Cold Disinfestation for ‘Red Globe’ Grape (Rhamnales: Vitaceae) Infested With Drosophila suzukii (Diptera: Drosophilidae)

Abstract The spotted wing drosophila, Drosophila suzukii Matsumura, which is widely spread in the main soft-skinned fruits production areas in China, presents a threat to importing countries. In order to develop a phytosanitary cold treatment measure for preventing the movement of this drosophila fly, cold tolerance of six immature life stages of D. suzukii was compared followed by time-mortality and large-scale confirmatory tests on the most tolerant stage in grape fruit. Egg was defined as the most cold-tolerant stage by comparing the mortality of all the immature stages (egg, first, second, and third instars, early and late pupa) treated at 0 and 2°C. The minimal lethal time (LT) for 99.9968% mortality (95% confidence level [CL]) estimated by the probit model was 10.47 d at 0°C and 11.92 d at 2°C, respectively. Hence, 11 d (at 0°C) and 12 d (at 2°C) were chosen as the target time to conduct the confirmatory tests. No survivors were found among the estimated 50,385 and 57,366 treated eggs, which resulted in the efficacy of 99.9941 and 99.9948% mortality (95% CL) at 0 and 2°C, respectively. Our study suggests a technical basis for cold disinfestation on D. suzukii in cage-infested Chinese ‘Red Globe’ (Vitis vinifera L.) grape, which could provide flexible phytosanitary treatment for control of D. suzukii in the international trade of grape.

after being treated (Aly et al. 2017). Third-instar feeding larvae were more cold-tolerant than third wandering larvae (Jakobs et al. 2017). Some of feeding larvae were still alive after 72-h exposure at 0°C, while wandering larvae all died. However, feeding larvae did not survive longer than 120 h (Jakobs et al. 2017). For phytosanitary cold treatment, Kim et al. (2018) studied the cold tolerance of D. suzukii fed on artificial diet. The result of time-response tests suggested that 8-d cold exposure at 1°C could be used for phytosanitary treatment of 'Campbell Early' grape.
Vitis vinifera L. (both table and wine grapes) are suitable hosts for D. suzukii and can be heavily infested in laboratory conditions (Baser et al. 2018, Shrader et al. 2019. The susceptibility of grape varieties to D. suzukii oviposition was related to penetrating resistance and Brix (Burrack et al. 2013, Ioriatti et al. 2015, Baser et al. 2018. 'Red Globe' grape with lower penetrate force and lower sugar content than other cultivars had certain susceptibility to infestation (Baser et al. 2018). In China, grape is one of the four major fruits, and the quantity of grape exported has gradually increased in recent years (Zhang and Li 2016). However, grape might be infested with D. suzukii, which will pose a potential threat to international trade. In addition, although grapes are usually stored and transported at low temperatures (Li et al. 2015, Wang et al. 2016, postharvest cold treatments have never been performed on 'Red Globe' grape for quarantine purposes. Therefore, it is necessary to develop phytosanitary cold treatment against D. suzukii to promote the export of grapes in China. Thus, the objectives of this study were to find out the most tolerant stage to cold and determine the minimum time required for probit 9 mortality (a mortality of 99.9968%) of D. suzukii in 'Red Globe' grapes to ensure quarantine security (NAPPO 2011).

Test Insects
The D. suzukii population was collected in an orchard near Yunnan Agricultural University, Kunming, China, from March to April 2017, and newly collected wild adults were introduced into the population once a year. Larvae and adults were reared together in insect cages (30 by 40 by 40 cm) with an artificial diet (wheat bran, sucrose, Brewer's yeast ≈ 9:3:1) modified from the larvae diet of papaya fruit fly, Bactrocera papayae (Liu et al. 2017). All stages were kept in the rearing room at 26 ± 1°C, 50-70% RH with the photoperiod of 14:10 (L:D) h.

Grape Infestation
Table grapes (cultivar of 'Red Globe') were bought from the local market near the Chinese Academy of Inspection and Quarantine, Beijing, China. Pesticides were not applied directly on the fruit and there were no pesticide residues. To prevent microbial contaminations, grapes with similar size, hardness, and color were washed with distilled water, dried at room temperature, and stored in the cold chamber at 0°C. For each time of egg collection, the grapes were taken out and warmed up to room temperature, and a longitudinal cut was made with a sterilized scalpel (Kim et al. 2018), and then the grapes were put into the cage for D. suzukii egg laying for 24 h. Every three to four infested grapes (40 ± 5 g for each) were put into one glass jar (0.42 liter) with filter paper at the bottom and covered with gauze to prevent contamination. Each jar was as one replication in the cold treatment and each treatment had three replications (three jars).

Developmental Time of D. suzukii in Grapes
The infested grapes in jars were placed in the rearing room for the development of the D. suzukii eggs. When larvae began to pupate, wheat bran was put under the grapes to prevent the larvae and pupae from drowning. Stages were checked every 12 h in the first 2 d and then at an interval of 24 h. At each time point, grapes were removed from three jars and dissected to count the number of each stages under a microscope.

Cold Tolerance Test
To determine the most tolerant stage against cold, grapes infested with different immature stages based on above rearing experiment of D. suzukii were treated at 0°C (0.027 ± 0.168) for 12 h and 2°C (2.034 ± 0.181) for 24 h via three replicates for each stage. Fruit pulp temperatures were recorded every 90 s by a temperature recorder (1586A SUPER-DAQ Precision Temperature Scanner, Fluke Co., WA) with six probes (Pt100, Chong Qing Well Co., Chongqing, China). All the probes were calibrated at the beginning and the end of each trial with a certified mercury glass thermometer. The starting time was recorded as all probes reached the target temperature. Twelve or 24 h later, the infested grapes were transferred to the rearing room for 4 d and the mortality of treated and controlled insects was checked under a microscope. These treatments were repeated three times each.
The criteria for mortality assessment in this research study were as follows: eggs: blacking or unhatched; larvae: no movement when prodded with a blunt probe; pupae: blacking, half-eclosion, deformation of eyes and wings.

Lethal Time Estimates
The grapes containing 1-d-old eggs, which were defined as the most tolerant stage after the cold tolerance test, were kept into the refrigerator chamber to be cold treated at 0 or 2°C. The core temperatures of the infesting grapes were monitored with six probes. Treatment was deemed to start after all the fruits had reached the target temperature. The grapes were transferred to the rearing room every day till 10-d treatment at 2°C and on the day of 0.5, 1, 1.5, 2, 3, 4, 5, 6, 7, and 8 for 0°C treatment. The mortality of all treatments and controls was checked 4 d later by using the methods mentioned above. The temperature monitored in the treatment was 0.0 ± 0.2 and 2.0 ± 0.2°C (mean ± SE) for the treatment at 0 and 2°C, respectively. These treatments were repeated three times each.

Confirmatory Test
For each trial of the confirmation test, grapes were put into the adult cages which contained about 2,000 adults from one batch of the colony and kept for 24 h. Then, the infested grapes were put into the cold refrigerate chamber for cold treatment at 0°C for 11 d and 2°C for 12 d according to the result of the efficacy test. The mortality was checked by using the method mentioned above. These treatments were repeated three times at each temperature.

Statistical Analysis
The mortality data were all adjusted with the Abbott's formula. The tolerance comparison and time-mortality tests were subject to one-way analyses of variance (ANOVAs) by SPSS (version 17, SPSS Inc., Chicago, IL) (Abbott 1925). Means were compared by Tukey's test. The time-mortality data of the most tolerant stage were analyzed by using the probit model and logit model (PoloPlus software program 2.0, LeOra Software, Berkeley, CA) to estimate the minimum lethal time for 99% (LT 99 ) and 99.9968% mortality (LT 99.9968 ) at the 95% confidence level (LeOra Software 2007).
The mortality proportion (1 − P u ) of the large-scale confirmatory tests was calculated by the equation (1): where P u is the acceptable level of survivorship, C is the confidence level (95%), and n is the number of treated D. suzukii with zero survivors (Couey and Chew 1986).

Developmental Time of D. suzukii in Grapes
The percentage of each immature stage at the specific time was calculated and listed in Fig. 1. The main developmental days (peak time) for egg, first, second, and third instars, and pupa stage were 0-2 (1), 1.5-2.5 (2), 2.5-3.5 (2.5-3), 3.5-7 (4-5), and 7-11 (9-11) days, respectively. The critical time with the similar percentage between stages was 2 d for egg and first instars, 3.5 d for second and third instars, and 7 d for third instars and pupa. Thus, the eggs developed in table grapes for 1, 2, 2.5, 4, 9, and 11 d were considered as the stage of egg, first, second, and third instars, early and late pupa, respectively, to conduct the following time-response test and large-scale confirmatory test. Each growth state observation of immature stage was done in triplication and each replication of eggs, first, second, and third instars included more than 100 test individuals. For pupae, each replication included more than 80 individuals.

Cold Tolerance Test
The mortality of each stage of D. suzukii was significantly different at 0°C for 12 h (F = 34.078, df = 5, 12; P < 0.001) and 2°C for 24 h (F = 36.354, df = 5, 12; P < 0.001) ( Table 1). The egg stage had the lowest mortality at 0°C (21.49 ± 1.25%) and 2°C (8.98 ± 0.41%) among the different immature stages indicating it was the most tolerant stage to cold. The first instar was the most sensitive (greatest mortality) at (81.46 ± 1.15%) at 2°C, while the second instar was the most sensitive at 0°C (72.65 ± 3.29%). Third instars, and early and late pupa had similar tolerance at 0°C as their developing time overlapped as shown in Fig. 1. Therefore, the sequence from most to least tolerance to 0°C was: egg > third instar, early and late pupa > first and second instars; to 2°C was: egg, third instars > early and late pupa > second and first instars. The 1-d-old egg was selected as the target stage for conducting the following dose-response test and large-scale confirmatory test.
Each growth state observation of immature stage was done in triplication and each replication of eggs and first instar included more than 100 test individuals. For second and third instars and early pupae, each replication included about 70 individuals, while late pupae contained more than 50 individuals.

Lethal Time Estimates
The mortality of 1-d-old eggs of D. suzukii during the time-response test increased significantly with the increase of time when treated at 0°C (F = 184.83, df = 10, 22; P < 0.001) and 2°C (F = 278.85, df = 10, 22; P < 0.001) for 8 d. The minimum days to 100% mortality were 6 d at 0°C and 7 d at 2°C, respectively (Table 2). Then, the data were subjected to probit analysis using the probit and logit model. The parameters including slope, heterogeneity (chi-square divided by degrees of freedom), LT 99 , and LT 99.9968 are presented in Table 3. The models were evaluated based on heterogeneity and LT 99.9968 by analyzing with the mortality of time-response test. The small value of heterogeneity obtained in the probit model for 0 and 2°C with log-transformed data (2.09-2.95) and the logit model without log-transformed data (2.34-2.73) indicated that the estimations matched the data. Both the LT 99 and the LT 99.9968 estimated by probit analysis (with log-transformation) were about 4 d longer than those estimated by the logit model (without log-transformation). The heterogeneity data in these two models were similar. Therefore, 7-and 11-d cold treatment at 0°C, and 9-and 12-d cold treatment at 2°C were all selected as the target lethal time for conducting the following confirmatory large-scale test (Table 4).

Confirmatory Test
Confirmatory tests were conducted to validate the estimated time to 99.9968% mortality of the 1-d-old eggs (the most tolerant stage in harvested grapes) of D. suzukii. For each trial, 238-539 grapes could collect more than 20,000 eggs. The temperature monitored during treatments was at 0.0 ± 0.2 or 2.0 ± 0.2°C (mean ± SE), respectively. For the treatment at 2°C for 9 d, eight living larvae were found in the treated grapes containing more than 24,500 eggs in the first trial, suggesting that the logit model was not appropriate. However, there were no survivors from a total of 57,366 eggs and 50,385 eggs treated at 0°C for 11 d and 2°C for 12 d, respectively (Table 4), indicating the probit model was validated. The disinfestation efficacy was 99.9948 and 99.9941% at 0 and 2°C, respectively.

Discussion
Cold treatment, which is considered a common method to eliminate pests in fresh fruits and other products, has been studied, analyzed, and incorporated into quarantine regulations for a number of years (Jessup et al. 1993, De Lima et al. 2011, Grout et al. 2011, Dohino et al. 2017, Fang et al. 2019. To develop the phytosanitary measure, the most tolerant stage should be compared and subjected to dose-response test and confirmatory test (IPPC 2007). Generally, pests in their natural habitat are optimal for the tolerance test (NAPPO 2011). In this study, developmental duration of the immature stages of D. suzukii was firstly observed in 'Red Globe' grapes to select the appropriate stages used in the cold tolerance test (Fig. 1). To obtain high levels of oviposition in grapes, a cut with a scalpel was made on the surface of 'Red Globe', according to the result of Shrader et al. (2019) that more flies emerging from injured grapes than intact grapes. Based on mortality estimates, the 1-d-old egg appeared to be the most cold-tolerant stage with the lowest mortality among different developing stages, both at 0 and 2°C (Table 1). The results are consistent with research on the model drosophila fruit fly, Drosophila melanogaster Meigen, whose eggs were the most tolerant in cold shock tests by assaying LT 50 over Means within a column followed by different letter a, b, or c were significantly difference (P < 0.05; Tukey's test). Means within a column followed by different letter a, b, or c were significantly difference (P < 0.05; Tukey's test).
−16 to 5°C for 1 h, followed by adults, pupae, and larvae (Jensena et al. 2007). However, Kim et al. (2018) reported that pupal stage of D. suzukii was more resistant than egg, and it was defined as the most tolerant stage with lab diet. Besides, to achieve complete mortality of D. suzukii immature stages reared on an artificial diet, the exposure durations required were 6 d at 1°C, and 8 d at 1.5 and 2°C (Kim et al. 2018). However, in our test, a few larvae emerged from the treatment at 2°C for 9 d. The reasons why our conclusions were not the same might be different infestation method, feeding conditions, temperature, fruit type, colony source, and the choice of mortality endpoint. In the present study, in order to get closer to the real-world environment, we chose to study cold response from host fruit of D. suzukii, rather than the lab diet, which might resulted in differences in fly physiology and development at low temperature. Furthermore, each stage used in cold-tolerate trial was actually a composite of multiple developmental stages, for example, 1-d-old egg containing approximately 92.0% egg and 8.0% first instars, early pupa containing 6.7% third instars, and 93.4% pupa. This condition was similar to the cold tolerance test of the Mediterranean fruit fly, Ceratitis capitata (Wiedemann), whose mixed stages were treated in host fruit (Armstrong et al. 1995, De Lima et al. 2011, Ware and Du Toit 2017. This could also lead to the different results of cold response in the same species. Usually, the probit model and logit model are used to estimate the minimal lethal time of 99.9968% mortality according to the time-response test (Follett et al. 2018, Al-Behadili et al. 2019, Fang et al. 2019). In the present study, four different models were selected separately for eggs at every temperature. Log on days with probit and no-log on days with logit were selected to estimate the LT 99 and LT 99.9968 as the best fitting models (Table 3). There was no survivor from a total of more than 50,000 eggs treated at 0°C for 11 d and 2°C for 12 d, respectively (Table 4). However, eight larvae emerged from the confirmatory tests when treated at 2°C for 9 d (based on the estimation of the logit model). The result suggested that the probit model was more suitable for analyzing the cold treatment of D. suzukii, which was similar to cold treatment of the C. capitata (Wiedemann), Bactrocera tryoni (Froggatt), B. dorsalis (Hendel), and B. zonata (Saunders) (De Lima et al. 2007, Hallman et al. 2013, Myers et al. 2016). In addition, the minimal time for probit 9 mortality tested in the large-scale confirmatory tests may also be affected by the host fruit, cold tolerance stage, endpoint checking and infesting method.
Besides the disinfestation of the infesting insects in the commodity, there should be no significant deleterious effects on the quality of the fruit. In general, the 'Red Globe' grape could be stored at an optimum temperature of −1 to 0°C to keep the quality for more than 1 month (Wu 2015). Several research studies have confirmed that a 12-d cold treatment at 0 and 1°C caused no significant damage to the quality of grapes (Wuernisha et al. 2010), and there is no significant negative effect on the taste of 'Red Globe' grape treatment at 0 and 2°C for 14 d in the primary test. Therefore, the schedules developed from this research could be used for the phytosanitary treatment of table grape 'Red Globe'.

Conclusions
In this research, the effects of phytosanitary cold treatment of D. suzukii in 'Red Globe' grape were examined. The results show that among six developmental stages (egg, first, second, and third instars, early and late pupa) of D. suzukii in grapes, egg was the most coldtolerant stage. The time-response test and large-scale confirmatory test suggest that a minimum of 11-d cold treatment at 0°C and 12-d at 2°C were recommended for disinfesting D. suzukii in grapes with efficacy of 99.9941 and 99.9948%, in order to provide postharvest pest control and quarantine security for international trade. All results suggest that cold treatment could be used as a methyl bromide alternative for quarantine control of D. suzukii in 'Red Globe' grape.