Changes in Bemisia tabaci feeding behaviors caused directly and indirectly by cucurbit chlorotic yellows virus

Cucumber (Cucumis sativus L. cv. Bojie-107) plants were grown in pots (d = 12.5 cm) in a greenhouse under a photoperiod 16:8 LD, temperature 26 ± 1 °C, and relative humidity 70 ± 0.5%. To obtain viruliferous cucumber plants, Agrobacterium tumefaciens-mediated CCYV clones were used to inoculate cucumber plants at one true-leaf stage [47]. About 25 days later, the infection status of cucumber plants was determined based on the symptom of yellowing and chlorotic leaf spots. Infection was further confirmed by reverse transcription-polymerase chain reaction [39]. All plants were maintained in separate insect-proof cages (60 cm × 40 cm × 80 cm) in greenhouse under the same conditions. Cucumber plants at 4 true-leaf stage were used for all experiments.

B. tabaci biotypes B and Q were maintained on non-viruliferous cucumber (Cucumis sativus L. cv. Bojie-107) plants for many years in insect-proof cages under conditions as above. The purity of biotypes B and Q populations was monitored every 1–2 generations by using the biomarkers of the mitochondrial cytochrome oxidase I (mtCOI) genes [48, 49].

Non-viruliferous and viruliferous B. tabaci colonies were established by transferring about 300 pairs of adult males and females of biotypes B and Q from the laboratory populations into insect-proof cages each with two virus-free or CCYV-infected cucumber plants, respectively. Non-viruliferous and viruliferous B. tabaci colonies were maintained for 2 generations in a greenhouse under the conditions described previously. Starting from the third generation, we randomly selected newly emerged B. tabaci male and female adults from each colony for use in the experiments.

A 4-channel direct-current EPG system (Wageningen University, the Netherlands) was used to monitor the feeding behaviors of B. tabaci. Prior to a recording, a gold wire (1.5 cm long and 12.5 μm in diameter) was attached to the pronotum of an insect using a drop of water-based silver glue. Each wired insect was starved for ca. 20 min before connected to the Giga-4 probe input and placed onto the abaxial surface of the third leaf of cucumber plant. Six hours of EPGs were continuously recorded for each replicate, which was defined as one adult B. tabaci feeding on one plant. All the recoding experiments were finished in an electrically grounded Faraday cage to block electric fields. All experiments were carried out in a quiet room under temperature 26 ± 1 °C, relative humidity 70 ± 0.5%, and 1000 lx artificial light. EPG signals were digitized with a DI-710-UL analogue-to-digital converter (Dataq Instruments, Akron, OH, the USA), and the output was acquired and stored with Stylet+ (d / a) for Windows software (Wageningen University, the Netherlands), and data were analyzed with this software after data conversion.

EPG waveforms were categorized as previously described [7, 50]. Four distinct waveforms were identified in this study: pathway [C, showing insect stylet activities from epidermis to the phloem, including intercellular penetration and sheath salivation, as well as, if occur, penetration difficulties (F waveform) and xylem sap ingestion (G waveform)]; potential drop (intracellular puncture) [pd], and the phloem phase salivation into a sieve element [E1] and ingestion of sieve element sap [E2]. The time from the start to the end of each waveform was recorded and exported by using Stylet+ software. Based on the information described above, we selected 6 non-phloem phase variables and 8 phloem phase variables for analysis and comparison of B. tabaci feeding behaviors of (1) non-viruliferous biotypes B and Q feeding on non-viruliferous cucumber plants, (2) viruliferous biotypes B and Q feeding on non-viruliferous cucumber plants, and (3) viruliferous biotypes B and Q feeding on viruliferous cucumber plants.

The direct effects of CCYV on feeding behaviors of B. tabaci biotypes B and Q were obtained by comparing data obtained with viruliferous B. tabaci (data from viruliferous B. tabaci on non-viruliferous plants) with data under control conditions (non-viruliferous B. tabaci on non-viruliferous plants).

The indirect effects of CCYV on feeding behaviors of B. tabaci biotypes B and Q were defined as the effect of CCYV on vector insects via viruliferous plants. Specifically, indirect effects were obtained by comparing data from viruliferous B. tabaci feeding on viruliferous cucumber plants (VW vs. VP) with data obtained with viruliferous B. tabaci feeding on non-viruliferous cucumber plants (VW vs. NVP).

Table 3 lists the multivariate statistics of comparative analyses on combinational EPG variables to examine possible interactions among various factors. First, EPG variables were combined to examine the effect of one factor in each analysis. For example, the ‘Biotype’ combination compares B. tabaci biotypes B and Q, without considering their sexes, insect infestation status and plant infestation status. B. tabaci biotypes exhibited major differences in 5 out of 14 variables, including total duration of pathway (Table 3, Variable 3), time to phloem from 1st probe (Table 3, Variable 5), total number of sap ingestion (Table 3, Variable 10), total duration of sap ingestion (Table 3, Variable 11) and total number of salivation after 1st sap ingestion (Table 3, Variable 12). Males and females exhibited differences in 2 out of 14 variables, including the total number of pathway (Table 3, Variable 2) and percentage of phloem phase (salivation + ingestion) (Table 3, Variable 14). Insect infestation status showed significant effect on 3 non-phloem variables [total number of pathway, total number of potential drop (intracellular puncture) and total number of probes before phloem] and 2 phloem variables (total duration of 1st salivation and total duration of salivation) (Table 3, Variables 2, 4, 6, 7 and 9). These changes in feeding behaviors could increase the probability of virus inoculation by viruliferous B. tabaci on non-viruliferous plants. Plants infestation status had significant effect in 8 out of 14 variables, including the duration of 1st probe (Table 3, Variable 1), total number of pathway (Table 3, Variable 2), total number of potential drop (intracellular puncture) (Table 3, Variable 4), time to phloem from 1st probe (Table 3, Variable 5), total duration of 1st salivation (Table 3, Variable 7), total duration of salivation (Table 3, Variable 9), total duration of salivation after 1st sap ingestion (Table 3, Variable 13) and percentage of phloem phase (salivation + ingestion) (Table 3, Variable 14).

Second, EPG variables were combined to examine the combinational effect of two factors in each analysis. As shown in Table 3, there was a significant combinational effect of biotypes and sexes on the duration of 1st probe (Table 3, Variable 1), time to 1st phloem from 1st probe (Table 3, Variable 5), total number of probes before phloem (Table 3, Variable 6) and percentage of phloem phase (salivation + ingestion) (Table 3, Variable 14). The combinational effect between biotype and insect infestation status was similar to the combinational effect between biotypes and plant infestation status, both showed a significant effect on total number of salivation (Table 3, Variable 8), total number of sap ingestion (Table 3, Variable 10) and total number of salivation after 1st sap ingestion (Table 3, Variable 12). The combinational effect between sexes and insect infestation status was on total duration of sap ingestion (Table 3, Variable 11), total duration of salivation after 1st sap ingestion (Table 3, Variable 13) and percentage of phloem phase (salivation + ingestion) (Table 3, Variable 14). The combinational effect between sex and plant infestation status was on total number of sap ingestion (Table 3, Variable 10), total number of salivation after 1st sap ingestion (Table 3, Variable 12), total duration of salivation after 1st sap ingestion (Table 3, Variable 13) and percentage of phloem phase (salivation + ingestion) (Table 3, Variable 14).

Combinational effects among three factors were also analyzed. The combinational effects among biotypes, sexes and insect infestation status were on the percentage of phloem phase (salivation + ingestion) (Table 3, Variable 14). The combinational effects among biotypes, sexes and plant infestation status were on total duration of salivation (Table 3, Variable 9), total number of sap ingestion (Table 3, Variable 10), total number of salivation after 1st sap ingestion (Table 3, Variable 12) and total duration of salivation after 1st sap ingestion (Table 3, Variable 13). Among the analyzed factors, CCYV on either insects or plants resulted in remarkable effects on B. tabaci feeding behaviors than other factors.

In this study, we found that the number of probes and the duration of phloem salivation significantly increased for both B. tabaci biotypes after individual vector insects were carried with CCYV. Since the increased number of probes and longer duration of phloem salivation can increase the spread of semipersistently transmitted viruses [53], our data suggest that CCYV can directly increase its vector host ability to spread the virus. This observation is consistent with previous reports that B. tabaci with TYLCV were more restless [7] and had more attempted probes and phloem salivation on virus-free plants [5]. Insect vectors of carrying CCYV also increased the frequency of insect feeding-site translocation coupled with larger number of short feeding bouts observed with B. tabaci biotype Q, resulting in a greatly increased rate of virus inoculation [13, 54]. Based on our observations that biotype Q was affected by CCYV to a greater degree than biotype B, we speculate that virus infection may result in a greater increase in the ability of biotype Q to spread this virus than biotype B on cucumber plants. This speculation is also supported by a previous report from our group, which has demonstrated that CCYV exerts stronger effects on feeding behaviors of biotype Q than those of biotype B on cotton plants [39]. Shi et al. [55] demonstrated biotype Q transmits Tomato chlorosis virus (ToCV), a semipersistently transmitted crinivirus virus, in tomato more efficiently than biotype B. Infected vector insects by another virus, TYLCV, also result in biotype Q with stronger ability of spreading virus than biotype B [13]. Currently, biotype Q has become the prevalent strain in the field in most regions of China due to heavy use of insecticides [56]. The replacement of biotype B by biotype Q as the prevalent strain in the field is coincidently with the rapid spread of CCYV in China. Further research is needed to determine whether the increase in the frequency of biotype Q is responsible for the rapid spread of CCYV in the field.

CCYV not only exerts differential impacts on the feeding behaviors of different B. tabaci biotypes, but also influences the feeding behaviors of males and females in biotype- and plant-dependent ways. In this study, we found that CCYV directly increases non-phloem probing and phloem salivation more on females than on males of biotype Q, but directly increased phloem salivation more on females than males of biotype B when B. tabaci feeds on cucumber plants. Previously, our group has found that CCYV enhances probing and saliva secretion of males more than females of both biotypes B and Q [39]. Ning and collaborators [57] also found TYLCV-infected biotype Q females were more efficient in transmitting virus than corresponding males. The exact factors that influence this complex relationship among CCYV, B. tabaci, and plant species and its biological implications remain to be determined. One possible reason for different impacts of CCYV on males and females feeding on different plant species is that host plants for B. tabaci and CCYV are not exactly the same. B. tabaci can feed on a wide range of plant species whereas CCYV has a narrower plant host range. For example, cotton is a very common host for B. tabaci, but is a non-host for CCYV. Differences in feeding behaviors of TYLCCNV-infected B. tabaci on TYLCCNV host tobacco plants and its non-host cotton plants have been also reported by other researchers [8]. During the long course of coevolution among insect vectors, virus, and plants, it is not surprising that a complex relationship is formed for best adaptation among the interactive species. In addition, CCYV increased attempt potential drops (pd) (intracellular probing) of biotype Q males and females. The intracellular probing represents stylet puncturing into the plant tissue cells and tasting the cytoplasm often related to the spread of non-persistent viruses [58], while ToCV can be transmitted when B. tabaci performed 6 pds (intracellular probing) [53], and intracellular probing (pd waveforms) may also play an important role in CCYV inoculation.

The indirect impact of viruliferous cucumber plants on B. tabaci feeding behaviors may also enhance viral transmission. Consistent with this possibility, we found that viruliferous cucumber plants could shorten phloem salivation of both B. tabaci biotypes, suggesting that viruliferous cucumber plants became more susceptible to B. tabaci feeding. Some previous studies have suggested that plant virus can improve the quality of resource for vectors by suppressing the jasmonic acid (JA) defense pathway [59‐61]. With weaker plant defense and better nutrients, insect vectors are more likely to be attracted to those viruliferous plants with better performance [12, 62]. B. tabaci feeding on TYLCCNV-infected plants have been found with reduced detoxification activity to reduce physiological cost, and can access a more balanced nutrition [63]. In contrast, ToCV-infected tomato reduced nymphal viability and prolonged duration of nymphal stage of biotype B [54]. We also observed differential impacts of CCYV on different B. tabaci biotypes indirectly via viruliferous cucumber plants. For example, biotype Q feeding on viruliferous cucumber plants had a shortened phloem sap ingestion. On the other hand, biotype B feeding on viruliferous cucumber plants had prolonged phloem sap ingestion. The shortening of phloem sap ingestion of biotype Q and prolonged duration of phloem sap ingestion of biotype B again suggest that biotype Q may possess greater ability for spreading virus when B. tabaci have already been infested with CCYV and biotype B may need more time to acquire virions from viruliferous cucumber plants. This speculation is based on the fact that both viruliferous biotypes B and Q became restless and urgent to transmit the virus out from their bodies. Our observation is consistent with previous reports that biotype Q has stronger ability to spread virus [52].

CCYV also had different impacts on feeding behaviors of males and females of both biotypes B and Q. For example, biotype B males had more non-phloem probing and feeding bouts than biotype B females on viruliferous cucumber plants, suggesting that biotype B males may contribute more to spread virus than those females in the field. On the other hand, biotype Q females were more resilient to the reduction of phloem sieve ingestion caused by viruliferous cucumber plants than biotype Q males. A previous study has reported that aster leafhopper, Macrosteles quadrilineatus, females spend more time feeding than males, and are also generally larger than males in order to obtain more nutrients needed to support ovarial development [64]. Pan and collaborators [65] have found that the endosymbiont Hamiltonella sp. is higher in TYLCV-infected biotype B and Q females than in males among field populations. Endosymbiotic bacteria may be another factor that affects virus transmission of B. tabaci males and females.

In this study, we were unable to examine non-viruliferous B. tabaci feeding on viruliferous cucumber plants. The difficulty to carry such an assay is that the time for B. tabaci to obtain saturated CCYV is only six hours. Within such a short time the stylet is hardly to have reached the phloem to start feeding. Because of this difficulty, the indirect effector of viruliferous cucumber plants on B. tabaci feeding behaviors was estimated using data obtained with viruliferous B. tabaci feeding on CCYV-infected cucumber plants. Therefore, the indirect impact of viruliferous plants on B. tabaci feeding behavior may have been underestimated in our study.