Distribution of Gonipterus Species and Their Egg Parasitoids in Australia: Implications for Biological Control

Gonipterus species are pests of Eucalyptus plantations worldwide. The egg parasitoid wasp Anaphes nitens is used in many countries for the biological control of Gonipterus spp. Recent taxonomic studies have shown that the three invasive Gonipterus spp., which were previously considered as G. scutellatus, form part of a cryptic species complex. These taxonomic changes have implications for the biological control of Gonipterus spp. The aims of this study were to understand the species composition and distribution of Gonipterus spp. and their egg parasitoids in Australia. Gonipterus spp. adults and egg capsules were collected in south-eastern Australia and Tasmania. Adult Gonipterus were identified using morphology and DNA barcoding. Parasitoids were reared from Gonipterus egg capsules and identified. Thirteen Gonipterus species were collected: twelve species were found on the Australian mainland and one species in Tasmania. These included three described species, four previously recognized but undescribed species, two undescribed species and four unidentified species. Five egg parasitoid species that attack Gonipterus spp. were identified. Anaphes nitens, Centrodora damoni and Euderus sp. were identified on the Australian mainland and A. tasmaniae and A. inexpectatus were identified in Tasmania. The results from this study will contribute to the improvement of Gonipterus biological control in the future.


Introduction
A number of strategies and protocols underpin the efficiency and success of biological control programs [1]. These include studies from the native ranges of the target pests aimed at understanding the diversity and ecology of natural enemies within the native range [2], searching for natural enemies in a region that is climatically similar to that of the introduced range [2,3] and collecting natural enemies from native populations of the pest that are genetically similar to those of the invasive population [2,4,5]. All of these approaches depend on a thorough understanding of the taxonomy, distribution and population structure of the target pest and its natural enemies in their native ranges [6][7][8][9][10][11].
The existence of cryptic species complexes can hamper the successful implementation of biological control. If the pest or biological control agent is part of a cryptic species complex, the misidentification of either of these components poses a significant risk. This can lead to the introduction of mismatched species of biological control agents, which in turn can lead to failure or inefficient control [12][13][14][15][16]. Combining ecological, behavioural and biosystematic studies with the morphological descriptions of species can contribute to and introduced range within Australia. This is especially in light of recent taxonomic understanding of the pest. The relevance of these results to biological control programmes of Gonipterus was also considered.

Insect and Plant Collections
Gonipterus adults and egg capsules were collected from mature and juvenile Eucalyptus trees in plantations, recreational parks and roadsides in South Australia (SA), Victoria (VIC), New South Wales (NSW), Australian Capital Territory (ACT), Queensland (QLD), Western Australia (WA) and Tasmania (TAS) (Figure 1). A total of 373 locations were inspected for the presence of Gonipterus. The sampling locations were selected by driving and stopping at regular intervals along the road, inspecting trees at recreational parks and plantation trees where possible. From these sampling points, Gonipterus was collected from 109 locations. Two collecting trips were done in early and late summer of 2017 in SA, VIC, NSW and ACT. Collections from TAS comprised only one sampling occasion, in early summer of 2017, but was deliberately limited because of recent surveys in that state [26,38]. Collections in QLD were done in late summer in 2017 and additional ad hoc sampling was done in 2018 and 2019. Collections in WA were done in 2018 from plantations. All collection areas other than WA include the known native distribution of the G. scutellatus species complex in Australia. Leaves and seed capsules were collected from mature Eucalyptus spp. from which beetles were collected and submitted to the Queensland Herbarium at the Brisbane Botanic Gardens, Mt. Coot-tha for identification. Collections were not limited to mature trees and included juvenile trees. No distinction was made between planted and native trees within the respective regions. GPS location data were recorded for each collection point. Adult beetles were killed by freezing and were stored in 99% ethanol prior to identification. Egg capsules were individually placed into gelatine capsules or 0.5 mL tubes in an incubator at 20 • C with 70% humidity. All parasitoids that emerged from the egg capsules were collected for identification.

Species Identification
Adult Gonipterus were identified using morphological characteristics and sequences of the Cytochrome Oxidase 1 (CO1) gene in the mitochondrial DNA. The most reliable morphological characteristic for the species identification for Gonipterus is the morphology of the male genitalia [14]. Therefore, only the males were identified by means of morphology (292 males). The females were identified by DNA barcoding. Between one and five females per collection site were used for identification. Sequence data from the DNA barcoding region were also included from identified males. Sequences from GenBank were used as reference samples.
Morphological identification of adult male Gonipterus individuals was done by incubating the male abdomens in a warm 10% solution of potassium hydroxide (KOH). The aedeagus was removed from the KOH solution, rinsed and dissected in 70% ethanol and examined in glycerine. The remains of the abdomen and the aedeagus were placed in genital vials with glycerine for temporary storage. Identifications of Gonipterus spp. were confirmed by R. Oberprieler and representative specimens were deposited with the Australian National Insect Collection (ANIC), CSIRO, Black Mountain, ACT.
Adult parasitoid wasps were placed in ethanol in a Petri dish and identified using a dissecting microscope (ZEIS) with 100× magnification and the keys by Huber and Prinsloo [39] and Prinsloo (Pers. comm.). Representative samples of each species were sent to G. Prinsloo (formerly affiliated with the Agricultural Research Council (ARC)-Biosystematics, Pretoria, Gauteng, South Africa) for species confirmation.

DNA Barcoding
DNA extraction. A total of 108 Gonipterus adults (17 male, 80 female and 11 unknown) were used for DNA barcoding. DNA was extracted from the wing muscles or hind leg of adult insects. A ZyGEM PrepGEM ® Insect DNA extraction kit and manufacturer's protocol were used. DNA extractions were diluted to approximately 50 ng/µL and stored at −20 • C.
Polymerase chain reaction (PCR). A 331 bp fragment of the CO1 gene region of the mitochondrial DNA was amplified using previously published primers (Table 1). PCR was performed for each sample using ProFlex PCR system (ThermoFischer Scientific Inc., Johannesburg, South Africa). Each 25 µL reaction contained 2.5 µL 10× PCR buffer (ROCHE, Roche Diagnostocs, Midrand, South Africa) 3 µL of 25 mM magnesium chloride (ROCHE), 0.5 µL Faststart Taq Polymerase (ROCHE), 2.5 µL DNTPs (ROCHE) (10 µM of each DNTP), 1 µL of each forward and reverse primer (Whitehead Scientific (Pty) Ltd. Johannesburg, South Africa; Inqaba Biotec TM , Pretoria, South Africa) diluted to 10 mM, 13 µL of distilled water and 1 µL of template DNA. The PCR thermal cycle program included a denaturation period of 5 min at 96 • C, 40 cycles of 1 min at 94 • C, 1 min at annealing temperature and 1 min and 30 s at 72 • C, followed by a final extension period of 10 min at 72 • C.  PCR products were visualized on a 1% agarose gel using electrophoresis (BioRad Gel DocTM Ez Imager and the software Image Lab v 4.0 build 16). PCR products were purified using the QIAquickTM PCR Purification Kit (QIAGEN) following the manufacturer's instructions and visualised on a 1% agarose gel as described above. Purified PCR products were used for sequencing. Each 10 mL sequence cycle reaction consisted of 2 µL sequencing buffer (Applied Biosystems TM , Thermo Fischer Scientific, Pretoria, South Africa), 0.5 µL BigDye (Applied Biosystems TM ), 0.5 µL primer for each forward and reverse primer separately, 6 µL distilled water and 2 µL purified PCR product. The thermal cycle reaction program included an initial denaturation of 2 min at 96 • C, 35 cycles of 30 s at 96 • C, 15 s at annealing temperature, and 4 min at 60 • C. Cycle sequencing products were cleaned using the Ethanol/NaAC precipitation of BigDye Terminator v3.1 DNA sequencing reactions protocol from the ABI manual. Samples were sequenced using an ABI PrismTM 3100 Genetic Analyzer (Applied Biosystems TM ).

Data Analysis
Phylogenetic analysis. Forward and reverse sequences were manually checked for base accuracy and trimmed using Chromas 2.6.2 [40]. Sequence contigs were created using BioEdit Sequence alignment editor version 7.2.5 [41] and aligned using ClustalW in MEGA 6 [42]. DNA sequences were translated to amino acids and searched for the presence of stop codons using AliView version 1.22 [43]. Seventy-three sequences of Gonipterus from Genbank (FJ88529.1-FJ88529.1, JN391478.1-JN391486) were added to the dataset before analysis. CO1 sequences (obtained from GenBank) for the closely related genus Haplonyx sp. were used as an outgroup. A neighbour joining tree using Maximum Composite Likelihood Model and 1000 bootstrap replicates was calculated using MEGA 6.
Species distribution: GPS and species identification data were used to map the geographical distribution of Gonipterus species and their egg parasitoids in Australia. Distribution maps were generated using GIS software (QGIS version 3.2.2, 2018).

Identification and Distribution of Gonipterus Species
Thirteen Gonipterus spp. were identified based on the genitalia sclerites of 292 adult males ( Figure 2). Five species, Gonipterus sp. n. 1-4 and G. pulverulentus are recognised members of the G. scutellatus species complex (Figure 2a-e), characterised by the squarely protruding apex of the male aedeagus [14]. Two described species not included in the G. scutellatus complex (G. notographus Boisduval and G. cinnamomeus Pascoe) were also collected. Furthermore, two species that have yet to be described were collected and are referred to here as Gonipterus sp. n. 6 and Gonipterus sp. n. 7. Three species, coded here as Gonipterus sp. 8, Gonipterus sp. 10 and Gonipterus sp. 11 could not be identified at the species level and further taxonomic studies are needed. A variety (synonym) of G. inconspicuus Lea was identified but needs taxonomic revision, here referred to as Gonipterus sp. 12 (Natalia M. de Souza pers. comm).

Identification and Distribution of Gonipterus Egg Parasitoids
Five egg parasitoids were identified, including Anaphes nitens, A. tasmaniae, A. inexpectatus, Centrodora damoni and Euderus sp. (Figure 5). All are known parasitoids of Gonipterus. Anaphes nitens, C. damoni and Euderus sp. were collected from all of the collecting regions on the Australian mainland other than WA where only A. nitens was collected. Anaphes tasmaniae and A. inexpectatus were only collected in Tasmania. An additional Hymenopteran species emerged from egg capsule samples collected from south-east QLD ( Figure 5). Identification at species level for this insect was not possible because only two individuals were collected, but they appeared to be species of Closterocerus Westwood (Hymenoptera: Eulophidae).

Gonipterus Host Plant Association
The association of Gonipterus spp. with Eucalyptus spp. differed between the regions surveyed ( Table 2). Gonipterus sp. n. 1 was collected from Eucalyptus conspicua, E. cinerea subsp. cinerea, E. crenulata, E. globulus and E. morrisbyi. Gonipterus sp. n. 2 was collected from the widest range of Eucalyptus spp., which includes commercially planted species such as E. globulus and E. dunnii. Gonipterus sp. n. 3 was collected from E. globulus and E. viminalis. Gonipterus sp. n. 4 was collected from E. propinqua and E. microcorys. Gonipterus notographus was collected from E. radiata subsp. radiata and Gonipterus sp. 11 from E. globulus. Gonipterus sp. n. 7 and G. cinnamomeus was collected from E. melanophloia and Gonipterus sp. 12 from E. propinqua. The host association of the Gonipterus species is not comprehensive, because many of the samples were collected from unidentified juvenile trees. In total, fifteen species of Eucalyptus were identified (Table 2), thirteen in the subgenus Symphyomyrtus and the remaining two species in the subgenera Eucalyptus and Alveolata [44].

Discussion
This study makes a significant contribution towards a better understanding of the distribution of Gonipterus species and their natural enemies on the Australian mainland. This study confirmed the native and introduced distribution range of the Gonipterus species known from the G. scutellatus species complex, extended the known distribution for two species of Gonipterus and provided distribution records of two previously undescribed species. In total, 13 species of Gonipterus were collected and identified. Nine species were phylogenetically grouped in the G. scutellatus species complex, but only five of these shared the squarely protruding apex of the aedeagus that is characteristic of the complex [14]. Six species of egg parasitoids were identified, five species that infest Gonipterus egg capsules and the status as Gonipterus parasitoid of the sixth species, Closterocerus sp., is unknown (Gerhard Prinsloo, pers. comm.). Species of Closterocerus parasitoids are known to parasitize leafminers and gall formers [45], for example, C. mirabilis Edwards & La Salle attacks Agromyzidae [46] and C. chamaeleon (Girault) attacks Ophelimus maskelli [47]. The known distribution range for three of the egg parasitoid species known to infest Gonipterus eggs was extended.
In this study, we confirmed that Gonipterus sp. n. 1-4, and G. pulverulentus belong to the G. scutellatus cryptic species complex as determined by Mapondera et al. [14]. Two additional undescribed species, Gonipterus sp. n. 6 and Gonipterus sp. n. 7 are identified in the G. scutellatus cryptic species complex based on current definition by morphological characteristics [14] and DNA barcoding. An additional three species, Gonipterus. sp. 12, Gonipterus sp. 10 and Gonipterus sp. 11 were also phylogenetically grouped within the G. scutellatus complex but did not have the squarely protruding apex of the aedeagus that is characteristic of the complex [14]. These species could also be separated based on their cuticular hydrocarbon profiles [48]. A taxonomic review of Gonipterus and the G. scutellatus complex is needed to describe the previously undescribed species and to determine whether the species without a squarely protruding apex is to be included in the G. scutellatus complex. Understanding these species boundaries and distributions of Gonipterus is important to understand species' interactions within the native range and to identify potential biological control agents.
The results of the present study confirm the native distribution of Gonipterus sp. n. 2-4, G. pulverulentus and G. cinnamomeus as determined by previous studies [14,26,38,49], and provide for the first time distribution records for Gonipterus sp. n. 6 and Gonipterus sp. n. 7. Importantly, they also expand the known distribution of Gonipterus sp. n. 1 and G. notographus to the Australian mainland, given that they were previously only known to be from Tasmania [14]. The Gonipterus sp. n. 1 population in Victoria, New South Wales and Australian Capital Territory was phylogenetically distinct from the Tasmania population, suggesting that the mainland population is not a recent introduction. In contrast, the G. notographus population in Victoria was not phylogenetically separated from the Tasmanian population, suggesting it could be a recent introduction. Other recent introductions of Gonipterus spp. part of the G. scutellatus cryptic species complex have also been recorded outside their native ranges within Australia. G. platensis and Gonipterus sp. n. 2 are invasive in Western Australia [14] and Gonipterus sp. n. 2 have recently been recorded in Tasmania [26].
This study confirms that the distribution of the invasive species differs in the native range, which indicates that natural enemies for the development of biological control programs should be collected from the respective distribution ranges within the native range. No new distribution records were recorded for the three invasive species apart from Gonipterus sp. n. 2, which was recorded further north-in south-eastern QLD in the present study than the study by Mapondera et al. [14]. G. pulverulentus was not collected on the Australian mainland in the present study despite the species being reported previously in NSW and QLD [14]. Thus, it can be inferred from the distribution records from the present and previous studies of the three invasive species that the biological control agent, A. nitens, was imported from the native range of Gonipterus sp. n. 2, as both the beetle and parasitoid species are present at Penola, from where A. nitens was originally collected [24].
The distribution records of the egg parasitoids emerging from the present study together with previous records [24,29,50,51] indicate that three of the five egg parasitoids of Gonipterus are present throughout the entire distribution range of Gonipterus species on the Australian mainland. Two species, A. nitens and C. damoni were previously recorded from the Australian mainland, but with limited distribution [24,29,51]. This study presents the first record of A. nitens in Queensland: previously, it was recorded from South Australia, Victoria, New South Wales, Australian Capital Territory, Western Australia and Tasmania [24,50,52]. Our results provide the first record of C. damoni in South Australia, Victoria and New South Wales. This species has previously been recorded from Queensland, Australian Capital Territory and Tasmania [26,38,51], Euderus sp. was recorded from Victoria, New South Wales and Queensland. Previously, Euderus sp. had been identified from Western Australia and Tasmania [26,38,50]. Despite the three species known to be parasitoids of Gonipterus spp., being present on the Australia mainland and Tasmania, further collections for the development and introduction of biological control agents should be focused on the native distribution of the respective invasive species and consider the local climatic conditions of the target region in comparison to the native range.
The native distribution range of A. nitens and Gonipterus sp. n. 2 includes five different states on the Australian mainland. The five states are divided into three major climatic zones, which are further subdivided based on regional rainfall [53]. Anaphes nitens used in the biological control program worldwide was originally collected from a single population in Penola, South Australia. The Mediterranean climate of this region is not the same across all the regions where the parasitoid is used as a biological control agent. For example, in South Africa, the climate in the Western Cape region is similar to that in Penola, with winter rainfall [24]. However, Gonipterus sp. n. 2 has established across a range of different climatic zones within South Africa, including temperate regions of the Highveld with winter temperatures much lower than that experienced at Penola, and sub-tropical regions of coastal KwaZulu-Natal that is more similar to Queensland, Australia [24]. Sub-optimal temperatures for A. nitens may therefore be influencing the high infestations of Gonipterus sp. n. 2 observed in these regions of South Africa (authors, pers. obs). The same is true for other regions of the world, such as Portugal, where low parasitism rates are associated with maximum winter temperatures below 10 • C [25]. Further introductions of A. nitens populations, or other parasitoid species, from the native range to these invaded regions may be required, where climate matching between collections in the native range, and the target range is considered.
The three egg parasitoid species, A. nitens, C. damoni and Euderus sp. identified on the Australian mainland occur throughout the known distribution range of Gonipterus sp. n. 2. This indicates that they do co-exist, but the nature of their interaction is not known. They are all endoparasitoids which share the same niche. It is thus expected that there will be some form of resource competition or niche separation. Niche separation can occur between species that co-exist by having different host and habitat preferences, temporal separation and varying degrees of adaptation to abiotic conditions [54][55][56]. In our study, Anaphes nitens was the dominant species collected at most sites on the Australian mainland, but this could temporally vary. Various aspects of A. nitens biology have been studied [24,[57][58][59][60][61][62], but little is known about its interactions with other parasitoid species within its native range. Exploitation and interference competition occurred between A. nitens and A. inexpectatus in laboratory studies using G. platensis, with lower temperatures favouring A. inexpectatus [62]. Factors such as the overall efficiency of the parasitoid species, the extent and form of niche separation, and the type of interaction between parasitoids will be important in understanding the final outcome of multiple species introductions on the Gonipterus population [63,64].
The present study identified five new host records not previously known from Gonipterus species. These include, G. notographus associated with E. radiata subsp. radiata, G. cinnamomeus collected from E. melanophloia and Gonipterus sp. n. 1 collected from E. conspicua and E. crenulata. E. radiata was previously considered resistant to Gonipterus attack [24], although these records were from South Africa, where only Gonipterus sp. n. 2 is present. Garcia et al. [26] also determined that the host associations for G. notographus and Gonipterus sp. n. 1 is different from the other species of Gonipterus present in Tasmania. Host plant records for the G. scutellatus cryptic species have been inconsistent in the past due to the three invasive species being identified as G. scutellatus [28]. The present study contributes to a better understanding of host association for ten of the Gonipterus species and 15 Eucalyptus host species present on the Australian mainland. Understanding host range and niche separation is valuable for planning future studies and more targeted approaches to collect biological control agents in the native range.
The number of Eucalyptus hosts identified was limited as most of the collections were from juvenile trees which are difficult to identify due to the absence of flowers and seeds [44]. Future studies could focus on the use of barcoding techniques to identify juvenile Eucalyptus trees. The study focused on the Australian mainland and did not include a big sampling effort in Tasmania because Valente et al. [52] and Garcia et al. [26] had recently extensively sampled Gonipterus species and their parasitoids in that state. Two of the species detected in their surveys but not in ours were G. platensis and G. scutellatus [14]. In this study, two collection trips were made at the end and beginning of summer in two consecutive years. Additional surveys in Queensland were undertaken by Souza et al. [65]. A temporal sampling strategy focused on specific regions, climatically matched with the target release area, will contribute to detecting the presence of Gonipterus and parasitoid species that were not abundant at the time collections were performed in the present study. This study and the recent surveys undertaken in the native distribution range of the G. scutellatus species complex significantly contribute to our understanding of Gonipterus biotic interactions in the native and introduced range in Australia and contribute to the development of more robust biological control programs for Gonipterus species in the introduced range.

Conclusions
This study is the first comprehensive survey of Gonipterus species and their parasitoids on the Australian mainland. This study contributes to our understanding of the distribution of Gonipterus species in the native range, host relationships and natural enemies. This is important information to underpin the further development of biological control, un-derstanding native-range interactions, and hasten identification in the event of additional Gonipterus invasions.
The results also highlight taxonomic uncertainties that present challenges for the development of biological control agents of Gonipterus spp. and in defining species boundaries in the genus. Obtaining clarity regarding the taxonomy of Gonipterus spp. will also play an important role in biosecurity. Closely related Gonipterus species feeding on commercially relevant Eucalyptus species, such as Gonipterus sp. 11 collected from E. globulus plantations, should be considered by biosecurity practitioners as there is considerable risk that these species escape the borders of Australia and become invasive pests in countries where Eucalyptus is commercially produced.