presence ofcalcium. In contrast, LPS-mediated adhesive aggregationcan be detected in the absence of calcium in aged gradientfractions by removing the aggregate with a centrifugationstep before adding a detergent to the supernatant. Underthese conditions, we observed that the LPS-inducedadhesive aggregation reaction was partly inhibited by theaddition of sugars (Fig. 3), which could indicate thatadhesive aggregation involved glycodeterminants andlectin-like pattern recognition proteins. Since laminarincauses lipophorin aggregation (Duvic and Brehelin, 1998)and given that LPS may be contaminated with peptidogly-can fragments, the adhesion may be caused by unknown

ARTICLE IN PRESS

Fig. 2. (a) Bacteria trigger aggregation of purified lipophorin particles.

E. coli (strain JM106) was grown overnight and aliquots incubated with

low-density gradient fractions from the high-density area (fraction 9) and

low-density area (fraction 15) containing similar amounts of protein. The

experiment was performed as described (Ma and Kanost, 2000). The

major protein bands in each of the two fractions are indicated as 230 kDa

apolipophorin I (apo I), 70 kDa apolipophorin II (apo II), 74 kDa

prophenoloxidase (PPO) and 72 kDa arylphorin (aryl). (b) LPS triggers

melanisation in purified lipophorin particles. Protein content measuring

2.4 OD (at 280 nm) from low-density gradient fraction S14 was mixed with

DOPA (final concentration 10mM) and CaCl (final conc. 1mM) and

optical density measured continuously (at 490 nm). When LPS (final conc.

10 mg/ml) was added after 40min (arrow), optical density increased with a

lag-phase of about 10min. In contrast, when DOPA and LPS was mixed

with lipophorin particles and CaCl2 was added after 40min (arrow),

optical density increased immediately. No increase in optical density was

observed in lipophorin-free fractions from the plasma (fraction S7) using

protein amounts up to 4.3OD (at 280 nm). Similar results were obtained

with corresponding fractions from immune-induced larvae (not shown).

0

20

40

60

80

100

120

1507537.518.89.44.72.31.10.5

LPS conc. (µg/ml)

Apo

lipop

horin

I (A

U)

LPS/Ca/Sugar

LPS/Ca

LPS

Fig. 3. LPS-triggered aggregation and covalent cross-linking of purified

lipoporin particles. LPS was added to aliquots of lipophorin fraction R15

in increasing amounts (0.5–150mg/ml) in the presence and absence of

CaCl2 (final conc. 1mM) and the relative amount of Coommassie blue-

stained apolipophorin I was measured. Similar results were obtained when

apolipophorin I was measured on Western blots after labelling with anti-

vWD antibodies (not shown). LPS-induced aggregation is partly inhibited

in the presence of mixtures of galactosamine, N-acetylgalactosamine and

N-acetylglucosamine (final conc. 10mM each). Note that sugar alone

appears to cause some aggregation of lipophorin particles for unknown

reasons.

M. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762 757

lectins or dimerisation of immulectins or b-glucan-bindingproteins as described for Drosophila peptidoglycan-bindingproteins (Mellroth et al., 2005). Alternatively, LPS-derivedsugar determinants may be cross-linked by the activationof an unknown lectin.

3.5. PNA-mediated self-assembly of lipophorin particles

Since Ca-dependent protease activation may triggeradhesive and covalent cross-linking, we added oligomericlectins to low-density purified particle fractions in the

presence and absence of calcium. When we used tetramericgalactose-specific peanut agglutinin (PNA), we detected aconcentration-dependent aggregation of lipophorin parti-cles (Fig. 4a), which could be dissolved when the reactionwas performed in the absence of calcium, but appears toinduce covalent cross-linking when performed in thepresence of calcium. When labelled PNA was used, thestaining was detected in the pellet (not shown), whichsuggests that the lectin became an integral part of theaggregate. This suggests that glycodeterminants from thelipophorin particles are cross-linked by putative particle-derived lectins, such as oligomerised pattern recognitionproteins (Mellroth et al., 2005), which can be emulated bythe addition of oligomeric lectins. To examine whetherPNA interaction with lipophorin particles is based onsugar-mediated binding, we incubated the mixture withvarious sugars to compete with the sugar determinants ofthe particle. Under these conditions, the PNA-mediatedaggregation was partially inhibited in the presence ofgalactosamine (Fig. 4b), while other sugars, such asGalNAc and GlcNAc had no effect. This suggests thatPNA oligomers interact with galactose-containing glyco-determinants on the lipophorin particles, causing aggrega-tion. This is consistent with previous observationssuggesting that glycodeterminants from bacterial glycoli-pids (Schmidt et al., 1993) and cell wall components areinvolved in the oligomerisation of pattern recognitionproteins (Mellroth et al., 2005). Given that PNA-mediatedaggregation occurred in the absence of LPS, it isconceivable that adhesive connections also include parti-cle-specific glycodeterminants. Since apolipophorin I/IIlack Gal-containing sugar modifications (Sundermeyer etal., 1996), the sugar may reside in associated glycoproteinsor glycolipids. Conceptually the activation of lipophorin

ARTICLE IN PRESS

0 0.5 9.31

PNA conc. (µg/ml)

PNA conc. (µg/ml)

Apo

lipop

horin

I (A

U)

100

80

60

40

20

0

Apo

lipop

horin

I (A

U)

100

80

60

40

20

0

PNA

PNA/Gal

PNA

PNA/Ca

15037.5 7518.72.3 4.6

100502512.56.23.1

(a)

(b)

Fig. 4. Aggregation of lipophorin particles by oligomeric lectins. (a) Gal-

specific lectin (peanut agglutinin, PNA) can induce aggregation and

covalent cross-linking of lipophorin particle proteins in the absence of

calcium. (b) PNA-induced aggregation is partly inhibited by galactosa-

mine (final concentration 50mM). Note that galactosamine alone appears

to cause some aggregation of lipophorin particles for unknown reasons.

Fig. 5. Lipophorin particle assemblies. Lipophorin particles form globular

structures. FITC-conjugated PNA was added to lipophorin particles from

G. mellonella (fraction 15) and inspected with a confocal microscope under

indirect UV-light. No labelled structures were obtained with GalNAc-

specific Helix pomatia lectin (FITC-conjugated HPL) or GlcNAc-specific

wheat germ agglutinin (FITC-conjugated WGA).

M. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762758

particles can be based on two scenarios: by elicitor-mediated enzymatic modification of particle glycodetermi-nants, which interact with pre-existing oligomeric lectins,or by elicitor-mediated oligomerisation of monomericlectins, which connect pre-existing sugar-specific particledeterminants.

3.6. Self-assembly into cage-like aggregates

This suggests that lectins, interact with lipophorinparticles to form a complex. It raises the question whetherlectins simply connect the lipophorin particles to thebacterial surface and LPS micelles, or whether theoligomeric lectin is involved in a LPS-triggered self-assembly of lipophorin particles. To answer this questionwe added FITC-conjugated PNA to purified lipophorinparticles from Galleria mellonella and inspected thecoagulation products under the confocal microscope.Under these conditions, we observed round or oval-shapedglobules less than 3 mm in size (Fig. 5). This suggests thatcoagulation products form regular-shaped cage-like assem-blies in the presence of oligomeric adhesion molecules withthe potential to sequester microbes and other objects by alipid-containing layer of lipophorin particles. This cell-freedefence reaction may form a first line of defence, such as inthe gut lumen, where inactivation of toxins can be effectiveeven in the absence of hemocytes (Rahman et al., 2004).

4. Discussion

Our observations suggest that lipid particles play animportant role in the regulatory process involved inactivating the LPS-mediated immune response. The pro-cess of cell-free immune activation by LPS involves at leasttwo steps (Fig. 6), a Ca-independent uptake of LPS into thelipid moiety of lipophorin particles, followed by a Ca-dependent ‘activation’ and sugar-dependent oligomerisa-tion of adhesive molecules that cause self-assembly oflipophorin particles into globular structures. These glo-bules effectively inactivate microbes and sequester objectswith a lipid-containing coating, which may precludebacterial growth and the spread of damaging toxins aswell as reactive oxygen intermediates.Whether the induction of melanisation and aggregation

of lipid particles (shown in Fig. 2) is caused by LPS orcontaminated peptidoglycan is not known from ourexperiments. We and others (Duvic and Brehelin, 1998)have observed that laminarin can induce the LPS-mediatedaggregation but not melanization, which indicates thatmore than one elicitor may be involved in some of thereactions.This raises the question, which type of adhesion

molecules may be responsible for the transition fromsoluble to adhesive particles? Given that adhesiveness oflipid particles may emerge as an essential feature of cell-freedefence reactions, the mode of activation of adhesiveproteins associated with lipid particles becomes a crucialstep in our understanding of innate immune recognition.

ARTICLE IN PRESS

Fig. 6. Schematic view of a two-step activation process, involving LPS uptake from lipid droplets into the lipid moiety of lipophorin particles

(Ca-independent) and subsequent oligomerisation of adhesion molecules (Ca-dependent) leading to the sugar-mediated cross-linking of particles into

globular structures. Note that lipophorin particles are depicted with LPS molecules being located on one of the two lipid bilayers only. While aggregation

is possible with apolar particles, globular structures are only expected in polarised particles for which there is no experimental evidence.

M. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762 759

Our analysis of lipid particles revealed several candidateproteins that can potentially contribute to the adhesivenessand each has a different mode of activation.

While it is not clear whether each adhesive protein isassociated with separate particles or whether more thanone adhesion molecule is associated with a single particle, anumber of potential proteins in lipophorin fractions havethe capacity to become adhesive in the presence of elicitors(LPS) or in response to environmental changes (lipidoxidisation etc). Firstly, the elicitor-binding proteinfamilies, immulectins and b-glucan-binding proteins con-tain two binding domains per protein (Yu et al., 2002).Similar to the Drosophila peptidoglycan recognitionprotein, these proteins become dimeric in the presence ofelicitors (Mellroth et al., 2005), which makes themadhesive. In fact, the observed immulectin dimers fromour gradient preparations are stable under denaturatingconditions and visible in Western blots (Ma et al., 2006).Secondly, we have identified an imaginal disc growth factorto be associated with lipid particles (Ma et al., 2006), whichis known to have lectin properties (Kawamura et al., 1999).Finally, it is known that exchangeable apolipoproteins,such as apolipoprotein III, have LPS-binding properties(Whitten et al., 2004), which cause the oligomeric proteinto attach to the surface of bacteria and induce phagocy-tosis. This suggests that lipid particles are associated withmore than one protein with the ability to become adhesivein response to different elicitors or under differentconditions.

Another question is how lipid particles become asso-ciated with immune proteins? Fig. 1 shows fraction 14,which has PAPs both in non-induced (S14) and in immune-induced (R14) larvae. Preliminary experiments suggest thatS21 does not show LPS-induced melanisation, whereas

R21 does (not shown). This is unexpected since particle-associated PPO requires PAPs for activation. One explana-tion is that PAPs are inactivated or dissociated fromparticles under high salt and in the presence of EDTAduring centrifugation. Alternatively this observation couldindicate that particles are assembled in a step-wise fashion,where PPO is recruited first but may not be inducible viaelicitors. Since PPO is associated with oxygen, particleslacking activating proteases may be involved in oxygencarrier functions (Decker and Terwilliger, 2000) transport-ing oxygen in the hemocoel and into eggs (Bai et al., 1996;Schmidt et al., 2005b). This is consistent with observationsdescribing lipophorin particles as multifunctional com-plexes that can be associated with growth factors (Ma etal., 2006) and morphogens (Panakova et al., 2005) toregulate the growth and shape of tissues.Our conclusions that phenoloxidase activity is linked to

lipid particles and regulated by particle-specific immuneproteins are consistent with very early observations thatcell-free defence reactions exist in insects (Boman andHultmark, 1987). Some species, such as mosquitoes withfew or no hemocytes in the hemolymph of adult insects arenevertheless capable of defending against intruding dama-ging organisms by humoral defence reactions (Goetz et al.,1987). Recent observations suggest that most of themalaria parasites encountered in the gut lumen ofmosquitoes are inactivated by cell-free defence reactionsincluding melanisation and other PO reactions (Kumar etal., 2003), which include reactive oxygen intermediates(Nappi and Ottaviani, 2000). Interestingly, specific proteinsinvolved in these reactions comprise C-type lectins, whichmay target parasite-specific but also insect-specific deter-minants (Osta et al., 2004), an observation that isconsistent with a general role of particle-associated

ARTICLE IN PRESS

Fig. 7. Schematic drawing depicting the removal from circulation of ‘adhesive’ lipophorin particles by cell-free sequestration and cellular clearance

reactions. Self-assembly into globular aggregates by adhesive lipid particles (a) may also occur on the cell surface if adhesion molecules interact with

receptors (b). Given the size differences of lectin and lipid particles, receptors interacting with oligomeric adhesion molecules are bend around the hinge-

like lipophorin particles, causing a curvature of the membrane (Schmidt and Theopold, 2004). Clustering of lipid particles on the cell surface may drive the

uptake of the toxin by a cellular clearance reaction based on dynamic adhesion processes on the cell surface (Schmidt and Schreiber, 2006).

M. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762760

adhesion molecules in lipid metabolism and cell home-ostasis.

In addition there is a large volume of literaturesuggesting an involvement of lipid particles in variousdefence and detoxification reactions. For example, lipo-phorin has toxin-binding properties (Vilcinskas et al., 1997)and binds and detoxifies LPS (Kato et al., 1994; Pratt andWeers, 2004). Lipophorin is regulated by immune induc-tion (Mullen et al., 2004). Many of the effects can be tracedto apoIII, which has opsonic properties, enhancingphagocytosis of bacteria (Whitten et al., 2004). Interest-ingly apoIII is also involved in immune activation (Kim etal., 2004; Wiesner et al., 1997) and humoral defence (Parket al., 2005) and induces the immune response in vivo(Niere et al., 2001).

The dual role of apoIII as a lectin-like protein in cellularreactions is quite interesting, since apoIII is able to detachspread hemocytes from the substrate when applied insoluble form but enhance adhesion and phagocytosis whenimmobilised on the surface of microbes (Whitten et al.,2004). Similar reports from other immune proteins, such ashemolin (Ladendorff and Kanost, 1991) and hemomucin-binding lectins (Schmidt and Schreiber, 2006) couldindicate that lipophorin is able to interfere with cellularadhesion by an unknown process. Whether this includes arole of apolipoproteins in preventing adhesion of hemo-cytes (Coodin and Caveney, 1992; Mandato et al., 1996)remains to be seen.

Lipophorin interacts with immune suppressors toinactivate hemocytes (Asgari and Schmidt, 2002). Im-mune-suppressed hemocytes from larvae parasitised byhymenopteran parasitoids have depolymerised actin-cytos-

keleton and round up due to lack of adhesive surfaceproteins (Schmidt et al., 2005a). A protein with thecapacity to inactivate hemocytes has been isolated (Asgariet al., 1997) and shown to form a complex with lipophorinbefore being taken up by hemocytes (Asgari and Schmidt,2002). Importantly, the immune suppressor alone does notbind to surface receptors, whereas the complex is clearedby massive cellular uptake reactions, which removeadhesive receptors from the hemocyte surface. Interest-ingly, there are no signs of coagulation in the hemocoeleven with the injection of large amounts of suppressor.This raises the interesting possibility that some interactionswith lipophorin particles generate adhesive particles thatengage in aggregation, whereas others lead to interactionswith the cell surface (Fig. 7). The implication of this modelis that the formation of a modified (adhesive) lipophorinparticle determines the interaction with other particles(coagulation or not) and with cells (lipid exchange oruptake). Self-assembly of particles into a cell-free cage-likesequestration product may resemble the assembly ofadhesive particles with membrane-bound receptors, whichinternalise the object by leverage-mediated uptake reac-tions (Schmidt and Theopold, 2004) creating a dynamicbalance with adhesion reactions (Schmidt and Schreiber,2006).

Acknowledgements

The authors thank Mike Kanost for antisera againstpattern recognition proteins and phenoloxidase activatingproteases (PAPs) and Jana Bradley, Caspar Jonker and

ARTICLE IN PRESSM. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762 761

Natasha McInnnes for help with the experiments. Thiswork was supported by a grant from BioInnovation SA.

References

Asgari, S., Schmidt, O., 2002. A coiled-coil region of an insect immune

suppressor protein is involved in binding and uptake by hemocytes.

Insect Biochemistry and Molecular Biology 32, 497–504.

Asgari, S., Schmidt, O., Theopold, U., 1997. A polydnavirus-encoded

protein of an endoparasitoid wasp is an immune suppressor. Journal of

General Virology 78, 3061–3070.

Babin, P.J., Bogerd, J., Kooiman, F.P., Van Marrewijk, W.J.A., Van der

Horst, D.J., 1999. Apolipophorin II/I, apolipoprotein B, vitellogenin,

and microsomal triglyceride transfer protein genes are derived from a

common ancestor. Journal of Molecular Evolution 49, 150–160.

Bai, G., Li, J., Christensen, B.M., Yoshino, T.P., 1996. Phenoloxidase

activity in the reproductive system and egg masses of the pulmonate

gastropod, Biomphalaria glabrata. Comparative Biochemistry and

Physiology Part B: Biochemistry and Molecular Biology 114, 353–359.

Boman, H.G., Hultmark, D., 1987. Cell-free immunity in insects. Annual

Review of Microbiology 41, 103–126.

Coodin, S., Caveney, S., 1992. Lipophorin inhibits the adhesion of

cockroach (Periplaneta americana) haemocytes in vitro. Journal of

Insect Physiology 38, 853–862.

Daffre, S., Faye, I., 1997. Lipopolysaccharide interaction with hemolin, an

insect member of the Ig-superfamily. FEBS Letters 408, 127–130.

Decker, H., Terwilliger, N., 2000. Cops and robbers: putative evolution of

copper oxygen-binding proteins. Journal of Experimental Biology 203,

1777–1782.

Duvic, B., Brehelin, M., 1998. Two major proteins from locust plasma are

involved in coagulation and are specifically precipitated by laminarin,

a beta-1,3-glucan. Insect Biochemistry and Molecular Biology 28,

959–967.

Goetz, P., Enderlein, G., Roettgen, I., 1987. Immune reactions of

Chironomus larvae (insecta; Diptera) against bacteria. Journal of

Insect Physiology 33, 993–1004.

Hall, M., Wang, R., van Antwerpen, R., Sottrup-Jensen, L., Soderhall, K.,

1999. The crayfish plasma clotting protein: a vitellogenin-related

protein responsible for clot formation in crustacean blood. Proceed-

ings of the National Academy of Sciences (USA) 96, 1965–1970.

Hoffmann, J.A., Kafatos, F.C., Janeway, C.A., Ezekowitz, R.A.B., 1999.

Phylogenetic perspectives in innate immunity. Science 284, 1313–1318.

Jiang, H.B., Wang, Y., Ma, C.C., Kanost, M.R., 1997. Subunit

composition of pro-phenol oxidase from Manduca sexta: molecular

cloning of subunit ProPO-P1. Insect Biochemistry and Molecular

Biology 27, 835–850.

Jiang, H., Wang, Y., Yu, X.-Q., Zhu, Y., Kanost, M., 2003. Propheno-

loxidase-activating proteinase-3 (PAP-3) from Manduca sexta hemo-

lymph: a clip-domain serine proteinase regulated by serpin-1J and

serine proteinase homologs. Insect Biochemistry and Molecular

Biology 33, 1049–1060.

Kanost, M., Jiang, H., Yu, X., 2004. Innate immune responses of a

lepidopteran insect, Manduca sexta. Immunology Review 198, 97–105.

Karlsson, C., Korayem, A.M., Scherfer, C., Loseva, O., Dushay, M.S.,

Theopold, U., 2004. Proteomic analysis of the Drosophila larval

hemolymph clot. Journal of Biological Chemistry 279, 52033–52041.

Kato, Y., Motoi, Y., Taniai, K., Kadonookuda, K., Yamamoto, M.,

Higashino, Y., Shimabukuro, M., Chowdhury, S., Xu, J.H., Sugiya-

ma, M., et al., 1994. Lipopolysaccharide–lipophorin complex forma-

tion in insect hemolymph—a common pathway of lipopolysaccharide

detoxification both in insects and in mammals. Insect Biochemistry

and Molecular Biology 24, 547–555.

Kawamura, K., Shibata, T., Saget, O., Peel, D., Bryant, P.J., 1999. A new

family of growth factors produced by the fat body and active on

Drosophila imaginal disc cells. Development 126, 211–219.

Kim, H.J., Je, H.J., Park, S.Y., Lee, I.H., Jin, B.R., Yun, H.K., Yun,

C.Y., Han, Y.S., Kang, Y.J., Seo, S.J., 2004. Immune activation of

apolipophorin-III and its distribution in hemocyte from Hyphantria

cunea. Insect Biochemistry and Molecular Biology 34, 1011–1023.

Kumar, S., Christophides, G.K., Cantera, R., Charles, B., Han, Y.S.,

Meister, S., Dimopoulos, G., Kafatos, F.C., Barillas-Mury, C., 2003.

The role of reactive oxygen species on Plasmodium melanotic

encapsulation in Anopheles gambiae. Proceedings of the National

Academy of Sciences (USA) 100, 14139–14144.

Ladendorff, N.E., Kanost, M.R., 1991. Bacteria-induced protein P4

(hemolin) from Manduca sexta: a member of the immunoglobulin

superfamily, which can inhibit hemocyte aggregation. Archieves of

Insect Biochemistry and Physiology 18, 285–300.

Lee, S.Y., Cho, M.Y., Hyun, J.H., Lee, K.M., Homma, K.I., Natori, S.,

Kawabata, S.I., Iwanaga, S., Lee, B.L., 1998. Molecular cloning of

cDNA for pro-phenol-oxidase-activating factor I, a serine protease is

induced by lipopolysaccharide or 1,3-beta-glucan in coleopteran insect,

Holotrichia diomphalia larvae. European Journal of Biochemistry 257,

615–621.

Li, D., Scherfer, C., Korayem, A.M., Zhao, Z., Schmidt, O., Theopold,

U., 2002. Insect hemolymph clotting: evidence for interaction between

the coagulation system and the prophenoloxidase activating cascade.

Insect Biochemistry and Molecular Biology 32, 919–928.

Ma, C.C., Kanost, M.R., 2000. A beta 1,3-glucan recognition protein

from an insect, Manduca sexta, agglutinates microorganisms and

activates the phenoloxidase cascade. Journal of Biological Chemistry

275, 7505–7514.

Ma, G., Hay, D., Li, D., Asgari, S., Schmidt, O., 2006. Recognition and

inactivation of LPS by lipophorin particles. Developmental and

Comparative Immunology 30, 619–626.

Mandato, C.A., Diehljones, W.L., Downer, R., 1996. Insect hemocyte

adhesion in vitro—inhibition by apoliphorin I and an artificial

substrate. Journal of Insect Physiology 42, 143–148.

Mellroth, P., Karlsson, J., Hakansson, J., Schultz, N., Goldman, W.E.,

Steiner, H., 2005. Ligand-induced dimerization of Drosophila pepti-

doglycan recognition proteins in vitro. Proceedings of the National

Academy of Sciences (USA) 102, 6455–6460.

Mullen, L.M., Lightfoot, M.E., Goldsworthy, G.J., 2004. Induced

hyperlipaemia and immune challenge in locusts. Journal of Insect

Physiology 50, 409–417.

Nagai, T., Kawabata, S., 2000. A link between blood coagulation and

prophenol oxidase activation in arthropod host defense. Journal of

Biological Chemistry 275, 29264–29267.

Nappi, A.J., Ottaviani, E., 2000. Cytotoxicity and cytotoxic molecules in

invertebrates. Bioessays 22, 469–480.

Niere, M., Dettloff, M., Maier, T., Ziegler, M., Wiesner, A., 2001. Insect

immune activation by apolipophorin III is correlated with the lipid-

binding properties of this protein. Biochemistry 40, 11502–11508.

Osta, M.A., Christophides, G.K., Kafatos, F.C., 2004. Effects of

mosquito genes on Plasmodium development. Science 303, 2030–2032.

Panakova, D., Sprong, H., Marois, E., Thiele, C., Eaton, S., 2005.

Lipoprotein particles are required for Hedgehog and Wingless

signalling. Nature 435, 58–65.

Park, S.Y., Kim, C.H., Jeong, W.H., Lee, J.H., Seo, S.J., Han, Y.S., Lee,

I.H., 2005. Effects of two hemolymph proteins on humoral defense

reactions in the wax moth, Galleria mellonella. Developmental and

Comparative Immunology 29, 43–51.

Pratt, C.C., Weers, P.M.M., 2004. Lipopolysaccharide binding of an

exchangeable apolipoprotein, apolipophorin III, from Galleria mello-

nella. Biological Chemistry, Hoppe-Seyler 385, 1113–1119.

Rahman, M.M., Roberts, H.L.S., Sarjan, M., Asgari, S., Schmidt, O.,

2004. Induction and transmission of Bacillus thuringiensis tolerance in

the flour moth Ephestia kuehniella. Proceedings of the National

Academy of Sciences (USA) 101, 2696–2699.

Schmidt, O., Faye, I., Lindstrom Dinnetz, I., Sun, S.C., 1993. Specific

immune recognition of insect hemolin. Developmental and Compara-

tive Immunology 17, 195–200.

Schmidt, O., Glatz, R.V., Asgari, S., Roberts, H.S.L., 2005a. Are insect

immune suppressors driving cellular uptake reactions? Archives of

Insect Biochemistry and Physiology 60, 153–158.

ARTICLE IN PRESSM. Mahbubur Rahman et al. / Journal of Insect Physiology 52 (2006) 754–762762

Schmidt, O., Li, D., Beck, M., Kinuthia, W., Bellati, J., Roberts, H.L.S.,

2005b. Phenoloxidase-like activities and the function of virus-like

particles in ovaries of the parthenogenetic parasitoid Venturia

canescens. Journal of Insect Physiology 51, 117–125.

Schmidt, O., Schreiber, A., 2006. Integration of cell adhesion reactions—a

balance of forces? Journal of Theoretical Biology 238, 608–615.

Schmidt, O., Theopold, U., 2004. An extracellular driving force of

endocytosis and cell-shape changes (Hypothesis). BioEssays 26,

1344–1350.

Sun, S.C., Lindstrom, I., Boman, H.G., Faye, I., Schmidt, O., 1990.

Hemolin: an insect-immune protein belonging to the immunoglobulin

superfamily. Science 250, 1729–1732.

Sundermeyer, K., Hendricks, J.K., Prasad, S.V., Wells, M.A., 1996. The

precursor protein of the structural apolipoproteins of lipophorin:

cDNA and deduced amino acid sequence. Insect Biochemistry and

Molecular Biology 26, 735–738.

Vilcinskas, A., Kopacek, P., Jegorov, A., Vey, A., Matha, V., 1997.

Detection of lipophorin as the major cyclosporin-binding protein in

the hemolymph of the greater wax moth Galleria mellonella.

Comparative Biochemistry & Physiology—C: Comparative Pharma-

cology & Toxicology 117, 41–45.

Whitten, M.M.A., Tew, I.F., Lee, B.L., Ratcliffe, N.A., 2004. A novel role

for an insect apolipoprotein (apolipophorin III) in {beta}-1,3-glucan

pattern recognition and cellular encapsulation reactions. Journal of

Immunology 172, 2177–2185.

Wiesner, A., Losen, S., Kopacek, P., Weise, C., Gotz, P., 1997. Isolated

apolipophorin III from Galleria mellonella stimulates the immune

reactions of this insect. Journal of Insect Physiology 43, 383–391.

Yamashita, S., Sakai, N., Hirano, K., Ishigami, M., Maruyama, T.,

Nakajima, N., Matsuzawa, Y., 2001. Roles of plasma lipid transfer

proteins in reverse cholesterol transport. Frontiers in Bioscience 6,

D366–D387.

Yu, X.-Q., Zhu, Y.-F., Ma, C., Fabrick, J.A., Kanost, M.R., 2002. Pattern

recognition proteins in Manduca sexta plasma. Insect Biochemistry

and Molecular Biology 32, 1287–1293.