SHORT COMMUNICATION

Hyperactivity and tree-top disease induced by the baculovirusAcMNPV in Spodoptera exigua larvae are governedby independent mechanisms

Stineke van Houte & Vera I. D. Ros & Monique M. van Oers

Received: 29 October 2013 /Revised: 4 February 2014 /Accepted: 5 February 2014# Springer-Verlag Berlin Heidelberg 2014

Abstract Although many parasites are known to manipulatethe behavior of their hosts, the mechanisms underlying suchmanipulations are largely unknown. Baculoviruses manipulatethe behavior of caterpillar hosts by inducing hyperactivity and byinducing climbing behavior leading to death at elevated positions(tree-top disease orWipfelkrankheit). Whether hyperactivity andtree-top disease are independent manipulative strategies of thevirus is unclear. Recently, we demonstrated the involvement ofthe protein tyrosine phosphatase (ptp) gene of the baculovirusAutographa californica multiple nucleopolyhedrovirus(AcMNPV) in the induction of hyperactivity in Spodopteraexigua larvae. Here we show that AcMNPV ptp is not requiredfor tree-top disease, indicating that in S. exigua baculovirus-induced hyperactivity and tree-top disease are independentlyinduced behaviors that are governed by distinct mechanisms.

Keywords Protein tyrosine phosphatase . Spodopteraexigua . Tree-top disease . AcMNPV . Baculovirus .

Behavioral manipulation

Introduction

Like many parasites, baculoviruses alter the behavior of theirhosts to enhance their transmission. Baculoviruses inducehyperactive behavior in lepidopteran larvae, which is thought

to increase virus dispersal over a larger area (Goulson 1997;Kamita et al. 2005; van Houte et al. 2012). In addition, theycan alter host climbing behavior leading to host death atelevated positions (Goulson 1997; Hoover et al. 2011), aphenomenon known as ‘tree-top disease’ or Wipfelkrankheit.This manipulation most likely aids in optimal virus dispersalon plant foliage after liquefaction of the host.

Recently, we showed that the baculovirus Autographacalifornica multiple nucleopolyhedrovirus (AcMNPV) inducestree-top disease inTrichoplusia ni larvae and inSpodoptera exigualarvae that underwent a larval molt during the infection, resultingin death at elevated positions (Ros et al., unpublished). We theninvestigated the involvement of the viral ecdysteroid UDP-glucosyl transferase (egt) gene in this behavior, as a homologueof this gene was previously shown to be required for tree-topdisease of Lymantria dispar larvae infected with the baculovirusL. disparMNPV (LdMNPV) (Hoover et al. 2011). However, noeffect of the AcMNPVegt gene on tree-top disease was found inboth T. ni and S. exigua larvae (Ros et al., unpublished). Thisindicates that egt does not play a conserved role in tree-top diseaseand that, at least in AcMNPV, a different viral gene might beresponsible for the observed behavioral manipulation.

Previously, the involvement of the AcMNPV protein tyro-sine phosphatase (ptp) gene in the induction of hyperactivebehavior in S. exigua was demonstrated (van Houte et al.2012). The ptp gene appears to play a conserved role in theinduction of hyperactivity in group I NPVs (Kamita et al.2005; van Houte et al. 2012). A homologue of the ptp geneis absent in lepidopteran infecting baculoviruses that belong togroup II NPVs, including LdMNPV, or to the genusBetabaculovirus, which harbors the granuloviruses (Jehleet al. 2006; van Houte et al. 2012). In the present study, wetested the hypothesis that the AcMNPV ptp gene, besidesbeing a key player in inducing hyperactive behavior, is alsoinvolved in tree-top disease in S. exigua. Our results revealthat virus-induced hyperactivity and tree-top disease are two

Communicated by: Sven Thatje

S. van Houte :V. I. D. Ros :M. M. van Oers (*)Laboratory of Virology,WageningenUniversity, Droevendaalsesteeg1, 6708 PB, Wageningen, The Netherlandse-mail: monique.vanoers@wur.nl

Present Address:S. van HouteCentre for Ecology and Conservation, Biosciences, University ofExeter, Penryn, Cornwall, UK

NaturwissenschaftenDOI 10.1007/s00114-014-1160-8

distinct strategies of baculoviruses to manipulate their cater-pillar hosts that are induced by different mechanisms.

Materials and methods

Insect larvae and virus amplification

Spodoptera exigua (Hübner) larvae were reared as describedin van Houte et al. (2012). Virus production, amplification and

purification were all done as described in van Houte et al.(2012). Stocks of occlusion bodies (OBs) were stored at 4 °C.

Behavioral assays

Early third instar S. exigua larvae were infected either withwild-type (WT) AcMNPVor with an AcMNPV ptp deletionmutant (Δptp) by droplet feeding as described in van Houteet al. (2012). Per treatment, 24–30 larvae were infected with aviral dose of 108 OBs/ml, which corresponds to a lethal

Fig. 1 AcMNPV-induced tree-top disease in S. exigua is independent ofthe ptp gene. a–c Replicate experiment 1. d–f Replicate experiment 2. a, dVertical position (mm) of mock-infected larvae (black squares, n=29 forboth replicates), WT-infected larvae (green dots, n=18 and n=28, respec-tively, for replicate experiments 1 and 2) and Δptp-infected larvae (orangetriangles, n=23 and n=26, respectively) at different time points afterinfection (hours post-infection (hpi)). Error bars represent standard errorof the mean (SEM). b, e Vertical position (mm) of WT-infected larvae thatdied as third instars (L3) (green dots, dotted line, n=10 and n=1, respec-tively), WT-infected larvae that died as fourth instars (L4) (green dots, solid

line, n=8 and n=27, respectively), Δptp-infected larvae that died as thirdinstars (L3) (orange triangles, dotted line, n=8 and n=12, respectively) andΔptp-infected larvae that died as fourth instars (L4) (orange triangles, solidline, n=15 and n=14, respectively). Error bars represent SEM. c, f Cumu-lative mortality (%) of WT- and Δptp-infected larvae for the two replicates.Striped green bars represent WT-infected larvae that died as third instars(†L3), and solid green bars represent WT-infected larvae that died as fourthinstars (†L4). Striped orange bars represent Δptp-infected larvae that diedas third instars (†L3), and solid orange bars represent Δptp-infected larvaethat died as fourth instars (†L4)

Naturwissenschaften

concentration of 90–95 % (LC90–95). Mock-infected larvaewere used as uninfected controls, and these were droplet-fedwith a virus-free sucrose solution as described in van Houteet al. (2012). Larvae were placed individually in sterile glassjars (120 mm tall × 71 mm wide) with metal lids (with smallholes for fresh air). The jars were lined with sterile mesh wireto facilitate climbing, and a block of artificial diet (approxi-mately 3.5 cm3) was placed at the bottom. The jars wereincubated in climate-controlled incubators under the sametemperature (27 °C) and light-dark conditions (14:10 light/dark) at which the larvae were reared. The vertical position ofthe larvae was monitored every morning and evening, startingfrom 1 day post-infection (dpi) until all larvae were either deador had pupated. The experiment was performed twice (indi-cated as replicate 1 and replicate 2). Larvae that did not die dueto virus infection (died of other causes or survived despitebeing droplet-fed with virus) were excluded from furtheranalyses. The height at death was analyzed using a linearregression model in the software package R (R CoreL3Team 2013) with treatment (WT or Δptp), stage (died as thirdor fourth instar) and experiment number as fixed effects.

Results and discussion

To determine whether the AcMNPV ptp gene plays a role intree-top disease in S. exigua larvae, we performed climbingstudies with mock-, AcMNPV WT- and AcMNPV Δptp--infected S. exigua larvae. Mock-infected control larvae startedclimbing upwards from day 1 (around 24 h post-infection(hpi)) (Fig. 1a, d). A small peak in height at day 2 (48 hpi)coincided with the larval molt from the third to the fourthinstar. From day 3 (70 hpi), control larvae again started toclimb upwards and had molted from the fourth to the fifthinstar at day 4 (100 hpi). After that, larvae tended to stay highup in the jar until the onset of pupation, corresponding toapproximately day 6 (140 hpi). At this moment, the larvaegradually descended and burrowed themselves in the piece ofdiet at the bottom of the jar for pupation. In contrast, virus-infected larvae showed a climbing pattern that was dependenton the developmental stage of the larvae (Fig. 1b, e). WT-infected third instar larvae that molted to fourth instar duringthe infection (so died as fourth instars) generally died atelevated positions in the jar (Exp. 1: 75 mm±15.3 (Fig. 1b),Exp. 2: 80 mm±7.1 (Fig. 1e)), while larvae that did notundergo a molt during infection (so died as fourth instars)showed downward movement during the infection and died atlow positions in the jar (Exp. 1: 13 mm±9.5 (Fig. 1b), Exp. 2:9 mm±0 (Fig. 1e)). WT-infected larvae started dying from85 hpi onwards, and at 140 hpi, all larvae had succumbed tobaculovirus infection (Fig. 1c, f).

Throughout the measurements, no differences were ob-served in the climbing behavior of WT- and Δptp-infected

larvae (Fig. 1). Similar to WT-infected larvae that died asfourth instars, Δptp-infected larvae that molted to the fourthinstar died at high positions (Exp. 1: 79 mm±9.4 (Fig. 1b),Exp. 2: 56 mm±9.4 (Fig. 1e)), while Δptp-infected larvae thatdied as fourth instars died at low positions (Exp. 1: 3 mm±1.4(Fig. 1b), Exp. 2: 21 mm±9.5 (Fig. 1e)). Like WT-infectedlarvae, Δptp-infected larvae started dying from 85 hpi on-wards, and at 140 hpi, all larvae had succumbed to the infec-tion (Fig. 1c, f).

When the results of the two replicate experiments werecombined, WT-infected fourth instar larvae died on average at79 mm±6.5 and Δptp-infected larvae fourth instars at 66 mm±6.9 (Fig. 2). WT-infected third instar larvae died on averageat 13mm±8.6, and Δptp-infected larvae third instars at 14mm±5.9 (Fig. 2). No significant difference was found between thetwo experiments (t test=−0.175; P=0.8612) nor between thetwo treatments (WTor Δptp; t test=−1.173;P=0.2440). How-ever, the interaction between height at death and larval stage atdeath (third or fourth instar) was highly significant (t test=6.955; P<0.001).

The results presented here indicate that the AcMNPV ptpgene does not play a role in baculovirus-induced tree-topdisease in S. exigua larvae. However, we found that tree-topdisease is dependent on whether hosts had molted during theinfection or not. The reason for this stage-dependent differ-ence in height at death is currently unclear and requires furtherinvestigation. Downward movement, as seen in this study forvirus-infected larvae that died as third instars, is also observedin another baculovirus-host system. Larvae of the winter mothOperophtera brumata infected with O. brumata NPV(OpbuNPV) descended from the foliage to the lower treestems to die there (Raymond et al. 2005). Since OBs wereshown to persist better on stems than on foliage, this behav-ioral change was hypothesized to contribute to virus persis-tence. The differences observed in this study in height at death

Fig. 2 Average height at death (mm) of WT- and Δptp-infected larvaethat died as either third or fourth instars. Combined data of two replicateexperiments. Striped green bars representWT-infected larvae that died asthird instars (n=11) (†L3), and solid green bars represent WT-infectedlarvae that died as fourth instars (n=35) (†L4). Striped orange barsrepresent Δptp-infected larvae that died as third instars (n=20) (†L3),and solid orange bars represent Δptp-infected larvae that died as fourthinstars (n=29) (†L4). Error bars represent SEM

Naturwissenschaften

between the stages may also be related to differences in theecology of third and fourth instar larvae. The first and secondinstars of S. exigua larvae are gregarious, while during thethird developmental stage, a transition occurs towards solitarybehavior (Smits et al. 1987). Further research is required tobetter understand this difference in behavior between thestages, for example, by studying tree-top disease in larvae thatare infected as fourth or fifth instars or by infecting larvae witha lower viral dose, which may allow larvae to undergo twolarval molts before death due to virus infection.

A recent study by Hoover et al. (2011) demonstrated thatlarval death at elevated positions involved the LdMNPVegtgene, which encodes ecdysteroid UDP-glucosyl transferase.This enzyme is known to inhibit host ecdysteroid hormones,thereby blocking the larval molt (O’Reilly 1995). WhileLdMNPV WT-infected L. dispar larvae died at an averageheight of 130 mm, larvae infected with an LdMNPV-Δegtvirus died on average at 70-mm height. However, forAcMNPV no effect of the egt gene has been observed ontree-top disease (Ros et al., unpublished), suggesting that adifferent viral gene might be responsible for the observedclimbing behavior. Given the conserved role that the ptp geneplays in hyperactivity in group I NPVs, this gene was anobvious candidate to be tested for a role in tree-top diseasein AcMNPV-infected S. exigua larvae. However, no effect ofAcMNPV ptp on tree-top disease was found, suggesting that,at least in S. exigua, the role of ptp in manipulating hostbehavior is restricted to inducing hyperactivity.

Together, these data indicate that hyperactivity and tree-topdisease are two independent behavioral manipulations that areinduced by distinct mechanisms. This implies thatbaculoviruses have evolved multiple strategies to alter hostbehavior, probably by manipulating distinct host signalingpathways. Although we can now exclude a role for ptp andegt in tree-top disease in AcMNPV-infected S. exigua hosts, itis as yet unknown whether a specific viral gene is responsible

for tree-top disease. However, it is also possible that tree-topdisease upon AcMNPV infection is not induced by a specificviral genetic factor per se. For example, virus infection couldindirectly affect host behavioral gene expression, e.g. as aresult of metabolic changes or immunological responses.

Acknowledgments The authors thank Linda Guarino of Texas A&MUniversity for kindly providing the AcMNPV Δptp bacmid. Just Vlak isacknowledged for reading the manuscript and for the useful discussion.SVH and VIDR were both supported by the Program Strategic Alliancesof the Royal Dutch Academy of Sciences (project 08-PSA-BD-01), andVIDR is supported by a VENI grant of the Netherlands Organisation forScientific Research (project 863.11.017).

References

Core Team R (2013) R: A language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna

Goulson D (1997) Wipfelkrankheit: modification of host behaviour dur-ing baculoviral infection. Oecologia 109:219–228

Hoover K, Grove M, Gardner M et al (2011) A gene for an extendedphenotype. Science 333:1401

Jehle JA, Blissard GW, Bonning BC et al (2006) On the classification andnomenclature of baculoviruses: a proposal for revision. Arch Virol151:1257–1266

Kamita SG, Nagasaka K, Chua JW et al (2005) A baculovirus-encodedprotein tyrosine phosphatase gene induces enhanced locomotoryactivity in a lepidopteran host. Proc Natl Acad Sci U S A 102:2584–2589

O’Reilly DR (1995) Baculovirus-encoded ecdysteroid UDP-glucosyltransferases. Insect Biochem Mol Biol 25:541–550

Raymond B, Hartley SE, Cory JS, Hails RS (2005) The role of food plantand pathogen-induced behaviour in the persistence of anucleopolyhedrovirus. J Invertebr Pathol 88:49–57

Smits PH, van Velden MC, van de Vrie M, Vlak JM (1987) Feeding anddispersion of Spodoptera exigua larvae and its relevance for controlwith a nuclear polyhedrosis virus. Entomol Exp Appl 43:67–72

van Houte S, Ros VID, Mastenbroek TG et al (2012) Protein tyrosinephosphatase-induced hyperactivity is a conserved strategy of asubset of baculoviruses to manipulate lepidopteran host behavior.PLoS ONE 7:e46933

Naturwissenschaften