﻿A new species of Diglyphus Walker (Hymenoptera, Eulophidae) from China, with morphological characterizations and molecular analysis

﻿Abstract Diglyphus Walker, 1844 (Hymenoptera: Eulophidae) is an economically important genus including species acting as biocontrol agents against agromyzid leafminer pests. A new species of Diglyphus, Diglyphusdifasciatus Liu, Hansson & Wan, sp. nov., was discovered during the identification of agromyzid leafminers and their associated parasitoid wasps collected from 2016 to 2022 in China, based on morphological characteristics and molecular analyses of COI, ITS2 and 28S genes. Diglyphusdifasciatus is similar to D.bimaculatus Zhu, LaSalle & Huang, distinguished by two interconnected infuscate vertical bands on the fore wing and the color of the scape. Molecular data support D.difasciatus and D.bimaculatus as two different species. The mean genetic distances between D.difasciatus and D.bimaculatus were 11.33%, 8.62%, and 0.18%, based on the COI, ITS2, and 28S genes, respectively.

Identification of Diglyphus species mainly depends on morphological data. However, combining analyses of the morphology with molecular data for species identification is essential owing to the morphological similarities among species (Bernardo et al. 2008;Gebiola et al. 2012;Hansson and Navone 2017;Ye et al. 2018). The cytochrome c oxidase I (COI) gene of the mitochondrial DNA and internal transcribed spacer II (ITS2) ribosomal DNA genes have previously been applied to enhance species identification (Hebert et al. 2003;Bernardo et al. 2008;Gebiola et al. 2009;Gebiola et al. 2012;Gebiola et al. 2015;Du et al. 2021). Although 28S ribosomal DNA (28S) has mostly used for phylogenetic studies at the genus level and above, it has also been used for species identification (Gauthier et al. 2000;Gebiola et al. 2009;Burks et al. 2011;Gebiola et al. 2015).
For this project we collected Diglyphus material from 33 sites in China during 2016 to 2022 (Fig. 11). The specimens were reared mainly from the agromyzid Chromatomyia horticola (Table 1). An undescribed species of Diglyphus, D. difasciatus Liu, Hansson & Wan, was discovered during the identification of the reared material. Altogether we recovered 125 female and 153 male specimens of D. difasciatus (Table 1).

Sampling
We collected the leaves of vegetables and ornamental plants infested with agromyzid leafminers in different provinces of China from 2016 to 2022. The leaves were placed in cages and each cage was labeled with collection date, locality, and host plant. The collected leaf material was maintained in climate chambers set at 25 ± 1 °C, 30-50% relative humidity, and a photoperiod of 14:10 h (light: dark) until agromyzid leafminers and their parasitoids emerged. All wasp specimens and their hosts were preserved in absolute ethanol and maintained at -20 °C at the Institute of Plant Protection (IPP), Chinese Academy of Agricultural Sciences (CAAS), Beijing, China. All data for D. difasciatus specimens are presented in Table 1.
Two males and two females of D. difasciatus reared from C. horticola were imaged and morphologically characterized. One male and one female (the holotype) were reared from leaves of Lactuca sativa Linn. and Brassica rapa var. glabra Regel in Hebei, China; one female was reared from leaves of Sonchus oleraceus Regel in Gansu, China; one male was reared from leaves of L. sativa in Hebei, China. Two female and one male of D. bimaculatus Zhu, LaSalle & Huang were used for imaging and morphological characterization, which were reared from leaves of Sonchus oleraceus in Tibet, China. Specimens used for molecular analyses included 67 specimens of D. difasciatus, 1♀ D. bimaculatus (Tibet), and 1♀ D. isaea (Walker) (Hubei). Diglyphus bimaculatus and D. isaea sequences were used as outgroups for analyzing the phylogenetic relationship of D. difasciatus. Furthermore, one female of D. bimaculatus used for molecular analyses was reared from leaves of Taraxacum mongolicum Hand.-Mazz., which was collected from Tibet, China (29°39'3"N, 91°08'41"E) in August 2020. The single D. isaea specimen was reared from C. horticola in Pisum sativum Linn. leaves, collected in Hubei, China (30°28'26"N, 114°21'17"E) in April 2017.

Morphological identification methods
The specimens were examined using a stereomicroscope (Olympus, SZX-16). Photographs were taken using an Olympus BX43 microscope equipped with a Helicon Focus 6.

Molecular diagnosis methods
Genomic DNA was extracted from the metasoma of each specimen. The extraction methods followed those described by De Barro and Driver (1997), with some modifications. The DNA extraction was performed using a 200-µL microcentrifuge tube (Bioevopeak, Shandong, China) and 200-µL pipette tip (Bioevopeak) sealed by heating to grind the metasoma into a homogenate. The homogenate was incubated at 65 °C, 25 °C, and 96 °C for 30, 2, and 10 min, respectively. After extraction, the genomic DNA was stored at -20 °C until molecular diagnosis. The primers used for amplification were in Table 2. Amplifications were performed as described by Hebert et al. (2003) and Du et al. (2021). Polymerase chain reaction (PCR) consisted of 0.4 µL Taq enzyme (2.5 UµL -1 ), 0.4 µL deoxynucleotide triphosphate (2.5 mM), 2.5 µL 10× buffer (containing Mg 2+ ), 0.4 µL forward primer, 0.4 µL reverse primer, 1 µL DNA template, and 19.9 µL double-distilled H 2 O. Next, the PCR cycling conditions consisted of an initial denaturation at 95 °C for 5 min, followed by 35 cycles of denaturation at 95 °C for 30 s, annealing for 45 s, extension at 72 °C for 60 s, and a single cycle of final extension at 72 °C for 5 min. The annealing temperatures for the COI, ITS2, and 28S genes were 50 °C, 52 °C, and 58 °C, respectively.
The unpurified PCR products were sent to Sangon Biotech Co., Ltd, Beijing, China, for bidirectional sequencing, and primers were designed by Sangon Biotech Co., Ltd, Beijing, China. The PCR instrument used was an ABI thermal cycler (Applied Biosystems Veriti 9902; Woburn, MA, USA).

Sequence analysis
The D. difasciatus sequences were analyzed using the National Center for Biotechnology Information (NCBI, https://www.ncbi.nlm.nih.gov/) and the Barcode of Life Data systems (BOLD, http://www.boldsystems.org/index.php). The phylogenetic relationships between D. difasciatus, D. bimaculatus, and D. isaea were also analyzed.   Campbell et al. 1993 All sequences were aligned following the default options of the CLUSTAL W tool (Kimura 1980); in Molecular Evolutionary Genetics Analysis (MEGA) X ver. 10.1.8 (Kimura 1980;Kumar et al. 2018). Pairwise and mean sequence divergence were estimated based on the Kimura-2 parameter (K2-P) (Kimura 1980). Gene haplotypes were calculated using DNA sequence polymorphism ver. 5 (Bioinformatics, Arlington, VA, USA) (Librado and Rozas 2009). The phylogenetic tree was constructed using the Neighbor-Joining method in MEGA X (Kimura 1980;Nei and Kumar 2000;Kumar et al. 2018). Bootstrap values were obtained after 1000 replications for sequence divergence and phylogenetic relationships. Bootstrap support values ≥75% is indicated above the branches of the phylogenetic tree. Diagnosis. Scape white with apical 1/3-1/2 dark brown (Figs 1-5). The yellow markings on the vertex and face, and those on the male are wider than those on the female. Fore wing with complete vertical infuscate bands below base of marginal and stigmal veins respectively, the two bands are interconnected medially (Figs 1-3, 7); speculum bare, without dense setae and postmarginal vein almost equal in length to stigmal vein (Figs 1-3, 7). Mid and hind femora black with apical 1/4 yellowish-white (Figs 1-3). Fore and mid tibia yellowish-white with a dark ring basally (Figs 1-3). Hind tibia black with apical 1/5 yellowish-white (Figs 1-3). Pretarsus on all legs black (Figs 1-3).
Head (Fig. 1). Head length 0.6× width in dorsal view, and length 0.9× width in frontal view. POL 1.8× OOL. Malar space 0.7× height of eye, and malar sulcus present. Frons and vertex with distinct reticulation. Eyes with sparse and short setae. Toruli situated below the level of lower margin of eyes. Maxillary palpus with two segments and labial palpus with one segment. Antennal flagellum with two funiculars and three clavomeres; scape 4.0× as long as broad and 2.8× as long as pedicel; pedicel 1.3× as long as broad; F1 1.4× and F2 0.9× as long as broad, F1 1.5× as long as F2; clava 2.4× as long as broad, 1.3× as long as scape, and 3.6× as long as F2.
Mesosoma (Fig. 1). Pronotum without transverse carina, reticulate, shorter than mesoscutum. Mesoscutum 1.2× as long as scutellum; mid lobe with two pairs of long setae; notauli incomplete and diverging posteriorly to meet anterior part of axillae. Setae on pronotum and mesoscutum pale. Scutellum as long as broad with straight sublateral grooves and two pairs of setae. Dorsellum with superficial reticulation with isodiametric meshes, posterior margin round. Propodeum shorter than scutellum and without median carina; callus with five setae. Fore wing with 5-7 setae on dorsal surface of submarginal vein; speculum mainly bare, with few scattered setae; costal cell with two rows of setae, including 15 setae at the base of costal cell and an incomplete row with eight setae in apical part; postmarginal vein almost equal in length to stigmal vein; Fore wing length 1.7× fore wing width. Petiole short and inconspicuous. Gaster subrotund, 1.9× as long as wide in dorsal view; apex acute. Tip of ovipositor sheaths visible in dorsal view. Male (Fig. 2). Similar to the female. Body length 1.4mm, Fore wing length 0.8mm. Head length 0.5× width in dorsal view, and length 0.8× width in frontal view. POL 1.1× OOL. Scape 4.7× as long as broad, 2.2× as long as pedicel. Pedicel 1.6× as long as broad. Antennal flagellum with two funiculars and three clavomeres, F1 0.8× and F2 0.7× as long as broad, F1 1.2× as long as F2. Clava 3.1× as long as broad, 1.1× as long as scape and 4.6× as long as F2. Mesoscutum 1.2× as long as scutellum. Scutellum as long as broad. Fore wing length 1.7× as long as fore wing width. Gaster 1.8× as long as wide in dorsal view.
Variation. Females are slightly larger than males (1.6 mm and 1.4 mm, respectively). Hosts and biology. Diglyphus difasciatus is a larval ectoparasitoid, primarily on Chromatomyia horticola, and occasionally on Liriomyza bryoniae (Kaltenbach), L. sativae, and L. trifolii (Burgess). The hosts are usually mining in leaves of Asteraceae, Brassicaceae and Fabaceae, especially on Ixeris polycephala Cass. ex DC. and Pisum sativum (Table 1). Diglyphus difasciatus occurs and reaches its highest occurrence period in May, and then disappears in October. Female Diglyphus exhibit three types of host-killing behavior (Zhu et al. 2000;Liu et al. 2013;Hansson and Navone 2017;Ye et al. 2018). The host-killing behavior of D. difasciatus is not known and requires further studies.
Etymology. The name is derived from a combination of the Latin di (double) and fascia (band) by referring to the two vertical infuscate bands in the fore wings.
Comments. Diglyphus difasciatus is very similar to D. bimaculatus (Figs 8-10), but has two complete vertical infuscate bands that are interconnected medially in the fore wing, whereas D. bimaculatus has two infuscate spots in the fore wing. In addition, the scape of D. difasciatus is white with apical 1/3-1/2 dark brown (Figs 1-5), which is less than the scape of D. bimaculatus with white upper surface (Fig. 9). Besides, molecular data support the separation of these two morphologically similar species as distinct species.

COI
The length of COI sequences from 67 D. difasciatus specimens was 514 bp, including 35 variable sites with 20 parsimony-informative sites, and 29 haplotypes were found (Fig. 12). The highest percentage similarity (89%) of sequences between D. difasciatus and other Diglyphus species in the NCBI and BOLD databases was with D. pulchripes (Erdös & Novicky) (NCBI accession number: MG442711).
The mean genetic distance of COI sequences between D. difasciatus/D. bimaculatus and D. difasciatus/D. isaea, based on the COI gene, was 11.33% and 13.37%, respectively (Table 3). The analyses of the intraspecific diversity in D. difasciatus showed that the mean genetic distance between the 29 haplotypes was 1.53% (Table 3). The genetic distance within D. difasciatus ranged from 0.19 to 3.42%.

ITS2
The length of the 25 D. difasciatus sequences was 415 bp; there were no variable sites. The highest percentage similarity of sequences in the NCBI and BOLD databases was between D. difasciatus and D. isaea (86%). The mean interspecific genetic distance between D. difasciatus/D. bimaculatus and D. difasciatus/D. isaea was 8.62% and 6.49%, respectively (Table 3).

28S
The length of the 11 sequences obtained from D. difasciatus was 547 bp; there were no variable sites. The highest percentage similarity of sequence in NCBI and BOLD between D. difasciatus and other Diglyphus species was with D. crassinervis (100% [NCBI accession number: MW393686]). The mean interspecific genetic distance between D. difasciatus, D. bimaculatus and D. isaea was 0.18% (Table 3).
All gene sequences are uploaded to GenBank with accession numbers OP933727-OP933732 and OP936054-OP936075.

Discussion
The new species, D. difasciatus, is defined by morphological data and molecular data from the genes COI, ITS2, and 28S. Morphologically, D. difasciatus is most similar to D. bimaculatus, from which it can be separated by a different wing pattern in the fore wing and the color of scape (Figs 3, 7-9). Molecular data from COI, ITS2, and 28S also show that D. difasciatus and D. bimaculatus are two different species.