A Comparison of Growth and Development of Three Major Agricultural Insect Pests Infected with Heliothis virescens ascovirus 3h (HvAV-3h)

Ascoviruses are double-stranded DNA viruses that are pathogenic to lepidopteran hosts, particularly noctuid larvae. Infection of a larva is characterized by retarded growth, reduced feeding and yellowish body color. In this paper, we reported the growth and development of three major agricultural noctuid insect pests, Helicoverpa armigera (Hübner), Spodoptera exigua (Hübner) and Spodoptera litura (Fabricius), infected with Heliothis virescens ascovirus 3h (HvAV-3h). Using 10-fold serial dilutions (0 to 7) of HvAV-3h-containing hemolymph to infect S. litura larvae, we found no significant difference in larval mortalities from 0 to 103-fold dilutions; however, significant differences were observed at 104-fold dilution and above. Using a 10-fold dilution of HvAV-3h-containing hemolymph to infect H. armigera, S. exigua and S. litura larvae, we found that the growth and development were significantly affected. All infected larvae could not pupate; the survival times of treated H. armigera, S. litura and S. exigua larvae were significantly longer than untreated control larvae. Body weight showed significant difference between treated and untreated control group from day 1 after inoculation in H. armigera and S. exigua, but day 2 in S. litura. Additionally, food intake also showed significant difference between treated and untreated control group from day 2 after inoculation in H. armigera and S. litura, but day 3 in S. exigua.

Once in a laboratory condition, larvae infected by ascovirus are easier to recognize through checking the hemolymph that is milky-white in color, which is opposite to the transparent hemolymph in healthy larvae [7,8]. This infection symptom in larvae is distinguishable from other viruses, such as baculovirus.
The milky-white hemolymph is attributed to the formation of virion-containing vesicles [10,12]. These vesicles contain large enveloped virions with 130×400 nm in size. The virion shape is similar to baculoviruses but more variable, from bacilliform to ovoid or allantoid [7,13]. Indeed, the virion shape and the hemolymph vesicles aid in ascovirus identification [14].
The effects of AV infection on larval growth and development were studied using Trichoplusia ni, Spodoptera frugiperda, Helicoverpa zea, and Heliothis virescens larvae in the early 1990s [19]. Although these studies reported that AV-infected larvae were observed stunted growth and gained little weight before dying of the disease, it remains unclear whether this is a general phenomenon [5,8]. Therefore, the physiological changes of the larvae were evaluated based on the recently discovered HvAV-3h isolate in this work.
In order to understand how the body weight changes with food intake, and how the HvAV-3h affects larval growth and development, we selected Helicoverpa armigera, S. exigua and Spodoptera litura to test the body weight, food intake of the larvae. Our results provide information for further investigation of how HvAV-3h infection regulates the growth and development of different insects.

Insects and viruses
The S. exigua eggs were kindly provided by Dr. Yue-Qin Song of Henan University of Science and Technology, the H. armigera larvae were donated by Dr. Zhu-Dong Liu of Chinese Academy of Sciences, and the S. litura larvae were collected from a vegetable field (an experimental field of our laboratory) near Hunan Agricultural University, Changsha, China. The S. exigua larvae were reared on artificial diets following Song's method [20], and the H. armigera and S. litura larvae were reared on pinto bean-based diets [21]. The larvae of the three species were maintained in an incubator with a controlled temperature at 27±1°C, relative humidity at 70%, and a light:dark photoperiod at 14:10-h [20]. The insect adults were supplied with 10% honey solution.

Ascovirus titer determination
The concentration of HvAV-3h genomes in the hemolymph obtained from infected insects was estimated by adapting the end-point dilution method [22,23]. A stock of purified HvAV-3h DNA (20 ng/µl) [10] was 10-fold serially diluted with sterile water, a total of seven solutions were made (10 0 -to 10 6 -fold dilution). The HvAV-3h-containing hemolymph was also diluted similarly. To investigate the detective performance of both the purified DNA sample and the HvAV-3h-containing hemolymph, 1 µl each of DNA and HvAV-3h containing-hemolymph from each sample was added to PCR with a pair of ascovirus polymerase gene primers (polF:  5'-CCAGGATCACCAACACAC-3',  polR:  5'-GCTAGAGGATCGCTAACG-3') using the conventional Taq  enzyme based amplification reaction (94°C-1 min, (94°C-1 min,  55°C-1 min, 72°C-1 min) ×30 cycles, 72°C-10 min). After the reaction, the PCR products were analyzed by agarose gel electrophoresis. The PCR provided an all-or-none end-point at a very low DNA. Then the HvAV-3h genome copy was determined based on the comparison of HvAV-3h-containing hemolymph dilutions with HvAV-3h DNA dilutions that could not produce detectable PCR band in the agarose gel electrophoresis.
Here, we considered that each copy of HvAV-3h DNA represents an infectious unit. The titer of HvAV-3h-containing hemolymph is a function of HvAV-3h genome copy number in hemolymph and the hemolymph volume, where the genome copy number is defined as a ratio of the total copy numbers over the genomic size (164 kbp) [10], based on the average molecular weight of one base-paired nucleotide (660 g/mol) and Avogadro's constant (6.022×10 23 ) [24].

Dose-response assay
Serial dilutions (10-fold) of HvAV-3h-containing hemolymph were used to prepare eight inocula (10 0 -to 10 7 -fold dilution) with sterile water. Thirty third-instar (L3) S. litura larvae were used for each dilution, with thirty L3 larvae used as an untreated control (healthy hemolymph with corresponding dilutions). The whole experiment was done in triplicate. Inoculation was performed by pinning a proleg of each larva with a mini-pin tip contaminated with the respective HvAV-3h dilution, and the virus-killed larvae were collected each day to calculate the mortality rate of each dilution.

Survival time and molting period
A 10-fold dilution was chosen to assess the effects of HvAV-3h infection on H. armigera, S. exigua, and S. litura larvae. To each species, a total of 60 L3 larvae of each species were randomly chosen: 30 were inoculated with 10-fold dilutions of HvAV-3h, and the other 30 were inoculated with 10fold dilutions of hemolymph from healthy larvae. Another two repeats were included to confirm the experiment, and a blank control without larvae was also included to calculate the evaporation rates.
After inoculation, all the tested larvae were transferred to test tubes (2×10 cm) containing a 1-cm 3 -diet cake. In order to analyze the survival time and ecdysis between groups, the larvae were monitored every day until die (treated larvae) or adults' emergence (untreated control larvae). Molting was recorded by checking exuviation and head capsule of the larva. Every day, the dead individuals were collected to calculate survival time and mortality rate.

Larval feeding and growth
Following the above, each larva was weighted separately using an electronic balance, and the diet cakes were weighted after removing the feces with a camel brush. The diet cakes were discarded and replaced with a fresh one every 3 days. The blank control containing only diet cakes were weighted every day as well. The evaporation rate was then calculated and used in the data. So, daily data of body weight and food intake was recorded for each larva in the experiment.

Statistical analyses
Larval mortality data were arcsine square-root-transformed to meet homoscedasticity assumptions [25,26], and the data of larval body weight and food intake were checked with the test of homogeneity. Then a one-way analysis of variance (ANOVA) was used to analyze the mortalities caused by different dilutions (with a post-hoc analysis) and the data of larval body weight and food intake between treated and untreated control groups [27]. Mortalities of the three species caused by HvAV-3h were checked by non-parametric test. Survival curves of H. armigera, S. exigua and S. litura larvae as well as median survival time (ST 50 , referring here as larval stage time) were obtained through Kaplan-Meier method (SPSS) [28], untreated control larvae data accounted from the day of inoculation to adults emergence.

HvAV-3h titration
Undiluted HvAV-3h-containing hemolymph did not result in PCR amplification, where 10 1 -and 10 2 -fold dilution showed discernible PCR bands ( Figure 1B). There was no detectable band in the lanes of 10 3 -fold dilution and above. This result indicated that undiluted HvAV-3h-containing hemolymph (probably including some PCR inhibitors) could not be directly used as template in a PCR for the viral DNA amplification, and the 10 2 -fold diluted HvAV-3h-containing hemolymph (maybe the PCR inhibitors being diluted away) could be used directly as template.
Using the purified HvAV-3h DNA dilutions as templates, the PCR had a detection limit at 0.2 ng ( Figure 1A), whereas the HvAV-3h-hemolymph dilution had a detection limit at 10 2 -fold ( Figure 1B). The result suggested that the viral DNA concentration in the 10 2 -fold purified HvAV-3h DNA sample was similar to the 10 2 -fold dilution of HvAV-3h-hemolymph. Based on this result, the titer of HvAV-3h-containing hemolymph was 1.1×10 11 genome copies/ml according to the calculation showed above. The titer was considered be a relative concentration, and was used for the dose-response experiment.

Survival time and molting period
To compare the immature stage (includes egg stage, larval stage, pre-pupal stage and pupal stage), the life spans of experimental individuals were recorded from the inoculating time point to adults' emergence or larval death. Median survival time (ST 50 ) of the infected H. armigera, S. exigua and S. litura larvae showed significant difference between treated and untreated groups ( Table 2). Survival curve of untreated H. armigera larvae dropped rapidly during day 14 to day 16 postinoculation. The curve indicated that nearly 70 percent untreated larvae finished larval stage (ending in adults'   Figure 3A). The curves of other two insect species untreated control S. eixgua larvae dropped during 8-12 days post-inoculation, suggesting that the larvae ceased feeding and finished pupating within 8-12 days post-inoculation. The curve of treated larvae declined gradually from day 4 to day 16 postinoculation. A straight line and a slight drop during day 16 to day 19 indicated the death of last HvAV-3h infected larvae. The two curves differed significantly (Log Rank χ 2 =19.96; df=1; P<0.0001; Figure 3B). Untreated (control) S. litura larvae pupated mainly between days 9 to 16 post-inoculation. The curve of treated larvae, on the other hand, dropped from day 9 to day 20, and the decrease rate (%) was similar among days. The two curves differed significantly as well (Log Rank χ 2 =20.24; df=1; P<0.0001; Figure 3 C).
We found that the longer larval stage in H. armigera was coherent to the longer molting period. To further test this explanation and to check whether the explanation could still be adjusted in S. exigua and S. litura, we analyzed days between 4 th instar (L4) and 5 th instar (L5) of the significantly longer-living larvae in the three species, the results showed that the days between L4 and L5 of treated S. litura and S. exigua larvae was similar to the untreated cohorts, no significant differences between treated and untreated control larvae (Table 3). On the other hand, the treated H. armigera larvae demonstrated increased length in instar duration (twice as many days as the untreated group in average), and differed significantly. To further understand the changes of larval growth and development among different species, we recorded data of the feeding and mass of larvae and made comparison between treated and untreated control groups.

Larval body weight and feeding
The changes in body weight and food intake (evaporation rate 4.0% was taken into consideration) integrated the impact of HvAV-3h on larval growth and development. Untreated control larvae completed development and ceased feeding by day 7 post-inoculation in H. armigera and S. exigua, while S. litura in day 10 depending on species. Accordingly, the body weight and the food intake of untreated control larvae grew significantly fast with the time. On the other hand, treated larvae grew more slowly and failed to pupate. The body weight of treated larvae slightly increased within a few days after inoculation, and did not continue to go up. The food intake of treated larvae changed a little with the time. Based on the significant difference analysis, the comparison between untreated and treated group was further investigated.
No difference was observed in the larval body weight of H. armigera treated and control groups at the time of inoculation (F 1, 177 =0.97, P=0.3257; Figure 4A). From day 1 to day 5, the differences in body weight increased between the untreated larvae, that develop normally, and the treated larvae, that gained little weight. The differences between the two groups were significant from day 1 ( Figure 4A) Figure 5A) and S litura ( Figure  6A). Although, in S. exigua, no difference in body weight was detected between the treated and untreated group (F 1, 160 =0.05, P=0.0828; Figure 5A), the differences between the treated and untreated control group continued to increase as incubation time extended (day 1: F  Figure  5A). In S. litura, the body weight of untreated control rose from 20 mg to 200 mg within the first four days, whereas the growth of treated group (rose from 20 mg to nearly 80 mg) was slow. Therefore, larval body weight showed a significant difference between the two treatments from day 2 to day 4 (day 2: F 1, 176 =4.57, P=0.0337; day 3: F 1, 176 =61.02, P<0.0001; day 4: F 1, 176 =148.63, P<0.0001; Figure 6A). This difference continued significantly from day 5 to day 9 (day 5: F The differences in growth are correlated with differences in food consumption for the three species. The data are summarized in Figures 4B, 5B and 6B for the three species H. armigera, S. exigua and S. litura, respectively. In H. armigera, the amount of food consumed by treated larvae, on the other hand, with the exception on the 1st day (F 1, 177 =6.47, P=0.0119), was significantly lower than the untreated control

Survival curves of three insect pests larvae obtained through Kaplan-Merier with log-rank test (SPSS). Helicoverpa armigera (A), Spodoptera exigua (B) and
Spodoptera litura (C). Control larvae were detected from inoculating time to pupating time (the end of larval stage). HvAV-3h treated larvae were detected from inoculating time to death. Treated larvae that pupated were not included in the plot.  Figure 4B). Untreated control larvae were gluttonous before pre-pupation (300 mg to 500 mg in average) but treated larvae showed no increment of food intake during this period of time. In S. exigua, the growth of food intake of untreated control group was related to body weight from day 1 to day 4 but not day 5 (control larvae digested less food in day 5 before pupation). The amounts of food consumed by treated larvae were not significantly different in day 1 and day 2 (day 1: F  Figure 5B). In S. litura, the food intakes of both the treated and untreated groups were in accord with body weight. The food intake of S. litura demonstrated a significant difference compared to untreated larvae on day 2 to day 9 (day 2: F  Figure 6B).

Discussion
The effects of ascovirus infection on noctuid larval physiology, including extended larval stages and difficulty in molting, have been documented in the literature [11]. However, the responses of different noctuid species to AV infection have not been systematically investigated. Accordingly, we assessed the different physiological responses to HvAV-3h infection in three noctuid species, providing information for further understanding of the possible physiological changes in larvae after AV infection.
HvAV-3h viral concentration was estimated from the total number of genomes using the end-point dilution method. The result obtained (1.1×10 11 infectious genomes/ml) are not very different from those previously published for other ascoviruses with the titers of 8.0×10 7 to 1.5×10 8 vesicles/ml in ascovirusinfected T. ni larvae and 5.0×10 8 vesicles/ml in S. frugiperda at day 5 post-inoculation [11]. The difference of viral concentration in the hemolymph was probably due to the fitness of the virus in the host. An end-point dilution approach coupled to PCR was used to determine virus concentration by the first time. But in fact, whether all virions are infectious or not was remained un-tested. It is possible some HvAV-3h form replication defective virus during cell infection, because missing of DNA polymerase in budded virus leading to defective baculovirus has been detected [29]. If this is true, our titer of HvAV-3h might be over-estimated. Also, although we determined the concentration of HvAV-3h of an inoculum, the variance of volume that entered the insect was remained undetectable.
We showed that low HvAV-3h doses (dilutions of 10 6 -10 7 of the stock) still caused mortality in the tested larvae. This suggests that a few viral particles would be enough to colonize a host ( Figure 2). Our data support earlier reports of low dose DpAV-4a caused mortality in Acrolepiopsis assectella [16,30]. Noctuid larval mortalities caused by different concentration of virus may vary among isolates. Here, the mortalities of other reported ascovirus isolates and S. litura induced by HvAV-3h were compared. The larval mortality (L3 and L4) caused by Sf 82-126 (then proven to be SfAV-1a) ranged from 90.0% to 96.0% at a concentration from 10 to 10 5 vesicles inoculated per larva without significant difference. The mortality obtained from the T. ni isolate (now refers to as TnAV-2a) [19], was lower, at 90.0-93.3% and varying with the vesicle content (10-10 5 vesicles inoculated per larva). Different AVs showed no significant difference in mortalities with different concentration of AV vesicles. Mortalities of S. litura larvae cause by HvAV-3h showed no significant difference among undiluted virioncontaining hemolymph, 10-, 10 2 -, and 10 3 -fold dilution. Our result supported the previous studies [19].
The three species, H. armigera, S. exigua and S. litura were with high susceptibility with HvAV-3h, and the mortalities of the three species caused by HvAV-3h did not differ significantly from each other. However, the mortalities caused by other HvAV-3 isolate varied in several noctuid larval species. To date, only HvAV-3g (isolated from Indonesia) has been evaluated for susceptibility to noctuid insects among HvAV-3 species [15]. HvAV-3g infections followed different patterns in five noctuid species, Pseudoplusia includens, H. virescens, T. ni, S. frugiperda, and also S. exigua. Except for H. virescens, the virus infected all these species. The mortality of S. exigua was the highest, at 100.0% in larval stage. In P. includens, mortalities were observed in the larval stage (28.3%) and the pre-pupal stage (46.7%). In T. ni, mortalities were also found at 45.4% in the larval stage, 13.6% in the pre-pupal stage, and in the pupal stage (41.0%). In S. frugiperda, mortalities were 86.2% in the larval stage and 13.8% in the pupal stage. The results were quite different from ours. The two Spodoptera species in our study were affected by HvAV-3h only in larval   stage, but S. frugiperda larvae (belongs to Spodoptera species) died of HvAV-3g in pupal stage [15]. The ST (survival time) ranges for S. litura, S. exigua caused by HvAV-3h were smaller than those reported for other virusinfected species (8-36 d for TnAV-2a, 20-50 d for SfAV-1a in L3 larvae and 13-37 d in L4 larvae) [6]. The different reactions among species may account for different infection rates, virion production and possibly different tissue tropism by different isolates. So, the relative longevity of H. armigera (6-36 d) might be caused by the virus replicating mainly in the tracheal epithelium tissue and epidermal cells, but not replicating in vital tissues [6]. Because the previous study did not address the virulence of HvAV-3a in other species, we cannot say in which tissue the virus replicates.
Two insect hormones, juvenile hormone and ecdysone regulate larval development, the former maintains the larval status and the latter promotes ecdysis [31][32][33]. Larval molting is another index reflecting changes in hormone balance induced by HvAV-3h or even reflecting larval metabolism during growing stages [19,34,35]. HvAV-3h is able to alter larval molting period, this is another important feature of AV infection in larvae [14,19]. In H. armigera, in contrast to the untreated control larvae, most H. armigera larvae showed an extended larval instar period, over 3 days in average ( Table 3). Infection of HvAV-3h may be similar to that caused by polydnaviruses, which reduce ecdysone synthesis, thus prolonging larval development [15,35]. Based on our experiments, HvAV-3h infection in H. armigera larvae may either produce a viral protein(s) that sequesters ecdysone to extend the instars or the infection reduces the production of ecdysone, thus leading to long period of instar development time. The validity of this explanation, however, remains to be tested.
The most conspicuous physiological change in noctuid larvae infected with HvAV-3h was the body weight and food intake over time comparing with control larvae. Body weight is a fundamental indication involving larval growth and development [36][37][38]. Specifically, the body weight in infected larvae rose very slowly and showed the significant difference with the untreated group from the first/second day postinoculation. The significant difference of food intake between groups, on the other hand, was observed from the second (third) day post-inoculation. Our results supported previous studies [14,19].
Over all, HvAV-3h has pronounced physiological impact on H. armigera, S. exigua and S. litura, and the three insect pest species cause significant crop loss in agriculture. Thereafter, our results suggest that HvAV-3h could play a role in the control of pest populations based on its function during ecological regulation, in other words, an important factor that help to maintain the stabilization and equilibrium of the ecosystem. Firstly, AVs, including HvAV-3h, can be transmitted in the field from ascovirus-infected larvae to healthy larvae by parasitic wasps during oviposition [10,17,19,[39][40][41]. Secondly, dose-response results indicated that successful infection could be obtained even when larvae are inoculated with low doses of virus; so, transmission via parasitoid wasps would not be limited to high concentration of virus [16]. These two reasons make transmission of HvAV-3h via parasitoid wasps an attractive hypothesis to be tested in the future. Furthermore, we found that HvAV-3h has a more rapid killing speed compared to other reported AVs [11,15,19] with the ability to immediately reduce larval feeding, weight and thus inhibit metamorphosis, namely, larvae infected with HvAV-3h would not have gluttony, as a result, the crop yield loss would be less with infected larvae. However, the technical barriers such as the viable infection method and the viral durable capability should be figured out before HvAV-3h could be put into use in the agricultural condition.