Introduction

Dinogamasus Kramer (Acari: Mesostigmata: Laelapidae) is known as a symbiotic genus of mites that includes 46 named species currently (Lundqvist 1999; Joharchi et al. 2016). They are strictly associated with African and Asian carpenter bees of the genus Xylocopa Latreille (Hymenoptera: Apidae: Xylocopini) (Lundqvist 1999). The adult females of Dinogamasus are transported by female carpenter bees in a metasomal chitinous pouch—a specialized acarinarium located inside the first metasomal segment (LeVeque 1930ab; Lundqvist 1999; Michener 2007; Makino et al. 2018) (see also Figures 1, 2).

LeVeque (1930a) proposed the perkinsi species group to accommodate three species—D. perkinsi, D. piperi LeVeque, and D. philippinensis LeVeque—based on the following shared characteristics: dorsal shield with lateral notch (LeVeque 1930a: figures 1a, 2a); anal shield with parallel lateral margins basally (LeVeque 1930a: figure 2d); peritrematral shield rudimentary or produced and fused with a projection from the dorsal shield anteriorly; fixed digit of chelicerae with prominent tooth (LeVeque 1930a: figure 2e); trochanter–genu I with unmodified setae ventrally (Lundqvist 1999: figures 83c, e); leg I with conical setae, although its chaetotaxy usually varies among species (LeVeque 1930a: figures 1a, 2a, 3a); and tarsus II with three blunt conical setae (Lundqvist 1999: figure 83g). A strongly-supported morphological phylogeny of the genus, however, indicated that the D. perkinsi group, as proposed by LeVeque (1930a), is not monophyletic: D. perkinsi and D. piperi formed the sister clade to all other species of the genus (except D. villosior (Berlese)), while D. philippinensis was placed within a relatively-derived clade (Lundqvist 1999). All three species are found exclusively in the Indo-Malay ecoregion (sensu Olson et al. 2001), which encompasses India and Southeast Asia (the latter including most of the Malay Archipelago; see Figure 3).

FIGURE. 1–2.

Posterodorsal view of the internal cavity of T1 of female Xylocopa tenuiscapa showing metasomal acarinarium (chitinous pouch). 1. Undissected acarinarium; 2. Acarinarium with most of its posterodorsal part removed to show individuals of Dinogamasus saengdaoae sp. nov.

FIGURE 3.

Distribution map of the species in the Dinogamasus perkinsi group based on published records of their respective hosts. D. perkinsi: (A) Oudemans (1901): X. tenuiscapa (Indonesia); (B) Vitzthum (1919): Xylocopa latipes and X. tenuiscapa (India, Indonesia, Vitenam); (C) Vitzthum (1930): X. tenuiscapa (Sri Lanka); X. latipes (Indonesia, Sri Lanka); X. auripennis (Indonesia); (D) LeVeque (1930a): X. latipes (Indonesia, South Thailand); (E) OConnor (1993): X. latipes (Indonesia). D. piperi: (F) LeVeque (1930a): X. tenuiscapa (India); (G) Lundqvist (1999): Xylocopa sp. (Myanmar). D. philippinensis: (H) LeVeque (1930a): X. latipes (Philippines). D. kerrianus: (I) Lundqvist (1999): X. kerri (South Thailand). D. saengdaoae sp. nov., (J) this study: X. tenuiscapa (Chiang Mai Prov., North Thailand.

Overall, the Indo-Malayan species of Dinogamasus are fairly distinct from their Afrotropical and middle-eastern congeners in morphology: tibia II of the Indo-Malayan species does not have any hook-shaped pv setae (Lundqvist 1999: figure 6), while such setae are present in most of the African and middle-eastern species; in addition, the av seta of femur II is setiform in the Indo-Malayan species, but modified in the African species (Joharchi et al. 2016). Both character states exhibited by African species are apomorphic within Dinogamasus (Lundqvist 1999). Species of the two regional groups of Dinogamasus are also known to use different geographically-restricted groups of Xylocopa as hosts (LeVeque 1930a).

The most comprehensive taxonomic treatment of Dinogamasus currently available is that of Lundqvist (1999), in which 38 species were recognized (including eight newly described). Of these, 21 species are known to occur in Asia, including a species recently described from the Middle East (Lundqvist 1999; Joharchi et al. 2016). Only two species of Dinogamasus have so far been recorded in Thailand—D. kerrianus LeVeque and D. perkinsi—which were found in Ranong and Trang Provinces (southern Thailand), respectively (see Figure 3).

The main goals of this paper are to describe a new species of Dinogamasus based on ten adult females found inside the metasomal pouch of six female X. tenuiscapa (Westwood), and to propose a phylogenetic hypothesis for its placement within the Dinogamasus' tree of life. To determine that the species described herein is indeed new, we proceeded though an integrative approach to taxonomy, combining standard morphological procedures and DNA barcodes (i.e. the 658-bp Folmer fragment of the cytochrome c oxidase subunit 1 (COI); see Hebert et al. 2003), a practice that has proven very successful when it comes to species delimitation within Acari (e.g. Beaulieu et al. 2009; Negm & Gotoh 2019; Young et al. 2019a). Dinogamasus saengdaoae sp. nov. is most similar to D. piperi and D. perkinsi, but it can be easily distinguished from both through the illustrated diagnosis and modified version of Lundqvist's key given below. We also provide DNA barcodes for X. tenuiscapa to confirm its identification (as in Sheffield et al. 2009; Packer & Ruz 2017) and to compare its intraspecific genetic variation with that of D. saengdaoae sp. nov.

Materials and Methods

Morphological methods. All specimens of the new Dinogamasus species described herein were obtained from six females of X. tenuiscapa, which were collected at two different localities in Chiang Mai Province, northern Thailand: the agricultural field of the Faculty of Agriculture of Chiang Mai University (18°47′ 41.3″ N, 98°57′ 34.3″ E) (Figure 4) and a private farm in Nong Chom Sub-district, San Sai District (18°50′ 27.8″ N, 99°01′ 31.0″ E). The mite holotype and bee specimens are deposited in the Queen Sirikit Botanical Garden, Thailand (QSBG), and the mite paratypes will be deposited in the following repositories: Attasopa Collection at Department of Entomology and Plant Pathology, Faculty of Agriculture, Chiang Mai University, Thailand (BCMU); Chulalongkorn University Museum of Natural History, Thailand (CUMZ); Natural History Museum, United Kingdom (NHMUK); Packer Collection at York University, Canada (PCYU).

FIGURE 4.

Type locality of Dinogamasus saengdaoae sp. nov.: the agricultural field of the Faculty of Agriculture of Chiang Mai University, Chiang Mai Prov., Thailand.

The mites were dissected from the metasomal acarinarium of the bees (Figures 1, 2) with extra fine forceps and then transferred to a container with 70% ethanol until they were further processed. Next, they were immersed in a 10% NaOH solution for 1–2 minutes at 90^oC for clearing and washed with distilled water. The mites were then mounted on microscopy slides with Hoyer's medium under a Nikon SMZ800 stereomicroscope to facilitate comparative morphological study. All slide-mounted mites were carefully examined under an Olympus CX31 compound microscope and measured with a calibrated ocular micrometer. All measurements are given in micrometers (µm); those of the holotype are shown in bold, followed by ranges (average±SD) calculated from the paratypes (n=9).

The terminology for mite morphology used in this paper is essentially that of Evans (1992), except that we followed Evans (1963a,b) for leg and palp chaetotaxy and Lindquist & Evans (1965) for dorsal and ventral chaetotaxy. Line drawings and maps were prepared using Adobe Illustrator CS6 and then mounted onto plates using Adobe Photoshop CS6.

Molecular procedures and tree estimation. Genomic DNA of mites and bees was extracted using the DNeasy Blood & Tissue Kit (Qiagen, USA) and then stored at -20°C until further processing. DNA concentration of each sample was estimated using a Nanodrop spectrophotometer at 260 nm (ThermoScientific, USA). PCR amplifications of COI were performed in 25 µl reactions using the forward primer LCO1490 (5′ GGTCAACAAATCATAAAGATATTGG-3′) and the reverse primer HCO2198 (5′-TAAACTTCAGGGTGACCAAAAAATCA-3′) (Folmer et al. 1994). Thermal cycling conditions were as follows: 94°C for 5 min; 40 cycles of 94°C for 30 sec, 55°C for 30 sec and 72°C for 45 sec; followed by a final extension of 72°C for 7 min. Five µl of PCR products previously mixed with 6× loading dye (New England BioLabs, USA) were loaded onto a 2% agarose gel and then electrophoresed for 40 min at 100 V in TBE buffer. Sequencing of high-quality PCR products was performed by Macrogen Inc. (South Korea).

The resulting COI sequences were compared with their respective trace files and edited manually in BioEdit v.7.2.5 (Hall 1999). They were then checked for stop codons and gaps and trimmed to equal size using MEGA v.7.0.26 (Kumar et al. 2016). Then, all sequences were uploaded into GenBank, where they were aligned using the algorithm MUSCLE v.3.6 (Edgar 2004). A maximum likelihood-based analysis of the aligned dataset was performed in PhyML v.3.0 (Guindon et al. 2010) through 10,000 bootstrap replicates. This analysis was carried out under a GTR+G model, which was previously suggested as the best one with the program SMS (Lefort et al. 2017). Androlaelaps sp. and Varroa destructor Anderson & Trueman were chosen as outgroups for the ML analysis of the mite dataset, whereas X. appendiculata Smith and X. violacea (Linnaeus) were chosen as outgroups for the ML analysis of the carpenter-bee dataset (see Table 1 for GenBank accession numbers). The resulting phylogenetic trees were visualized with FigTree v.1.4.3 (Rambaut 2012) and then redrawn using Adobe Illustrator CS6. Pairwise distances within the COI datasets of mites and bees were estimated through the Kimura 2-parameter model (Kimura 1980) with a gamma distribution in MEGA. COI haplotype networks of mites and bees (outgroups excluded) were constructed using TCS network methods (Clement et al. 2002) in the program PopArt (Leigh & Bryant 2015).

Parsimony analysis. We added D. saengdaoae sp. nov., as well as the recently described D. kazerunensis Joharchi, Khodaparast & Moghadam, to the morphological dataset matrix (Table 3) provided by Lundqvist (1999) and then conducted a parsimony analysis in the program TNT v.1.5 (Goloboff et al. 2003, 2008; Goloboff & Catalano 2016). Trees were estimated through ‘New Technology search' using the Ratchet, Drift and Tree fusing. We performed 1,000 replications with ‘driven search' with the option ‘find minimum length' set to 100. Consistency and retention indexes (CI and RI, respectively) were calculated with the ‘stats’ script (available at  http://tnt.insectmuseum.org/index.php/scripts). As in Lundqvist (1999), all characters were treated as non-additive (except characters 1, 9 and 12) and group support was estimated by the percentage of times that a given clade was found among the most parsimonious trees. Final cladogram was edited in WinClada v.1.0.8 (Nixon 2002) and Adobe Photoshop v.13.0.1.

TABLE 1.

Voucher and GenBank accession numbers of the taxa included in our ML analyses of COI data.

Results

FIGURE. 5–8.

Dinogamasus saengdaoae sp. nov., gnathosoma of the female holotype. 5. Palp trochanter—tibia, dorsal view; 6. Palp tarsus, dorsal view; 7. Subcapitulum; 8. Chelicerae, lateral view.

FIGURE. 9–10.

Dinogamasus saengdaoae sp. nov., holotype female. 9. Dorsal view of idiosoma; 10. Ventral view of idiosoma. MDS: margin of dorsal shield.

FIGURE. 11–14.

Dinogamasus saengdaoae sp. nov., female holotype. 11. Coxae I–IV and trochanters I–IV, ventral view; 12–14. Femora I–III, dorsal view.

FIGURE. 15–21.

Dinogamasus saengdaoae sp. nov., female holotype, left legs. 15–17. Genua I–III, dorsal view; 18–20. Tibae I–III, dorsal view; 21. Tarsus II, ventral view.

FIGURE. 22–24.

Resulting trees from ML analyses of the COI datasets of Dinogamasus saengdaoae sp. nov.(red line) and Xylocopa tenuiscapa (blue line). 22. D. saengdaoae sp. nov. 23. X. tenuiscapa. 24. Tangled plot constructed from the clades within the dashed line rectangles of Figures 22, 23. Numbers above internodes are bootstrap values (values < 50% not shown).

Discussion

In this paper, we resorted to traditional methods in alpha-taxonomy and a ML analysis of newly-generated COI sequence data to describe a new species of Dinogamasus from northern Thailand. Dinogamasus saengdaoae sp. nov. was unequivocally an undescribed species as shown by a unique set of morphological features exhibited by the female (regrettably, no males were available for study; see ‘Diagnosis’ above). Moreover, the sequenced specimens of D. saengdaoae sp. nov. were recovered as a well-supported monophyletic group in our ML analysis (Figures 22–24). On the other hand, we must emphasize that the outgroups (Varroa and Androlaelaps) used in the analysis are both relatively distant allies of Dinogamasus, and therefore the monophyly of D. saengdaoae sp. nov. in the resulting ML tree is but an obvious outcome and thus should be interpreted with caution. A rigorous assessment of whether this new species comprises a monophyletic taxonomic unit or not would necessarily need to include a more comprehensive sampling of Dinogamasus, with special attention to the D. perkinsi group (as understood herein). However, there unfortunately is no COI sequence of Dinogamasus available on GenBank currently, which renders the dataset provided herein an important step towards the construction of a more comprehensive genetic library for the genus.

TABLE 2.

Pairwise-distance estimations based on COI data. 1–7: Dinogamasus saengdaoae sp. nov.; a–g: Xylocopa tenuiscapa; 8–9 and h–i: outgroups. Distances and standard errors are shown below and above the blank diagonals, respectively.

The re-analysis of Lundqvist's (1999) dataset with the inclusion of D. saengdaoae sp. nov. allowed us to conclude that this new species is phylogenetically closely related to both D. perkinsi and D. piperi, although this is not surprising, given the notable morphological similarities shared by the three species. Nor it is surprising the fact that they herein formed a clade without the fourth member of the D. perkinsi group, D. phillipinensis, which was placed in a clade of relatively late divergence, as previously found by Lundqvist (1999). This result implies that the D. perkinsi group, as originally proposed by LeVeque (1930a), is paraphyletic as far as Lundqvist's data are concerned. Although LeVeque provided the morphological basis for the establishment of the group, the author was also able to find a series of characteristics that clearly differentiates D. phillipinensis from both D. perkinsi and D. piperi (e.g. dorsal shield covered with moderately long hairs in the former but covered with short hairs in the two latter; see also LeVeque 1930a: 3), which are herein shown to also apply to D. saengdaoae sp. nov.

FIGURE 25.

COI haplotype networks of Dinogamasus saengdaoae sp. nov. (red circles) and their hosts of Xylocopa tenuiscapa (blue circles).

Although the peculiar metasomal acarinarium found in some species of Xylocopa was firstly noted almost two centuries ago (Brilman in 1839; see LeVeque 1930b), the relationships between mites and their carpenter bee hosts remain little understood. Various authors have suggested commensalism, mutualism, or even parasitism (see below). Whereas this may be partly due to a lack of consensus regarding the terminology commonly used in the field, the main reason seems to rest in the complexity of the associations and the difficulty in studying them. LeVeque (1930b) assumed that Dinogamasus mites feed upon excess pollen in the nest galleries of Xylocopa in South Africa, thus implying they are commensals. Skaife (1952) showed that larvae of D. braunsi Vitzthum feed off the exudations of pupae of the carpenter bee X. caffra (Linnaeus) and concluded that the host bees are neither harmed nor benefit from such association. However, it has been shown that Dinogamasus may also be mildly parasitic on the immature bees (Watmough 1974; Madel 1975; Gerling et al. 1989). It appears safe to assume that the development of such a specialized structure, as the acarinarium, must be advantageous for the hosts, hence some form of mutualism should be expected. In a notable case involving Allodynerus delphinalis (Giraud), a wasp that also has an acarinarium, it was demonstrated that the mites Ensliniella parasitica (Vitzthum) protect their hosts' immature forms by attacking the parasitoid wasp Melittobia acasta (Walker) (Okabe & Makino 2008). In another study, laelapid mites were shown to keep nests of halictid bees clean by controlling fungal contamination (Biani et al. 2009), although in this case the host bees do not have an acarinarium.

Interestingly, it has long been known (e.g. Lundström 1887) that plants also developed analogous structures to bee acarinaria to shelter mites, termed acarodomatia. They consist of depressions partly covered with hairs, which are found on the vein axils located on the abaxial side of leaves. For example, predatory mites have been found in acarodomatia of Viburnum tinus L. (Parolin et al. 2011). A review of positive, neutral and negative associations between vertebrate and invertebrates, plants, fungi and microorganisms in Central European ecosystems has been recently published (Gigon 2020).

FIGURE 26.

Majority-rule consensus tree obtained from the parsimony analysis of Lundqvist's (1999) morphological dataset with the inclusion of Dinogamasus saengdaoae sp. nov. (red branch). and D. kazerunensis. Numbers above internodes represent the percentage of times that corresponding clades were found among the most parsimonious trees.

TABLE 3.

Character matrix used in our parsimony analysis of Dinogamasus. ? = missing data.

(Continued)

It has been demonstrated that the same carpenter bee species may host multiple species of Dinogamasus over its geographical range (Vitzthum 1919; LaVeque 1930a; see also Figure 3). Actually, today we know that synhospitality (i.e. the association of two or more parasites with one host species; Eichler 1966) is a relatively common phenomenon in the relationship between Dinogamasus and Xylocopa. For instance, X. latipes is a known host of both D. perkinsi and D. philippinensis, while X. tenuiscapa, which was previously found in association with both D. piperi and D. perkinsi (Vitzthum 1919, 1930; LeVeque 1930a), is herein shown to also host D. saengdaoae sp. nov. (Figure 3). In particular, the relationship between X. tenuiscapa and its associated mite complex represents a clear case of phylogenetic synhospitality (sensu Bochkov & Mironov 2008), since the three Dinogamasus species comprise a monophyletic group (Figures 22, 24). Phylogenetic synhospitality occurs because mites typically evolve faster than their hosts, mainly due to shorter generation times (Kaltz & Shykoff 1998; Paterson & Banks 2001). In fact, we showed herein that the intraspecific variation in COI is greater in D. saengdaoae sp. nov. than in X. tenuiscapa, as shown by overall higher pairwise distances (Table 2) and greater number of haplotypes (Figures 22, 24, 25), which in turn point to a faster evolution rate. Phylogenetic synhospitality could in theory have occurred as a result of a disjunct distribution of X. tenuiscapa, which would have allowed for allopatric speciation within the D. perkinsi group; however, the distribution of the three mite species overlap with each other (Figure 3), rendering this hypothesis inapplicable to the case. It thus seems more likely that speciation may have taken place in sympatry in response to different evolutionary adaptations to the host body (Bochkov & Mironov 2008).

Surprisingly, males of only two Dinogamasus species (both from Africa) have so far been found and described. One of them is D. amaniensis Vitzthum whose two known deutonymph males were re-examined by Lundqvist (1999). The same author also described a single adult male of D. occidentalis Lundqvist, and noted that in males of both D. amaniensis and D. occidentalis the genu and tibia of leg I lack modified ad3 and setae pd3, a remarkable sexual dimorphism as such setae are found in females of the two species (Lundqvist 1999). Given the specialized morphology of setae ad3 and pd3, it is possible that they are used by females of Dinogamasus to keep themselves more firmly attached to female carpenter bees. This would also explain, at least to some extent, the reason why males of the genus are almost never found in association with the bees.

It would be most desirable that research on Dinogamasus be expanded to include, besides taxonomy, also study of the biology of the nest inquilines of Xylocopa bees. Results may offer important insights into a wide range of roles that might be played by mites in hymenopterans, with possible applications to beekeeping and wild bee conservation.