Baculovirus PTP2 Functions as a Pro-Apoptotic Protein

The family Baculoviridae encompasses a large number of invertebrate viruses, mainly infecting caterpillars of the order Lepidoptera. The baculovirus Spodoptera exigua multiple nucleopolyhedrovirus (SeMNPV) induces physiological and behavioral changes in its host Spodoptera exigua, as well as immunological responses, which may affect virus transmission. Here we show that the SeMNPV-encoded protein tyrosine phosphatase 2 (PTP2) induces mild apoptosis in Spodoptera frugiperda (Sf) 21 cells upon transient expression. Transient expression of a catalytic-site mutant of ptp2 did not lead to apoptosis, indicating that the phosphatase activity of PTP2 is needed to induce apoptosis. We also found that the caspase level (indicator of apoptosis) was higher in cells transfected with the ptp2 gene than in cells transfected with the catalytic mutant. Adding a caspase inhibitor reduced the level of ptp2-induced apoptosis. Moreover, deletion of the ptp2 gene from the viral genome prevented the induction of apoptosis in S. exigua hemocytes. The virus titer and virulence indices (the viral infectivity and the time to death) were not affected by deletion of the ptp2 gene. However, the viral occlusion body yield from S. exigua larvae infected with the mutant virus lacking the ptp2 gene was much lower than the yield from larvae infected with the wild-type (WT) virus. We hypothesize that the observed pro-apoptotic effects of PTP2 are the result of PTP2-mediated immune suppression in larvae, which consequently leads to higher viral occlusion body yields.


Introduction
Apoptosis is an active process of programmed cell death that is involved in immunity, normal development and cell differentiation [1]. It is characterized by morphological changes, including cell blebbing and cell shrinkage, chromosomal DNA fragmentation, as well as energy-dependent biochemical changes [1,2]. Apoptosis is part of the insect's innate immune response against pathogens, including viruses, triggering premature cell death of infected host cells [3]. Viruses have developed a range of strategies to counteract host-induced apoptosis [3], including the expression of anti-apoptotic genes. On the other hand, some viruses have been found to actively induce apoptosis in their host in order to suppress the host immune system and/or enhance virus dissemination [4,5]. For example, it was found that the polydnavirus Microplitis demolitor bracovirus (MdBV) induced apoptosis in immune cells of Spodoptera frugiperda caterpillars and that this activity contributed to the immunosuppression of the host [5]. Several viral genes and proteins have been found to induce pro-apoptotic effects in host-derived cell lines [4][5][6][7][8]. For example, Suderman et al. [5] found that the protein tyrosine phosphatase-H2 (PTP-H2) from MdBV induced apoptosis in S. frugiperda

Construction of Plasmids for Transient Expression Assays
The pIB-DEST expression vector (Invitrogen) was used for transient expression assays in Sf21 cells. The vector contains the constitutively expressed early OpIE2 promoter, derived from Orgyia pseudotsugata (Op) MNPV, to drive the expression of the gene of interest. Each gene to be expressed was cloned downstream of the egfp open reading frame (ORF), from which it was separated by the foot and mouth disease virus 2A ribosome skipping element (FMDV2A). EGFP was used to monitor the transfection efficiency. FMDV2A is a seventeen amino acids long (NFDLLKLAGDVESNPGP) element that allows co-translational cleavage between the marker protein (EGFP) and the protein to be analyzed. In total five pIB-DEST constructs were made: (i) modified pIB-DEST vector (pIB-DESTmod) in which the chloramphenicol resistance (cat) gene and the ccdB gene were replaced by a 888 bp long internal DNA segment from the AcMNPV gp64 ORF (nt509-1396) to have a negative control vector that could be amplified in the E. coli strain DH5α (the ccdB gene in the original pIB-DEST is lethal to E. coli); (ii) pIB-EGFP encoding only EGFP; (iii) pIB-SePTP2 encoding EGFP and SeMNPV PTP2; (iv) pIB-SePTP2mut encoding EGFP and SeMNPV PTP2 containing a mutation in its catalytic site (mutation of cysteine to serine at position 110 (C110S)); and (v) pIB-AcPTP encoding EGFP and the AcMNPV PTP protein (distantly related to PTP2) [23], to allow comparison with a different viral protein tyrosine phosphatase. To construct these plasmids, the AcMNPV gp64 segment and the ptp ORF were amplified from the AcMNPV E2 bacmid [24] using primer pairs 1 and 2, and 3 and 4, respectively (Table S1). The Septp2 ORF was amplified from the SeMNPV bacmid SeBac10 using primers 5 and 6 (Table S1). To obtain the C110S catalytic mutant of Septp2 a ptp2 C110S mutation was created by a two-step PCR reaction using SeBac10 as template. First, primer 7 (Table S1), which introduced two point mutations at nucleotides 328 and 330 relative to the ATG start codon, was combined with primer 8 (Table S1), which annealed to the 3 end of the ptp2 ORF, to create a 3 segment with the required mutations. The resulting 198 bp PCR product was used as reverse primer and combined with primer 9 (Table S1), annealing to the 5 end of the ptp2 ORF, to create the full length ptp2 C110S insert, again with SeBac10 as a template. Then the ptp2 C110S insert was verified by sequence analysis after cloning into the pJET1.2 cloning vector (Fermentas, Merelbeke, Belgium). The resulting plasmid pJET-ptp2 C110S was then used as template for PCR using the primers 5 and 6 to allow Gateway cloning (see below).
The forward primers were designed to introduce an attB1 site and a HindIII site, while the reverse primers introduced an attB2 site (Table S1) to enable Gateway ® cloning (Invitrogen). All PCRs were performed with the proofreading Phusion polymerase (Finnzymes, Waltham, MA, USA). The resulting PCR amplicons were first cloned into the pDONR207 donor plasmid (Invitrogen) and sequenced and subsequently cloned into the pIB-DEST plasmid. The EGFP-FMDV2A was obtained as HindIII fragment from the plasmid CHIKrep-pac2AEGFP as described before [25] and inserted into HindIII-linearized pIB-DEST plasmids containing the genes of interest.

Transient Expression Assay
In total, two transient expression assays were conducted and each assay was performed twice as two independent replicates. In the first assay the apoptotic effects of the above-mentioned pIB-DEST-derived plasmids on Sf21 cells was assessed. In total, seven treatments were included in this assay: (1) pIB-EGFP + Actinomycin D (ActD, Sigma-Aldrich), a chemical inducer of apoptosis, as a positive control; (2) pIB-SePTP2; (3) pIB-SePTP2mut; (4) pIB-AcPTP; (5) pIB-EGFP; (6) pIB-DESTmod, to monitor whether the pIB-DEST vector itself induces apoptosis on Sf21 cells; and (7) mock transfected cells (no vector used). The second transient expression assay was conducted to assess whether apoptosis inhibitors were able to inhibit Septp2 induced apoptosis in Sf21 cells. In this case cells were transfected with pIB-SePTP2, in the presence or absence of the pan-caspase inhibitor carbobenzoxy-valyl-alanyl-aspartyl-(O-methyl)-fluoromethylketone (Z-VAD-FMK, Promega, Leiden, The Netherlands).
For both transfection assays Sf21 cells were seeded at a confluency of 25-30% in 6-well plates. The cells were incubated overnight and two h prior to transfection, the culture medium was replaced with Grace's medium without serum. Transfections were performed with 4 µg of the pIB-DEST plasmids containing the genes of interest. Cellfectin II (Invitrogen) was used as the transfection reagent according to the manufacturers' protocol. ActD was added to the medium 40 h post transfection (hpt) at a final concentration of 0.25 µg/mL. The caspase inhibitor Z-VAD-FMK was supplied in the normal medium at a concentration of 20 µM at 5 hpt. Cells were monitored daily using a Zeiss Axio Observer inverted fluorescence microscope to analyze EGFP expression and apoptotic body formation.

Caspase Assay
A caspase assay (treatments) were performed using the experimental set-up for transient expression assay 1 described above; the assay was performed twice as two independent replicates. The transfection and apoptosis induction procedures were the same as described above, with the exception that all transfections were performed in 24-well plates, making each of the reagent volumes four times smaller. At 48 hpt cells were homogeneously resuspended in culture medium. For each sample, 50 µL of cell suspension was mixed with 50 µL of caspase-glo 3/7 substrate (Promega) in white 96-well plates. All samples were prepared in duplo per replicate. The plate was incubated in the dark for one hour at room temperature. Luminescence was measured in a fluorometer (Optima; settings: 5 flashes with maximal gain and top optic measurement). Luminescence was measured as relative luminescent units (RLUs). Differences in RLUs between the treatments were tested using the software package SPSS, version 22.0 (Armonk, NY, USA) [26]. Data were first tested for normality (Kolmogorov-Smirnov test) and homogeneity of group variance (Levene's test). Where possible, logarithmic transformations were performed to attain normality and homogeneity of variances. A one-way analysis of variance (ANOVA) was performed to determine whether there were significant differences in RLUs among the different treatments. When a significant difference was found, pairwise comparisons were performed using Tukey post hoc tests.

Construction of Recombinant Virus
An SeBac10-derived bacmid with a partial deletion of the ptp2 ORF (∆ptp2) was constructed using the combined lambda RED and Cre-recombination methods described before [27]. The mutant had a deletion of the major part (354 bp) of the ptp2 ORF (498 bp), ranging from nucleotide 48 to 401. Briefly, PCR products of the chloramphenicol (cat) resistance gene flanked by modified loxP sites and with 50 bp overhangs homologous to the flanking regions of the region to be deleted were generated with Phusion polymerase using primer 10 and 11 (Table S1). Subsequently, the major part of the ptp2 ORF was replaced by the cat gene flanked by modified loxP sites. The cat gene was then removed by Cre-recombinase, leaving an inserted segment of 162 bp that contained the recombined loxP site. The deletion of the cat gene was checked by PCR using primers 12 and 13 (Table S1) that annealed to the 5 and 3 areas outside the deleted region. To enable oral infection of S. exigua larvae, the SeMNPV polyhedrin promoter and ORF were introduced into the ∆ptp2 SeBac10, which, similar to the ancestral SeBac10, lacks the polyhedrin gene, as described before [20].

Apoptosis in S. exigua Larvae Hemocytes
Late 2nd instars of S. exigua were starved overnight for 16 h and allowed to molt. The next morning, newly molted 3rd instars were infected with WT SeMNPV or ∆ptp2 SeMNPV using droplet feeding as described before [27]. Viral concentrations of 10 6 OBs/mL, known to kill at least 90% of WT SeMNPV-infected larvae, were used for infection. A virus-free sucrose solution was used for mock infections. For each treatment, 48 larvae were infected and the experiment was performed twice as two independent replicates.
Apoptosis of hemocytes was measured by bleeding 3rd instar S. exigua larvae from a cut proleg 48 h post infection (hpi). Two microliter hemolymph was collected from each infected larva and the hemolymph samples from 10 larvae were pooled in 80 µL PBS containing 1-phenyl-2-thiourea (PTU, 0.1% final concentration) to prevent melanization. Next, 5 µL of Annexin V-EGFP (Annexin V-EGFP Apoptosis Detection Kit, BioVision, Milpitas, CA, USA) was added to stain any apoptotic cells, while Hoechst dye was added at a concentration of 10 µM to stain the nuclei of all cells. The mixtures were incubated at RT for 10 minutes and fluorescence was studied with a Zeiss Axio Observer inverted microscope.

RNA Extraction and RT-PCR
To confirm the expression of ptp2 in WT SeMNPV-infected larvae, and to confirm the absence of ptp2 expression in ∆ptp2 SeMNPV-infected larvae, an RT-PCR amplification was performed on total RNA extracted from infected single whole larvae at two days post infection (dpi). Infections of 3rd instar S. exigua larvae were performed using the same procedure and treatment as described above. At two dpi, a single larva was homogenized in 250 µL Trizol reagent (Invitrogen). Total RNA purification and subsequent cDNA synthesis were performed as described before [20]. RT-PCR was performed using primer pairs to amplify (i) a 486 bp sequence within the ORF of the S. exigua host translation initiation factor eIF5A [28] to verify that larval RNA was successfully extracted and that cDNA was synthesized (primer 14 and 15; Table S1); (ii) a 492 bp sequence within the coding sequence of the SeMNPV immediate early (ie1) gene to confirm successful virus infection (primer 16 and 17, Table S1); (iii) a 365 bp sequence encompassing the recombined loxP site to check for the correct deletion (the forward primer stretched from 27 bp upstream of the ptp2 start codon to the first 9 bp of the ptp2 ORF, the reverse primer annealed from 109 to 129 bp downstream of the ptp2 stop codon) (primer 12 and 13; Table S1). For each sample, a non-RT control sample was included, in which water was added instead of reverse transcriptase in the RT step. In addition, for each PCR a negative control with only water and a positive control with purified WT SeBac10 as template were processed.

Virus Titrations
To confirm whether deletion of the ptp2 gene affect BVs production in larval hemocytes, BVs titers in WT SeMNPV-and ∆ptp2 SeMNPV-infected larvae were compared at two dpi. Infections were performed using the same procedures and treatments as described above. The experiment was performed three times as three independent replicates and per replicate 70 larvae were infected for each treatment. Two microliter of hemolymph was collected at 48 hpi from each larva as described above. Samples from 50 larvae were pooled in a total volume of 4 mL PBS containing 0.1% PTU and filtered through a 0.45 µm non-pyrogenic filter to remove hemocytes and potential microbial contaminants. The filtered sample was used for virus titration. The 50% tissue culture infectious dose (TCID 50 ) was determined using an endpoint dilution assay (EPDA) on Se301 cells [29] and scored at 7 dpi for the presence of SeMNPV by looking at the cytopathic effect. The results were analyzed for significant differences in virus titers by a t test in GraphPad Prism 5 using a 95% confidence interval.

Infectivity Assays
To see whether removal of the ptp2 gene of SeMNPV affected viral infectivity, we performed infectivity assays to determine the infectivity of WT and ∆ptp2 SeMNPV in 3rd instar S. exigua larvae as described before [20]. Five different concentrations were included for each virus, in 6-fold serial dilutions: 1.3 × 10 6 , 2.0 × 10 5 , 3.6 × 10 4 , 6.0 × 10 3 , and 1.0 × 10 3 OBs/mL. Infections were performed as described above. Mock-infected larvae fed with a virus-free solution were used as controls. Larvae were scored for mortality from three dpi onwards until all larvae had died or pupated. Larvae that died for other reasons than virus infection were excluded from the analysis. The assays were performed three times. The program R v3.0.0 [30] was used for analyzing the data using a logistic regression model as described before [27]. Treatment was used as a fixed effect and the model followed a binomial distribution.
To compare the time to death for WT and ∆ptp2 SeMNPV in 3rd instar S. exigua larvae, 48 larvae were infected with an concentration of 10 6 OBs/mL, known to kill at least 90% of infected larvae as described above. Mock-infected larvae were included as controls. Larvae were checked for mortality twice per day as described above and three independent replicates were performed. The effects of treatment and experiment on time to death were analyzed using Cox's proportional model in the program R as described before [27]. Since most larvae died as 3rd instars, larval stage was excluded as a factor in the model. LT 50 values were calculated in R.

Determination of OB Yield
Five larvae were randomly selected from each LT 50 assay to determine the OB yield/larva. In total, 15 larvae were used for each virus (WT and ∆ptp2 SeMNPV). The cadavers were individually homogenized in 0.5 mL sterile water and then filtered through a double layer of cheese cloth. The filtrate was centrifuged at 6000 rpm for 5 min. The supernatant was discarded and the pellet was resuspended in 0.5 mL sterile water. OB yield/larva was calculated by counting the number of OBs in 10 µL virus solutions using a Bürker-Türk hemocytometer. The counting was performed three times and the average was used for calculating the OB yield per larva. The OB yield was analyzed (virus treatment was used as a fixed factor) by a t test in GraphPad Prism 5 using a 95% confidence interval.

Transient Expression of SeMNPV Ptp2 in Sf21 Cells Induces Mild Apoptosis
Since the protein tyrosine phosphatase gene (ptp-H2) from the polydnavirus MdBV showed pro-apoptotic effects on host cells, we hypothesized that SeMNPV ptp2 might function in a similar way. To investigate the pro-apoptotic effects of the SeMNPV ptp2 gene, we performed transient expression assays in Sf21 cells, using plasmids from which ptp2 was expressed together with EGFP (see Figure 1A). Based on the number of EGFP-expressing cells, transfection efficiencies were estimated to be 40-50% for each tested plasmid. The negative controls included mock-transfected cells, cells transfected with the modified 'empty' expression vector pIB-DESTmod (to check whether expression vector itself induces apoptosis in transfected cells), and cells transfected with pIB-EGFP expressing only EGFP, and in these treatments the cells did not show signs of apoptosis ( Figure 1B; Figure S1). As a positive control, cells transfected with pIB-EGFP were treated with the strong apoptosis inducer ActD and morphological changes characteristic for apoptotic cells (cell blebbing and formation of apoptotic bodies) were observed ( Figure 1B; Figure S1). Similarly, cell blebbing and apoptotic bodies were observed in cells transfected with pIB-SePTP2, but to a lesser extent compared with cells treated with ActD ( Figure 1B; Figure S1). In contrast, cells transfected with pIB-SePTP2mut, expressing a catalytic mutant of PTP2, did not show the formation of apoptosis and appeared healthy ( Figure 1B; Figure S1). Signs of apoptosis were not observed in cells transfected with pIB-AcPTP, coding for the distantly related AcMNPV PTP gene ( Figure S1). These results indicate that transient expression of SeMNPV PTP2 caused apoptosis in Sf21 cells. Since transfection with the catalytic mutant did not result in apoptotic cells, we conclude that the phosphatase activity of SePTP2 is needed to induce apoptosis.  (Table S2).

SePTP2 Induces Apoptosis by Caspase Activation
Caspases are a family of cysteine proteases that play an essential role during the induction of apoptosis. To study whether caspases are activated in the presence of PTP2, we measured the activity of the effector caspases 3 and 7 at 48 hpt. This activity was quantified by interaction of the effector caspases 3 and 7 with caspase-glo 3/7 substrates, which generates a luminescent signal (expressed in RLUs). The positive control expressing EGFP in the presence of ActD (ActD + EGFP) showed the highest caspase activity, corresponding to approximately 430,000 RLUs ( Figure 1C). This level of activity was significantly higher compared to all other treatments (P < 0.001 for all comparisons; Table S2). Cells expressing SePTP2 showed caspase activity levels of approximately 120,000 RLUs ( Figure 1C), which was significantly lower than for the ActD + EGFP treatment (P < 0.001), but significantly higher than for the other treatments (P < 0.005 for all comparisons; Table S2), except for AcPTP (P = 0.149). Mutation of the SePTP2 catalytic domain decreased the caspase activity level significantly to approximately 54,000 RLUs (P < 0.001). Cells transfected with pIB-EGFP and pIB-DESTmod, showed similar low levels of caspase activity as cells transfected with pIB-SePTP2mut ( Figure 1C; P = 1.000 for both comparisons). For both replicates, cells transfected with a pIB-DEST vector (ActD + EGFP, SePTP2, SePTP2mut, AcPTP, EGFP, or DESTmod) showed higher levels of caspase activity than the mock transfection ( Figure 1C; P < 0.001 for all comparisons; Table S2). The results indicate that caspase activities were significantly higher in cells transfected with pIB-SePTP2 than cells transfected with other plasmids (except for cells transfected with pIB-AcPTP), supporting the conclusion that SePTP2, but also AcPTP, induces mild apoptosis upon transient expression in Sf21 cells. However, for AcPTP, no visual signs of apoptosis (apoptotic bodies or cell blebbing) were observed ( Figure S1).
To confirm the involvement of caspases in SePTP2-induced apoptosis, we repeated the transfection experiment with pIB-SePTP2 in the presence or absence of the caspase inhibitor Z-VAD-FMK. In the absence of caspase inhibitor, cells transfected with pIB-SePTP2 again showed induction of apoptosis ( Figure 2A). However, the presence of the caspase inhibitor blocked the induction of apoptosis: cells appeared to be healthy and cell blebbing and apoptotic bodies were not observed ( Figure 2B). These data provide further evidence that SePTP2 induces apoptosis in Sf21 cells via caspase activation.

Deletion of the Ptp2 Gene Reduces SeMNPV-Induced Apoptosis in S. exigua Hemocytes
To investigate whether SePTP2 also has a pro-apoptotic effect in hemocytes of S. exigua larvae, a mutant SeMNPV virus lacking the ptp2 gene (∆ptp2 SeMNPV) was created. Hemocytes were collected from WT and ∆ptp2 SeMNPV-infected larvae at 48 hpi and then stained with Annexin V-EGFP (to stain apoptotic cells) and Hoechst (to stain the nuclei of all collected cells). Fluorescence microscopy analysis showed that WT SeMNPV also induced apoptosis in S. exigua hemocytes: The green Annexin V-EGFP signal was observed in many cells, indicating that these cells went into apoptosis (Figure 3). In contrast, hemocytes isolated from ∆ptp2 SeMNPV-infected and mock-infected larvae showed a much lower number of green cells, indicating apoptosis was induced at a much lower level by these treatments (Figure 3). To confirm the expression of ptp2 in WT SeMNPV-infected larvae and to verify the absence of a full-length ptp2 transcript in ∆ptp2 SeMNPV-infected larvae, an RT-PCR was performed on total RNA extracted from mock-and virus-infected individual larvae at two dpi. As expected, a full-length ptp2 transcript was present in WT SeMNPV-infected larvae and absent in mock-infected larvae (Figure 4,  right panel, lanes 1 and 2). Partial deletion of 354 bp within the ptp2 ORF in ∆ptp2 SeMNPV resulted in a smaller product (Figure 4, right panel, lane 3) The SeMNPV ie1 gene, included as a control to check for successful virus infection (Figure 4, middle panel), was expressed in all virus-infected larvae and absent in mock-infected larvae. An RT-PCR targeting the mRNA of the S. exigua eIF5A gene was included as a control for correct RNA extraction and cDNA synthesis, and as expected eIF5A was expressed in both virus-and mock-infected larvae (Figure 4, left panel).

Virus Infectivity Was Not Affected by Deleting the Ptp2 Gene
To determine whether deletion of the ptp2 gene affected the production of infectious budded viruses (BVs), hemolymph of WT and ∆ptp2 SeMNPV-infected larvae was collected at 48 hpi and the infectious BV titer was measured as the TCID 50 value for Se301 cells for three biological replicates. The BV titers were not significantly different between the two treatments at 48 hpi (t test = 1.314; d.f. = 4; P = 0.2592), indicating that deletion of the ptp2 gene did not affect the production of infectious BVs (Table S3).
To study whether deletion of the viral ptp2 gene affected viral infectivity, we performed a logistic regression on the mortality data obtained after infecting S. exigua third instar larvae with WT and ∆ptp2 SeMNPV, for each replicate separately. The ratio of infectivity of ∆ptp2 SeMNPV to WT SeMNPV (odds ratio) was determined. In all three replicates, the infectivity of WT and ∆ptp2 SeMNPV were not significantly different (judged by overlap of 95% confidence interval of the odds ratio; Table S4). There were no significant differences in mortality between the three replicates.
We then investigated whether deletion of the viral ptp2 gene affected the time to death for infected third instars of S. exigua, using a survival analysis. The Cox's proportional hazards model was used to determine the mortality rate (hazard rate, rate at which larvae died) for WT and ∆ptp2 SeMNPV. Treatment and experiment were included as factors. The mortality rate for larvae infected with mutant virus was similar to that of WT SeMNPV-infected larvae (ratio of 1.05; z = 0.375; P = 0.721). There were no significant differences between the three replicates.

Deletion of the Ptp2 Gene Decreases SeMNPV OB Yield
To determine the effect of deletion of the ptp2 gene on the total OB yield, we randomly selected 15 cadavers from WT and ∆ptp2 SeMNPV-deceased larvae. The OB yield of individual larvae was counted using a Bürker-Türk hemocytometer. The OB yield from larvae killed by WT SeMNPV was significantly higher that from larvae killed by ∆ptp2 SeMNPV, in which the OB yield was decreased by 34% ( Figure 5; t test = 2.583; d.f. = 28; P < 0.05). Though deletion of the ptp2 gene did not affect virus infectivity nor the time to death, the OB yield was lower for larvae infected with the mutant SeMNPV lacking ptp2.

Discussion
Many viruses carry genes that are involved in inducing apoptosis in host-derived cell lines [4,5,13,15,31], such as the ptp-h2 gene from MdBV, which induces apoptosis in Sf21 cells [5]. The baculovirus SeMNPV carries a protein tyrosine phosphatase gene (ptp2) which contains the same catalytic domain as ptp-h2, therefore, we tested whether SeMNPV ptp2 has a pro-apoptotic function as well. To test this hypothesis we took advantage of Sf21 cells, which are known to undergo apoptosis upon different stimuli [32]. Sf21 cells transfected with a plasmid expressing SePTP2 showed morphological signs of apoptosis and increased levels of caspase activity. However, the pro-apoptotic effect of SePTP2 might be restricted to certain cell types. Similar results were found for MdBV PTP-H2 as well; PTP-H2 induced apoptosis in Sf21 cells, but not in granulocytes or plasmatocytes of S. frugiperda larvae, T. ni High Five cells or Drosophila S2 cells [5]. More interestingly, some viruses even induce apoptosis in one cell type, but block apoptosis formation in another cell type. For example, Herpes simplex virus (HSV) induces apoptosis in Jurkat cells, a T-cell leukemia line, but protected HEp-2 cells (a carcinoma cell line) from apoptosis triggered by tumor necrosis factor alpha [31]. To our knowledge, there is no explanation so far why a viral gene induces apoptosis in one cell type, but not in another cell type. A possible explanation is that certain regulatory pathways are not active in cell types in which apoptosis is not induced. As a result, the pro-apoptotic effect of viral genes is restricted to certain cell types or host tissues.
Our data showed a significantly increased caspase activity in transfected cells expressing SePTP2. In addition, the phosphatase activity of SePTP2 was needed to activate the effector caspases. Cells expressing the distantly related baculovirus PTP (AcPTP) also showed increased caspase activity; however, apoptosis was not observed in these cells (absence of cell blebbing and apoptotic bodies), indicating that the pro-apoptotic effect was not an overall effect of transient expression of baculovirus PTPs. Similar results were also observed for MdBV; MdBV encodes 13 different PTPs, but only PTP-H2 showed a pro-apoptotic effect in Sf21 cells [5,33]. SePTP2 protein possesses a C-terminal consensus sequence that is characteristic for MAPK phosphatases, which remove the phospho-residue from MAPKs. There are three major MAPK pathways: extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK) and p38. All three pathways have been reported to be involved in regulating apoptosis [34][35][36]. Therefore, SePTP2 may induce apoptosis by regulating MAPK activity in cells. Further studies are needed to identify the substrates of SePTP2 to confirm this.
In our study, deletion of the ptp2 gene severely reduced SeMNPV-induced apoptosis in S. exigua hemocytes, while the SeMNPV BV titer, virus infectivity and time to death were not affected upon deletion of ptp2. However, the OB yield was significantly higher from larvae infected with WT SeMNPV than from larvae infected with ∆ptp2 SeMNPV. A previous study on BmNPV showed that ERK-and JNK-dependent signaling pathways contribute to BmNPV virus yield; knocking down the Bombyx mori erk or jnk genes reduced the production of both OBs and BVs [37]. However, whether (and which) viral genes are involved in activating the host ERK-and JNK-pathways during a BmNPV infection is not known. SePTP2 may function as a MAPK phosphatase that regulates host MAPK pathways that eventually affect OB yield in infected hosts. Virus-induced apoptosis is normally regarded as a defense mechanism against virus infection; however, some viruses stimulate apoptosis in immune cells to maximize viral fitness and transmission [38,39]. For example, the influenza A virus (IAV) induces apoptosis in innate immune cells at the early stage of infection to subvert host immunity [39]. Similar results were also found in MdBV infected S. frugiperda larvae [5]. We hypothesize that the induction of apoptosis by SeMNPV PTP2 suppresses the S. exigua larval immune system and by doing so the infected larvae accumulate more OBs. These higher OB yields most likely benefit virus dissemination, and hence transmission.
It seems contradictory that viruses carry both anti-apoptotic and pro-apoptotic genes. Baculoviruses are known to produce anti-apoptotic proteins (e.g., P35 and IAP) to suppress or delay apoptosis [15]. For example, besides SePTP2, SeMNPV also encodes the anti-apoptotic proteins IAP2 and IAP3 [17,18]. However, anti-apoptotic and pro-apoptotic genes may function via different pathways. While anti-apoptotic genes counteract host-induced apoptosis (which interferes with viral replication in general), allowing the virus to replicate and eventually increasing virus progeny, pro-apoptotic genes might serve to suppress the host immune system by targeting specific cell types. Future studies should show whether the induction of apoptosis in S. exigua larvae is restricted to hemocytes.
Overall, we conclude that the SeMNPV ptp2 gene functions as a pro-apoptotic gene in cultured Sf21 insect cells, inducing mild apoptosis, and that the phosphatase activity of SePTP2 is needed for this process. Furthermore, we showed that the SeMNPV ptp2 gene is involved in inducing apoptosis in host hemocytes. Viral suppression of host immunity eventually contributes to a higher OB yield.
Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4915/10/4/181/s1, Figure S1: Transient expression of SePTP2 induced apoptosis in Sf21 cells, Table S1: Primers used in this study, Table S2: P values of the one-way ANOVA, Table S3: TCID50 values of BVs, Table S4: Outcome of the logistic regression analysis of the infectivity assays.