Attachment and germination of Entomophaga maimaiga conidiaon host and non-host larval cuticle

Ann E. Hajek* and Callie C. Eastburn

Department of Entomology, Cornell University, Comstock Hall, Ithaca, NY 14853-0901, USA

Received 15 January 2002; accepted 20 November 2002

Abstract

The lepidopteran-specific fungal pathogen Entomophaga maimaiga is highly virulent against Lymantria dispar (gypsy moth)

larvae, and other members of the family Lymantriidae. Numerous species in the subfamily Cuculliinae (Family Noctuidae) are not

susceptible to E. maimaiga due to the inability of this fungus to penetrate the larval cuticle. Conidial attachment and germination

were compared among five cuculliine species and L. dispar using bioassays and scanning electron microscopy. Although conidia

were showered evenly across larvae during bioassays, on L. dispar conidia were most abundant on segments, where they adhered

well to the cuticle and germinated at high percentages. Conidia on cuculliine cuticles were predominantly found in large, loose

aggregations in intersegmental areas. Few conidia on cuculliine cuticle germinated and scanning electron microscopy revealed a

thick film of mucous enveloping conidia. We hypothesize that the conidia on cuculliines become coated by this film and were only

loosely attached to the larval cuticle. No such film was seen on L. dispar larvae where individual conidia appeared well attached. On

L. dispar larvae many conidia also adhered to setae. To determine if hydrophobicity affected the ability of E. maimaiga conidia to

attach and germinate on a substrate, a goniometer was used to determine relative hydrophobicity of larval cuticles. L. dispar cuticle

was more hydrophobic than cuculliine cuticle, suggesting that a high level of hydrophobicity could be a required characteristic for

hosts. Cuticles from four cuculliine species and L. dispar were sequentially extracted using hexane, chloroform, and methanol.

Conidia were showered onto glass slides coated with the different extracts and germination was quantified. Methanol extracts of

cuculliine cuticle consistently decreased germination, compared to all extracts of L. dispar cuticle. For all L. dispar extracts, the

majority of conidia produced germ tubes, which is a normal prerequisite for cuticular penetration. For the cuculliines, conidia

exposed to hexane and chloroform extracts produced secondary conidia as did all controls, but the conidia exposed to cuculliine

methanol extracts that germinated produced germ tubes. These studies demonstrated that a range of factors act in concert to prevent

E. maimaiga infection of the cuculliine species investigated.

� 2003 Elsevier Science (USA). All rights reserved.

Keywords: Entomopathogenic fungi; Entomophthorales; Germination; Adhesion; Cuticle hydrophobicity; Non-target effects; Biological control

1. Introduction

Insect cuticle forms the first formidable barrier to

pathogens. There are a number of discrete stages leading

to breaching the cuticle, including spore attachment,

spore germination, spore differentiation (with or without

formation of appressoria) and then penetration of the

cuticle by enzymatic and mechanical means (St. Leger,

1991). A diversity of factors have been identified as in-fluencing the activity of conidia at the cuticular level.

Conidia of many species of entomopathogenic fungi arethought to initially attach nonspecifically (Boucias and

Pendland, 1991). The hydrophilic conidia of Entom-

ophthorales are covered by a mucilaginous coat that is

released upon cuticular attachment and is thought to act

as a glue (see Boucias and Pendland, 1991). Surface

hydrophobicity has been associated with adhesion of

plant pathogenic fungi (Kuo and Hoch, 1996; Terhune

and Hoch, 1993). Conidial distribution on host cuticlecan be region specific (e.g., Sosa-Gomez et al., 1997) and

surface topography has been shown to influence growth

of deuteromycetes after adhesion (Boucias and Pend-

land, 1991). Chemical components of the cuticle can also

affect conidial development after adhesion by either

Journal of Invertebrate Pathology 82 (2003) 12–22

Journal ofINVERTEBRATE

PATHOLOGY

www.elsevier.com/locate/yjipa

*Corresponding author. Fax: +1-607-255-0939.

E-mail address: aeh4@cornell.edu (A.E. Hajek).

0022-2011/03/$ - see front matter � 2003 Elsevier Science (USA). All rights reserved.

doi:10.1016/S0022-2011(02)00198-2

causing production of nonpenetrant germ tubes or byinhibiting germination altogether (Boucias and Pend-

land, 1991).

The host range of the entomophthoralean pathogen

Entomophaga maimaiga (Zygomycetes: Entomophtho-

rales) is limited. During laboratory bioassays evaluating

larvae of a diversity of forest-dwelling Lepidoptera,

35.6% of 78 species assayed became infected but only

8.2% of susceptible species were infected at levels > 50%(Hajek et al., 1995). Consistently high levels of infection

were only observed among members of the Lymantrii-

dae, the tussock moth family that includes the principal

host of E. maimaiga, Lymantria dispar (gypsy moth)

(Hajek et al., 1995).

Many species of owlet moths (Noctuidae) inhabit the

same or similar forest habitats as L. dispar. One of the

common subfamilies in the Noctuidae is the Cuculliinae,whose larvae have been referred to as arboreal cut-

worms since they feed off the ground and have few setae.

While earlier instars of cuculliines feed in trees, late in-

stars can also be found on the forest floor. High levels of

infection can occur when late instar L. dispar rest in the

leaf litter (Hajek et al., 2000) or when larvae are caged

on top of the soil (Hajek, 2001). However, only one of

47 larvae of the cuculliine Sunira bicolorago collected inthe leaf litter (Hajek et al., 2000) was infected by E.

maimaiga. Curiously, studies comparing infection by

injection of fungal cells with infection requiring cuticu-

lar penetration had demonstrated that cuculliines were

unique among the many species of Lepidoptera tested

(Hajek et al., 1995). When E. maimaiga cells were in-

jected into the hemocoels of six species of cuculliine

larvae, five species became infected although none ofthese were infected when they were challenged by de-

positing E. maimaiga conidia onto their cuticles.

Based on the results of these bioassays and field

studies, we hypothesized that successful infection of

cuculliine larvae by E. maimaiga is prevented at the

cuticle. We compared attachment and germination of

E. maimaiga conidia on the cuticles of L. dispar and

five cuculliine species. Cuticle hydrophobitities and theeffects of cuticular extracts on conidial germination

were also compared among these species.

2. Materials and methods

2.1. Experimental insects

Lymantria dispar larvae used in this study were ob-

tained as neonates from the USDA, APHIS, Otis

Methods Development Center, Massachusetts. They

were reared on high wheat germ diet (Bell et al., 1981) in

236ml plastic cups at 23 �C, 14:10 (L:D).

Noctuids in the subfamily Cuculliinae, Chaetaglaea

sericea, Eupsilia vinulenta, Sericaglaea signata, S. bi-

colorago, and Xylotype capax, were collected in NewJersey. Adult females of species that overwinter as adults

(S. signata and E. vinulenta) were collected in February

or March, when they mate and the other cuculliine

species were collected in late October or November.

Adult females were collected using sugar baits and if

they had not yet mated, they were confined in outdoor

cages with males. Once mated, females were placed

singly in 200ml glass jars. White paper towels wereprovided for oviposition and honey or maple syrup,

diluted 1:2 with water, was provided for food. Eggs

produced were placed in cups outdoors until hatch or

were placed directly into sleeves slightly before hatching.

Larvae from cups were reared on cuttings until late first

or second instar when they were placed in sleeves on

branches of trees or shrubs. Larvae of C. sericea, E.

vinulenta, S. signata, and X. capax were reared on blackcherry (Prunus serotina) and larvae of S. bicolorago were

fed apple (Pyrus malus). Larvae were removed from

sleeves at the penultimate instar and were subsequently

reared in humid, ventilated plastic containers (10.5 cm

diameter� 13 cm height) at 23 �C, 14:10 (L:D) where

they were provided with fresh host plant foliage until use

in experiments.

2.2. Fungal inoculum and infection methodology

The isolate of E. maimaiga used in this study was

ARSEF 6162 (USDA, ARSEF Agricultural Research

Service Collection of Entomopathogenic Fungal Cul-

tures, Ithaca, NY). Hyphal bodies and naturally oc-

curring protoplasts of E. maimaiga were grown in 95%

Grace�s insect tissue culture media (Gibco-BRL, Gai-thersburg, MD) supplemented with 5% fetal bovine se-

rum (Gibco-BRL) at 20 �C in the dark.

To obtain dense showers of conidia over a relatively

short time period, we infected L. dispar larvae and used

resulting cadavers to produce inoculum. Fourth instars

were injected with 10 ll of 1:0� 105 protoplasts/ml in

Grace�s insect tissue culture media and were then

maintained at 20 �C in the dark until death. Insects weremonitored daily from 4 to 6 days post-injection and then

transferred to 100% RH at 15 �C in the dark to initiate

fungal outgrowth and conidial discharge.

2.3. Larval infection

To deliver a consistent dose within each bioassay,

larvae were inoculated by showering conidia on them astheir containers rotated beneath sporulating cadavers

within a ‘‘showering box.’’ Each showering box con-

sisted of an 11.3 L polyethylene storage container

(Rubbermaid Incorporated, Wooster, OH) in which a

15 cm diameter platform was mounted. The platform

was attached to a 0.75 rpm gear motor powered by a

regulated 13.8V power supply. High humidity was

A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22 13

maintained within tightly sealed showering boxesby lining the bottom and sides with moistened paper

towels.

To inoculate larvae, 10–15 cadavers that were dis-

charging conidia were placed dorsal side down on net-

ting covering a platform of 1.2 cm mesh hardware cloth

(15� 15 cm). The hardware cloth was suspended 7–8 cm

above test larvae in the showering box. Conidia from

sporulating cadavers fell through the mesh onto testlarvae that were on the rotating platform directly below.

To prevent larvae from escaping during the conidial

shower, they were placed individually in 29.6ml plastic

cups coated internally with Sigmacote (Sigma Chemical,

St. Louis, MD). Seven larvae at a time could be posi-

tioned in cups beneath sporulating cadavers.

2.4. Conidial attachment and germination studies

Studies of attachment and germination of E.

maimaiga conidia to host versus non-host larval cuticle

used 3 fourth instar L. dispar larvae and 3 non-target

larvae between penultimate and ultimate instars for each

of three replicates for each cuculliine species. For each

replicate, conidia were showered onto test insects for a

4-h period. Insects were then placed singly in 29.6mlplastic cups containing moistened filter paper to create a

100% RH environment and leaves from a host plant for

cuculliines or artificial diet for L. dispar. Lids were

placed on cups and larvae remained in these conditions

for a 12-h germination period. Larvae were then trans-

ferred to new 29.6ml plastic cups containing approxi-

mately 3 g silica gel desiccant; this reduced subsequent

condensation on cadavers that could disrupt conidia onthe cuticle. Larvae with desiccant remained at room

temperature for 20min prior to being frozen at )80 �C.After thawing, conidia on dorsal surfaces of cadavers

were counted at 64� using a dissecting microscope. An

ocular grid (750� 750lm) was used to quantify conidial

density and production of germ tubes for ten grids per

larva. Conidia were not evenly distributed on larvae so

we evaluated conidial density and activity by cuticularregion. However, cuculliine and L. dispar larvae have

very different topography of their bodies so we could not

use the same sampling plan for both. L. dispar larvae

have abundant setae attached to segmental areas and,

on many segments, heavily sclerotized dorsal and lateral

elevations that bear setae, called verrucae. Cuculliine

larvae have few setae and no unusual cuticular topog-

raphy. For each cuculliine, we evaluated conidia in fourintersegmental regions and six areas on segments. For

each L. dispar larva, we evaluated conidia on four ver-

rucae, four regions on segments surrounding the verru-

cae and two intersegmental regions. For all larvae,

conidia on setae within the ocular grids were counted

separately, while counting conidia on segments or ver-

rucae.

2.5. Goniometer studies

Hydrophobicity of larval cuticles was evaluated by

measuring the contact angle that a 1mm diameter

droplet of distilled water made with each surface using

an NRC Contact Goniometer (Model 100-00, Ram�ee-Hart, Mountain Lakes, NJ) (Neumann and Good,

1979). The formula used to determine average contact

angle is: cos�1ððcos h1 þ cos h2Þ=2Þ; h1, the advancingangle and h2, the receding angle (Nishino et al., 1999).

Larval cuticles were prepared by cutting the head off of a

larva with a sharp razor and cutting a 1.5� 2.5 cm sec-

tion of cuticle from the dorsum. Larger internal struc-

tures, e.g., tracheae, were carefully removed and the

cuticle was flattened, cuticle-side up, on hardened glue

and secured at the corners with minuten pins. Removal

of the setae from the cuticle of L. dispar was requiredbefore a reading could be taken because the large sec-

ondary setae either prevented the water from touching

the cuticular surface or burst the water droplets so that

readings could not be obtained. For removal, the distal

ends of secondary setae were gently touched with sticky

tape; this removed secondary setae without disturbing

the cuticle surface. Samples were kept at 100% RH until

readings were conducted, generally within 1 h of prep-aration. Two to six cuticles were quantified for each

species, based on availability of larvae.

2.6. SEM preparation

Larvae were showered with E. maimaiga conidia as

described above, followed by a 12-h germination period.

Cuticles were removed as described for goniometer tests,with care being taken to avoid touching the dorsal sur-

face and to prevent internal fluids from washing over the

dorsal surface. Cuticles were stretched and held in place

using minutens inserted in hardened hot glue on top of a

2.54 cm diameter scanning electron microscopy stub,

sputter-coated with 30 nm gold/palladium and subse-

quently observed using a Hitachi 4500.

2.7. Cuticular extract studies

Late instar larvae of four species of cuculliines (E.

vinulenta, S. signata, S. bicolorago, and X. capax) and L.

dispar were starved for 24 h and then frozen at )80 �C.Due to the difficulty in obtaining large numbers of these

cuculliines, groups of twenty were considered represen-

tative for each species. Twenty larvae were thawed andthen each individual was placed in 15ml quantities of

hexane then chloroform, followed by methanol with 15-s

exposures in each solvent. Hexane extracted only

non-polar lipids, chloroform extracted lipids of greater

polarity and methanol extracted more polar organic

constituents. Extracts were concentrated to 1ml using a

rotary evaporator. Larvae of the different species were

14 A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22

similar in size and surface areas of larvae of each specieswere approximately estimated as the surface area of a

cylinder. We determined that by spreading 50 ll of anextract on an 18� 18mm cover slip, we would be ap-

proximating the concentration found on the cuticle of

one insect. For each insect species, we transferred 50 llof each extract onto a cover slip, with individual sol-

vents or solely glass slides used as controls. After sol-

vents had dried, cover slips were placed under conidialshowers as described above for 2–3 h to achieve a den-

sity of >10 conidia/mm2, because previous studies

demonstrated that conidial density can affect germina-

tion (Hajek et al., 2002). Cover slips were then inverted

and incubated under humid conditions at 22� 1 �C for

22–24 h. Inversion of cover slips prevented any second-

ary conidia produced from landing on the cover slip.

Cover slips were then placed over drops of water onmicroscope slides. Conidial activity was recorded for 10

conidia in each of 10 fields of view at 400�, using phase

contrast on a compound microscope. Conidial activity

was scored as producing a germ tube or secondary co-

nidium or not germinating. Each type of solvent plus the

four controls was tested three times for each species

(three test solvents plus four controls� three replicates).

2.8. Data analysis

Conidial density and percent germination on L. dis-

par larvae were compared by cuticular region using

analysis of variance followed by post hoc comparisons

using the Bonferroni inequality to partition the overall aof 0.05. Data for conidial density and percent germi-

nation on cuculliines were not normally distributed. Tocompare conidial densities by intersegmental versus

segmental regions across cuculliine species, data were

transformed using logðxþ 1Þ and analyzed using a hi-

erarchical linear model with post hoc Tukey tests with

bioassay as a random effect. Comparisons of conid-ial densities on L. dispar larvae versus cuculliines by

cuticular region were made using non-parametric

Mann–Whitney tests. Percent conidial germination was

compared by region among L. dispar and cuculli-

ine species, or across regions for S. bicolorago, using

Mann–Whitney tests.

v2 tests were used to compare results from extract

studies. For each type of extract, germination andgrowth form were compared among species, with a

partitioned overall a of 0.05.

3. Results

Densities of conidia by body region varied dramati-

cally after inoculated larvae had been allowed to walk inexperimental cups for 12 h (Table 1). On L. dispar lar-

vae, E. maimaiga conidia were always more abundant

on the segmental cuticle and setae, with progressively

lower densities on the verrucae and the lowest densities

were found on intersegmental regions.

For four of the five cuculliine species tested, the dis-

tribution of conidia was reversed; the highest conidial

densities occurred in the intersegmental areas with sig-nificantly lower densities on the segmental cuticle (Table

1). The remaining cuculliine species, S. bicolorago, did

not follow this trend and conidial densities on inter-

segmental and segmental cuticle were the same. Cucul-

liines had no setae (S. signata and X. capax) or if setae

were present, they were relatively smaller than many of

the setae on L. dispar and were found in 6 37% of the

fields of view. Densities of conidia on setae of cuculliineswere extremely low (Table 1).

Lymantria dispar larvae always had more conidia on

setae than cuculliines (Table 1), at least partly because

they have many more setae. The hardened, raised ver-

Table 1

Mean densities (�SE) of E. maimaiga conidia by regions on the dorsal surfaces of lepidopteran larvae

Lepidopteran family Larval species No. of larvaeb Cuticular regiona

Intersegmental

(no./mm2)

Segmental

(no./mm2)

On setae

(no./mm2)

On verrucaed

(no./mm2)

Lymantriidae Lymantria dispar 44 0.9� 0.2 c 13.5� 0.7 a 11.7� 0.7 a 7:7� 0:4 b

Noctuidae, Cuculliinae Chaetaglaea sericea 9 12.5� 3.1 a* 2.6� 1.8 b* 0.3� 0.2* –

Eupsilia vinulenta 9 20.0� 2.7 a* 2.0� 0.3 b* 0.1� 0.0* –

Sericaglaea signata 9 12.2� 2.5 a* 1.0� 0.1 b* c –

Sunira bicolorago 9 11.5� 2.7 a* 8.8� 1.5 a 1.0� 0.4* –

Xylotype capax 8 17.8� 2.7 a* 2.8� 1.0 b* c –

a Letters following densities denote significant differences in conidial densities between cuticular regions for the species in that row (p < 0:05).

Asterisks following letters indicate that the conidial density for that region on that species differed from the density in the same region on L. dispar

larvae during the same bioassay (p < 0:05).b Total data for L. dispar are summarized in this table although only data from L. dispar assayed simultaneously with cuculliines were used for

separate tests.c Larvae of these species have no setae.dCuculliines do not have verrucae.

A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22 15

rucae bearing setae on the segmental areas of L. disparalso had high conidial densities, but cuculliines do not

have these structures. Conidial densities on segmental

cuticle of L. dispar were greater than densities in this

region on cuculliines for four of the five species

(p < 0:05). Densities were equivalent comparing conidia

on segmental cuticle of L. dispar and S. bicolorago. The

relative densities were reversed between L. dispar and

the cuculliines for the intersegmental cuticle; for allcuculliines, intersegmental conidial densities were much

higher than for L. dispar.

None of the conidia germinated on the cuticles or

setae of E. vinulenta, S. signata, and X. capax (Table 2).

A few conidia germinated on S. bicolorago cuticle in

different regions and in the segmental region of C.

sericea (e.g., 2/369). For S. bicolorago, percent germi-

nation was greatest in the segmental region and on setae.However, due to the low numbers of conidia, this meant

that only 2 of the 38 conidia on setae germinated. On L.

dispar, where conidial densities were higher in the seg-

mental region and on the setae, highest percent germi-

nation occurred on the verrucae and in the segmental

region and germination on setae was lower. The lowest

germination was seen in the intersegmental region, as

with cuculliines. In fact, overall the percent of conidiathat germinated never approached 50%.

Scanning electron micrographs showed that cuculli-

ine cuticles were smooth, with a moist, mucous-like film

covering the entire cuticle. E. maimaiga conidia on

cuculliine cuticles were enveloped by this film (Fig. 1). L.

dispar cuticle appeared much more varied and did not

appear smooth, with many very small projections that

appeared to be covered by exudate. When drying cuc-ulliine cuticle for SEM, the film cracked in places (Fig.

1A), revealing underlying cuticle that was similar in

appearance to L. dispar cuticle (Fig. 2). On cuculliine

cuticles, conidia in intersegmental areas adhered to each

other, forming large aggregates and then many of the

conidia were not directly in contact with the cuticlesurface (Fig. 3). Conidia on the surface of L. dispar were

in direct contact with the cuticle, not covered by any

films over the host surface, and conidial aggregations

were not seen. On L. dispar, conidia on verrucae or in

segmental areas produced both short and long germ

tubes (Fig. 4). Conidia on setae that germinated were

often seen attached to germ tubes extending toward the

cuticle.The cuticle of L. dispar is much more hydrophobic

than cuculliine cuticle with contact angles ranging from

92.6� to 118.7� (Table 3). The difference in hydropho-

bicity among the species tested agreed with both the

attachment and germination results; conidia attached

and germinated on the more hydrophobic cuticle but

were dislodged and rarely germinated on the more hy-

drophilic cuticle.

3.1. Extract studies

Regardless of the extract used, P95% of the conidia

on cover slips with L. dispar extracts germinated (Table

4). Conidia on cover slips with hexane and chloroform

extracts from cuculliines also consistently germinated.

Fewer conidia exposed to methanol extracts from thefour cuculliine species germinated relative to conidia

exposed to methanol extracts from L. dispar (v2 tests;

p < 0:005).Greater than 90% of conidia on cover slips without

any extract or with solvents alone produced secondary

conidia. Most conidia that germinated on cover slips

with extracts from L. dispar developed germ tubes

(Table 4). Conidia exposed to hexane and chloroformextracts from cuculliines produced secondary conidia, as

did controls. Many of the germinating conidia exposed

to methanol extracts from cuculliines produced germ

tubes, although always fewer than conidia exposed to

L. dispar methanol extracts (v2 tests; p < 0:005).

Table 2

Percent germination (�SE) of E. maimaiga conidia on different regions of the dorsal surfaces of lepidopteran larvae

Lepidopteran family Larval species No. of larvaeb Cuticular regiona

Intersegmental (%) Segmental (%) On setae (%) On verrucaec (%)

Lymantriidae Lymantria dispar 44 2:3� 2:3 b 13.0� 1.4 a 12.1� 1.3 a 15:9� 1:6 a

Noctuidae, Cuculliinae Chaetaglaea sericea 9 0.0 0:7� 0:7* 0.0 —

Eupsilia vinulenta 9 0.0 0.0 0.0 —

Sericaglaea signata 9 0.0 0.0 d —

Sunira bicolorago 9 0:2� 0:2 b 5.5� 2.2 a* 2.2� 2.2 ab* —

Xylotype capax 8 0.0 0.0 d —

aLetters following densities denote significant differences among cuticular regions for that species (p < 0:05). Asterisks following letters indicate

that percent germination in that region for that species differed from percent germination in the same region on L. dispar larvae during the same

bioassay.b Total data for L. dispar are summarized in this table although only data from L. dispar assayed simultaneously with cuculliines were used for

separate tests.c Cuculliines do not have verrucae.d Larvae of these species have no setae.

16 A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22

4. Discussion

We conducted conidial showers so that both host and

non-host larvae would receive the same concentrations

of conidia over their entire bodies. However, we did not

find equivalent densities on the bodies of hosts (L. dis-par) and non-hosts (cuculliines) 12 h later. For the five

cuculliine species, we consistently found the majority of

Fig. 1. Dorsal cuticle of X. capax larva with conidia of E. maimaiga. (A) Film covering a conidium plus underlying cuticle seen through cracks in the

film caused during processing for SEM. Bar, 20lm. (B) Group of film-covered conidia on cuticle. Bar, 50 lm.

A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22 17

conidia in aggregations in intersegmental membranes.

Likewise, conidia of hyphomycetes have been found in

aggregations in intersegmental folds in beetle larvae

(Fernandez et al., 2001; Vey et al., 1982), grasshoppers

(Inglis et al., 1995) and thrips (Schreiter et al., 1994). Incontrast, Wraight et al. (1990) found no difference in

densities of conidia of the entomophthoralean Z. radi-

cans across regions of leafhopper bodies. SEM showed

that E. maimaiga conidia in the intersegmental areas on

cuculliines were enveloped by a film that covered the

cuculliine cuticle and conidia were found in large ag-

gregations in intersegmental areas adhering to each

other. We hypothesize that the accordion-like move-ment when larvae walked moved the poorly attached

conidia to the intersegmental areas where they became

aggregated. Conidia of other entomopathogens are

known to be moved after initial deposition on cuticles

but generally they are reported as being lost. Conidia of

Paecilomyces fumosoroseus were dislodged from ventral

surfaces when larvae of the diamondback moth, Plutella

xylostella walked (Altre et al., 1999). Many conidia ofVerticillium lecanii were dislodged from dorsal surfaces

after 24 h on an ant, a springtail and two carabid species,

but fewer conidia were dislodged from aphids, the typ-

ical hosts (Sitch and Jackson, 1997). A high percentage

of conidia of Conidiobolus obscurus were lost from the

cuticles of host aphids within 2 h of application (Latg�eeet al., 1982). In the extreme, conidia of the entomoph-

thoralean Z. radicans did not adhere when applied tonon-host leafhoppers (McGuire, 1985).

In contrast to the locations of conidia on the cuculli-

ines, conidia on L. dispar larvae were found individually,

often on segmental areas, setae and verrucae and these

conidia did not appear to have been moved during larval

walking. In agreement, studies with entomopathogenicHyphomycetes also found higher conidial densities in

areas of the cuticle with spines (Boucias et al., 1988) or

setae (Pekrul and Grula, 1979; Sosa-Gomez et al., 1997).

It has been suggested that conidia are trapped in these

areas, although with L. dispar it also seems likely that

areas with setae do not move as much to dislodge conidia

when larvae walk. On L. dispar, conidia were also fre-

quently found on the abundant setae but this did notprevent conidia from germinating and growing as germ

tubes toward the cuticle. Based on previous observations,

we believe that the low number of conidia in interseg-

mental regions is possibly due to rupture of these well-

attached conidia when larvae walked and intersegmental

regions folded inwards (Hajek et al., 2002). Better adhe-

sion of conidia to L. dispar cuticle that is more hydro-

phobic than cuculliine cuticle agrees with findings ofincreased adhesion of plant pathogenic fungi associated

with higher surface hydrophobicity (Kuo and Hoch,

1996;Terhune andHoch, 1993).Our results are consistent

with results from previous studies finding only low per-

centages of entomophthoralean conidia that germinate

produce germ tubes for subsequent penetration (Brey

et al., 1986; Nadeau et al., 1996; Wraight et al., 1990);

evenunder ideal circumstances, themajority of conidia onL. dispar larval cuticle did not germinate.

Fig. 2. Dorsal cuticle of L. dispar larva with conidium of E. maimaiga. Bar, 20 lm.

18 A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22

We hypothesize that the mucous film on the cuculli-

ines acts as an external protection for these larvae thathave virtually no setation and live part of their lives at

the soil surface. This film acts as a physical barrier,

preventing adhesion of conidia to the underlying larval

cuticle and it deserves further study. The film-engulfedconidia on cuculliines rarely germinated. These results

would agree with suggestions that germination is greater

Fig. 3. Large aggregations of E. maimaiga conidia on the dorsal cuticle of cuculliine noctuids. (A) X. capax. Bar, 230lm. (B) S. bicolorago. Bar,

165lm. For both species, the film covering the larval surface became dry and cracked when processed for SEM.

A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22 19

with increased cuticular-conidial contact (St. Leger et al.,

1991). In contrast, the many setae of L. dispar that arethought to have evolved for protection seem also to

increase the surface area for attachment of E. maimaiga

conidia; to further evaluate whether setae protect

against or increase chance of infection, the numbers ofconidia on setae that actually are responsible for infec-

tions should be evaluated.

Fig. 4. Germination of E. maimaiga on L. dispar cuticle. (A) Penetration close to conidium. Bar, 20lm. (B) Growth of long germ tubes across the

cuticle surface. g, germ tube; c, conidium. Bar, 85lm.

20 A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22

Methanol extracts of the cuticle of cuculliine species

reduced germination, suggesting that a methanol-solu-

ble component of the cuculliine cuticle may be fungi-

static or fungitoxic. Compounds present on cuticles of

some insects are known to inhibit attachment, germi-

nation and appressorium formation, steps that are pre-

cursors to cuticular penetration (St. Leger, 1991). In this

study, all extracts from L. dispar induced germ tubeformation. Curiously, conidia exposed to cuculliine

methanol extracts that were able to germinate produced

germ tubes rather than secondary conidia, suggesting a

possible stimulatory effect. Studies with the entomoph-

thoralean Erynia variabilis demonstrated that exposure

to chloroform/methanol extracts of host fly cuticle

produced germ tubes and not secondary conidia (Ker-

win, 1984). Exposure of C. obscurus conidia to chloro-form/methanol extracts from hosts or non-hosts also

induced germ tube production (Boucias and Latg�ee,1988); these authors suggested that this stimulatory re-

sponse is probably relatively nonspecific. Our findings

suggest that there is some component in methanol ex-

tracts from cuculliine cuticle that induces germ tube

formation but some other chemical components are

present that inhibit germination.In the field one of the few species that has been col-

lected infected by E. maimaiga was one individual of S.

bicolorago, suggesting that sometimes this pathogen in-

fects this species in nature (Hajek et al., 2000). Whether

that infection was initiated by penetration directly

through the cuticle or through a wound is not known.Results from this study agree with field collections be-

cause this is the one cuculliine species studied that had

the highest chance of becoming infected based on bio-

assay results.

Knowledge of host specificity is necessary for identi-

fying effective control agents but also for reducing non-

target effects. Effects of biological control agents on

non-target organisms are presently under close scrutiny(Follett and Duan, 2000; Wajnberg et al., 2001). Al-

though not always available, effective agents with nar-

row host specificity are desired for classical biological

control. Because we cannot test all organisms that are

present in the environment, the ability to predict whe-

ther non-target species might be affected is critically

important. One of our goals in undertaking this study

was to investigate whether species within the samesubfamily had similar determinants of host specificity;

can we use host taxonomy to help predict susceptibility?

Studies with Entomophthorales have shown that infec-

tion can occur in different families within an insect order

but there is a trend for greater levels of infection in

specific groups. For E. muscae isolated from a calli-

phorid fly but maintained in the laboratory on Musci-

dae, the greatest infection was in one species of muscidtested (Steinkraus and Kramer, 1987). For Z. radicans

isolated from cicadellid leafhoppers, the greatest infec-

tion was seen in cicadellids, although only in 3 of the 4

species tested (McGuire et al., 1987). For E. maimaiga

the greatest infection was consistently found only for

species in the same host family from which the fungus

had been isolated, the Lymantriidae (Hajek et al., 1995).

In this study, we found that species within the noctuidsubfamily Cuculliinae consistently displayed similar at-

tributes that reduced risk of infection by E. maimaiga;

these results would suggest that perhaps lower level

taxonomic relationships, i.e., subfamily in this case,

could help in predicting patterns of host specificity.

However, further studies with other systems should be

conducted before relying on phylogenetic relatedness to

determine safety.

Table 3

Average contact angles for cuticles of lepidopteran larvae

Lepidopteran

family

Species Contact angle

(mean� SE)

Lymantriidae Lymantria dispar 170.8� 6.9

Noctuidae,

Cuculliinae

Chaetaglaea sericea 92.6� 3.0

Eupsilia vinulenta 118.7� 0.7

Sericaglaea signata 95.1� 5.1

Sunira bicolorago 111.6� 12.1

Xylotype capax 112.7� 2.4

Table 4

Percent germination and production of germ tubes (�SE) by E. maimaiga on extracts of lepidopteran larval cuticlesa

Species Germinationb (%) Germ tubesb ;c

Methanol Chloroform Hexane Methanol Chloroform Hexane

Lymantria dispar 96:0� 1:1 a 95:0� 2:9 a 96:0� 1:0 a 98:6� 1:4 a 72.8� 21.6 a 82.8� 12.8 a

Eupsilia vinulenta 50:0� 5:9 c 97:7� 0:3 a 92:5� 0:4 a 60:3� 13:4 d 0.7� 0.7 d 0.5� 0.4 d

Sericaglaea signata 62:5� 3:8 b 94:3� 1:2 a 94:3� 1:3 a 91:2� 1:9 bc 4.6� 0.7 c 16.9� 3.1 b

Sunira bicolorago 61:0� 5:4 bc 98:0� 0:6 a 97:0� 1:5 a 91:8� 5:6 b 29.2� 14.3 b 19.8� 6.3 b

Xylotype capax 59:0� 1:5 bc 94:7� 1:5 a 95:5� 2:0 a 83:8� 9:7 c 3.9� 2.3 cd 6.8� 1.1 c

aCover slips with individual extracts or no extract were used for controls and >90% of conidia on control cover slips always germinated and

produced secondary conidia.bDifferent letters within a column demonstrate significant differences between percentages (v2 tests each tested at a ¼ 0:005).c Percent production of germ tubes versus secondary conidia.

A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22 21

Acknowledgments

We thank C. Daugherty for her excellent eye, deft

touch, and patience with the scanning electron micro-

scope. Thanks to D. Schweitzer and T. Hupf for col-

lection and initial rearing of cuculliines. A. Savage

helped rear larger cuculliines and M. Bertoia, K. Poole,

K.-L. Tsai, and K. Zuniga helped rear L. dispar larvae

and infect larvae. J.L. Kerwin and J.A.A. Renwickhelped with suggestions for procedures during cuticular

chemistry studies and J.L. Kerwin provided helpful

comments on the manuscript. F.M. Vermeylen assisted

with the use of hierarchical linear models. Y. Ando,

S.-H. Kang, and X. Li helped with use of the goniometer

belonging to the Cornell University, Department of

Materials Science and Engineering. This study was

funded by USDA, NRICGP #96-04343.

References

Altre, J.A., Vandenberg, J.D., Cantone, F.A., 1999. Pathogenicity of

Paecilomyces fumosoroseus isolates to diamondback moth, Plutella

xylostella: correlation of spore size, germination speed, and

attachment to cuticle. J. Invertebr. Pathol. 73, 332–338.

Bell, R.A., Owens, C.D., Shapiro, M., Tardif, J.R., 1981. Development

of mass-rearing technology. In: Doane, C.C., McManus, M.L.

(Eds.), The Gypsy Moth: Research Toward Integrated Pest

Management. USDA. For. Serv. Tech. Bull. 1584, 599–633.

Boucias, D.G., Latg�ee, J.P., 1988. Nonspecific induction of germination

of Conidiobolus obscurus and Nomuraea rileyi with host and non-

host cuticle extracts. J. Invertebr. Pathol. 51, 168–171.

Boucias, D.G., Pendland, J.C., 1991. Attachment of mycopathogens to

cuticle, the initial event of mycoses in arthropod hosts. In: Cole,

G.T., Hoch, H.C. (Eds.), The Fungal Spore and Disease Initiation

in Plants and Animals. Plenum, New York, pp. 101–127.

Boucias, D.G., Pendland, J.C., Latg�ee, J.P., 1988. Nonspecific factors

involved in the attachment of entomopathogenic deuteromycetes to

host insect cuticle. Appl. Entomol. Microbiol. 54, 1795–1805.

Brey, P.T., Latg�ee, J.-P., Prevost, M.C., 1986. Integumental penetration

of the pea aphid , Acyrthosiphon pisum, by Conidiobolus obscurus

(Entomophthoraceae). J. Invertebr. Pathol. 48, 34–41.

Fernandez, S., Groden, E., Vandenberg, J.D., Furlong, M.J., 2001.

The effect of mode of exposure to Beauveria bassiana on conidia

acquisition and host mortality of Colorado potato beetle, Lepti-

notarsa decemlineata. J. Invertebr. Pathol. 77, 217–226.

Follett, P.A., Duan, J.J. (Eds.), 2000. Non-target Effects of Biological

Control. Kluwer Academic Publishers, Dordrecht, NL.

Hajek, A.E., 2001. Larval behavior in Lymantria dispar increases risk

of fungal infection. Oecologia 12, 285–291.

Hajek, A.E., Butler, L., Wheeler, M.M., 1995. Laboratory bioassays

testing the host range of the gypsy moth fungal pathogen

Entomophaga maimaiga. Biol. Contr. 5, 530–544.

Hajek, A.E., Butler, L., Liebherr, J.K., Wheeler, M.M., 2000. Risk of

infection by the fungal pathogen Entomophaga maimaiga among

Lepidoptera on the forest floor. Environ. Entomol. 29, 645–650.

Hajek, A.E., Eastburn, C.C., Davis, C.I., Vermeylen, F.M., 2002.

Deposition and germination of conidia of the entomopathogen

Entomophaga maimaiga infecting larvae of gypsy moth, Lymantria

dispar. J. Invertebr. Pathol. 79, 37–43.

Inglis, G.D., Feniuk, R.P., Goettel, M.S., Johnson, D.L., 1995.

Mortality of grasshoppers exposed to Beauveria bassiana during

oviposition and nymphal emergence. J. Invertebr. Pathol . 65, 139–

146.

Kerwin, J.L., 1984. Fatty acid regulation of the germination of Erynia

variabilis conidia on adults and puparia of the lesser housefly,

Fannia canicularis. Can. J. Microbiol. 30, 158–161.

Kuo, K., Hoch, H.C., 1996. Germination of Phyllosticta ampelicida

pycnidiospores: prerequisite of adhesion to the substratum and

the relationship of substratumwettability. Fungal Gen. Biol. 20, 18–

29.

Latg�ee, J.P., Papierok, B., Sampedro, L., 1982. Aggresivit�ee de Conid-

iobolus obscurus vis-�aa-vis du puceron du pois, Acyrthosiphon pisum

(Hem.: Aphididae).I. Comportement des conidies sur la cuticule

avant la p�een�eetration du tube germinatif dans l�insecte. Entomoph-

aga 27, 323–330.

McGuire, M.R., 1985. Erynia radicans: studies on its distribution,

pathogenicity, and host range in relation to potato leafhopper,

Empoasca fabae. Ph.D. Thesis, Univ. IL, Urbana.

McGuire, M.R., Maddox, J.V., Armbrust, E.J., 1987. Host range

studies of an Erynia radicans strain (Zygomycetes: Entomophthor-

aceae) isolated from Empoasca fabae (Homoptera: Cicadellidae). J.

Invertebr. Pathol. 50, 75–77.

Nadeau, M.P., Dunphy, G.B., Boisvert, J.L., 1996. Development of

Erynia conica (Zygomycetes: Entomophthorales) on the cuticle of

the adult black flies Simulium rostratum and Simulium decorum

(Diptera: Simuliidae). J. Invertebr. Pathol. 68, 50–58.

Neumann, A.W., Good, R.J., 1979. Techniques of measuring contact

angles. In: Good, R.J., Stromberg, R.A. (Eds.), Surface and

Colloid Science, vol. 11. Plenum Press, New York, pp. 39–91.

Nishino, T., Meguro, M., Nakamae, K., Matsushita, M., Ueda, Y.,

1999. The lowest surface free energy based on –CF3 alignment.

Langmuir 15, 4321–4323.

Pekrul, S., Grula, E.A., 1979. Mode of infection of the corn earworm

(Heliothis zea) by Beauveria bassiana as revealed by scanning

electron microscopy. J. Invertebr. Pathol. 34, 238–247.

St. Leger, R.J., 1991. Integument as a barrier to microbial infections.

In: Binnington, K., Retnakaran, A. (Eds.), Physiology of the Insect

Epidermis. CSIRO Publications, Melbourne, pp. 284–306.

St. Leger, R.J., Goettel, M., Roberts, D.W., Staples, R.C., 1991.

Prepenetration events during infection of host cuticle by Metarhiz-

ium anisopliae. J. Invertebr. Pathol. 58, 168–179.

Schreiter, G., Butt, T.M., Beckett, A., Vestergaard, S., Moritz, G.,

1994. Invasion and development of Verticillium lecanii in western

flower thrips, Frankliniella occidentalis. Mycol. Res. 98, 1025–

1034.

Sitch, J.C., Jackson, C.W., 1997. Pre-penetration events affecting host

specificity of Verticillium lecanii. Mycol. Res. 101, 535–541.

Sosa-Gomez, D.R., Boucias, D.G., Nation, J.L., 1997. Attachment of

Metarhizium anisopliae to the southern green stink bug Nezara

viridula cuticle and fungistatic effect of cuticular lipids and

aldehydes. J. Invertebr. Pathol. 69, 31–39.

Steinkraus, D.C., Kramer, J.P., 1987. Susceptibility of sixteen species

of Diptera to the fungal pathogen Entomophthora muscae (Zygo-

mycetes: Entomophthoraceae). Mycopathology 100, 55–63.

Terhune, B.T., Hoch, H.C., 1993. Substrate hydrophobicity and

adhesion of Uromyces urediospores and germlings. Exper. Mycol.

17, 241–252.

Vey, A., Fargues, J., Robert, P., 1982. Histological and ultrastructural

studies of factors determining the specificity of pathotypes of the

fungus Metarhizium anisopliae for scarabeid larvae. Entomophaga

27, 387–397.

Wajnberg, E., Scott, J.K., Quimby, P.C. (Eds.), 2001. Evaluating

Indirect Ecological Effects of Biological Control. CABI Publica-

tions, Wallingford, Oxon, UK.

Wraight, S.P., Butt, T.M., Galaini-Wraight, S., Allee, L.L., Roberts,

D.W., 1990. Germination and infection processes of the entom-

ophthoralean fungus, Erynia radicans on the potato leafhopper,

Empoasca fabae. J. Invertebr. Pathol. 56, 157–174.

22 A.E. Hajek, C.C. Eastburn / Journal of Invertebrate Pathology 82 (2003) 12–22