First Descriptions of Larva and Pupa of Bagous claudicans Boheman, 1845 (Curculionidae, Bagoinae) and Systematic Position of the Species Based on Molecular and Morphological Data

In this paper, the mature larva and pupa of Bagous claudicans are described and illustrated for the first time. Measurements of younger larval instars are also given. The biology of the species is discussed in association with larval morphology and feeding habits. Overall larval and pupal morphological characters of the genus Bagous are presented. Confirmation of the larva identification as Bagous claudicans species was conducted by cytochrome oxidase I (COI) sequencing. DNA barcoding was useful for specimen identification of larval stages. The systematic position of the species within the Bagous collignensis-group, based on morphological and molecular results, is also discussed.


Introduction
The globally distributed (except for Central and South America) weevil genus Bagous Germar, 1817 includes about 300 valid species, of which approximately 130 occur in the Palaearctic region, 82 in the Western Palaearctic, and 31 in Central Europe [1][2][3][4][5][6][7]. This makes it one of the most numerous among the weevil genera and the largest group of hydrophilous beetles, which comprise less than 1% of all known coleopteran species [8]. The genus has received much taxonomic revision (e.g., [4,[9][10][11][12]), and a comprehensive classification of the tribe Bagoini Thomson, 1859 was recently proposed following a global phylogenetic analysis performed by Caldara et al. [13].
Only several of the Bagoini can be regarded as well-known with reference to their biology and plant association (e.g., B. nodulosus Gyllenhal, 1836, B. glabrirostris (Herbst, 1795)). Based on records collected to date, most are regarded as nocturnal, inhabiting aquatic biotopes, e.g., ponds, old riverbeds, swamps, floodplain meadows and forests, where many of them develop on submerged vegetation. But several species (such as B. tempestivus (Herbst, 1795), B. lutulosus (Gyllenhal, 1827) or B. diglyptus of the head capsule (HW). In pupae, body length (BL), body width (BW) (at the level of middle legs), and width of pronotum (THW) were given.
The observations of chaetotaxy and measurements were conducted using a light compound microscope with calibrated ocular graticules. Drawings and outlines were made using a drawing tube (MNR-1) installed on a stereomicroscope (Biolar, Polskie Zakłady Optyczne, Warsaw, Poland) and processed by computer software (Corel Photo-Paint X7, Corel Draw X7). Photos were taken using an Olympus BX63 microscope and processed by Olympus cellSens Dimension software. The larvae selected for pictures using SEM (scanning electron microscope) were at first dried in absolute ethyl alcohol (99.8%), rinsed in acetone, critical-point dried and then gold-plated. TESCAN Vega 3 SEM was used for the examination of selected structures. General terminology and chaetotaxy follow Anderson [40], May [32], Marvaldi [41][42][43][44], and Skuhrovec et al. [45], with antennae terminology following Zaharuk [46].

DNA Extraction, Amplification and Sequencing
Eight specimens (one larva and seven imagines) were used for molecular analysis. Before DNA extraction, all specimens were cleaned using ethanol and distilled water in order to reduce the risk of contamination. DNA was extracted from whole insect body (without any infraction in case of two specimens, they were retained as morphological vouchers, see above), as well as without protocol modification using the NucleoSpin Tissue kit (Macherey-Nagel, Düren, Germany). To elute purified DNA, 100 µL of Elution Buffer were applied onto the silica membrane. To amplify the barcode fragment of cytochrome oxidase gene, the following primer pair was used: LepF1 and LepR1 [47].
Polymerase chain reaction (PCR) amplification for all DNA fragments analyzed was carried out in a final volume of 20 µL containing 30 ng of DNA, 1.25 U GoTaq G2 Flexi (Promega, Madison, WI, USA), 0.8 µL of 20 µM of each primer, 4 µL of 5x PCR buffer, and 0.4 µL of 10 mM dNTPs in a Mastercycler ep system (Eppendorf, Hamburg, Germany). The cycling profile for the PCR was as follows: 95 • C for 2 min, 35 cycles of 95 • C for 1 min, 50 • C for 1 min, 72 • C for 1 min, and a final extension period of 72 • C for 7 min.
In order to assess the quality of the amplification, PCR products were electrophoresed in 1% agarose gel for 45 min at 85 V with a DNA molecular weight marker (Mass Ruler Low Range DNA Ladder, Thermo Fisher Scientific, Waltham, MA, USA). PCR products were purified using Exo-BAP Mix (EURx, Gdańsk, Poland).
Samples were sequenced in both directions using the same primers as for PCR reactions in combination with the Bright Dye Terminator Reaction Ready Mix v. 3.1 (Nimagen, Nijmegen, the Netherlands) using the chain termination reaction method [48]. The sequencing reaction was conducted using the PCR product in a total volume of 10 µL, containing 1 µL Bright Dye Terminator Reaction Ready Mix v. 3.1 (Nimagen), 1.5 µL 5× sequencing buffer (Nimagen), 3.2 mol/ µL primer solution, and 3 µL purified PCR product. The cycle-sequencing profile was 3 min at 94 • C followed by 30 cycles of 10 s at 96 • C; 5 s at 50 • C; and 2 min at 60 • C.
Sequencing products were precipitated using ExTerminator (A&A Biotechnology, Poland), and were separated on an ABI PRISM 377 DNA Sequencer (Applied Biosystems, USA). Sequences for Bagous claudicans are available in the GenBank database under the following accession numbers MK533683-MK533690.

Sequence and Data Analysis
Raw chromatograms were evaluated and corrected in Geneious R10 (https://www.geneious.com). The possibility of having sequenced numts [49,50] was rejected by translating DNA data into amino acid sequences using invertebrate genetic code within Geneious R10.
The nucleotide sequences were verified using BLAST (Basic Local Alignment Search Tool) searches of NCBI (National Center for Biotechnology Information) (http://blast.ncbi.nlm.nih.gov/Blast. cgi). The alignment of the studied sequences was performed using the MAFFT [51] plugin within Geneious R10.
The mtCOI sequences for the remaining Bagous species were retrieved from GenBank to perform the phylogenetic analysis.
Phylogenetic trees were reconstructed using Bayesian inference (BI) and maximum likelihood (ML). The most appropriate substitution model of evolution was inferred using jModelTest [52]. The Akaike information criterion was preferred over the hierarchical likelihood ratio test to compare the various models as recommended by Posada and Buckley [53]. Bayesian analyses were performed with MrBayes 3.2 [54], consisting of two Markov chain Monte Carlo (MCMC) analyses run for 1,000,000 generations, with trees sampled every 100 generations and using four chains and default priors. Convergence of each run was also visually inspected using Tracer [55]. An initial10% of sampled trees were discarded as burn-in and a majority-rule consensus tree was obtained in TreeAnnotator v1.10.1 [56] ML analysis was performed using PhyML3.0 [57]. Support values were computed with 1000 bootstrap replications. All trees were visualized with TreeView v.1.6.6 (http://tree.bio.ed.ac.uk/software/figtree/) and graphically edited in CorelDraw X8 (CorelDraw Corporation, Ottawa, ON, Canada).
Pairwise distances were calculated using MEGA v.7 [58]. Analysis of other metrics (such as haplotype and nucleotide diversities) was not performed due to the low number of available sequences per species.

Description of the Larva of Bagous claudicans
General morphology ( Figure 1A-F, Figure 2A-F). All thoracic and abdominal segments were white-yellow ( Figure 1A). Cuticle densely covered with asperities ( Figure 2E,F). Pronotal area of first thoracic segment feebly sclerotized, light yellow. Body very slender, elongated, slightly curved ( Figure 1A), round in cross section. Prothorax slightly shorter than the subequal meso-and metathorax. Abd. 1-6 of almost equal length, slightly longer than metathorax; Abd.7-9 decreasing gradually to the terminal parts of the body; Abd. 10 reduced to four anal lobes of unequal size (dorsal the biggest, ventral the smallest, lateral lobes slightly smaller. Dorsal parts of Abd. 1-7 divided into three lobes, Abd. 8 with two dorsal lobes. Lateral folds of Abd. 1-9 well isolated, on segments 6-9 developed into conical protuberances. Nine pairs of bicameral spiracles, first pair placed on anterior margin of pronotum, next seven pairs latero-medial, the last pair placed dorso-laterally on Abd. 8 ( Figure 1B-F, Figure 2A-D).
Chaetotaxy ( Figure 1B-F, Figure 2A-D). Setae light yellow, fine, trichiform, varying in size, feebly developed, sometimes poorly distinguished from asperities ( Figure 2E, F). Prothorax on each side ( Figure 1B) with seven prns of unequal length (three long and four short, all located on sclerotized shield), two ps of various length, and one short eus. Mesothorax ( Figure 1B) on each side have one minute prs, three pds (first and third minute, second relatively long), one medium as, one long eps, one medium ps and one minute eus. Chaetotaxy of metathorax ( Figure 1B) similar to mesothorax. Each pedal area of thoracic segments has one long pda. Abd. 1-8 with one minute prs, four pds (first, third, and fourth minute, second long), one minute ss, one long eps, one medium lsts and two minute eus. Abd. 9 on each side has one very long ds, one long ps, and two minute sts ( Figure 1C-F). Each lateral lobe of Abd. 10 has two minute ts ( Figure 1F).
Head and antenna ( Figure 3A-F). Head light yellow, slightly narrowed bilaterally, frontal suture distinct, Y-shaped ( Figure 3A). Setae on head trichiform, various in length. Des 1 , des 3 and des 5 elongated, equal length, des 2 very short, des 4 absent. Des 1 and des 2 placed in central part of epicranium, des 3 on frontal suture, and des 5 located anterio-laterally ( Figure 3A-C). Fs 4 as long as des 1 , placed antero-laterally, close to epistoma. Les 1 very short, les 2 slightly shorter than des 1 . Post epicranial area has five very short pes 1-5 ( Figure 3A). Frons with two pairs of pores placed medially; epicranium with two pairs of pores: one placed near to des 1 , next near to des 5 . Antenna located on end of frontal suture; antennal segment with Se elongated, located medially; basal membranous article with five sb and two sa ( Figure 3D-F).
Mouth parts ( Figure 4A-E, Figure 5A-E). Labrum approximately 3.0 times as wide as long, with three pairs of lrs 1-3 of various length; lrms 1 medium, lrs 2 very long, lrs 3 very short, all lrs placed medially ( Figure 4B). Anterior margin of labrum slightly round emarginate. Clypeus 3.5 times as wide as long, with two short, triangular, and equally long cls 1,2 and one clss between them; all localized posterio-medially. Anterior margin of clypeus gently arcuated inwards. Epipharynx ( Figure 4A, C-E) with three pairs of als 1-3 of various length (first and second finger-like, third the longest more hair-like); two pairs of short ams 1,2 (first short, finger-like; second medium length, hair-like) and two pairs of conical mes 1,2 . Lr elongated, slightly converging posteriorly ( Figure 4A, C). Mandibles ( Figure 5A) relatively broad, slightly truncate, bifid, teeth of unequal length, the cutting edge almost straight. Both mds 1,2 very short. Maxillary stipes ( Figure 5B) have one short stps and two pfs 1,2 (first very short, second elongated); mala with one minute mbs, seven bacilliform or finger-like dms 1-7 of various length ( Figure 5C); vms 1-5 various in length and shape ( Figure 5D); vms always shorter than dms. Maxillary palpi with two palpomeres almost of equal length; basal with minute mps and a pore, apical with single pore. Apical part of distal palpomeres with a group of 5-6 elongated, conical, sensillae. Praelabium ( Figure 5B,D) rounded, with a pair of short prms located medially. Ligula with two pairs of hairform, micro ligs 1,2 of equal length. Premental sclerite well visible, in a form of complete ring, with elongated proximal part. Labial palpi one-segmented; each palpomeres with a pore and some short, sensillae apically. Postlabium ( Figure 5B,D) on ventral part densely covered by asperities and with three pairs of various in length pslbs 1-3 (first pair short, localized medially, second very long placed latero-medially, third very short, situated antero-laterally). Insects 2019, 10, x FOR PEER REVIEW 6 of 18
Chaetotaxy. Setae hair-like, of unequal length, yellow or light brown, on rostrum, head and pronotum based on small protuberances. Rostrum with two pairs of rs 1,2 (equal in length); head capsule bearing a pair of vs, two pairs of sos 1,2 (almost equal in length), two pairs of os 1 , 2 (various in length), and a pair of pas ( Figure 6C, D). Vs distinctly bigger than remaining setae of head and rostrum. Pronotum on each side with two pairs of as 1,2 , a pair of ls, two pairs of ds 1,2 , and three pairs of pls 1-3 . All setae of pronotum almost equal in size ( Figure 6C,D). Chaetotaxy of metathorax as on mesothorax, consisting of paired triplets d 1-3 ( Figure 6C). Setae of meso-and metathorax very long, distinctly longer than setae of abdominal segments. Each femoral apex with a pair of fes 1,2 of almost equal length ( Figure 6B-D). Each tergal part of Abd. 1-8 with four pairs of d 1-4 located medially, respective sternal parts of Abd. 1-8 with 2 pairs of l 1,2 located, close to the border with tergum. Dorsal and lateral setae of Abd. 1-8 very short, equal in length. Abd. 9 with two pairs of setae: first placed close to gonotheca, second on urogompia ( Figure 6B-D).

Genetic Results
Phylogenetic inferences were obtained for COI gene using Bayesian inference (BI) and Maximum likelihood (ML). The GTR+G+I model with gamma correction of 0.478 and invariable sites of 0.456 was selected by the AIC in jModelTest for the matrix. Heuristic searches resulted in one ML tree (−ln = (−6064.16432). The Bayesian inference and maximum likelihood analyses resulted in similar trees, the only differences between them being the degree of statistical support for the recovered nodes ( Figure 7). Nodal supports were generally poor across all backbone nodes. Tychius schneideri (Herbst) was used to root the topologies. The phylogram shows the existence of two clades. The first one formed by Bagous limosus, B. frit, B. longitarsis, B. collignensis and B. claudicans. The second had B. claudicans from Germany (based on data from BOLD Systems). The second clade consists of remaining Bagous species (Figure 7).
The B. claudicans forms in the first clade two phylogenetic lines, one with all newly investigated samples, both larva and imago from the Polish locality and the second with sample from Germany (based on data from BOLD Systems).
Based on the appropriate nucleotide substitution model, genetic distances between specimens ranged from 0% to 45%, while mean genetic distances between species ranged from 5% to 45% (Table 1).       [34] for Bagous.
Interestingly, the larvae, pupa and imagines of B. claudicans have been collected not from Equisetum, but instead from the stems and roots of the decorative variety of Sedum maximum, growing in the center of a city (Katowice). Sedum Linnaeus, 1753 plantings grew under unfavorable conditions: on a thin, dry substrate (insulated with a geomembrane), in an exposed area, and under strong insolation. This site was more akin to xerothermic environments than to humid habitats settled by most species from the genus. Thus, the treatment of B. claudicans as a monophagous of Equisetum is certainly not based on full knowledge of its biology and requires more detailed study. Detailed observations of this species are, however, very difficult due to its rarity and similarity to closely related B.collignensis.
Results from phylogenetic analysis confirm that investigated larva and imagines belonged to Bagous claudicans species, which was additionally supported by comparing them with the known sequence from BOLD Systems (Figure 7).
Moreover, the performed phylogenetic analysis indicates that Bagous longitarsis is a sister group to B. collignensis and B. claudicans, which confirms the previous results obtained by Caldara et al. [13].
Despite the value of the mean genetic distance between B. claudicans and B. collignensis (5% ,  Table 1), which may indicate that the investigated species can belong to the same species presented in this study, morphological and ecological data suggest the existence of two separate biological species. Also B. longitarsis are genetically close to B. collignensis (8%) and B. claudicans (12%), all these species are grouped in close related collignensis-group distinguished by Caldara [13] (Table 1).
Therefore, the hypothesis of Scherf [18] that B. claudicans is a synonym of B. collignensis must be further investigated, especially using a large number of specimens from different localities as well as additional markers, as the genetic distance within B. claudicans can range to 3% on specimens from different localities (Germany vs. Poland). Moreover, further phylogenetic studies of all the known Bagous species are required in order to understand the relationships within the Bagoini tribe.

Conclusions
Morphological characteristic of immatures of Bagous claudicans (in both, larval and pupal stages) are typical for the genus Bagous, especially for B. colligensis group. At the same time, some original features make possible distinguishing of B. claudicans from other known Bagous species. Additionally, host plant and ecological preferences of B. claudicans, meticulously analyzed during presented study, emphasize taxonomical distance between B. claudicans and B. collignensis (which was previously questioned). Furthermore, DNA barcoding confirms not only larval identification but also its usefulness in specimen identification of larval stages Moreover, it seems, that advanced study on morphology of developmental stages and ecology of Bagous are required in order to clarification some of systematic ambiguities and efficient protection of the genus. Acknowledgments: The authors are grateful to Anna Gruszka (Voivodeship Plant Health and Seed Inspection Service, Katowice, Poland) who discovered the site of B. claudicans, and then collected and reared most specimens used in this study.

Conflicts of Interest:
The authors declare no conflict of interest.