Previous Article | Next Article ![]()
Journal of Virology, February 2003, p. 2227-2232, Vol. 77, No. 3
0022-538X/03/$08.00+0 DOI: 10.1128/JVI.77.3.2227-2232.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Molecular, Cellular, and Developmental Biology Program, Division of Biology, Kansas State University, Manhattan, Kansas 66506
Received 5 August 2002/ Accepted 24 October 2002
|
|
|---|
|
|
|---|
Baculoviruses are known to possess at least two different types of antiapoptotic genes, namely, the p35 and iap (inhibitor of apoptosis) genes (4). The baculovirus Autographa californica M nucleopolyhedrovirus (AcMNPV) is capable of preventing apoptosis through the action of the antiapoptotic protein P35, which blocks apoptosis by binding to and inactivating caspases (1). Infection of Sf21 cells, derived from the fall armyworm caterpillar Spodoptera frugiperda, with AcMNPV mutants lacking the p35 gene results in the apoptotic death of the infected cells within 24 h postinfection (6). Apoptotic Sf21 cells show very little production of late and very late viral transcripts and proteins, and they yield significantly less progeny than cells infected with wild-type AcMNPV (7, 12). In addition, S. frugiperda caterpillars are highly resistant to infection by AcMNPV mutants lacking the p35 gene, requiring 1,000-fold more mutant virus than the wild type to achieve the same degree of lethality by intrahemocoelic injection, although only 25-fold more p35 mutant virus is required to orally infect S. frugiperda neonate larva (7, 8). In contrast, cells derived from the caterpillar Trichoplusia ni are resistant to a variety of apoptotic stimuli, including infection with p35 deletion strains of AcMNPV, and the T. ni caterpillar itself is equally susceptible to infection by the p35 deletion strains and the wild-type strains of AcMNPV (7, 8). Finally, a role for apoptosis in the attenuated behavior of p35 mutant viruses in S. frugiperda is further suggested by the observation that the infectivity of a p35 mutant virus was completely restored by expression of Cp-iap, an unrelated antiapoptotic iap gene from a different baculovirus (8).
Homologues of the AcMNPV p35 gene have been found in the related viruses Bombyx mori NPV (13) and Spodoptera littoralis NPV (9). Although the B. mori NPV p35 gene is significantly less capable of blocking caspase activity than the AcMNPV p35 gene (14) and does not appear to play a substantial role in inhibiting apoptosis during infection (13), the S. littoralis NPV p35 homolog, p49, is able to replace the AcMNPV p35 gene in AcMNPV, allowing productive infection of Sf9 cells (9). In addition, wild-type AcMNPV has been shown to induce apoptosis in cultured cells from S. littoralis (2) and Choristoneura fumiferana (16). However, to date there have been no reports of in vivo apoptosis induced by baculovirus infection.
Virus construction. In order to study in detail the role of apoptosis as an antiviral mechanism in the insect immune system, we performed a series of experiments in which S. frugiperda and T. ni larvae were injected with measured doses of a p35 deletion virus and a revertant virus, both expressing enhanced green fluorescent protein (eGFP). The p35 deletion virus vHSGFP/P35del was constructed in TN-368 cells by using a p35 deletion mutant of the L1 strain of AcMNPV, vP35delBsu36IGus (obtained from Lois Miller), which contained the ß-glucuronidase (GUS) gene under control of the Drosophila hsp70 promoter at a site immediately upstream of, and in opposite orientation to, the wild-type polyhedrin gene. The vP35delBsu36IGus virus itself was constructed by inserting the hsp70-GUS cassette into the previously described virus vP35del (8). In vHSGFP/P35del, the GUS gene was replaced with the eGFP gene. Construction of the revertant virus vP35delRev involved replacing the p35 gene back into the vHSGFP/P35del virus by cotransfecting DNA extracted from vHSGFP/P35del with the plasmid pEcoRI-S (10), containing wild-type p35, into Sf21 cells and isolating an occlusion-positive, eGFP-positive plaque. Both viruses were subjected to three rounds of plaque purification, and their correct genomic structure was verified by restriction enzyme analysis. Virus stocks were grown and the titers were determined on TN-368 cells by plaque assay.
Viral infectivity. These recombinant viruses were used to infect late-instar S. frugiperda (obtained from Agripest, Zebulon, N.C.) and T. ni (obtained from Entopath Inc., Easton, Pa.) larvae by intrahemocoelic injection. The caterpillars were injected with measured doses of budded virus diluted in complete TC-100 medium within 12 h of molting to the fifth instar (the penultimate instar in S. frugiperda and the final instar in T. ni). Injection-related mortality was less than 1% in each species. Intrahemocoelic injections were performed with a Hamilton 26-gauge syringe by inserting the needle between the first and second abdominal segment of the caterpillar and depositing between 1 and 4 µl of virus-containing medium at the rear of the abdomen to avoid excessive bleeding. Caterpillars that survived to pupation were considered to have survived viral infection. Caterpillars that died exhibited eGFP expression, confirming viral infection. In both species of caterpillars, initial experiments demonstrated that the eGFP-expressing viruses each had 50% lethal dose values that were similar to those previously reported for a p35 mutant and revertant viruses lacking eGFP (7), with 50% lethal dose values of 0.1 PFU for vP35delRev and 1 PFU for vHSGFP/P35del in T. ni and 20 PFU for vP35delRev and 105 PFU for vHSGFP/P35del in S. frugiperda (data not shown). Thus, similar to previously published results, deletion of p35 results in greatly decreased infectivity in S. frugiperda larvae.
Hemolymph assays. In order to better assess the pathogenesis of virus infection, T. ni and S. frugiperda caterpillars were injected with either vHSGFP/P35del or vP35delRev (4 x 103 PFU in T. ni and 2 x 105 PFU in S. frugiperda; the dose injected was chosen on the basis of the lethality of vHSGFP/P35del to each caterpillar species), and hemolymph was extracted daily from cohorts of 10 caterpillars into ice-cold anticoagulant buffer [4 mM NaCL, 40 mM KCl, 1.7 mM piperazine-N,NN-bis(2-ethanesulfonic acid), 146 mM sucrose, 0.1% polyvinyl pyrrolidine, 8 mM EDTA, 9.5 mM citric acid, 27 mM sodium citrate]. Hemocytes were separated from the hemolymph by centrifugation and were examined for eGFP expression. A minimum of 100 cells were scored from each sample, and the means ± 95% confidence limits are shown in Fig. 1A. Levels of virus in the hemolymph were assessed by a modified end point dilution assay. Cell-free hemolymph was serially diluted in anticoagulant buffer and was added to TN-368 cells, and for each set of wells the lowest dilution displaying eGFP expression from infected cells was recorded. For the purposes of calculating the mean, samples that yielded no infection even when undiluted were assigned a dilution value of 10-1. Three replicates were scored for each sample, and the mean dilutions (± 95% confidence limits) at which virus could be detected in the hemolymph were calculated and are shown in Fig. 1B. The pathogenesis of wild-type AcMNPV in both T. ni and S. frugiperda has been described previously (3), and the revertant virus vP35delRev did not differ significantly from the wild type in buildup of infected hemocytes or viral hemolymph titers in either species (Fig. 1). In addition, the progression of vHSGFP/P35del infection was similar to that of vP35delRev in T. ni. However, in S. frugiperda the pattern of infection with vHSGFP/P35del virus differed dramatically from that of vP35delRev (Fig. 1). S. frugiperda caterpillars injected with 2 x 105 PFU of vHSGFP/P35del showed very little increase in the levels of infected hemocytes or the titers of virus in the hemolymph, with many caterpillars having levels of virus in the hemolymph below the threshold of detection in the assay used.
![]() View larger version (18K): [in a new window] |
FIG. 1. The progress of infection following injection with vHSGFP/P35del (closed symbols) or vP35delRev (open symbols). Larvae of T. ni and S. frugiperda were injected with virus, and hemolymph from cohorts of 10 caterpillars was collected every 24 h. (A) Infection of hemocytes as assessed by eGFP fluorescence. (B) Levels of budded virus in the hemolymph.
|
Internal fluorescence. Dissection and internal assessment of infection was also performed on S. frugiperda caterpillars injected with vHSGFP/P35del. Despite a large degree of variation in the extent of the infection among the caterpillars, some general observations could be made. At the 24-h time point, caterpillars showing external signs of infection (n = 13) showed a distribution of eGFP expression and individual foci of infection concentrated in the ventral and lateral peripheral fat body adjacent to the body wall, with fewer foci in the peripheral fat body of the dorsal midline and very little eGFP expression in the perivisceral fat body within the hemocoel. In addition, approximately half of the caterpillars showing no external sign of infection (n = 12) contained several small eGFP-positive foci of infection in the fat body. By 48 h, all examined caterpillars lacking external eGFP expression were also free of eGFP expression internally.
The strong phenotype. Over the course of the infection, strong-phenotype caterpillars showed a pattern of increasing eGFP expression in the peripheral, and to a lesser extent the perivisceral, fat body coupled with very large reduction in fat body volume. Infection of the tracheal epithelia in strong-phenotype caterpillars was first observed at 48 h as widely scattered individual eGFP-positive cells, in a pattern inconsistent with direct cell-to-cell transmission of the virus through the extent of the trachea. The number of eGFP-positive cells increased as the infection progressed, with bare regions of chitin appearing along the major trachea in most strong-phenotype caterpillars by 96 h postinfection. With the exception of the epithelia of the appendages and surrounding the spiracles (in which infection in the former could be detected as early as 24 h), body wall epithelia did not typically show signs of infection in these caterpillars until 96 h and then did so only in regions immediately adjacent to infected fat body. Caterpillars displaying the strong phenotype lived between 6 and 8 days following injection, at which point they died and melanized, although they did not undergo the melting process commonly associated with fatal baculovirus infection, as was previously reported (7).
The weak phenotype. In contrast to the strong phenotype, weak-phenotype caterpillars showed a pattern of decreasing eGFP expression in all tissues with the eventual reduction of eGFP expression to a limited number of persistent foci. The numbers and distribution of eGFP-positive foci of infection in the fat body and tracheal epithelia of weak-phenotype caterpillars were low and widely scattered by 48 h, and by 96 h these tissues were free of eGFP expression in almost all of the weak-phenotype caterpillars examined. In addition, fat body and tracheal epithelia in weak-phenotype caterpillars did not show any of the reduction in volume or tissue destruction observed in the strong-phenotype caterpillars. At time points 96 h and beyond, weak-phenotype caterpillars showed eGFP expression localized in the appendages and spiracles, with the frequent appearance of eGFP fluorescence in the brain, ganglia, and in isolated muscle bands. Caterpillars displaying the weak-infection phenotype survived up to 14 days after injection, during which time they did not molt. The actual cause of death of these caterpillars is uncertain; however, the reduction in the fat body tissue observed very late in the infection suggested that the caterpillars had ceased feeding and had died of starvation. Weak-phenotype caterpillars that did not die of starvation attempted a precocial molt to pupa despite being in the penultimate instar and died as partially melanized prepupa.
Histology and terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) assay. Tissue from vP35delRev-infected and vHSGFP/P35del-infected (strong phenotype) S. frugiperda caterpillars at 96 h postinjection was fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned, and the sections were stained with hematoxylin and eosin (Fig. 2). Uninfected tissue showed normal fat body and epithelial morphology, while vP35delRev-infected epithelia displayed typical baculovirus pathology, with the tissues being distinctly swollen and the nuclei of both epithelial and fat body cells having increased in diameter (occlusion bodies present in the nuclei of vP35delRev-infected cells before staining were dissolved by alkaline conditions during the staining procedure). In contrast, vHSGFP/P35del-infected fat body from caterpillars displaying the strong phenotype showed a highly disrupted morphology consisting of both condensed regions of darkly staining tissue and the presence of cavities in the normally contiguous fat body (Fig. 2C). Epithelia from vHSGFP/P35del-infected insects had a similar disrupted morphology (not pictured).
![]() View larger version (67K): [in a new window] |
FIG. 2. Hematoxylin-and-eosin-stained peripheral tissue sections from abdomens of uninfected (A), vP35delRev-infected (B), or vHSGFP/P35del-infected (strong phenotype) (C) S. frugiperda caterpillars sacrificed 96 h postinjection. m, muscle; fb, fat body; cu, cuticle; ep, epidermis. Additional images can be viewed at http://www.ksu.edu/virology/Clem_lab/Supplemental_figures/.
|
![]() View larger version (69K): [in a new window] |
FIG. 3. Tissue sections from peripheral regions of S. frugiperda caterpillar abdomens. Caterpillars were either uninfected (A to D), infected with vP35delRev (E to H), or infected with vHSGFP/P35del (strong phenotype) (I to L) and were sacrificed 96 h postinjection. Sections were stained with DAPI to visualize nuclei, while apoptotic cells were identified by TUNEL assay with rhodamine-labeled dUTP. Photographs were taken under normal light (A, E, and I), DAPI fluorescence (B, F, and J), eGFP fluorescence (C, G, and K), or rhodamine fluorescence for TUNEL (D, H, and L). fb, fat body; tr, trachea; bw, body wall (cuticle and epidermis); ob, occlusion bodies. Additional images can be viewed at http://www.ksu.edu/virology/Clem_lab/Supplemental_figures/.
|
Conclusions. It was previously hypothesized that apoptosis was responsible for the decrease that is seen in the infectivity of p35 mutant viruses in S. frugiperda larvae, based on the following correlative pieces of evidence (reviewed in reference 5): (i) the induction of apoptosis in Sf21 cells by p35 mutant viruses and the decreased infectivity of these mutant viruses in S. frugiperda larvae; (ii) the lack of apoptosis in TN-368 cells infected with p35 mutant viruses and the normal infectivity of these mutant viruses in T. ni larvae; (iii) the reduced levels of viral replication associated with apoptosis in cultured cells; and (iv) the ability of an unrelated antiapoptotic gene, Cp-iap, to rescue both the in vitro replication and in vivo infectivity defects of p35 mutant viruses. However, up to now apoptosis occurring in the cells of p35 mutant-infected larvae had not actually been documented. In this study, we have demonstrated for the first time a strong correlation between viral induction of apoptosis in vivo and a deficiency in viral infectivity following injection of virus directly into the hemocoel. In S. frugiperda caterpillars infected with vHSGFP/P35del and displaying the strong phenotype, we observed large numbers of apoptotic cells corresponding to regions of high eGFP expression in caterpillars that became strongly infected. Extensive tissue destruction occurred in the fat body and epithelium of these insects, which correlated strongly with areas of TUNEL staining and viral replication. Although we did not find significant numbers of apoptotic cells in caterpillars displaying weak or no eGFP expression, this was presumably due to the small number of these cells, their scattered nature, and the propensity for isolated apoptotic cells to be rapidly phagocytosed in vivo (18). Given the available evidence, we conclude that apoptosis can play an important role in insect immunity to baculovirus infection, and thus should be recognized as an arm of the insect immune system. It will be interesting to determine whether a similar effect can occur during oral infection, when the infection must first pass through the bottleneck imposed by the midgut, associated basal lamina, and the tracheal epithelia. While a previous study found that there was less of a reduction in mortality when S. frugiperda larvae were infected orally with p35 mutant AcMNPV than when they were infected by injection (25-fold versus 1,000-fold) (7, 8), it is difficult to directly compare these results, because the feeding studies used neonates while the injection studies used late-instar larvae. Whether apoptosis can be an effective barrier to oral infection is presently under investigation.
ADDENDUM IN PROOF After this paper was accepted for publication, a report describing in vivo apoptosis induction by wild-type ACMNPV in Spodoptera litura larvae was published (P. Zhang, K. Yang, X. Dai, Y. Pang, and D. Su, J. Gen. Virol. 83:3003-3011).
This material is based upon work supported by the Cooperative State Research, Education, and Extension Service, U.S. Department of Agriculture, under agreement no. 2001-35302-09972. Support was also provided by the Kansas Agricultural Experiment Station.
Contribution number 02-352-J from the Kansas Agricultural Experiment Station. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»