Histopathological Effects of the Yen-Tc Toxin Complex from Yersiniaentomophaga MH96 (Enterobacteriaceae) on the Costelytra zealandica(Coleoptera: Scarabaeidae) Larval Midgut

Sean D. G. Marshall,a Michelle C. Hares,b Sandra A. Jones,a Lincoln A. Harper,a James R. Vernon,a Duane P. Harland,a

Trevor A. Jackson,a and Mark R. H. Hursta

AgResearch, Innovative Farm Systems, Lincoln Research Centre, Christchurch, New Zealand,a and Centre of Ecology and Conservation, University of Exeter, CornwallCampus, Penryn, Cornwall, United Kingdomb

Yersinia entomophaga MH96, which was originally isolated from the New Zealand grass grub, Costelytra zealandica, pro-duces an orally active proteinaceous toxin complex (Yen-Tc), and this toxin is responsible for mortality in a range of insectspecies, mainly within the Coleoptera and Lepidoptera. The genes encoding Yen-Tc are members of the toxin complex (Tc)family, with orthologs identified in several other bacterial species. As the mechanism of Yen-Tc activity remains unknown,a histopathological examination of C. zealandica larvae was undertaken in conjunction with cultured cells to identify theeffects of Yen-Tc and to distinguish the contributions that its individual subunit components make upon intoxication. Aprogressive series of events that led to the deterioration of the midgut epithelium was observed. Additionally, experimentsusing a cell culture assay system were carried out to determine the cellular effects of intoxication on cells after topical ap-plication and the transient expression of Yen-Tc and its individual components. While observations were broadly consis-tent with those previously reported for other Tc family members, some differences were noted. In particular, the distinctstepwise disintegration of the midgut shared features associated with both apoptosis and necrotic programmed cell deathpathways. Second, we observed, for the first time, a contribution of toxicity from two chitinases associated with the Yen-Tccomplex. Our findings were suggestive of the activities encoded within the subunit components of Yen-Tc targeting differ-ent sites along putative programmed cell death pathways. Given the observed broad host range for Yen-Tc, these targetedloci are likely to be widely shared among insects.

The toxin complex (Tc) family is a bacterially produced familyof multisubunit supramolecular “ABC” protein complexes,

with the individual Tc protein subunits defined as subunits A, B,and C; all subunit classes are required for full insecticidal activity(10). Since their original discovery, orthologs of the Tc compo-nents have since been identified in several bacterial species from awide variety of Gram-negative bacteria (e.g., Serratia entomophilaSep; Photorhabdus luminescens Tca, Tcb, Tcc, and Tcd; Xenorhab-dus nematophilus Xpt; and Yersinia pseudotuberculosis Tca-like)(5, 15, 18, 31) and at least two Gram-positive bacteria (e.g., Paeni-bacillus nematophila Tca-like and Bacillus thuringiensis Tca-like)(4, 11).

Currently, the mode of action of Tc is not well described. Thehistopathological effects of Tc complexes from P. luminescensagainst Manduca sexta (hawk moth) and Leptinotarsa decemlin-eata (Colorado potato beetle) demonstrated that exposure to theTca Tc complex triggered the destruction of the midgut throughthe lysis of midgut epithelia (2, 3). In contrast to this, no discern-ible effect of the S. entomophila Sep Tc complex on larvae of theNew Zealand grass grub, Costelytra zealandica, was observed aftertreatment; the amber disease caused by the Sep proteins is chronic,with the larvae taking 3 to 4 months to die after the ingestion of S.entomophila (21).

Work to date on the understanding of the role of individual Tcsubunits in the observed toxicity has provided evidence that the Asubunit component contributes to host range specificity, with theXenorhabdus A protein subunit (XptA1) being shown to bind tobrush border membranes of Pieris brassicae (cabbage butterfly)(28). Second, host toxicity can be changed by interchanging the A

subunit component from different Tc loci within or between bac-terial species with that of the BC subcomplex (32, 34). Subse-quently, Lang et al. (27) showed that the C subunits function as themajor toxin components, with the C3 and C5 subunits from P.luminescens Tca encoding ADP-ribosylate activities against actinand RhoA, respectively. To date, no specific function has beendemonstrated for the B subunit component.

The bacterium Yersinia entomophaga MH96 was originally iso-lated from a diseased C. zealandica larva (16). Subsequent hostrange testing revealed that the bacterium is able to cause mortalityacross a broad range of insect species, including those of the ordersColeoptera (e.g., C. zealandica) and Lepidoptera (e.g., Plutellaxylostella, the diamondback moth) (19). The major pathogenicitydeterminant for Y. entomophaga has since been identified as a2.46-MDa multisubunit protein Tc and is referred to as Yen-Tc(10). The genes that encode the individual A, B, and C subunits ofthe Yen-Tc complex are located on a 31.8-kb pathogenicity island,with the Tc operon containing two tcA-like genes (yenA1 andyenA2), a tcB-like gene (yenB), and two tcC-like genes (yenC1 and

Received 12 February 2012 Accepted 21 April 2012

Published ahead of print 27 April 2012

Address correspondence to Sean D. G. Marshall, sean.marshall@agresearch.co.nz.

S.D.G.M. and M.C.H. contributed equally to this work.

Supplemental material for this article may be found at http://aem.asm.org/.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AEM.00431-12

July 2012 Volume 78 Number 14 Applied and Environmental Microbiology p. 4835–4847 aem.asm.org 4835

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

yenC2) (19). Interestingly, a point of difference with Yen-Tc com-pared to other members of the Tc family is the required coexpres-sion of two chitinase genes (chi1 and chi2 [both protein productsare enzymatically active]), whose positions flank the yenA1 andyenA2 genes (6, 19, 25). The structure of Yen-Tc has recently beenresolved by negative-stain reconstruction to a resolution of 17 Å(25). The A1 and A2 components were found to form a pentam-eric pyramidal cage-like structure with the chitinase protein sub-units decorating the outer shell. The B and C components re-mained unresolved but were found through density mapping tobe positioned at the pyramidal point of the pentamer.

Larvae of the New Zealand grass grub, C. zealandica, are theoriginal hosts of two different Tc-bearing bacteria, with distinctTc variants (Sep and Yen-Tc) producing different pathologies.This places C. zealandica as a model to delineate potential differ-ences or commonalities between these Tc clusters. The chronic(but relatively benign) effects of Tcs from pADAP-bearing strainsof S. entomophila on C. zealandica have been well studied (17, 18,20, 21); however, the effects of the highly potent Yen-Tc toxinhave yet to be fully elucidated (19). We present here a detailedhistological study of the effects of the ingestion of Yen-Tc on C.zealandica larvae to determine the histopathology of this toxin.Additionally, we have extended our study to cultured cell lines toinvestigate the effects that individual Yen-Tc subunit componentshave on the observed histopathology and cytology.

MATERIALS AND METHODSSource of bacteria and production of toxin complex proteins. Stock cul-tures of Y. entomophaga (strain MH96T [ATCC BAA-1678T]) are main-tained at AgResearch, Lincoln, New Zealand (AgResearch Bacteria Cul-ture Collection) (16). Methods for the production and purification ofYen-Tc were described previously by Landsberg et al. (25). Briefly, 50-mlcultures of Y. entomophaga were grown overnight at 25°C in LB broth andspun down to collect the Yen-Tc-containing supernatant. This superna-tant was subjected to ammonium sulfate precipitation and size-exclusionchromatography. After fractions containing pure Yen-Tc were identified,pooled fractions were spun through an Amicon Ultracel-30K device toconcentrate purified Yen-Tc in Tris-buffered saline (TBS) (25 mM Tris,130 mM NaCl [pH 7.5]). The protein concentration of purified Yen-Tcwas determined by using a Bradford assay (Bio-Rad).

Collection, handling, and treatment of C. zealandica larvae. Livethird-instar C. zealandica larvae were collected from pastures in the Can-terbury region of South Island, New Zealand, and held in soil at 4°C. Priorto the experimental treatments being set up, larvae were removed fromsoil, fed carrot (3-mm cube), and incubated at 15°C. After 3 days of as-sessment for feeding activity and the absence of disease symptoms, healthylarvae were selected, randomly separated into groups, and starved of foodfor 24 h prior to experimental treatment. Bioassays were carried out intriplicate, with each replicate consisting of 4 larvae per treatment over atime course (0-, 16-, 24-, 48-, 72-, and 96-h intervals). The starved,healthy larvae were placed into 24-well culture trays (1 larva per well) andincubated at 15°C. Experimental treatments were administered as a single5-�l dose of purified Yen-Tc (250 ng) or 5 �l of TBS (control treatment)via direct oral injection into the larval foregut. After treatment, larvaewere fed carrot (3-mm cube), which was replenished daily, and symptomphenotypes were recorded for each time interval.

Histopathology studies. Larvae were removed from trays at appropri-ate time intervals and prepared for histological processing. For the samplefrom each time point, the larval cuticle was opened to expose the digestivetract, and the grub was fixed for 48 h in a 10% neutral buffered formalinsolution. Standard paraffin embedding, serial sectioning, and hematoxy-lin and eosin (H&E) staining methods (23) were carried out on samples byGribbles Veterinary Services (Christchurch, New Zealand). Stained slides

of the alimentary tract were examined under bright-field and differentialinterference contrast (DIC) optics by using an Olympus BX50 uprightmicroscope and photographed with an Olympus DP-12 digital camera.

Electron microscopy. Transmission electron microscopy (TEM) wascarried out on midgut samples collected from third-instar C. zealandicalarvae 24 h after treatment with either TBS or Yen-Tc. Tissue was dissectedin ice-cold primary fixative (2% formaldehyde, 2.5% glutaraldehyde, and2 mM CaCl2 in 0.1 M cacodylate buffer [pH 7.2] with hydrogen peroxidedrops added just prior to dissection). The dissected midgut was thentransferred into fresh fixative for 1 to 1.5 h at room temperature (�22°C).Following three 15-min washes in buffer (0.1 M cacodylate, 2 mM CaCl2[pH 7.2]), samples were stored overnight at 4°C and then placed in 4°Csecondary fixative {1% OsO4, 1.5% K3[Fe(CN)6]} for 1 h, followed bythree washes (5 min each) with ultrapure water at room temperature.Samples were soaked in mordant (1% tannic acid in phosphate buffer [pH7.0]) for 2 h and washed for 15 min in a solution of 1% Na2SO4 in phos-phate buffer (pH 7.0), followed by three 5-min water washes, before beingplaced into a tertiary fixation solution (1% uranyl acetate) at room tem-perature for 1 h in the dark. Samples were washed three times (5 min each)in water and stored overnight at 4°C. Dehydration was done via a gradedethanol series (70, 80, and 90% ethanol for 15 min each at room temper-ature), twice in 100% ethanol (20 min each), and then in dry acetone (20min). Samples were embedded in Procure 812 (ProSciTech Pty., Austra-lia) epoxy resin (50% resin–50% acetone for 1 h, three changes of 100%resin for 1 h, and 100% resin overnight with rotation). Resin blocks werepolymerized at 60°C for 22 h. Sections (100 nm thick) were cut by using anUltracut UCT ultramicrotome (Leica, Germany) fitted with a 45° dia-mond knife (Diatome, Switzerland). Sections were cut onto ultrapurewater, collected onto 100-mesh copper Formvar-coated grids, and exam-ined by using an FEI Morgagni 268D transmission electron microscope.Micrographs were captured by using an SIS/Olympus Megapixel III digi-tal camera mounted above the phosphor screen.

Programmed cell death assays. Sectioned larval samples for terminaldeoxynucleotidyltransferase-mediated dUTP nick end labeling (TUNEL)staining were prepared from paraffin blocks as described above but with-out H&E staining. TUNEL reactions were conducted according to themanufacturer’s instructions, using an in situ cell death detection kit(tetramethylrhodamine [TMR] red; Roche). Sections were counter-stained to visualize nuclei by incubating slides at room temperature for 5min in phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, and 10mM phosphate buffer solution [pH 7.4]) containing 100 ng ml�1 4=,6=-di-amidino-2-phenylindole hydrochloride (DAPI; Roche), followed by three5-min rinses in PBS. Prior to coverslipping, the stained sections weremounted with a drop of antifade mountant consisting of 1% (wt/vol)Mowiol 4-88 (Calbiochem) and 2.5% (wt/vol) 1,4-diazabicyclo[2.2.2]octane(DABCO; Sigma) in Tris-buffered glycerol (pH 8.5) (1). Fluorescence stain-ing was observed by using Olympus filter sets U-MWIG2 (for TMR) (520- to550-excitation wavelength [ex]/588-nm emission wavelength [em]) and U-MWU2 (for DAPI) (330- to 385-nm ex/420-nm em).

Samples for the monitoring of caspase-3/7 activity were collected bydissecting midguts from TBS- and Yen-Tc-treated grubs at various timepoints, with three replications of six larvae per treatment. Following dis-section, midguts were rinsed in 1 to 2 ml PBS and ground with mi-cropestles in 1.7-ml tubes containing 100 �l of ice-cold extraction buffer(20 mM Tris-HCl, 5 mM EDTA, 5 mM EGTA, 1 mM dithiothreitol[DTT], 150 mM NaCl [pH 7.5]). Samples were centrifuged for 10 min at13,000 � g, and supernatants were used for enzyme assays. Caspase-3/7activity was measured by the addition of 5 �l of midgut extract to 100 �lcaspase reaction buffer (25 mM HEPES-KOH, 10% glycerol, 0.25 mMEDTA, 2.5 mM DTT [pH 7.5]), with the reaction initiated by the additionof N-acetyl–Asp–Glu–Val–Asp–p-nitroanilide (DEVDpNA) (caspase-3/7substrate) to a final concentration of 0.1 mM. For inhibitor experiments,prior to the addition of DEVDpNA, the midgut sample was incubated for30 min in caspase reaction buffer supplemented with either 0.1 mM N-acetyl–Asp–Glu–Val–Asp–Val (DEVD-CHO) or 0.1 mM (each) leupep-

Marshall et al.

4836 aem.asm.org Applied and Environmental Microbiology

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

tin and aprotinin. The reactions were monitored at 410 nm by using aBMG Fluorostar Optima microplate reader, and initial velocities weredetermined.

Topical application of the native Yen-Tc complex. Clonal Spodopterafrugiperda IPLB-Sf-21-AE (Sf9) cells (Invitrogen) were seeded into 6-wellplates containing borosilicate glass coverslips (BDH Merck Ltd.) andgrown for 24 h in Grace’s medium, supplemented with 3.3 g liter�1 lact-albumin hydrolysate hydrosolate (Invitrogen), 3.3 g liter�1 yeastolate L-glutamine (Invitrogen), and gentamicin (25 �g ml�1) (Sigma) and incu-bated in air at 28°C until cells were 80% confluent. Purified Yen-Tc in asolution containing 20 mM Tris-HCl (pH 7.5) and 130 mM NaCl wasadded to 1 ml of fresh medium to a final concentration of 5 ng ml�1 andincubated for 24 h at 28°C. After 24 h, cells were washed with PBS andfixed with 4% paraformaldehyde in PBS for 20 min at room temperaturebefore being permeabilized with 0.02% Tween 20 in PBS for 20 min. Cellswere washed three times in PBS before being incubated with 50 �g ml�1

fluorescein isothiocyanate (FITC) phalloidin (Sigma) for 1 h at roomtemperature in the dark, followed by a 10-min incubation with 0.12 mgml�1 Hoechst 33258 DAPI stain (Sigma). Cells were subsequently washedtwice with PBS, followed by two washes with distilled water. Coverslipswere mounted with Mowiol mounting medium (0.12 M Tris-HCl [pH6.8], 30% glycerol, 12% Mowiol, 2.5% DABCO). Slides were examinedand photographed with an Olympus BX61 microscope. Each experimentwas conducted in triplicate and repeated twice.

The human Caucasian colon adenocarcinoma cell line (Caco-2) (12)was treated as described above, except that cells were incubated with Dul-becco’s modified Eagle’s medium (DMEM) (Invitrogen) supplementedwith 10% fetal calf serum, 1% nonessential amino acids (Sigma), and 1%penicillin and streptomycin at 37°C with 5% CO2.

Construction of Y. entomophaga Tc Myc fusion expression vectors.Full-length Yen-Tc-associated coding sequences were PCR amplified byusing primers Chi1, A1 (yenA1), A2 (yenA2), Chi2, B (yenB), C1 (yenC1),and C2 (yenC2), which contain appropriate Myc (forward) and Myc-STOP (reverse) primer combinations (Table 1), to allow the introductionof an N-terminal Myc fusion tag. PCRs were performed by using Ac-cuprime Pfx DNA polymerase (Invitrogen) according to the manufactur-er’s instructions, with amplified products subsequently being cloned intothe BamHI and XbaI restriction sites of the cytomegalovirus (CMV) pro-moter-based eukaryotic expression vector pRK5myc (Clontech). PCRproducts and plasmid templates were purified by using the High Pure

PCR product purification kit and the High Pure plasmid isolation kit(Roche Diagnostics GmbH), respectively. All pRK5myc-Tc subunit-de-rived clones were sequenced to confirm the integrity of the expressionconstructs.

Transient expression of individual Yen-Tc components in cell cul-ture. Caco-2 cells were seeded into 6-well plates containing borosilicateglass coverslips (BDH) and grown for 24 h in 2 ml DMEM (Invitrogen)supplemented with 10% fetal calf serum, 1% nonessential amino acids(Sigma), and 1% penicillin and streptomycin at 37°C in 95% air–5% CO2

(vol/vol) until cells were 80% confluent. One microgram of eachRK5myc-Tc subunit clone was cotransfected with 0.5 �g pEGFP-actin byusing Nanojuice (Invitrogen). Each experiment was performed in tripli-cate and repeated independently three times. After 24 h, the cells werefixed with 4% paraformaldehyde in PBS, permeabilized with 0.2% TritonX-100 in PBS, and blocked with Image-iT FX signal enhancer (Invitrogen)according to the manufacturer’s instructions. Cells were labeled withmonoclonal mouse anti-Myc primary antibody (Invitrogen) and visual-ized with Alexa Fluor 594 goat anti-mouse dye-labeled secondary anti-body (Molecular Probes) to detect the Myc-labeled protein. Nuclei werevisualized by incubation with Hoechst 33258 DAPI stain (Sigma) at a finalconcentration of 0.12 mg ml�1 in PBS for 10 min, followed by two washeseach with PBS and double-distilled water. Coverslips were mounted inMowiol mounting medium and subsequently examined and photo-graphed by using an Olympus BX61 microscope.

RESULTSGross external phenotypic effects of Yen-Tc exposure on C. zea-landica larvae. C. zealandica larvae exposed to 250 ng of purifiedYen-Tc toxin displayed the same progression of symptoms thatwas observed previously for larvae that ingested live Y. ento-mophaga cells (19). Prior to treatment, healthy larvae exhibitedactive movement, a darkened gut filled with ingested food/soilmatter, and an expulsion of discrete frass pellets. Within 1 to 2days following the ingestion of Yen-Tc, C. zealandica larvae dis-played an amber coloration and clearance of gut contents (see Fig.S1 in the supplemental material), which was similar to the pheno-type described previously by Hurst et al. (19). This was accompa-nied by notable fluid loss through vomiting (larvae also became

TABLE 1 Primer sequences and their resultant pRK5myc clone designations

Primer Primer sequence (5=¡3=)a Plasmid

Chi1-Myc AAAGGATCCATGGAAAAAGAAGAAAAAAGCAATCTCATCTACG pRK5myc-Chi1Chi1-Myc-STOP AAATCTAGATTATTTGATATTACATGCTTCGCACTGTGGC

A1-Myc AAAGGATCCATGGATAAATATAATAATTATTCTAATGTAA pRK5myc-A1A1-Myc-STOP AAATCTAGATTAGACATTATCGCTATCTCCTTGTGCCC

A2-Myc AAAGGATCCATGTCTAATTCTATTGAAGCGAAACTACAGG pRK5myc-A2A2-Myc-STOP AAATCTAGATTATTCATCGTTGTTTATATTGGCCAATATTTTCC

Chi2-Myc AAAGGATCCATGGTCAATAAATATACTTACACCTCATC pRK5myc-Chi2Chi2-Myc-STOP AAATCTAGATTTACTTGCTCTTAATTTCAGAAATTGATTTTAAG

B-Myc AAAGGATCCATGCAAAATTCACAGGAAATGGCTATCACG pRK5myc-BB-Myc-STOP AAATCTAGATCTAAATCGCCAGATTTGCTGCCGTATCG

C1-Myc AAAGGATCCATGAACCAGTTTGATTCTGCGCTCCATC pRK5myc-C1C1-Myc-STOP AAATCTAGATCAGCGTAATTTAACGACGGTATTCTTTC

C2-Myc AAAGGATCCATGGACATACAGCTGTTTAGTAAAACCCC pRK5myc-C2C2-Myc-STOP AAATCTAGACTAAAATGACTTTCGACGCTTTAGTCTTACATCa Underlining indicates the restriction enzyme site used for cloning; boldface type denotes the Tc-derived termination codon (STOP).

Histopathological and Cytological Effects of Yen-Tc Toxin

July 2012 Volume 78 Number 14 aem.asm.org 4837

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

increasingly flaccid), an excessive expulsion of discrete frass ma-terial, and the progressive development of lethargy and browningcoloration. The average time to death was 5.7 � 0.4 days posttreat-ment, which was estimated by using the Kaplan-Meier method(22), with the observed mortality rates being 41.7% by day 4 and80.6% by day 9 posttreatment (see Table S1 in the supplementalmaterial).

Histopathological examination of Yen-Tc treatment of C.zealandica larvae. All third-instar C. zealandica tissues were ini-tially examined for damage caused by Yen-Tc exposure. However,prior to larvae reaching a moribund stage, only the midgut dis-played major alterations in morphology. Therefore, we focused onthe changes occurring in the larval midgut as a result of Yen-Tcintoxication. In conjunction with an examination of the histo-pathological effects of purified Yen-Tc on the C. zealandicamidgut, we conducted control oral injections with TBS buffer (thecarrier buffer for purified Yen-Tc) to ensure that the detection ofany effects was not caused by the delivery procedure itself. Nodamage or alteration to the midgut was observed for larvae in-jected with TBS.

In healthy third-instar C. zealandica larvae, the anteriormidgut is comprised of densely packed elongated columnar cellsneatly arranged in a typical invaginated configuration, with largeovoid nuclei lined up in a chain-like fashion, while the posteriormidgut is similarly arranged but lacks the invaginations promi-nent in the anterior midgut (Fig. 1A and B). The tissue folds in theanterior midgut are thought to be caused by the high abundanceof underlying regenerative nidi (RN) and presumptive endocrinecells; note that goblet cells, which are typical of other holometabo-lous insects, are absent in the C. zealandica larval midgut. Histo-logical observations also revealed well-defined microvilli liningthe midgut lumen (Fig. 1A and B). The gut lumens from TBS-treated larvae were filled with ingested food/soil material enclosedby a peritrophic membrane to form a discrete food bolus (Fig. 1Aand B). Other features observed included the occasional presenceof nucleus-free vesicles, which are thought to be either putativeaprocrine secretory vesicles or “old/damaged” columnar cellssloughed off from the midgut epithelium (data not shown).

For the purposes of comparison against both control-treated(TBS, as described above) and Yen-Tc-treated (see below) larvae,histology sections from midgut tissues of larvae orally treated withthe Serratia entomophila Sep toxin (another member of the Tcfamily which also induces a gut-clearing, amber coloration phe-notype in C. zealandica larvae) were also examined (see Fig. S2 inthe supplemental material). Although the Sep complex itself elic-ited the clearance of the gut contents (resulting in an amber gutcoloration), no other obvious changes in the midgut epithelialstructure were observed after the administration of the Sep pro-tein complex. Although the reasons for this remain unknown, thelack of an obvious effect was not unexpected, since the ingestion ofthe amber disease-causing bacterium, S. entomophila, was previ-ously demonstrated to have limited toxicity to C. zealandica larvae(20, 21).

During the early stages following the oral injection of purifiedYen-Tc (prior to approximately 16 h posttreatment), any ob-served change or damage was inconsistent and difficult to detect,with only a few affected cells seen among a mass of relativelyhealthy-looking midgut cells. Beginning approximately 16 h afterthe oral administration of the Yen-Tc dose, an anterior-to-poste-rior sequence of events was consistently observed for the C. zea-

landica larval midgut (Fig. 1C and D). Due to the anterior-poste-rior progression of the observed effects, we have arbitrarilydivided the midgut into two portions. The anterior section beginsat the cardiac valve and ends just ahead of the first cecal row, whilethe posterior section of the midgut extends from the two rows ofcecae and continues to the pyloric sphincter.

From 16 to 48 h after treatment (Fig. 1C to H), the most con-spicuous histological changes were the clearing of gut contentsaccompanied by the progressive deterioration and eventual loss ofthe peritrophic membrane. Concurrently, patches of microvillibecame less distinct and eventually disappeared throughout themidgut by 48 h.

Between approximately 16 and 24 h post-Yen-Tc treatment(Fig. 1C to F), the anterior midgut columnar cells displayed prom-inent vacuolization and began to lose their normal neatly orderedrectangular form, with the cells taking on an oval and slightlyshrunken and ragged appearance and with the arrangement of thenuclei now appearing disarrayed. The loss of columnar cell integ-rity coincided with the appearance in the gut lumen of roundedvesicles containing putative condensed pyknotic nuclei (“nuclearvesicles”). The posterior midgut columnar cells displayed a similarprogressive disorganization and extrusion of these cell-like bod-ies; however, the timing of events was shifted approximately 4 to 8h after the events of the anterior midgut (Fig. 1D and F).

By 48 h, a large number of vesicles had already been sloughedoff into the midgut lumen; this was accompanied by the large-scale deterioration of the gut epithelium (Fig. 1G and H). At 72 hpost-Yen-Tc treatment, the epithelial cells of the midgut had rap-idly degenerated to a point where it was difficult to discern anyspecific pattern of events (Fig. 1I and J). By 96 h, the midgutepithelium had completely broken down, with remnants of theputative midgut being observed in only a few sections from thespecimens examined (Fig. 1K and L).

Ultrastructural changes observed for C. zealandica midgutcells exposed to Yen-Tc. To examine more closely the subcellulareffects of Yen-Tc intoxication, ultrathin sections of midgut tissuewere examined 24 h after treatment (Fig. 2). For larvae treatedwith TBS as a control treatment (Fig. 2A and B), a wide array ofstructures (presumptive secretory vesicles, mitochondria, andother unidentified structures) were observed, and the nucleus ap-peared to be intact, with chromatin material being well dispersed.Toward the basal end of the cell, the mitochondria were packed inan anisotropic configuration, being oriented perpendicular to thenearby basal membrane (Fig. 2B). However, 24 h after Yen-Tctreatment (Fig. 2C to E), many changes were observed. Large,roughly circular patches of condensed chromatin became appar-ent within the nuclei of several cells (although this was not ob-served for all cells), and breakages in one layer of the nuclearenvelope double layers were also detected (Fig. 2C). The cells’basal mitochondria had a more isotropic appearance than those ofthe control cells, and throughout the cells, most mitochondriaappeared slightly swollen, with dark spots now being easily ob-served within them (Fig. 2C and E). As shown in Fig. 2D, cells thatdisplayed a cytoplasm containing numerous vacuoles and darkelectron-dense staining material were also observed, and there wasalso evidence for the cytoplasm being released from cells withbroken membranes. Finally, as indicated by histology sections(Fig. 1), not all cells responded uniformly to toxin treatment (asshown in Fig. 2E).

Programmed cell death assays on C. zealandica larvae. From

Marshall et al.

4838 aem.asm.org Applied and Environmental Microbiology

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

the histopathological observations made subsequent to Yen-Tcexposure, it appeared that an ordered series of cellular events ledto the eventual breakdown of the midgut tissue. In particular, theappearance of cell body-like vesicles containing condensed nu-clear material displayed a striking similarity to one of the hallmarkfeatures described previously for cells undergoing an apoptosis-

like processes (29). To determine if the condensed chromatin ma-terial was fragmenting (another characteristic of apoptotic nu-clei), in situ TUNEL assays were carried out on midgut sectionscollected 24 h after either Yen-Tc or TBS buffer treatment (Fig. 3).TUNEL-positive cells were observed only in Yen-Tc treated larvae(Fig. 3D to F), indicative of nicked or broken DNA being present

FIG 1 Longitudinal midgut sections from third-instar C. zealandica larvae stained with hematoxylin and eosin following oral ingestion of either a TBSbuffer-only (control) treatment (24 h posttreatment) (A and B) or the purified Y. entomophaga Tc protein complex at the indicated time points (C to L). Allimages are representative of midgut tissue located either anteriorly or posteriorly (as indicated and in relation to the first and second cecal rows of the C.zealandica digestive tract). (A and B) For the control treatment, note the well-formed columnar cells, an intact microvillus surface lining the luminal surface ofthe midgut, and the presence of ingested soil matter enclosed by a peritrophic membrane. (C and D) At 16 h post-Yen-Tc treatment (C), food matter hadgenerally disappeared, the anterior columnar cells had shrunk, and a sloughing of nucleus-containing vesicles into the gut lumen was observed; during this time,the columnar cells in the posterior gut (D) were only just beginning to show signs of disorganization and shrinkage. (E and F) By 24 h, both the anterior (E) andposterior (F) regions of the gut epithelium showed obvious signs of intoxication, with nucleus-containing vesicles being easily observed within the luminal space.(G to L) From 48 h onwards, the entire gut was in the process of disintegration, and by 96 h (K and L), remnants of the midgut were only occasionally detected.Arrows point to highlighted structures labeled as follows: BM, basement membrane; CC, columnar cell; L, gut lumen; M, muscle; MV, microvilli; N, nucleus; NV,nucleus-containing vesicles; PM, peritrophic membrane; RN, regenerative nidi; V, vesicle. Scale bars represent 50 �m.

Histopathological and Cytological Effects of Yen-Tc Toxin

July 2012 Volume 78 Number 14 aem.asm.org 4839

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

in those cells. Although positively stained cells were not as numer-ous as expected, nor were they as brightly labeled (compared to thepositive-control DNase I-treated sections [data not shown]), theTUNEL negative-control sections (data not shown) and healthy

midgut sections (Fig. 2A to C) were almost completely devoid ofany staining in the TMR channel. Associated with the sloughedcell-like bodies within the gut lumen, the DAPI-counterstainedsections revealed a pattern of densely staining DNA material,

FIG 2 Transmission electron micrographs of ultrathin sections (stained with 1% uranyl acetate) of midgut tissue from third-instar C. zealandica larvae 24 hfollowing oral ingestion of either control treatment (TBS buffer only) (A and B) or purified Yen-Tc toxin (C to E). (A) The apical portion of a healthy activemidgut cell showing structures involved in secretion and absorption. (B) The basal portion of a healthy midgut cell showing elongated mitochondria orientedalong a basal-apical axis, surrounded by small smooth vesicles. (C) At 24 h post-toxin treatment, the region around the nucleus had variable-sized, irregular-shaped vesicles and swollen mitochondria with prominent dark spots, while the nucleus displayed a globular electron-dense condensation of nuclear material andevidence of breaks in the double membrane of the nuclear envelope. (D) Presence of both apoptotic-like cells (upper cell showing a condensed cytoplasm withevidence of organelles being broken down in lysosomes) and necrotic-like cells (lower cell showing cytoplasm escaping from broken membranes). (E) The twowhite triangles highlight the junction of two cells, revealing cell-to-cell variation, with the left-hand cell displaying variable-sized vesicles with irregular profilesand less-well-oriented mitochondria (both swollen and elongated mitochondria are present) and the right-hand cell showing less dramatic changes. Arrows pointto highlighted structures labeled with as follows: B, basement membrane; cN, nucleus with condensed chromatin; M, mitochondria; Ne, nuclear envelopedisrupted; SV, secretory vesicle; Va, vacuole. Paired white triangles indicate the boundary of two adjoining cells. Scale bars represent 2 �m.

FIG 3 TUNEL staining of longitudinal sections from third-instar C. zealandica larvae 24 h post-oral treatment with either the control (TBS buffer only) (A to C) orpurified Yen-Tc (D to F). Arrows indicate examples of either TUNEL-positive nuclei (TMR label appears as red fluorescence) (A and D), nuclei stained with DAPI (bluefluorescence) (B and E), or nuclei displaying both TUNEL-positive and DAPI staining (purple appearance in merged images) (C and F). Scale bars represent 50 �m.

Marshall et al.

4840 aem.asm.org Applied and Environmental Microbiology

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

which corroborated the concentrated purple eosin pigmentationobserved for the H&E-stained histology sections and the con-densed chromatin in the TEM observations.

DNA damage can be caused by other processes, includingother cell death pathways, such as programmed necrosis (24).Generally speaking, two principal mechanisms of apoptotic pro-grammed cell death have been identified: the intrinsic pathway,where agents interact with the mitochondria, and the extrinsicpathway, where agents directly activate the family of cell deathreceptors. This results in the activation of a defined cascade ofproteins known as the caspases, which initiate (caspase-2, -8, -9,and -10) and induce (caspase-3, -6, and -7) a progressive series ofevents which ultimately leads to cell death (24). To investigate ifanother component part of an apoptosis-like response was beingactivated, the appearance of cysteine aspartate-specific proteinase(caspase) activity in C. zealandica larval midguts exposed to thepurified Yen-Tc protein complex was monitored (Fig. 4). Thehydrolysis of the substrate DEVDpNA is generally accepted as ameasure of caspase-3/7 activity being present. As shown in Fig. 4A,a negligible cleavage of DEVDpNA was observed for the mock-treated larvae (0 h), but by 16 h post-Yen-Tc treatment, a 43-foldincrease in DEVDpNA hydrolysis was measured (0.694 versus

0.016 nmol min�1 gut�1); significant levels of caspase activitywere also observed over time points from 24 to 48 h posttreat-ment. While a decline in the level of caspase activity over thistime was observed, the levels were still manyfold higher thanthose observed for healthy larvae (with 34-, 19-, and 16-foldincreases over the mock-treated larvae at 24, 36, and 48 h,respectively). The specific inhibition of this activity was dem-onstrated by the addition of the reversible inhibitor DEVD-CHO prior to the measurement of DEVDpNA cleavage (Fig.4A). As C. zealandica midguts are known to contain a high levelof serine protease activity, a second experiment was conducted,whereby extracts were preincubated with the serine proteaseinhibitors leupeptin and aprotinin (Fig. 4B) prior to the mea-surement of DEVDpNA hydrolysis, to ensure that promiscu-ous substrate utilization was not contributing to the activitylevels observed. No significant inhibition was recorded whenextracts were preincubated with the serine protease inhibitorsleupeptin and aprotinin (Fig. 4B).

Effects of Yen-Tc and individual Yen-Tc components on cul-tured cell lines. To delineate the effects observed with the histo-logical examinations described above, Yen-Tc was topically ap-plied to the Spodoptera frugiperda clonal insect cell line Sf9.Strikingly, after 24 h, 34% (standard error [SE] � 0.9) of cellsshowed apoptosis, as demonstrated by pyknotic condensed nuclei(NC) (Fig. 5B and C, red arrows), compared to 10% (SE � 1.5) ofTBS-treated control cells (Fig. 5E and F), and was associated withactin condensation (AC) (Fig. 5A, C, and G). Yen-Tc-treated Sf9cells either were observed to be normal compared to the controlcells or showed nuclear and actin condensation. Few cells showedany intermediate stages, such as actin vacuoles (14%), and theseeffects were comparable to those on the control cells (13.5%).Interestingly, a small but not significant effect, an increase in thenumbers of multinucleated cells (M) (5% of cells [SE � 0.9]) (Fig.5B, C, and G), compared to 3% (SE � 3.25) of TBS-treated controlcells, was also observed.

When Yen-Tc was applied to Caco-2 cells, only 0.8% of cellsshowed pyknotic condensed nuclei, which is considerably lowerthan that observed for the Sf9 cells topically treated with Yen-Tc(34%), while 9% (SE � 3.15) of Caco-2 cells showed AC (control,0.25% [SE � 0.25]) (Fig. 6A, C, and G). However, increases innumbers of intermediate stages such as actin vacuoles (V) (9%[SE � 3.57]) (Fig. 6A and C) compared to the control (4.3% [SE �3.15]) (Fig. 6D and F) as well as enlarged cells (7% [SE � 3.4])compared to the control (2.4% [SE � 2.4]) were observed. Anincrease in numbers of multinucleated cells (7% [SE � 3.0]) (Fig.6B and C), compared to 3% (SE � 0.4) of TBS-treated controlcells (Fig. 6E and F), was also observed.

To further ascertain the importance of the individual compo-nents of Yen-Tc as well as the significance of Chi1 and Chi2,pRK5myc constructs of the components of the Yen-Tc complex(A, B, and C subunits as well as the two chitinases) were trans-fected into cells. As described above, similar phenotypic effectswere observed for Caco-2 and Sf9 cells. However, the observedphenotypic effects of Yen-Tc on Sf9 cells (as shown in Fig. 5)occurred very rapidly (i.e., cell death within 24 h) in comparisonto the events observed for Caco-2 cells (Fig. 6). Indeed, an analysisof Caco-2 cell viability by trypan blue staining or 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)carbonyl]-hydroxide (XTT) colorimetric analysis showed an increased num-ber of cells at 24 h; only after 48 h did the number of viable cells

FIG 4 Caspase activity present in gut extracts collected from third-instarC. zealandica larvae at 0, 16, 24, 36, or 48 h after control or Yen-Tc treat-ment (as indicated). (A) Hydrolysis of the caspase-3/7 substrate (DEVD-pNA at a 0.1 mM final concentration) measured after incubation in eitherthe absence or the presence of the caspase-3/7 inhibitor (DEVD-CHO at a0.1 mM final concentration). (B) Hydrolysis of the caspase-3/7 substrate(DEVDpNA) measured after incubation in either the absence or the pres-ence of the serine protease inhibitors leupeptin and aprotinin (0.1 mMeach) (�Leu�Apro). The cleavage of DEVDpNA was measured as therelease of p-nitroaniline (pNA).

Histopathological and Cytological Effects of Yen-Tc Toxin

July 2012 Volume 78 Number 14 aem.asm.org 4841

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

decrease (data not shown). The slower progression of Yen-Tc tox-icity in Caco-2 cells likely reflects differences based on relative cellsize, with Sf9 cells being smaller than Caco-2 cells, although adifference in host specificity may also contribute. Based on thisslower progression and to allow comparisons with data from pre-viously reported Tc-related studies undertaken by Waterfield et al.(34) and Hares et al. (15), Caco-2 cells were used for transfection

studies. Additionally, Caco-2 cells are significantly larger than Sf9cells and therefore easier to study for phenotypic variation.

Surprisingly, after 24 h, 20% (SE � 10.2) of Caco-2 cells wereobserved to have actin blebs when treated with pRK5myc-Chi1(Fig. 7A and D), and increases in numbers of cells with nuclearcondensation (17% [SE � 4.6]) and nuclear blebbing (16% [SE �3.7]) compared to the control (11% [SE � 1.6] and 10% [SE �

FIG 5 (A to F) Fluorescence micrographs of Sf9 cells topically treated with the native Y. entomophaga toxin complex (A to C) or the Tris buffer control(D to F). Cells were stained for F-actin with FITC-phalloidin (A and D) and for nuclei with Hoechst 33258 DAPI stain (B and E); merged images are alsopresented (C and F). Red arrows indicate actin and nuclear condensation (A to C), and yellow arrows indicate multinucleated cells (B and C). (G) Graphpresenting percentages of cells (minimum of 110 cells counted) scored that displayed actin condensation (AC), nuclear condensation (NC), andmultinucleation (M). Asterisks indicate high numbers of cells showing actin and nuclear condensation compared to the control. Error bars representstandard errors. Scale bars represent 20 �m.

Marshall et al.

4842 aem.asm.org Applied and Environmental Microbiology

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

0.7], respectively) were also observed (Fig. 7AC and AF), whilepRK5myc-Chi2 was observed to have an increase in the number ofcells showing actin condensation (24% [SE � 5]), actin vacuoles(19% [SE � 3.8]), and multinucleation (10% [SE � 7.4]) com-pared to control cells (12.5% [SE � 2.3], 10% [SE � 6.3], and 6%[SE � 3], respectively) (Fig. 7AC to AF). Both pRK5myc-Chi1 andpRK5myc-Chi2 were observed to colocalize to the cytoplasm (Fig.7B, D, F, and H). While effects were observed with Chi1 and Chi2,these were not as striking as those observed with the Yen-Tc A, B,or C construct.

Transfections with pRK5myc-A1 and pRK5myc-A2 bothshowed significant increases in the numbers of cells with actincondensation (pRK5myc-A1, 29% [SE � 6.8]; pRK5myc-A2, 23%[SE � 10.3]) and actin vacuoles (pRK5myc-A1, 18% [SE � 6.3];pRK5myc-A2, 36% [SE � 8.7]) (Fig. 7I, L, M, and P) compared to

the pRK5myc control (12.5% [SE � 2.3] and 10% [SE � 3.8],respectively) (Fig. 7AC and AF), consistent with the phenotypeobserved with Yen-Tc topically applied to Sf9 (Fig. 5) and Caco-2(Fig. 6) cells. Interestingly, only pRK5myc-A1 showed increases innuclear condensation (27.5% [SE � 3.3]) and numbers of multi-nucleated cells (10% [SE � 8.5]) (Fig. 7K and L), whereaspRK5myc-A2 (Fig. 7O and P) showed 18% (SE � 0.8) of cells withnuclear condensation (controls displayed a nuclear condensationof 11% of cells [SE � 5.0] and multinucleation of 6% of cells [SE �3.0]) (Fig. 7AE and AF). As can be seen in Fig. 7J, N, L, and O,pRK5myc-A1 and pRK5myc-A2 localized to the cytoplasm.

The transfection of cells with pRK5myc-B also resulted in in-creases in the numbers of cells with actin condensation (31%[SE � 16.4]) and large actin vacuoles (24% [SE � 2.8]) (Fig. 7Qand T). Nuclear fragmentation and blebbing were also observed

FIG 6 (A to F) Fluorescence micrographs of Caco-2 cells topically treated with the native Y. entomophaga toxin complex (A to C) or the Tris buffer control(D to F). Cells were stained for F-actin with FITC-phalloidin (A and D) and for nuclei with Hoechst 33258 DAPI stain (B and E); merged images are alsopresented (C and F). Red arrows indicate vacuoles (A and C), and yellow arrows indicate multinucleated cells (B and C). (G) Graph presenting percentagesof cells (minimum of 360 cells) with multinucleation (M), actin condensation (AC), or vacuole (V) phenotypes observed. Asterisks indicate high numbersof cells showing multinucleation and actin condensation compared to the control. Error bars represent standard errors. Scale bars represent 50 �m.

Histopathological and Cytological Effects of Yen-Tc Toxin

July 2012 Volume 78 Number 14 aem.asm.org 4843

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

FIG 7 Intracellular transient expression of tc toxin genes from Y. entomophaga. Expressed cMyc fusion proteins were visualized with Alexa Fluor 594 goatanti-mouse dye-labeled secondary antibody (Molecular Probes) (B, F, J, N, R, V, and Z, compared to AD [control]). Actin was detected with an expressedactin-enhanced green fluorescent protein (EGFP) fusion protein (A, E, I, M, Q, U, Y, and AC), while nuclei were highlighted with Hoechst 33258 DAPI stain (C,G, K, O, S, W, AA, and AE). Merged fluorescence signals are presented in panels D, H, L, P, T, X, AB, and AF. (A to D) Cells transfected with pRK5myc-Chi1 andpEGFP-actin. These cells show actin blebs (yellow arrows) (A and D) and are localized to the cytoplasm (B and D) (red arrows). (E to H) Cells transfected withpRK5myc-Chi2 and pEGFP-actin. As can be observed, pRK5myc-Chi2 causes actin vacuoles (E and H) (yellow arrows) although to a lesser degree than thoseobserved for RK5myc-A1 (I and L) and RK5myc-B (Q and T). Again, pRK5myc-Chi2 is localized to the cytoplasm (F and H). (I to L) Cells transfected withpRK5myc-A1 and pEGFP-actin. pEGFP-actin in RK5myc-A1 shows vacuoles (I and L) (yellow arrows) and nuclear (K and L) and actin (I and L) condensation(white arrows). The RK5myc-A1 tag shows localization to the cytoplasm (J and L) (red arrows). (M to P) Cells transfected with pRK5myc-A2 and pEGFP-actin.pEGFP-actin in RK5myc-A2 shows vacuoles (M and P) (yellow arrows). The RK5myc-A2 tag shows localization to the cytoplasm (N and P) (red arrows). (Q to

Marshall et al.

4844 aem.asm.org Applied and Environmental Microbiology

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

for 21.5% (SE � 21.5) of cells (Fig. 7S, white arrow) compared to0.7% (SE � 0.7) of control cells, as observed by the histologicalexamination of C. zealandica larvae fed Yen-Tc (Fig. 1), and werelocalized to the nucleus (Fig. 7Q and S).

Both pRK5myc-C1 and pRK5myc-C2 caused increases in num-bers of actin aberrations, including actin blebs (32% [SE � 3.7] and23% [SE � 2.3], respectively), actin vacuoles (27% [SE � 3.3] and30% [SE � 8.8], respectively), and actin condensation (34% [SE �9.9] and 24% [SE � 12.2], respectively) (Fig. 7U, X, Y, and AB),compared with the control cells (2% [SE � 1.6], 10% [SE � 6.3], and12.5% [SE � 2.3], respectively) (Fig. 7AC and AF). As observed forthe Sf9 cells with topically applied treatment (Fig. 5), there was also anincrease in the number of cells with condensed nuclei (18% [SE �0.8] and 20% [SE � 11.5]), indicative of apoptosis. pRK5myc-C2 wasalso observed to have an increase in the number of multinucleatedcells (16% [SE � 8.3]) compared to control cells (6% [SE � 3.0]).Both pRK5myc-C1 and pRK5myc-C2 were observed to colocalize tothe cytoplasm (Fig. 7V, X, Z, and AB).

DISCUSSION

The main disease determinant of the insecticidal bacterium Y.entomophaga has been identified as Yen-Tc, which is a novel mem-ber of the toxin complex family (19). To better understand thelethal effects that Yen-Tc has on insects and to determine its directcontribution to insect mortality from that of the entire Y. ento-mophaga bacterium, a histopathology study of the C. zealandicalarval midgut after exposure to purified Yen-Tc protein was con-ducted. This study demonstrated that purified Yen-Tc alone wasable to cause a series of sequential and predictable events thateventually led to death of the intoxicated insects. Following theoral ingestion of the Yen-Tc toxin, a distinctive clearing of gutcontents was coupled with a progressive loss of the peritrophicmembrane and the disappearance of microvilli. Prior to the even-tual complete deterioration of the gut epithelium, midgut colum-nar cells became progressively more disorganized in appearance,which was coupled with the extrusion and sloughing of cell-likebodies (containing condensed pyknotic nuclei and fragmentednuclear material) into the gut lumen. The observed ultrastructuralchanges included the appearance of grainy patches and prominentelectron-dense material, a reorganization of basally located mito-chondria, and a rippling of the basal membrane.

The destruction of the C. zealandica larval midgut epitheliumby Yen-Tc displayed a histopathology similar to those observedpreviously for other Tc-intoxicated insects, including that for P.xylostella intoxicated by Yen-Tc (19, 25). The oral exposure of M.sexta larvae to the P. luminescens Tca toxin complex was demon-strated previously to cause an accelerated release of vesicles fromthe midgut epithelium into the gut lumen, followed by the even-tual disintegration of the entire gut epithelial wall (2). Furthertesting of P. luminescens Tca against L. decemlineata showed thatthe toxin effects were equally as disruptive to the gut epithelium(3). While such similarities seem obvious, given the shared se-quence similarity of key regions within each of the A, B, and C

subunit classes, there are also regions of amino acid sequence vari-ability between the orthologous subunits encoding the various Tccomponents (13, 15, 19, 26). These observed differences may ex-plain, at least in part, the spectrum of phenotypes observed,whereby different host effector molecules are being targeted by arange of activities encoded by the variety of Tc subunit combina-tions.

Interestingly, the destruction of the gut epithelium caused bythe Tc toxins described above is in marked contrast to the absenceof any observed histopathological effects on C. zealandica larvaeinoculated with S. entomophila strains carrying the Sep Tc toxin(the causative agent of amber disease) (21) or semipurified Septoxin (presented here in the supplemental material). While we donot yet understand the exact significance, it is interesting that theamber coloration and gut clearance phenotypes observed after theearly stages of Yen-Tc intoxication share similarity to the observedphenology after the ingestion of S. entomophila (strain AM1O2 isthe causative agent of amber disease). In amber disease, the gutcolor turns from dark brown-black to amber, and this change isaccompanied by a dramatic clearance of gut contents through thedischarge of frass pellets; however, the gut epithelium remainsintact (20, 21). This contrasts with observations of Yen-Tc,whereby the complete dissolution of the gut is preceded by an“amber-like” physical appearance of the gut (prior to the forma-tion of a moribund brown coloration) and is accompanied byvomiting and the discharge of frass material. That the Sep complexhas only ever shown activity against C. zealandica points to a spe-cific coevolution, and this “mild” effect could possibly provide anas-yet-unidentified ecological advantage to Serratia species carry-ing the Sep Tc genes. This concept has some support from theobservation that the Tc activity associated with Yersinia pestis hasbeen tailored so that it is highly active against mammalian cells butis not active against its insect host (9). Within the C. zealandicasystem, a reduced dose of either S. entomophila Sep or Yen-Tcresults in a reversion-like effect, whereby intoxicated C. zealandicalarvae that began to display disease symptoms would then revertto a healthy phenotype (17, 19). These observations suggest that acontinuous dose of Tc is likely to be required throughout thedisease process.

The apparent sloughing of vesicle-like structures and the pu-tative condensation of DNA material (pyknotic nuclei) are remi-niscent of events often described for cells undergoing apoptosis(24). The presence of TUNEL-positive cells and the activation ofcaspase-3/7 enzyme activity in Yen-Tc-treated larvae provide ev-idence that apoptosis-like programmed cell death is activated.Whether this is a direct or indirect effect of the toxin has not yetbeen determined. The apparent Tc-induced induction of an apop-tosis-like event has also been described for other Tc family mem-bers (e.g., P. luminescens and other Yersinia species), and the re-sults for Yen-Tc presented here are consistent with this view (15,34). However, a closer inspection of the data also reveals an ap-parent shrinkage of columnar cells within the deteriorating gutepithelium and the presence of large amounts of cell debris (pu-

T) Cells transfected with pRK5myc-B and pEGFP-actin. pEGFP-actin in RK5myc-B shows vacuoles (Q and T) (yellow arrows). Here the RK5myc-B tag showslocalization to the nucleus (R and T) (red arrows) and causes nuclear blebbing (S and T) (white arrow). (U to X) Cells transfected with pRK5myc-C1 andpEGFP-actin. As can be seen, pRK5myc-C1 shows actin blebs (U and X) (yellow arrows) and is localized to the cytoplasm (V and X) (red arrows). (Y to AB) Cellstransfected with pRK5myc-C2 and pEGFP-actin. As observed for pRK5myc-C1, pRK5myc-C2 shows actin blebs (Y and AB) (yellow arrows) and is localized tothe cytoplasm (Z and AB) (red arrows). (AC to AF) Control cells transfected with pEGFP alone. The control cells show normal actin and nuclear morphology andno cMyc colocalization. Scale bars represent 20 �m.

Histopathological and Cytological Effects of Yen-Tc Toxin

July 2012 Volume 78 Number 14 aem.asm.org 4845

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

tatively from cell lysis) within the gut lumen. Both observationshave been associated with programmed necrosis (7, 30). Given thecurrent uncertainty over how to specifically define and character-ize apoptotic and necrotic programmed cell death pathways (14,24), it is interesting to speculate that perhaps a combination ofboth pathways, or, indeed, a novel pathway, is triggered by Yen-Tcactivity. Although the current study was unable to definitivelydifferentiate between the activation of a single apoptotic ornecrotic programmed cell death pathway, the predictable se-quence of observed events after Yen-Tc intoxication coupled withthe above-described evidence supports the idea that Yen-Tc doesactivate a programmed cell death pathway. As the mode of actionof the majority of Tc family members remains largely unknown,an experimental reevaluation of other Tc members may shed lighton what programmed cell death mechanisms are involved, as itappears that the interconnectedness of the various death pathwaysis more common than previously believed.

The ectopic application of Yen-Tc onto Sf9 and Caco-2 cellsestablished that Yen-Tc was able to elicit a cytopathic effect (i.e.,vacuole formation, cell actin contraction, and the formation ofpyknotic nuclei), and these indicators are consistent with pheno-types observed previously for programmed cell death (8, 24). ThatYen-Tc was active against these cell lines allowed their use to fur-ther examine the contribution of each individual Yen-Tc subunitcomponent (A, B, and C subunits, plus the two associated chiti-nases) via transient expression.

Unexpectedly, the transient expression of the Chi componentsin Caco-2 cells provided observable effects. The intracellular ex-pression of either of the two Chi components was able to cause a1.5- to 2-fold increase in levels of actin reorganization, vacuoleformation, and multinucleation (compared with mock-treatedcells). Although reasons for such effects are not immediately ob-vious, it is known that the presence of both Chi subunits in theYen-Tc ABC complex is required for insecticidal activity (19),both Chi1 and Chi2 possess endochitinase activity (6, 25), and thepresence of chitinase genes in close proximity to Tc loci was notedpreviously for P. luminescens and Xenorhabdus nematophila Tcloci (although the were not reported to be associated with the Tccomplex) (6, 35). Interestingly, specific actin-binding activity hasbeen demonstrated for a 32-kDa chitinase found in STDR-1 po-tato cells (33), which leaves the potential for such a nonobviousfunction being present within the Yen-Tc-associated Chi sub-units.

The cellular effects caused by the individual Yen-Tc A, B, and Csubunits are broadly consistent with observations previously re-ported for other members of the Tc family, such as other Yersiniaspecies and P. luminescens (27). The transient expressions of thethree main Yen-Tc subunit components (A, B, and C subunits)revealed that they all have a role in actin organization (blebbing,condensation, and/or vacuole formation). In addition, Yen-Tcsubunits A1 and C1 were both associated with nuclear condensa-tion, and the intracellular expression of subunits A1 and C2caused multinucleation in some cells, which is consistent with anapoptotic phenotype. Yen-Tc B subunit expression was shown tobe responsible for nuclear fragmentation and blebbing, which isconsistent with necrosis, with the protein being localized to thenucleus, as opposed to the cytoplasm for the Tc A and Tc C sub-units. From this, it would appear that the Yen-Tc A and C subunitshave a role initiating an apoptosis-like programmed cell death

pathway, while the effects of the Yen-Tc B subunit correlate morewith a necrosis-type programmed cell death event.

The finding that all three of the A, B, and C subunits and thetwo Chi proteins affect the actin cytoskeleton is both interestingand somewhat puzzling. The effects on actin of the topical appli-cation of native Yen-Tc toxin onto Sf9 and Caco-2 cells demon-strate a role in cytoskeleton remodeling (as previously shown forother Tc members); however, the specific mechanisms responsi-ble for this remain to be determined. The discovery that P. lumi-nescens Tc-mediated actin clustering was a direct result of theADP-ribosylation activity contained within two different Photo-rhabdus Tc complexes provides some direction for future investi-gations into the mechanisms of Yen-Tc-mediated cytoskeletalchanges (26).

The C3 and C5 subunits of the P. luminescens Tc complexeswere shown previously to possess ADP-ribosylation activities,which directly modify ADP-ribosylate residue T148 of actin andresidues Q-61 and Q-63 of the RhoA GTPase, respectively (26,27). This activity was predicted from sequence comparisons withknown protein domains, but the demonstration of activity hasbeen a breakthrough in the identification of specific moleculartargets for Tc toxins. The Yen-Tc C1 and C2 subunits, by sequencecomparison, are predicted to have a Rho-activating function sim-ilar to that of cytotoxic necrotizing factor and deaminase activity,respectively (19), while the Y. pseudotuberculosis Tc C subunit(from the Tcc complex) displays similarity to tyrosine phospha-tase (15). Assuming that these predicted activities are present, thisdiversity of activity is suggestive of the Tc family members beingcomposed of subunits encoding domains that are able to targetdifferent sites along putative programmed cell death pathwaysand other as-yet-unknown pathways.

ACKNOWLEDGMENTS

This research has been funded by the Ministry of Science and Innovation,New Zealand (contract number C10X0804).

We thank Richard Townsend (AgResearch) for field collection of C.zealandica larvae, Richard Walls (AgResearch) for assistance with prepar-ing samples for electron microscopy studies, Chikako van Koten(AgResearch) for statistical assistance, Jennifer Lucas at Gribbles Veteri-nary Pathology (Christchurch, New Zealand) for providing histology ser-vices, and Richard ffrench-Constant (University of Exeter) for use of cellculture facilities.

REFERENCES1. Allan VJ. 2000. Basic immunofluorescence, p 1–26. In Allan VJ (ed),

Protein localization by fluorescence microscopy: a practical approach.Oxford University Press, Oxford, United Kingdom.

2. Blackburn M, Golubeva E, Bowen D, ffrench-Constant RH. 1998. Anovel insecticidal toxin from Photorhabdus luminescens, toxin complex a(Tca), and its histopathological effects on the midgut of Manduca sexta.Appl. Environ. Microbiol. 64:3036 –3041.

3. Blackburn MB, Domek JM, Gelman DB, Hu JS. 2005. The broadlyinsecticidal Photorhabdus luminescens toxin complex a (Tca): activityagainst the Colorado potato beetle, Leptinotarsa decemlineata, and sweetpotato whitefly, Bemisia tabaci. J. Insect Sci. 5:32.

4. Blackburn MB, Martin PAW, Kuhar D, Farrar RR, Jr, Gundersen-Rindal DE. 2011. The occurrence of Photorhabdus-like toxin complexes inBacillus thuringiensis. PLoS One 6:e18122. doi:10.1371/journal-.pone.0018122.

5. Bowen D, et al. 1998. Insecticidal toxins from the bacterium Photorhab-dus luminescens. Science 280:2129 –2132.

6. Busby JN, et al. 2012. Structural analysis of Chi1 chitinase from Yen-Tc:the multisubunit insecticidal ABC toxin complex of Yersinia ento-mophaga. J. Mol. Biol. 415:359 –371.

Marshall et al.

4846 aem.asm.org Applied and Environmental Microbiology

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

7. Christofferson DE, Yuan J. 2010. Necroptosis as an alternative form ofprogrammed cell death. Curr. Opin. Cell Biol. 22:263–268.

8. Coleman ML, Olson MF. 2002. Rho GTPase signalling pathways in themorphological changes associated with apoptosis. Cell Death Differ.9:493–504.

9. Erickson DL, et al. 2007. Acute oral toxicity of Yersinia pseudotuberculosisto fleas: implications for the evolution of vector-borne transmission ofplague. Cell. Microbiol. 9:2658 –2666.

10. ffrench-Constant R, Waterfield N. 2005. An ABC guide to the bacterialtoxin complexes, p 169 –183. In Laskin AI, Bennett JW, Gadd GM, Sari-aslani S (ed), Advances in applied microbiology, vol 58. Elsevier AcademicPress Inc, San Diego, CA.

11. ffrench-Constant RH, Dowling A, Waterfield NR. 2007. Insecticidaltoxins from Photorhabdus bacteria and their potential use in agriculture.Toxicon 49:436 – 451.

12. Fogh J, Wright WC, Loveless JD. 1977. Absence of HeLa cell contami-nation in 169 cell lines derived from human tumors. J. Natl. Cancer Inst.58:209 –214.

13. Fuchs TM, Bresolin G, Marcinowski L, Schachtner J, Scherer S. 2008.Insecticidal genes of Yersinia spp.: taxonomical distribution, contributionto toxicity towards Manduca sexta and Galleria mellonella, and evolution.BMC Microbiol. 8:214. doi:10.1186/1471-2180-8-214.

14. Galluzzi L, et al. 2012. Molecular definitions of cell death subroutines:recommendations of the Nomenclature Committee on Cell Death 2012.Cell Death Differ. 19:107–120.

15. Hares MC, et al. 2008. The Yersinia pseudotuberculosis and Yersinia pestistoxin complex is active against cultured mammalian cells. Microbiology154:3503–3517.

16. Hurst MR, Becher SA, Young SD, Nelson TL, Glare TR. 2011. Yersiniaentomophaga sp. nov. isolated from the New Zealand grass grub Costelytrazealandica. Int. J. Syst. Evol. Microbiol. 61:844 – 849.

17. Hurst MR, Jones SM, Tan B, Jackson TA. 2007. Induced expression ofthe Serratia entomophila Sep proteins shows activity towards the larvae ofthe New Zealand grass grub Costelytra zealandica. FEMS Microbiol. Lett.275:160 –167.

18. Hurst MRH, Glare TR, Jackson TA, Ronson CW. 2000. Plasmid-locatedpathogenicity determinants of Serratia entomophila, the causal agent ofamber disease of grass grub, show similarity to the insecticidal toxins ofPhotorhabdus luminescens. J. Bacteriol. 182:5127–5138.

19. Hurst MRH, et al. 2011. The main virulence determinant of Yersiniaentomophaga MH96 is a broad-host-range toxin complex active againstinsects. J. Bacteriol. 193:1966 –1980.

20. Jackson TA, Boucias DG, Thaler JO. 2001. Pathobiology of amber dis-

ease, caused by Serratia spp., in the New Zealand grass grub, Costelytrazealandica. J. Invertebr. Pathol. 78:232–243.

21. Jackson TA, Huger AM, Glare TR. 1993. Pathology of amber disease inthe New Zealand grass grub Costelytra zealandica (Coleoptera: Scarabaei-dae). J. Invertebr. Pathol. 61:123–130.

22. Kaplan EL, Meier P. 1958. Nonparametric estimation from incompleteobservations. J. Am. Stat. Assoc. 53:457– 481.

23. Kiernan JA. 1990. Histological and histochemical methods: theory andpractice, 2nd ed. Pergamon Press, Toronto, Ontario, Canada.

24. Kroemer G, et al. 2009. Classification of cell death: recommendations ofthe Nomenclature Committee on Cell Death 2009. Cell Death Differ. 16:3–11.

25. Landsberg MJ, et al. 2011. 3D structure of the Yersinia entomophaga toxincomplex and implications for insecticidal activity. Proc. Natl. Acad. Sci.U. S. A. 108:20544 –20549.

26. Lang AE, et al. 2010. Photorhabdus luminescens toxins ADP-ribosylateactin and RhoA to force actin clustering. Science 327:1139 –1142.

27. Lang AE, Schmidt G, Sheets JJ, Aktories K. 2011. Targeting of the actincytoskeleton by insecticidal toxins from Photorhabdus luminescens. Nau-nyn Schmiedebergs Arch. Pharmacol. 383:227–235.

28. Lee SC, et al. 2007. Structural characterisation of the insecticidal toxinXptA1, reveals a 1.15 MDa tetramer with a cage-like structure. J. Mol. Biol.366:1558 –1568.

29. Martinez MM, Reif RD, Pappas D. 2010. Detection of apoptosis: a reviewof conventional and novel techniques. Anal. Methods 2:996 –1004.

30. McCall K. 2010. Genetic control of necrosis—another type of pro-grammed cell death. Curr. Opin. Cell Biol. 22:882– 888.

31. Sergeant M, et al. 2006. Identification, typing, and insecticidal activity ofXenorhabdus isolates from entomopathogenic nematodes in United King-dom soil and characterization of the xpt toxin loci. Appl. Environ. Micro-biol. 72:5895–5907.

32. Sergeant M, Jarrett P, Ousley M, Morgan JAW. 2003. Interactions ofinsecticidal toxin gene products from Xenorhabdus nematophilusPMFI296. Appl. Environ. Microbiol. 69:3344 –3349.

33. Takemoto D, Furuse K, Doke N, Kawakita K. 1997. Identification ofchitinase and osmotin-like protein as actin-binding proteins in suspen-sion-cultured potato cells. Plant Cell Physiol. 38:441– 448.

34. Waterfield N, Hares M, Yang G, Dowling A, ffrench-Constant R. 2005.Potentiation and cellular phenotypes of the insecticidal toxin complexesof Photorhabdus bacteria. Cell. Microbiol. 7:373–382.

35. Waterfield NR, Bowen DJ, Fetherston JD, Perry RD, ffrench-ConstantRH. 2001. The tc genes of Photorhabdus: a growing family. Trends Micro-biol. 9:185–191.

Histopathological and Cytological Effects of Yen-Tc Toxin

July 2012 Volume 78 Number 14 aem.asm.org 4847

on April 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from