Synopeas maximum Awad & Talamas (Hymenoptera, Platygastridae): a new species of parasitoid associated with soybean gall midge, Resseliella maxima Gagné (Diptera, Cecidomyiidae)

Hymenoptera


Introduction
Gall midges (Diptera: Cecidomyiidae) are a hyper-diverse lineage, representing 30% of dipteran diversity in some ecosystems (Huang et al. 2022). Similarly diverse are parasitoids in the subfamily Platygastrinae (Hymenoptera: Platygastridae), (Chen et al. 2021). Platygastrines are larval or egg-larval parasitoids of Cecidomyiidae, and development of the parasitoid is often suspended until the host is in the last instar or prepupal stage (Kim et al. 2011;Abram et al. 2012;Chen et al. 2021). For both Cecidomyiidae and Platygastrinae, the diversity of species far exceeds what has been described (Srivathsan et al. 2022). However, a recent treatment of the genus Synopeas from Papua New Guinea, using morphology, DNA barcoding, and available data on host associations, has set a new standard for making inroads into both the taxonomy and ecology of gall midge parasitoids (Awad et al. 2021). The diversity of species and trophic interactions between Cecidomyiidae and Platygastrinae are large and complex. However, agricultural ecosystems can provide excellent opportunities to investigate these relationships within much simpler systems.
In 2018, larvae of an unknown species of cecidomyiid were found associated with dying soybean plants, Glycine max (L.) Merr., in the midwestern United States (Gagné et al. 2019). In 2019, this species of unknown origin was described as Resseliella maxima Gagné, the soybean gall midge (Gagné et al. 2019). Soybean plants become susceptible to R. maxima attack when the plants have two or three expanded trifoliate leaves, which is when the plant stems present natural fissures below the cotyledonary node . Adult female midges oviposit in these fissures and the larvae feed within the stem at the base of the plant, resulting in necrotic lesions that can cause lodged, wilted, and dead plants Helton et al. 2022).
Efforts to reduce R. maxima injury to soybeans have focused primarily on chemical control, but have so far provided insufficient protection of soybean plants (Hodgson and Helton 2021;McMechan 2021). Therefore, additional tactics for more effective and sustainable management are needed. One potential management approach is biological control, which has yet to be examined for R. maxima. Other pestiferous Cecidomyiidae have been successfully managed using biological control agents, including Aprostocetus epicharmus (Walker) (Hymenoptera: Eulophidae) that parasitized up to 38% of raspberry cane midge, Resseliella theobaldi (Barnes) (Vétek et al. 2006), and Synopeas myles (Walker) (Hymenoptera: Platygastridae) that parasitized up to 28% of swede midge, Contarinia nasturtii (Kieffer) (Abram et al. 2012b;Ferland 2020).
The present work investigates potential parasitism of R. maxima by rearing parasitoids from field-collected soybean stems. We present a taxonomic and molecular description of the R. maxima-associated parasitoid, Synopeas maximum Awad & Talamas, sp. nov., and a phylogenetic analysis of Synopeas sequences available on the Barcode of Life Data System (BOLD).

Field collection and emergence cages
Soybean stems presenting symptoms of infestation by R. maxima (i.e., darkened swollen lesions at the base of the stems) were collected during the summer of 2021 in two fields on one farm near the city of Luverne (Rock County), Minnesota, USA. Field collection started on 30 June 2021, when soybean plants started to show symptoms of infestation, and continued every other week until R. maxima larval infestation was no longer detected on 01 September 2021. On each sampling date, 10 randomly selected symptomatic plants were collected per eight sampling locations per field by pulling the entire plant from the soil. These plants were trimmed above the first pair of unifoliate leaves, placed in zipper-locking plastic bags (17.7 × 18.8 cm, Ziploc), and held in coolers until brought to the laboratory (approximately five hours).
In the laboratory, the stems were prepared for placement in emergence cages. The cut end of each stem was wrapped with a small piece of PARAFILM to slow plant dehydration. Soybean roots were trimmed to a length of five centimeters. Ten trimmed stems were placed together in one emergence cage per location. Emergence cages consisted of 5-liter clear plastic paint-mixing buckets with lids (TCP Global Corporation, Lakeside, California, USA). A 6-cm diameter hole was cut in the side of each bucket approximately 6 cm from the bottom of the bucket. A white fine-mesh (0.02 cm mesh size, 100% polyester, Quest Outfitters, Sarasota, Florida, USA) sleeve 30 cm long was attached to the hole with hot glue to allow access to the contents of the cages. The sleeves were tied to prevent insects from escaping. In each cage, the stems were placed vertically into a 3 cm deep layer of potting soil (BM2 Seed Germination and Propagation Mix, Berger, Saint-Modeste, Quebec, CA). The emergence cages were maintained at room temperature in 16h light:8h dark, watered as needed to maintain soil moisture, and checked daily for emergence of insect adults. Adult insects were collected manually into microcentrifuge tubes, freeze-killed in -20 °C for 24 hours and preserved in 95% ethanol for taxonomic and molecular identification.

DNA extraction
Non-destructive DNA extraction from individual specimens followed a modified Hot-SHOT protocol (Truett et al. 2000). Each specimen was placed in a 0.2 mL PCR tube (Olympus plastic, Cat# 27-125) with 100 μL of the lysis reagent (25 mM NaOH: 0.2 mM disodium EDTA) and incubated at 95 °C for 30 minutes on a Mastercycler nexus PCR cycler (Eppendorf ). Samples were cooled to 4 °C and 100 μL of neutralizing reagent (40 mM Tris-HCl) was added to each sample (a final volume of 200 μL). The aqueous solution containing DNA was moved to a fresh tube and 95% Ethanol was added to the specimen for preservation.

DNA barcoding
We barcoded all Synopeas (n=16) as well as six adult specimens of R. maxima randomly selected from the emergence cages. The cytochrome oxidase subunit I (COI) gene was amplified alongside negative controls using the universal primer pair LCO-1490/HCO-2198 for S. maximum and COIA/J-1718 for R. maxima (Folmer et al. 1994;Simon et al. 1994;Funk et al. 1995;Gagné et al. 2019). PCR reactions were prepared in a final volume of 20 μL with 1 μL of DNA template, Q5 Hot Start High-Fidelity 2X Master Mix (New England BioLabs), and 500 nM of each primer. Thermalcycling was performed on a Mastercycler nexus PCR cycler (Eppendorf) with an initial denaturation of 2 min at 98 °C, followed by 40 cycles of amplification (10 s at 98 °C, 30 s at 60 °C, and 20 s at 72 °C), and a final elongation of 2 min at 72 °C. The annealing temperature was determined using NEB Tm calculator (version 1.15.0, https://tmcalculator.neb.com/). PCRproducts were separated by gel electrophoresis on a 1% agarose gel and imaged under ultraviolet light after staining with GelRed 3X in water (Biotium). PCR-products were cleaned with the Exo-CIP Rapid PCR Cleanup Kit (New England BioLabs) according to manufacturer's instructions and Sanger sequenced in both directions at the University of Minnesota Genomics Center (Saint Paul, Minnesota, USA). Sequences were inspected for peak quality, aligned, and trimmed of priming regions in SnapGene (SnapGene5.3.2).

Synopeas maximum-specific primer design and DNA barcoding
We designed a Synopeas maximum-specific forward primer after finding that (1) not all Synopeas samples were compatible with the LCO-1490/HCO-2198 primer set, and (2) Synopeas samples that were amplified with LCO-1490/HCO-2198 primers typically generated low sequence quality when sequencing from the LCO-1490 primer (these sequences were ~50% of the amplicon length and were low quality, with average quality scores around 20, rather than a typical mean of ~40). Using the preliminary data from the reverse reads, we selected a new forward primer targeting a region that was conserved across our samples (SYN_F: 5'-CGATTAGAAGTTGGAACTCC-3') and generated a 550-bp amplicon when combined with the HCO-2198 reverse primer. A new PCR reaction mix was prepared for all wasps (n=16) as described in the section above, using the new primer pair (SYN_F/HCO-2198). Thermal cycling was performed on a Mastercycler nexus PCR cycler (Eppendorf) with an initial denaturation of 2 min at 98 °C, followed by 35 cycles of amplification (10 s at 98 °C, 30 s at 61 °C, and 20 s at 72 °C), and a final elongation of 2 min at 72 °C.New PCR-products were separated, purified, and sequenced as described above. Resulting sequences were uploaded to BOLD and are listed in Table 1.

Phylogenetic reconstruction
A phylogenetic analysis was performed with Synopeas sequences available on BOLD (Suppl. material 1) along with S. maximum sequences, and selected outgroups (n=2412). Amino acid sequences were aligned with MAFFT version 7.475 (Katoh and Standley 2013) using default parameters for downstream use in phylogenetic analyses.
The alignment was trimmed in SnapGene (SnapGene5.3.2) to remove sequences from outside of the SYN_F/HCO-2198 amplicon. A reduced version of the alignment, which excluded identical sequences (n=1459, 518 bp in length), was submitted to maximum likelihood analysis in RAxML version 8.2.11 using a GTRGAMMA substitution model and 1000 bootstrap replicates (Stamatakis 2014). Tree topology was rooted and visualized in FigTree version 1.4.4 and annotated in Inkscape version 1.2.1. Species delimitation was performed with the ASAP web server and default settings (https://bioinfo.mnhn.fr/abi/public/asap/) including the full aligned FASTA file containing all sequences (n=2412) and a JC69 Jukes-Cantor substitution model (Puillandre et al. 2021). The best scoring partition was selected for downstream processing as per ASAP guidance.

Imaging
Photography was performed using a Macropod microphotography system (Macroscopic Solutions) using 10X and 20X Mitutoyo objective lenses, with image stacks rendered in Helicon Focus. Images of primary types were deposited in Zenodo (Table 2), and images of molecular voucher specimens were deposited in BOLD.

Institutional abbreviations
Specimens examined during this study are deposited in the following institutions and abbreviated as follows:

CNCI
Canadian National Collection of Insects, Ottawa, Canada; FSCA Florida State Collection of Arthropods, Gainesville, Florida, USA;

Emergence cages
We collected 2221 adults of R. maxima. Other cecidomyiids collected from the cages included two individuals of Lestodiplosis spp. Two taxa of parasitoids were collected from the cages, including 16 individuals of S. maximum and 4 individuals of Aphanogmus sp. (Ceraphronidae).

DNA Barcoding
We performed DNA barcoding on each of the 16 Synopeas adults recovered from emergence cages (Table 1). COI sequences from these specimens had greater than 98% sequence similarity to each other but no close matches on GenBank. The most similar was a specimen identified as Platygastrinae sp., (GenBank ID MG501619.1, 97% query cover, 89.6% nucleotide identity), collected in Banff National Park, Alberta, Canada. Unfortunately, there are no photographic records of this specimen and a morphological comparison was not possible. We also verified the identity of randomly selected gall midges that emerged from the cages. The specimens were morphologically identified as R. maxima, had identical nucleotide sequences to each other, and a 97.67% sequence similarity to a R. maxima specimen from Nebraska, USA (accession number LC437340.1).

Systematics of Synopeas
The generic concept of Synopeas is rather straightforward, and it can be separated from other platygastrines by the fusion of T1-T2 and S1-S2 (Jackson 1969;Awad et al. 2021). However, taxonomic structure within the genus is essentially nonexistent. Given that Synopeas includes over 350 species, this presents a significant challenge for species identification (Awad et al. 2021). This situation is exacerbated by the spread of numerous species by human activities, requiring revision of the world fauna to be certain that a species was not previously described. Our ability to ascertain if S. maximum is adventive is presently limited to the use of DNA barcode libraries, and comparison to type specimens that are in the morphological vicinity of S. maximum. So far, neither of these approaches have provided a match. Our efforts are not exhaustive, nor is it currently feasible to examine type material for all described species of Synopeas. Fortunately, S. maximum has a distinctive characteristic shared by a minority of species: a deep scuto-scutellar sulcus (i.e., deep divide between the mesoscutum and mesoscutellum) (Fig. 6C, D). This eliminates the need for comparison to the vast majority of described Synopeas. Within the Nearctic region, primary type images provided by Talamas et al. (2017) yielded only two species, S. cynipsiphilum (Ashmead) and S. flavicorne (Ashmead), with the mesosomal structure found in S. maximum. Given that the soybean gall midge is an emerging pest, we also considered it possible that S. maximum represents an adventive population, derived from a distribution outside of the United States. Our literature search and examination of type specimens found other species with the deeply divided mesosoma, but not a species-level match. Because the mesosomal divide creates such a distinctive "hunchbacked" appearance, we consider that it has high value as a diagnostic character and use it to define the rhanis-group of Synopeas, named for S. rhanis (Walker, 1835), which was the earliest described species with this character. The rhanis-group includes 26 species ( Table 2) that we consider to be worth comparing to S. maximum. We selected five species from this group for closer examination and comparison, based on morphological similarity and occurrence in cold temperate climates. However, all species of the rhanis-group were considered and compared to the new species using type images or published descriptions. Fig. 1 Leptacis cynipsiphila Fig. 4 Platygaster Rhanis Walker, 1836: 225 (original description). Platygaster Acco Walker, 1836: 229 (original description. Synonymized by Vlug (1985); Vlug 1985: 209 (junior synonym of Synopeas rhanis (Walker) Description. Females. Body length: 1.4-1.7 mm. Body color: black. Color of legs: coxae dark brown, otherwise yellow to dark brown. Color of mesoscutellar spine: concolorous with mesoscutellar disc.
Wing. Length of setae on disc of fore wing: shorter than distance between setal bases. Density of setae on disc of fore wing: moderately dense. Arrangement of setae on disc of fore wing: uniformly setose distally, proximally glabrous with linea setosa. Fore wing marginal setae: uniformly very short.
Males. Body length: 1.1 to 1.3 mm. Identical to females except for metasoma and antenna.
Antenna. Setation: A1 and A2 with few scattered setae, A3 to A10 with long, uniformly dense setae. A2 in lateral view: slightly longer than wide, distally widened forming a "teardrop" shape. A3: round, about half the size of A2 or A4. A4: roughly cylindrical, about twice as long as wide. A5 in lateral view: about half as long as A4, proximally widened. A6 to A9: roughly ovoid, wider in lateral view than in anterior view, A6 slightly smaller than following antennomeres. A10: about twice as long as wide. Etymology. The species epithet refers to the ecological association with soybean gall midge, Resseliella maxima Gagné, and soybean, Glycine max (L.) Merr.
Diagnosis. Synopeas maximum can be separated from other species in the rhanis group by the following combination of characters: scuto-scutellar sulcus deep, causing mesoscutum to be elevated relative to mesoscutellum; hyperoccipital carina present between lateral ocelli, laterally weakened; mesoscutellar spine short, pointing posteriorly, sometimes with a slight upturn at the tip, but always originating from below the dorsal apex of the mesoscutellum (separating it from S. gibberosum, S. prospectum, and S. rhanis); female S2 expanded ventromedially, with microsculpture absent or very faint; female S6 and T6 entirely sculptured, triangular, about 2 times as long as wide. The latter character is very useful for separating S. maximum from S. cynipsiphilum and S. flavicorne, in which female T6 is wider than long. Figure 7. Simplified phylogenetic tree of the genus Synopeas. Maximum likelihood analyses were used to reconstruct a Synopeas phylogeny. This reduced version focuses on S. maximum (red box) and its closest relatives in clade A (gray box). Node circles are color coded to indicate bootstrap support. Specimens are named with sequence ID and taxon originated from BOLD. The full tree can be found in Suppl. material 3.

Phylogenetic reconstruction
The tree in Fig. 7 is a simplified version of the full consensus tree (Suppl. materials 3, 4) that comprises all the COI sequences derived from putative Synopeas sp. available in BOLD (excluding duplicates) wherein clades not relevant to species-level treatment of S. maximum were collapsed. While the backbone largely has lower bootstrap support (<50%), many species or putative species groups are strongly supported. ASAP analysis for species delimitation defined a total of 279 putative Synopeas species with 0.024 threshold distance (Suppl. material 5). Additionally, all S. maximum sequences were clustered into a single putative species by ASAP (Suppl. material 2).
We used the phylogenetic reconstruction (Fig. 7) to infer S. maximum relationships and origin. Specimens in clade A are all from Canada (n=37) and the northern United States (n=7), suggesting that S. maximum is also native to North America. However, not all specimens in clade A belong to the rhanis group, and some representatives of the rhanis group are outside of clade A. Our analysis does not support the monophyly of the rhanis group, although the deep scuto-scutellar sulcus is still useful for identification.

Discussion
The goal of the emergence cages was to obtain R. maxima and potential parasitoids of cecidomyiid agricultural pests. In addition to R. maxima, the only other cecidiomyiids collected were two individuals identified as Lestodiplosis spp., which are known to be predaceous (Gagné and Jaschhof 2017). Although we detected emergence of two hymenopteran parasitoid species (i.e., 4 individuals of Aphanogmus sp. and 16 individuals of S. maximum) in our emergence cages, we focused our investigations only on S. maximum for several reasons. Species from the genus Aphanogmus are primarily hyperparasitoids (Polaszek and Dessart 1996;Jaramillo and Vega 2009;Hofsvang et al. 2014;Pérez-Rodríguez et al. 2019) or parasitoids of predaceous cecidomyiids (Gilkeson et al. 1993;Matsuo et al. 2016), while Synopeas is a genus that is known to parasitize only cecidomyiids (Awad et al. 2021). However, the role of Aphanogmus sp. in this system needs further investigation.
Platygastrinae are important natural enemies of cecidomyiids (Austin 1984;Kim et al. 2011;Johnson et al. 2013;Chavalle et al. 2015). The recent emergence and spread of R. maxima and several other cecidomyiids, such as Contarinia nasturtii Kieffer (Philips et al. 2017), Contarinia brassicola Rondani (Mori et al. 2019) and Chilophaga virgati Gagné (Calles Torrez et al. 2014), poses a threat to agriculture in the northern U.S. and Canada. To improve management for these and other cecidomyiids that may emerge as pests in the future, further work on the ecological relationships and taxonomy of Synopeas and other Platygastrinae is required.
Here, morphological assessments grouped S. maximum with the rhanis-group (i.e., "hunchbacked" appearance) which facilitates its identification. Although the rhanisgroup did not form a monophyletic clade in molecular analyses, this feature appears to be useful for diagnostics. Additionally, we discovered that S. maximum clustered with other putative Synopeas species collected from Canada and the United States, suggesting that S. maximum may be native to North America.
Different methods for assessing parasitism of midges have been described, including rearing of field-collected hosts in the laboratory (Abram et al. 2012b), PCR-based molecular methods (Greenstone 2006;Magagnoli et al. 2022), and host dissections (Roubos and Liburd 2013). Here, we opted to rear out putative natural enemies of R. maxima because host dissections would not provide us with adults for taxonomic identification, and PCR-based molecular methods were impractical since we did not want to make assumptions about putative parasitoids that might be present. Our results suggest that S. maximum is likely a primary parasitoid of R. maxima because (1) it was reared out of emergence cages with field-collected soybean stems heavily infested with R. maxima and (2) the genus Synopeas is known to exclusively parasitize Cecidomyiidae. However, additional research is needed to confirm that S. maximum is indeed a parasitoid of R. maxima. Furthermore, the potential impact of S. maximum as a biological control agent of R. maxima is still unknown and more research needs to be performed in this area.
With the known geographic range of R. maxima expanding , sustainable methods to manage this pest, such as biological control, need to be explored. As taxonomic work is foundational to the introduction, conservation, and augmentation of natural enemies (de Moraes 1987), the description of S. maximum will facilitate future research on the biological control of R. maxima.