Description of larval morphology and phylogenetic relationships of Heterotemna tenuicornis (Silphidae)

Providing clear and detailed morphological descriptions of endemic species in limited areas enables new knowledge of their biology and ecology to be obtained through citizen science. This information can be further used for their protection. Our study presents the first morphological description of the larvae of all three instars of Heterotemna tenuicornis (Brullé, 1836), an endemic species of the Canary Islands that, together with H. britoi García & Pérez, 1996 and H. figurata (Brullé, 1839), belongs to the peculiar genus Heterotemna Wollaston, 1864. Furthermore, we present the first sequences of two mitochondrial genes (COI, 16S) obtained from larval specimens, and cross reference them with sequences from an adult specimen. Phylogenetic analysis of molecular data placed the genus Heterotemna within the genus Silpha Linnaeus, 1758, suggesting paraphyly of Silpha. In our study, we underline the importance of using a combination of morphological description and molecular data, that can be used for barcoding developmental stages which could not otherwise be definitely associated.

forest (laurisilva), pine forests, and alpine scrub 21 together with relatively recent volcanic activity and diverse landscapes created suitable conditions for Tenerife's high endemism and unique composition of habitats [21][22][23] .
The subfamily Silphinae consists of 18 genera or subgenera 24 and Heterotemna Wollaston, 1864 is the last genus of Silphinae lacking formal larval description and a clear phylogenetic placement within the subfamily. Only a single photograph of the dorsal side of a probably 3rd instar larva has been published so far, but without any further comment 25 . Relationships and taxonomic position of genera inside the internal group of the subfamily Silphinae 26 were established based on morphological resemblances of adults, and they are not generally agreed upon. For example, the genera Phosphuga Leach, 1817 and Ablattaria Reitter, 1885 are often considered to be subgenera of the genus Silpha 24,27 . Detailed study coupled with molecular phylogenetic analysis would contribute towards our understanding of the evolutionary relationships within the subfamily Silphinae.
All three recognized Heterotemna species are considered to be forest inhabitants. However, H. britoi García & Pérez, 1996 and H. figurata (Brullé, 1839) can be also found in more open environments. Heterotemna tenuicornis is a common species that can be found in the interior of the laurel forest in the Teno and Anaga regions (Fig. 2b) under logs and in the litter throughout the year 25 . Based on its limited geographical range, confined only to Tenerife and La Palma, the whole genus Heterotemna could face several challenges. Until recently, Heterotemna were the only Silphidae known to inhabit the Canary Islands. However, the closely related Silpha puncticollis Lucas, 1846 was reported from Tenerife in 2010, suggesting its recent introduction; it has formed a viable population in La Laguna city 28 , and could pose a threat to the native species. Furthermore, climatic change is a serious threat to species with limited range and confined to small islands, as they have very limited options of evasion of unsuitable conditions. Knowledge of larval morphology is thus crucial for making the species available for citizen science-based studies and consequently for conservation measures.
The aims of this study are to produce the first morphological description of all instars of larvae of H. tenuicornis, and to investigate the phylogenetic placement of the genus at the subfamily level, and its association with the genus Silpha, based on molecular data (16S and COI).

Results
In total 48 larval specimens of H. tenuicornis were obtained and analysed. We identified 30 larvae of the first instar, 14 of the second instar and 4 of the third instar. Two larvae and one adult specimen of H. tenuicornis were used for molecular phylogenetic placement of the genus within the subfamily Silphinae. The phylogenetic tree was obtained using Bayesian analysis from the concatenated partial 16S (434 bp) and COI (609 bp) sequences (Fig. 1).
Species identification based on genetic distances. The calculated p-distances between concatenated sequences of 16S and COI of larval and adult specimens of H. tenuicornis were between 0.0029 and 0.0078 (the mean calculated p-distance within Heterotemna specimens was 0.01). Conversely, the distance between different species of Silpha was shown to be higher (mean calculated p-distance within the Silpha species was 0.08), thus the larval specimens were confirmed as belonging to the same species as the adult specimen, H. tenuicornis (SM1).  Morphometry. The two commonly used measurements for instar identification, head width and width of protergum , are applicable in the case of H. tenuicornis (Fig. 2c, d) as these two measurements do not overlap between the instars and show significant differences. More specifically, the following measurements were very different between instars; head width (F statistic = 231 on 2, df = 45, p value < 2.2e−16), protergal width (F statistic = 4.109 on 2, df = 45, p value < 2.2e−16). Significant difference was also observed in the length of the first segment of the maxillary palpus (F statistic = 9.181 on 2, df = 44 , p value < 0.0004653), all three antennomeres (AI F statistic = 112. 3 (Fig. 3a, d). Short median unsclerotized line extending beyond epicranial stem anteriorly (Fig. 6e). Epipharynx anterolaterally with two pairs of sensory pegs on its heavily sclerotized margin (Fig. 7b, lsp). Ventral epicranial ridges present, extending past the posterior edge of the hypostomal ridge (Fig. 4d). Antenna very slender and elongated (Figs. 4a, 5a). Antennomere II with large, round and flattened www.nature.com/scientificreports/ sensorium, bearing several sclerotized pores surrounded by narrow, sclerotized ring (Fig. 6f), externally with a wide unsclerotized area. More apically, antennomere II with three additional smaller sensilla (Figs. 4a, 5a, 6f). Anterior margin of protergum medially with narrow emargination (Fig. 8i). Rudimentary spiraculum present on metasternum (Fig. 3b). Paratergites of meso-and metathorax and abdominal segments I-VIII resembling the shape of pig ears, constricted posteriorly between tergite and paratergite, with apex pointed posteriorly (Figs. 3a, 4c, g). Ventrite II on abdomen entire, not subdivided into three sclerites (Fig. 3b).   (Fig. 3a-g), dorsoventrally distinctly flattened (Fig. 3c). Terga well-sclerotized and sparsely and regularly covered with short, recumbent setae (Fig. 4b, c). Dorsal surface dull, sparsely granulate, with regular isodiametric microsculpture (Fig. 4b). Paratergites of meso-and metathorax and abdominal segments I-VIII wide, resembling the shape of pig ears, constricted posteriorly between tergite and paratergite, with apex pointed posteriorly (Figs. 3a, 4c, g). Abdominal ventrites more glossy, without distinct microsculpture, posterior margin of ventrites with row of long, semierect setae (Fig. 4e). Dark brown pigmentation present on thoracic and abdominal sclerites, as well as on bases of paratergites. Remaining area of paratergites and lateral edges of protergum ochre (Fig. 4c, g). Laterally, tergites with darker brown spots and shallow depressions, arranged in two, slightly irregular, longitudinal rows; medially, dorsum with pale ecdysial line, reaching posteriorly to abdominal tergite IV. Additionally, pair of lighter subcircular translucent spots present on the frontal medial area of protergum, duller and not clearly delimited www.nature.com/scientificreports/ ( Fig. 3a, b). Instar II. Pair of lighter subcircular translucent spots on protergum not clearly delimited ( Fig. 3f-g). Mean total body length: 15.55 mm ± 1.16 mm. Instar I. Two lighter subcircular spots present on protergum more apparent, well-delimited ( Fig. 3d-e). Mean total body length: 13.02 mm ± 0.77 mm.
Maxilla (Fig. 7c-e) Instar III. Attached closely laterally to labium. Cardo broad (Fig. 7e, cd), separated in ventral view into two parts by narrow, heavily sclerotized, longitudinal apodeme (joined anteriorly to inner margin of stipes). Base of cardo with one large seta placed on outer ventral margin and two smaller setae, one placed dorsally and another ventrally. Stipes elongate, subrectangular; ventral side regularly sclerotized, with more distinctly sclerotized apodeme on inner margin, joining cardo with inner base of lacinia (Fig. 7e, st). Ventral surface with group of several short setae and one large seta in the middle of the length. Outer lateral margin with another small seta and additional single large seta, placed more anteriorly. Dorsal surface only narrowly sclerotized along outer margin, with irregular row of short setae and pores (Fig. 7d, st). Most of dorsal surface weakly sclerotized. Lacinia and galea basally fused, separated only apically (Fig. 7d-e). Sclerotized basal fused part ca. 1.4 as long as wide. Distinct cuticular spines present on inner lateral and dorsal base of lacinia; base in dorsal view with two distinctly sclerotized, slightly dorsally elevated regions (Fig. 7d). Fused basal part on outer margin in ventral view with two large, laterally pointing setae, and several small setae and pores. Another two large setae present at base. Free apical part of lacinia distinctly sclerotized, with elongate longitudinal apodeme in ventral view. Inner margin with heavily sclerotized lobes, with 9-11 large, stout teeth. Lacinia elongate apically, apex shorter than galea. Galea with a compact, relatively small brush of setae (Fig. 7d, e, ga). Maxillary palpus trimerous, short basal palpifer present (Fig. 7c, mpf). Third segment cylindrical, apically with group of sensilla. Proportions of maxillary palpomeres: MPI 0.248 mm ± 0.009 mm, MPII 0.277 mm ± 0.057 mm, MPIII 0.468 mm ± 0.026 mm. Instar II: Basal fused part of lacinia and galea shorter and wider, sclerotized part only slightly longer than wide. Galea with large brush of setae. Proportions of maxillary palpomeres: MPI 0.189 mm ± 0.032 mm, MPII 0.245 mm ± 0.024 mm, MPIII 0.396 mm ± 0.087 mm. Instar I: Basal fused part of lacinia and galea short and wide, sclerotized part as long as wide. Galea with large brush of setae. Proportions of maxillary palpomeres: MPI 0.158 mm ± 0.041 mm, MPII 0.199 mm ± 0.029 mm, MPIII 0.384 mm ± 0.067 mm.
Labium (Fig. 8e-g) Instar III. Submentum present ventrally at base of maxillo-labial complex aslarge, broadly trapezoidal sclerite, posterolateral corners extending to base of cardo, and to transversely oriented posterior tentorial pits (Fig. 4d). Anterior part weakly sclerotized, distinctly separated from mentum. Surface postero-laterally with two pairs of large setae, surface irregularly covered by several additional short setae. Mentum wide, sclerotization extended also laterally, with two pairs of very large setae positioned laterally, surrounded by additional several short setae and pores (Fig. 8e, mt). Prementum transverse, narrowing anteriad, irregularly sclerotized. Laterally with one pair of large setae, medially with several small setae (Fig. 8e, pm). Ligula bilobed, heart-shaped, sclerotized laterally and basally, each lobe along the sagittal plane covered dorsally by numerous longitudinal lines of fine short setation and dense bulbous projections apically and centrally between the two lobes (as on Fig. 8f, li). Labial palpus bimerous; basally with longitudinal, laterally sclerotized palpifer (as on Fig. 8f, lpf). Basal palpomere club-shaped, laterally with four and ventrally with two short setae, distal palpomere conical, with several very short setae and several sensilla laterally and apically. Instar II. Mentum ventrally with one pair of very long and one pair of slightly shorter setae laterally (Fig. 8g). Instar I. Setae on mentum as in instar II.

Discussion
Our study reports the first DNA sequences of the genus Heterotemna, allowing molecular identification of the genus. This can be useful, especially if dealing with incomplete specimens or stages that do not allow morphological identification, such as eggs and pupae. Based on the DNA sequences, we were able to infer the phylogenetic placement of the genus within the subfamily Silphinae. The phylogenetic tree presented in this study is mostly in agreement with the current molecular phylogeny of the subfamily 26,29,30 . The aim of our study was not to reconstruct the phylogeny of the entire Silphinae, but provide a tentative placement of the genus Heterotemna. For this purpose, we used representatives of several recognized genera-Ablattaria, Necrodes, Oiceoptoma, Phosphuga, Silpha, and Thanatophilus. Our results show that the monophyly of the genera Necrodes, Oiceoptoma, and Thanatophilus is well supported and in accordance with previous works. In our study we also observed a well-supported clade that consists of the genera Ablattaria, Heterotemna, Phosphuga, and Silpha. As the general topology of the presented phylogenetic tree is in agreement with previous studies, we assume that it can be considered a reliable estimate of the evolutionary relationships of the genus Heterotemna to other silphine genera. The monophyly of the clade containing Ablattaria, Aclypea Reiter, 1885, Dendroxena Motschulsky, 1858, Silpha, and Phosphuga is well supported 26,29,30 . However, relationships and taxonomic positions of genera inside the internal group of the subfamily Silphinae 26 are unresolved and there is an ongoing discussion regarding the genera Phosphuga and Ablattaria which are either considered subgenera of the genus Silpha or separate monophyletic genera 24,27 . Furthermore, evidence for paraphyly of the genus Silpha with respect to the genus Aclypea was indicated by the previous phylogenetic study 29 . Our results add up more evidence toward suggested paraphyly also with the respect to the genus Heterotemna. The most pragmatic taxonomic solution would to treat Heterotemna as a junior synonym of Silpha. To fully resolve the issue is out of scope of our study as more robust phylogeny at the genus level is needed. Such a step will require the inclusion of additional Silpha species, e.g., possibly closely related S. puncticollis, and other species of the internal group. Based on the phylogenetic tree presented in this study, Silpha tristis was placed as a sister species to H. tenuicornis. Silpha tristis is a widespread species in the Western Palaearctic region, also recorded in North Africa (Morocco) 31 but not present on the Canary Islands. The only Silpha species occurring in Tenerife is the introduced Silpha puncticollis, a western Mediterranean species 28 possibly closely related to S. tristis. The effect of the presence of S. puncticollis on endemic species of Heterotemna has not been studied. However, as there is very limited information on the ecology and biology of both species, we believe, that it is crucial to be able to recognize all life stages of the endemic species from the newly introduced one. Adults of these two species are easily distinguished from each other by dense, fine and uniformly punctured pronotum with thick and reflexed borders, and the absence of protuberances or carina on its disk, and elytra with elevated costae in S. puncticollis 28,32 , and disk with a pair of median longitudinal carinae and elytra with much finer costae in Heterotemna 25 (Fig. 2a). However, the morphology of larvae of genus Heterotemna has not been previously described and we hereby provide the first clear detailed description.
The larvae fit the general body plan and features typical for the larva of Silphinae, summarized as 33 : relatively large (usually over 12 mm in L3); body slightly to strongly flattened, body surface heavily pigmented and sclerotized; head with 6 pairs of stemmata; mandible without a molar lobe or prostheca; maxilla with broad, Scientific Reports | (2021) 11:16973 | https://doi.org/10.1038/s41598-021-94744-x www.nature.com/scientificreports/ apically cleft basal fused part bearing galea with dense setal brush on outer lobe; urogomphi articulated, usually 2-segmented. The larva of Heterotemna differs from other known larvae of Silpha as follows: (1) protergum with two lighter subcircular spots anterolaterally on disc; anterior margin medially with narrow emargination (in Silpha, protergum differently coloured: unicolour or with paler postero-lateral part; anterior margin simple, regularly rounded or very widely emarginated-see 34 for details). However, the larva of S. puncticollis is unknown.
In our study we did not observe an overlap in head width and protergal width between all three instars. Variation in larval head and protergal width was observed across Silphinae 35,36 for example in Thanatophilus 38 , therefore, it seems that by using these two measurements, instars of H. tenuicornis can be reliably distinguished. In addition, the pair of lighter spots on the protergum seemed to be also one of the characteristics to distinguish between instars (being most contrasted and sharply delimited in first instar larvae than in second and third instar). However, larvae of all three instars that were available for this study varied in date of collecting and some of the specimens were collected as long as 13 years ago, therefore, we are cautious as the colours may have changed during storage, as can be seen in other species of Silphinae (M. Novak, unpubl. data). Therefore, we focused mainly on morphological structures and use the coloration as secondary trait of instar determination. Larvae of H. tenuicornis seem to share some morphometric characters with other species of the subfamily Silphinae. There is a general pattern of abrupt increase of the ratio between the length of the first and the second segment of the urogomphi as well as between the second segment of the urogomphi and the terminal seta when comparing first and second instars. A similar pattern was also observed in two species of the genus Thanatophilus 37,38 and our preliminary observations also confirm this for larvae of Diamesus osculans (Vigors, 1825).
In order to obtain more information regarding the biology and ecology of the genus Heterotemna we created a basic identification key, that could encourage data collecting. Identification based on morphology of focal species is a crucial tool when it comes to collecting information in the field as well as studying the ecology and adaptations of the species. It is still an irreplaceable discipline that cannot be neglected or replaced, even by fast-evolving molecular methods.

Materials and methods
Specimen sampling. Specimens were collected in 2007, 2011 and 2017 (SM3) with pitfall traps, and killed and stored in 75% or 96% ethanol. In total, we obtained 48 individuals for further morphological examination. Only the specimens stored in 96% ethanol were used for phylogenetic analysis. As the larvae were collected in the wild and were not raised to adults, the species identification could not be confirmed morphologically. Therefore, we obtained a single adult specimen of H. tenuicornis to confirm the molecular identification of the larvae as the same species. Genetic distance was calculated using the COI and 16S genes among all taxa used for the phylogenetic analysis, and the adult specimen of H. tenuicornis was used to confirm the conspecificity of the larval H. tenuicornis using DNADIST version 3.5c implemented in BioEdit v7.0.5.3 39 .
Morphological analysis. The morphological terminology used in this paper follows Lawrence and Ślipiński 40 and Novák et al. 38 . The terminology of the epipharynx follows Dorsey 41 and Anderson 42 . Morphological characters were measured, documented, and examined using a Keyence VHX-6000 digital microscope. To observe detailed structures located within the head capsule, the head was detached and submerged in hot (90 °C) 10% potassium hydroxide (KOH) for 3 min. The head capsule was subsequently dissected and detailed structures of the epipharynx and the head appendages were observed.
To observe very fine structures such as setae and pores, we used a scanning electron microscope. Preparation of samples follows the methodology of Novák et al. 38 . Selected specimens were dehydrated using a graded series of ethanol (75%, 80%, 90%, 95%, 100%) and left in each concentration for approximately 30 min. before transferring to acetone overnight. Dehydrated samples were dried using the critical point drying method. Dry samples were then attached to an aluminium disk target using copper foil tape and coated with gold in Bal-Tec Sputter Coater SCD 050. Samples were observed and documented with a JSM-6380LV (JEOL) scanning electron microscope.
The following morphological characters were measured: The length of the first antennomere (AI), the length of the second antennomere (AII), protergal width (N1W), head width (at the widest point) (HW), length of the first urogomphal segment (UI), length of the second urogomphal segment (UII), length of urogomphal seta (when present) (US II), length of all three palpomeres (MPI, MPII, MPIII). The morphological characters are described based on the third instar (L3) larvae followed by observed differences in second (L2) and first instar larvae (L1). The mean and standard deviation (± SD) of characteristics measured is indicated in the text.
The changes in the sizes of the measured morphological characters and their ratios throughout the developmental stages were tested using a linear model with a normal distribution of errors. The significance level was set at 5%. The analysis was carried out in R program (R Core Team 2020). Graphical outputs were created using ggplot2 and sjPlot packages 43,44 . Molecular analysis. Genomic DNA was extracted using commercial Tissue & Blood Kit (Geneaid, New Taipei City, Taiwan) following the protocol provided. Two mitochondrial genes were partially amplified-Cytochrome oxidase I (COI) using the primer pair "Jerry" 5′-CAA CAT TTA TTT TGA TTT TTTGG-3′ and "Pat" 5′-TCC AAT GCA CTA ATC TGC CAT ATT A-3′ 45 and 16S (rDNA) using following primer pairs "LR-J-12887" 5′-CTC CGG TTT GAA CTC AGA TCA-3′ and "LR-N-13398" 5′-CGC CTG TTT ATC AAA AAC AT-3′ 45 and "16SL" 5′-ATT CTA AAT YYA WNG CAC TAW TCT GCC AAA-3′ 46 and "16SAH" 5′-YGC CTG TTT AWY AAA AAC ATG-3′ 47  Phylogenetic analyses. The electropherograms obtained were proofread and corrected for miss-called bases in Chromas 2.6.6. (Technelysium Pty Ltd, South Brisbane, Australia). Additional sequences used for the phylogenetic analyses were obtained from GenBank (National Centre for Biotechnology Information, https:// www. ncbi. nlm. nih. gov/ genba nk/) (SM2). Multiple sequence alignments were generated with MAFFT version 7 48 using The Guidance2 Server 49 . Aligned sequences were further manually edited in BioEdit 7.0.5.3 39 . Concatenated sequences consisting of 16S and COI were analysed under the criterion of maximum parsimony (MP) using PAUP 4.0a 50 . The MP analysis was conducted with heuristic search and 10,000 bootstrap replicates. The GTR + I + G evolutionary model was selected in jModelTest 2 51,52 for both genes (16S and COI) using the Akaike Information Criterion (AIC) 53 . The tree topology was estimated using Bayesian phylogenetic inference (BI) based on selected evolutionary model (GRT + I + G; invgamma) for both partitions using MrBayes 3.1.2. software 54,55 .
The search was conducted for two simultaneous runs with four independent chains for 10,000,000 generations, sampled every 1000 generations, the average standard deviation of split frequencies reached 0.002093. The first 25% generated trees from both runs were discarded as burnin. Maximum likelihood (ML) analysis was conducted using IQ-TREE web server 56 based on GTR + I + G model and 10,000 bootstrap replicates. Additionally, COI and 16S sequences obtained from an adult specimen were compared to relevant sequences from larval specimens in MEGA 10.1.07 57 , sequences were analysed among and within groups (group Silpha: S. tristis, S. obscura, S. perforata, S. carinata, S. olivieri and group Heterotemna: H. tenuicornis (Larva) 1, H. tenuicornis (Larva) 2, H. tenuicornis (Adult)) using K2P model. The outgroups, two species of the family Staphylinidae, were selected based on previous studies: Scaphidium quadrimaculatum (subfamily Scaphidiinae) 30 and Aleochara curtula (subfamily Aleocharinae) 29 .

Data availability
We provide following data used in our study-Measurements and ratios of all three larval stages of H. tenuicornis (in millimetres) (SM4), GenBank accession numbers of all sequences used in our study (SM2), including newly generated sequences that have been submitted in GenBank.