Short-term thermal changes affect parasitism on Asian chestnut gall wasp

Simple Summary: The Asian cynipid gall wasp (ACGW ) “ Dryocosmus kuriphilus ” has be come widespread in Europe, and in all invaded areas it is parasitized by native parasitoids associated to oak galls, for which ACGW represents a non-saturated adaptation space. Taking into account the last two-decade increase in the frequency of extreme climatic events (i.e. low temperature during the vegetative period of the chestnut tree), this research aimed to elucidate the thermic stress effects (i.e. cold treatment) on both ACGW biology and parasitism by native and introduced parasitoids. The ACGW-parasitoid system represents an ideal subject to evaluate the effect of sudden extreme climatic events due to their biological characteristics that include the ability to complete the development even in galls detached from the plants. We have shown that parasitism and mortality of ACGW in three chestnut fields were affected by the cold treatment. Our results revealed species-specific differences in the abundance and performance of parasitoids associated with ACGW in response to cold stress. For example, the frequency of Eupelmus spp. and Mesopolobus tibialis doubled in the cold treatment in all three chestnut fields in both years. Therefore, plasticity in response to short-term thermal changes may be associated with individual fitness in some parasitoid species. Abstract: Thermal variations affect tri-trophic interactions involving plants, herbivores, and parasitoids, causing a mismatch between their phenological cycles. In the current environmental context of climate change, climatic factors can undergo profound and sudden changes, such as sudden hot or cold snaps. Here we show that episodes of short but sustained low temperatures have increased, mainly in the month of May, in the last two decades. We subjected galls induced by Asian chestnut gall wasp (ACGW) Dryocosmus kuriphilus to cold stress to assess if and how the pest and its parasitoids were affected. Over the course of two years seasonal parasitism, parasitism rates, the relative abundance of each parasitoid species, and mortality of ACGW were calculated. We found that the cold treatment affected both pest and parasitoids, resulting in a reduction in emerging ACGWs and differing ratios of species within the parasitoid community. The striking example was the change in the relative frequency of three species of Eupelmus spp. and of Mesopolobus tibialis , which doubled in thermally-stressed galls in all chestnut fields. The thermal effects on the development of the host and the direct effects of cold temperatures on the surface of galls (i.e., humidity or hardness of the galls) warrant further research in this direction. as M. tibialis . Furthermore, our results confirmed that some parasitoids, such as Torymus auratus (Geoffroy in Fourcroy, 1785) and B. dorsalis , prefer to attack later during pupal formation. Graphical Abstract. We examined the effect of thermal cold stress on Dryocosmus kuriphilus Yasumatsu (Hymenoptera: Cynipidae) and its parasitoids. We found that cold treatment affected parasitism, resulting in a general reduction in emerging ACGWs and parasitoids under field conditions and differing ratios of parasitoid species compared to untreated galls. A striking example was the relative frequency of Eupelmus spp. and Mesopolobus tibialis , which doubled in thermally-stressed galls. Experimental design. An overview of the experimental setup is illustrated (not drawn to scale). Galled shoots were stressed at 8°C for 7 days and placed back into their chestnut field of origin (below) to compare insect emergence with that of non-thermally-treated galls (above).


Introduction
In the context of the current global climate change, climatic factors can undergo profound and sudden changes. The increase of the mean temperature is associated with climatic oscillations such as heat and cold waves [1][2][3]. Modifications in abiotic factors such as thermal variations can act on different biological and evolutionary aspects [4,5], and may disrupt the seasonal phenology patterns of organisms and determine unpredictable changes in ecological niches of different insect groups [6,7], affecting, for example, the tri-trophic interactions involving plants, herbivores, and parasitoids [8]. The role of thermal variations in the trophic mismatch is increasingly understood, although nearly every type of species interaction could be modified by climate changes [6,9,10]. Anomalous and unpredictable cold spells can be observed more and more frequently in a short period of the year [11], thus are among the rapid thermal fluctuations worth investigating to assess possible effects on the abundance, distribution, and function of species in a food web. Temperatures can affect the phenology of both hosts and parasitoids [7,12,13] and asymmetric changes in the seasonal activities of species between closely interacting species are likely responsible for desynchronization in their life cycles [8].
Insects at higher trophic levels are expected to be more strongly affected by environmental changes than organisms from low trophic levels due to cascading effects in the food chain [14]. Parasitism rate can be affected by the phenology of parasitoid species associated with oak gall wasps, which are often multivoltine, as parasitoids must parasitize their hosts during the appropriate "window of vulnerability" [13,15]. Therefore, variation in parasitism rates is an immediate response to environmental factors, including seasonal temperature variations, climate change, and variations in habitat [13,[16][17][18]. However, it is difficult to pinpoint what factors determine variations in the composition of parasitoid species and why some species are more successful after thermal stress.
In this study, we used the Asian cynipid gall wasp (ACGW) Dryocosmus kuriphilus Yasumatsu and its parasitoids to investigate their response to thermal stress. The ACGW is a serious pest that affects all chestnut tree species in the genus Castanea (Fagales: Fagaceae) [19][20][21][22][23][24][25]. The species also represents a perturbation to the natural trophic relationships between the community of native parasitoids and oak gall wasps [26][27][28]. ACGW was first detected in the Calabria region (southern Italy) in 2009 [29] about seven years after its first record in Northern Italy (Piedmont), and heavy infestations of chestnut fields in the Aspromonte National Park have since been reported [18]. ACGW is a univoltine species whose females lay their eggs into chestnut buds, which induces the formation of galls on the growing shoots that can inhibit shoot development and flowering [30,31]. Populations are composed entirely of females that reproduce by thelytokous parthenogenesis and can lay more than 100 eggs during their 10 days of life [32,33]. Approximately 30-40 days after oviposition, the first-instar larvae emerge and overwinter in the buds [33,34]. Larval development continues during the following spring with two more instars developing inside the galls [35]. [36] observed that modification of tissues surrounding D. kuriphilus eggs that differentiate to form the larval chamber, started approximately one month after oviposition. The larval chambers continued to increase slightly in size throughout the autumn months until January followed by rapid growth from March to May, during the period from bud swelling to bud break [36]. Before adult emergence, between June and August, females remain inside galls in the pharate stage for ovarian maturation so that they are ready to lay eggs in the new buds upon emergence [37]. Recent research on galls morphology has shown that gall mass and volume follow a pattern that can be associated with a climatic cline [38].
In recent years, ACGW has become widespread in Europe [39], and in all invaded areas it is parasitized by native parasitoids [40][41][42][43] associated to oak galls, for which ACGW represents a nonsaturated adaptation space. A large community of generalist parasitoids related to oak galls rapidly shifted to ACGW [28,41], with local variations in the composition that depends on phenological overlap and habitat features [13,18,44,45]. The composition of this parasitoid community depends on morphological traits of galls (i.e., size, shape, hardness) [30,46], on the developmental stage of the gall (i.e., young or fully formed), cline, and on the number of gall chambers [47]. All newly-associated indigenous parasitoids attack concealed ACWGs through gall tissues, and the larger fully developed galls with a higher number of chambers and thicker sclerenchyma layers surrounding the larval chamber are negatively correlated with parasitoids, indicating that these gall traits are important defense mechanisms. The different lengths of the parasitoids' ovipositors can, therefore, limit or favor some species over others [30].
Other factors that affect the composition of the parasitoid community include the phenology of the host and the parasitoid, the establishment date of the new invasive host species, and the characteristics of the habitat [18,48]. The successful adaptation of a parasitoid to a new host requires synchrony between the presence of adult parasitoids and of suitable host life stages [44].
Since 2005, the parasitoid Torymus sinensis Kamijo (Hymenoptera: Torymidae), a generalist parasitoid of galling insect native to China and univoltine, has been released in several European countries to successfully control ACGW populations [49][50][51][52][53][54][55][56][57]. Early reports relating to the introduction of T. sinensis highlighted an altered community structures of parasitoids which recently shifted toward the ACGW in multiple ways: utilizing some oak galls in addition to ACGW galls, and displacing some generalist oak galls parasitoids able to attack ACGW [28,47,[58][59] but to date, the magnitude of its impact appears minimal [60] or not well known.
Taking into account the last two-decade increase in frequency of extreme climatic events (i. e. low temperature during the vegetative period of the chestnut tree), this research aimed to elucidate the thermic stress effects (i.e. cold treatment) on both ACGW biology and parasitism by native and introduced parasitoids.
The ACGW-parasitoid system represents an ideal subject to evaluate the effect of sudden extreme climatic events due to their biological characteristics that include the ability to complete the development even in galls detached from the plants. The results obtained can guide further experimental studies on the effects of sudden and extreme temperature events.

Study area
Shoots with galls from three chestnut fields were collected weekly between May and July 2017-2018. The chestnut trees were located in an upland within the Aspromonte Mountains (Reggio Calabria, southern Italy) (latitude 38°3', longitude 15°44'; 980 m asl). The fields are adjacent to farmland where other fruit crops (e.g., cherries, apples) and vegetable crops (e.g., cereals and potatoes) are grown. Downy oak (Quercus pubescens Willd), European oak (Quercus robur L., etc.), and holm oak (Quercus ilex L.) are widespread in the study area. The chestnut fields used for data collection included two orchards (with 25 and 50-year old trees, respectively) and one coppice (with 15-year old trees), which were named A, B, and C, respectively. The chestnut orchards are made up of wild plants grafted with the local Nzerta cultivar while the coppice is made up of trees obtained from seedlings of the same cultivar. Fields were managed without pesticides, growth regulators, or fertilizers.

Study Design
The experimental design took into account both the natural parasitization of different parasitoids species on ACGW and the effects of the thermal stress artificially caused on parasitization and on ACGW emergence. The sampling was weekly with thermal stress tests repeated during the spring-summer period to ensure the presence of all parasitoid species involved in biological control of ACGW.

Field sampling and seasonal parasitism description
For two years, an observational study design with sequential sampling throughout the larva-toadult life history period was adopted. Shoots with galls from fields A, B, and C were collected from May until adult ACGWs emerged, which occurs in the study area in July. Approximately 24 shoots were collected weekly from less than 5 m above the ground in each chestnut field (A, B, and C), with one shoot collected from every ordinal direction: NE, SE, SW, and NW, from 6 different trees. Thirtysix galls (12 for each chestnut field) were randomly chosen from the shoots (nine galls in total per ordinal direction) and carefully dissected under a stereomicroscope (SZX9, Olympus ® , Tokyo, Japan) within 24 h of collection. All chambers and vital and non-vital stages of gall wasps and parasitoids (i.e., larva, pupae, and pharate adults) were recorded. The parasitism rate was calculated as: number of chambers with living parasitoids (larvae, pupae, adults, emerged) / total number of chambers, excluding chambers damaged, with rot and without an identifiable host.

Cold stress experiment
To evaluate the effect of thermal stress on parasitoid species composition, 30 galls (from now on referred to as non-thermally-treated, NTT), chosen at random among those present on the collected shoots, were individually placed in closed alveolar containers containing cotton and stored at lab room temperature. NTT was a baseline group that did not receive any treatment, and that provided information on parasitoids that have emerged under natural conditions. Other 15 shoots per field were coated with wax and the leaves were removed to prevent wilting. Each shoot was wrapped in plastic film and stored at a controlled temperature (8°C) in a laboratory fridge for 7 days. This temperature is lower than the theoretical value at which ACGW stops their development but does not die; hence it is a stress factor [13]. These shoots were then repositioned on trees in the field where they were collected to evaluate the response of ACGW and parasitoids to cold stress (lab treatment). After one week, the shoots were collected and carried back to the laboratory, where 30 thermally treated (TT) galls were isolated in alveolar containers to allow for the emergence of ACGW adults and/or parasitoids. NTT and TT galls were kept in a walk-in laboratory room at temperatures ranging from 14º to 25°C. All galls were inspected weekly from May to September and every two weeks from October to April.

Emergence of the ACGW and its parasitoids
All emerged individuals from NTT and TT galls were separated, identified, and stored in alcohol. Identification was conducted using the following taxonomic keys: [27,[61][62][63][64][65]. Specimens were then compared with individuals identified by the Institute for Sustainable Plant Protection based on molecular protocols (COI, 28S, and ITS2) following [66] and [67].
Due to the difficulty of separating the cryptic species present in the Torymus flavipes Walker complex and some species of the genus Eupelmus [65,68], some analyses were performed based on aggregated data at the genus level for Eupelmus spp. and at the morphospecies level for T. flavipes. Voucher specimens have been deposited in the LEEA Laboratory (Laboratorio di Entomologia ed Ecologia Applicata), Università Mediterranea di Reggio Calabria, Italy.

Historical thermal data
To identify the frequency of the intervals in which thermal stress acts on ACGW and its parasitoids, historical thermal data were collected in the same area and altitude of fields studied for the twenty years (2000-2019) in Arpacal regional environmental monitoring station (see study area).

Statistical analysis
Data on the number of chambers in galls were tested for normality using Kolmogorov-Smirnov (K-S) tests (P = 0.05). The number of chambers in dissected galls in each year was compared using analysis of variance (ANOVA). The data were log transformed to meet the assumption of normality and model included chestnut fields (n = 3; A, B, or C), type of gall (n = 2; simple or compound, sensu [33], the ordinal direction (n = 4; NE, SE, SW, or NW) and chestnut fields -type of gall interaction as variables. We used a generalized linear model of binomial data (0 = chambers not parasitized; 1 = chambers parasitized) to compare differences in parasitism rates of dissected galls collected in each year. The following categorical variables were included in the model: chestnut field (A, B, or C), ordinal direction of galls (NE, SE, SW, or NW), and gall type (simple or compound). Lastly, the date of collection was added as a continuous variable. We restricted our analysis to the distinguishable stages of the gall wasp and its parasitoids.
The number of ACGWs and parasitoids emerging from NTT and TT galls was compared using a chi-squared (χ 2 ) test of the equality of distributions in each year. For species in the genus Eupelmus, another chi-squared (χ 2 ) test was used to assess the differences between species in the different treatment groups. The -2Log Likelihood value was used to assess the distribution of emerging adults according to the sampling date of the TT and NTT treatments. This last test was performed only for more significant species that emerged in this study.
The historical thermal data for the twenty years (2000-2019) of the monitoring station (n° 2465) was used to find the frequency of periods of at least 4 consecutive days in which the minimum daily temperature was less or equal than 8.5 °C. Historical thermal variations in the cold interval periods from April to August over the twenty years were analysed with the linear regression method.
We used SPSS v. 23 [69] for all data analyses and [70] to produce graphs. All data are expressed as untransformed mean values ± standard error (SE).

Results
This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation as well as the experimental conclusions that can be drawn.

Field sampling and seasonal parasitism description
Overall, 7,423 chambers were recorded in galls (3, (Table 1). Therefore, in the second year of monitoring simple galls increased while compound galls decreased. The ACGW parasitism rate was affected by several variables ( Table 2). The parasitism rate varied between chestnut fields and sampling dates in each year but not between ordinal directions or gall typologies ( Table 2). The lowest parasitism rate of galls was recorded during 2017 in all monitored fields (Table 3), with about half of hosts parasitized, whereas ratio was ≃ 0.8 in 2018 (Figure 1). The data collected each year in the three chestnut fields were pooled. The parasitism rate was significantly higher in field C (coppice chestnut) than in fields A and B (orchard chestnut) in both sampling years (Table 3)
Both the increase in the frequency of (Eupelmus sp. and M. tibialis) and the decrease observed for some species including B. dorsalis, were not concentrated on some samples but were observed during the entire sampling period.

Historical thermal data
Historical thermal data of the study area showed that the time intervals of at least 4 days with minimum daily temperatures of 8.5°C or less mainly occurred between the months of May and June (69 and 11 intervals). In the twenty years analyzed, only shorter intervals (1-3 days) were detected in July at temperatures close to 8.5°C (Figure. A3, A4). Time intervals with minimum daily temperatures lower than or equal to 8.5°C for at least 4 consecutive days are increasingly frequent in the month of May (R square 0.569; df = 2.18; P <0.001; Figure S3, S4).

Discussion
In our experimental approach, we simulated thermal stress events at low temperatures, which caused a variation in the ratio of parasitoids species. The analysis of temperatures of our study area over the past twenty years highlighted that cold snaps are occurring more frequently in May. Unpredictable cold waves have a direct and more easily verifiable effect on plants. [71] highlighted that, despite an increase in the mean annual temperature, unpredictable cold waves may retard the expansion of plants towards higher altitudes and latitudes.
We have shown that parasitism and mortality of ACGW in three chestnut fields were affected by the cold treatment. The ACGW population was smaller in 2018 than in 2017, and this was linked with higher parasitism rates in galls. The parasitism rate varied throughout single seasons (from May to July), and different mean values of parasitism were found between years and between chestnut fields. In particular, higher parasitism rates were found in the coppice field, which is consistent with previous results from studies in the same chestnut fields (2013-2015) [18].
T. sinensis was introduced to the study area in 2014, but not in the experimental fields, and just recently it is becoming the dominant parasitoid (Bonsignore, unpublished data). The higher parasitism rates recorded in 2018 were associated with a strong increase in T. sinensis, which, however, resulted very variable in the different samplings indicating a non-uniform distribution on the trees and hence that dynamic parasitoid adaptation was still in progress.
This proliferation is a clear indication of adaptation of this species to ACGW in the study area and confirms results recorded in Italy and other European countries [54,56,57,[72][73][74].
The cold treatment did not affect the composition of the parasitoid community in any chestnut field, which is congruent with what reported by [75], who did not found differences in diversity, richness, or evenness of ACGW-associated species despite differences in their habitat. Conversely, [76] showed that native parasitoid communities emerging from galls differed between pure and mixed chestnut stands at the same altitude, even if there was no difference in their abundance.
Our results revealed species-specific differences in the abundance and performance of parasitoids associated with ACGW in response to cold stress. In particular, the frequency of Eupelmus spp. and M. tibialis doubled in the cold treatment in all three chestnut fields in both years. Typically, M. tibialis parasitizes ACGW in early May (see Figure A2) and completes its development quickly. In contrast, Eupelmus spp. females are active later in June and are usually the last parasitoids to emerge. The high proportion of specimens belonging to the genus Eupelmus recorded at all sample sites (TT) may be related to the ability of Eupelmus spp. to develop as hyper-parasitoids [20,77]. However, the higher presence of Eupelmus spp. could also be linked to variations in emissions of herbivore-induced plant volatiles in response to thermal stress [78].
The three detected species of Eupelmus were present in all fields, but while the frequency of E. azureus increased after thermal stress treatment, E. urozonus and E. kiefferi were less affected. The increase of E. azureus can be related either to the greater ability of this species to develop also on decaying hosts (a reduction in the emergence and therefore an increase in ACGW mortality was observed in 2018) or to better resistance to cold stress. The influence of temperature on both ACGW and its parasitoids was also highlighted in a previous study, although in that case, the differences in the temperatures occurred in fields at different altitudes [13].
Thermal stress, although limited to one week, prolonged the development of the juvenile stages of ACGW and its parasitoids inside galls and delayed the emergence of parasitoids. This prolonged phenology may have ensured a wider temporal window for parasitism by some parasitoid species. Furthermore, re-exposure of galls in trees could have caused excessive drying and possibly resulted in the emission of plant volatile substances that are attractive to Eupelmus spp. and M. tibialis, as well as the variations in the water content of the galls, could have affected parasitoid success.
Thermal stress negatively affected the abundance of T. flavipes, T. sinensis, S. biguttata, and Ormyrus pomaceus (Geoffroy). Interestingly, all of these species can develop in the same larval stages of ACGW as M. tibialis. Furthermore, our results confirmed that some parasitoids, such as Torymus auratus (Geoffroy in Fourcroy, 1785) and B. dorsalis, prefer to attack later during pupal formation.
These parasitoid species, along with S. variegata (Curtis), were the last to emerge during the last few days of June and throughout July [42] and their number was also reduced by thermal stress.

Conclusions
In conclusion, the present study contributes to our understanding of the thermic stress on insect growth and host stage succession, and provides information on parasitoid adaptation to new host phenologies. A new host can share parasitoids that already exist in the area, but the processes that influence their adaptation are far more complex than the simultaneous presence of the adult parasitoid and the susceptible stage of the host. Therefore, this research focused on variations that occur in the parasitoid community due to simple phenological variation and stress conditions. We have shown that exposure of galls to low temperatures affects the structure of the parasitoid community. Understanding which processes affect parasitoid community structure is fundamental, particularly when a phytophagous insect shifts towards a new host, as occurred with ACGW and C. sativa.