Article | Open | Published:

Hypothesis on monochromatic vision in scorpionflies questioned by new transcriptomic data

Scientific Reportsvolume 8, Article number: 9872 (2018) | Download Citation


In the scorpionfly Panorpa, a recent study suggested monochromatic vision due to evidence of only a single opsin found in transcriptome data. To reconsider this hypothesis, the present study investigates opsin expression using transcriptome data of 21 species including representatives of all major lineages of scorpionflies (Mecoptera) and of three families of their closest relatives, the fleas (Siphonaptera). In most mecopteran species investigated, transcripts encode two opsins with predicted peak absorbances in the green, two in the blue, and one in the ultraviolet spectral region. Only in groups with reduced or absent ocelli, like Caurinus and Apteropanorpa, less than four visual opsin messenger RNAs have been identified. In addition, we found a Rh7-like opsin in transcriptome data derived from larvae of the mecopteran Nannochorista, and in two flea species. Peropsin expression was observed in two mecopterans. In light of these new data, we question the hypothesis on monochromatic vision in the genus Panorpa. In a broader phylogenetic perspective, it is suggested that the common ancestor of the monophyletic taxon Antliophora (Diptera, Mecoptera and Siphonaptera) possessed the full set of visual opsins, a Rh7-like opsin, and in addition a pteropsin as well as a peropsin. In the course of evolution individual opsins were likely lost in several lineages of this clade.


Colour vision has two prerequisites1,2: receptors with different spectral responses and a neural system that can process their output in a way that preserves colour information. Ultimately the presence of colour vision in a species can only be verified by behavioural experiments. Yet, showing that photoreceptors express different opsins already gives valuable clues regarding the possibility of colour vision.

Opsins belong to the large family of G protein coupled receptors (GPCRs) which form a photoreceptive complex when covalently bound to their chromophore retinal. An incident photon causes the cis-retinal isomer to change its conformation to all-trans which leads to an interaction of the opsin with G proteins and ultimately to a de- or hyperpolarization of the receptor cell3. The wavelength of peak absorption depends on relatively few spectral tuning sites in the amino acid sequence of the opsin protein4,5,6. Choice of chromophore or use of special sensitizing pigments can further influence spectral sensitivity7. Opsins can be differentiated according to the morphology of the receptor cells they are expressed in (ciliary or rhabdomeric), the G protein they interact with, and by criteria based on gene phylogenies8. Another useful distinction is between visual and non-visual opsins, the latter being expressed in various tissues not directly involved in visual functions9. Opsin-based photoreception is universal in animals and it has been concluded that the last common ancestor of Bilateria already possessed at least nine different opsins10. Five of them were likely still present in the ancestral arthropod, with one being used for visual tasks11. This single opsin gave rise to all visual opsins observed in extant arthropods by multiple duplication events8,12.

Possession of multiple opsins is neither necessary nor sufficient for colour vision:1,13 in case only a single opsin is expressed, colour vision could still be possible by means of differential optical filtering along the light path, e.g. by screening pigments, and multiple opsins can be expressed in a single receptor to enable a broadband response. Insect vision generally is trichromatic with green, blue and UV receptors4,7. Yet the number of opsins, the ommatidial types and their arrangement differ14. For example duplication led to a total number of over 20 opsins in some dragonflies15,16,17, but it has not yet been explored how these different opsins are put to use on the receptor and neural level. Loss of visual opsin classes has been found mostly as an adaptation to special photic environments, like nocturnal, subterranean or aquatic life18,19,20. Nonetheless, a recent study21 recovered only a single long wavelength sensitive opsin in two scorpionfly species of the genus Panorpa (Mecoptera: Panorpidae) using transcriptome data. This is astonishing considering the well-developed compound eyes and ocelli in this genus and the fact that the same study found an opsin with a long- and two opsins with a short wavelength peak absorbance in the mecopteran genus Boreus21. While compound eye, ocellar and larval eye ultrastructure has been investigated in some depth22,23,24,25,26,27, only two studies so far dealt with opsins in scorpionflies (Mecoptera)21,28 and their closest relative order29, the fleas (Siphonaptera)28.

Therefore we generated transcriptomic data from the head of a female specimen of Panorpa communis and sampled published29,30 and unpublished transcriptome data generated by the 1KITE consortium ( Upon screening all transcriptome data, we found four visual opsins with different predicted spectral sensitivities, expressed in almost all investigated mecopterans. Furthermore, we report opsins of unknown, likely non-visual, function in both Mecoptera and Siphonaptera.


Animal collection

A female Panorpa communis was collected in Vienna, Austria (08/06/2016). Discrimination from the very closely related species Panorpa vulgaris31,32,33 was based on differences in the wing pattern34. The specimen was kept under natural illumination for one day, decapitated under carbon dioxide anaesthesia and the head was immediately used for RNA extraction.

RNA extraction, cDNA library preparation and sequencing

Total RNA of the whole head (including antennae) of the aforementioned Panorpa communis specimen was extracted using a RNeasy Plus Universal Mini Kit (Qiagen) according to the manufacturers protocol, with further cleanup by a column based kit (RNA clean and concentrator 5, Zymo Research). cDNA library preparation and next generation paired-end sequencing were performed by VBCF (Vienna Biocenter Core Facilities; NEB bead based mRNA Enrichment and Illumina HiSeq v4 reagent kit). Slightly over 12 million paired end reads with a nominal read length of 125 bp and a GC content of 37% were sequenced on an Illumina HiSeq. 2500 device.

Sequence data processing

Access has been granted for unpublished raw sequence reads and current assemblies generated within the 1KITE project. 1KITE data will be released only upon publication of the associated phylogenetic papers but are available upon request. Unless stated otherwise in Table 1, the 1KITE project processed whole animals. All transcriptome data provided by 1KITE were sampled, sequenced and processed like previously described35. Transcriptome assemblies, raw sequencing data or single sequences of all other species were downloaded from NCBI. Raw sequence reads of the Panorpa communis head transcriptome generated for this study is deposited at the NCBI SRA archive. The NCBI accession numbers or 1KITE library identifiers for all data used in this study are listed in Supplementary Table S1.

Table 1 Data mining results.

Before assembly the raw read data of Panorpa communis and Xenopsylla cheopis were quality trimmed using fastq-mcf 36. Thereafter de novo assembly was performed with IDBA-Tran37 using a maximum k-mer length of 115 and otherwise default settings. For Panorpa pryeri and Apteropanorpa tasmanica the cDNA libraries have been sequenced twice and the contigs from the two respective assemblies have been pooled for further analysis. The assemblies were searched for transcripts encoding opsins using custom Perl scripts for partial automation: first, a hidden Markov model profile was generated from an alignment of visual opsin protein sequences used in a previous study38. Transcripts were translated to all six reading frames and candidate sequences identified by HMMER 3.1b239 were subsequently searched using blastp 2.2.31+40 against the manually curated UniProtKB/Swiss-Prot database (release 2013–04)41. We only included sequences for further analysis if they either had a HMMER search E-value equal or better than 1E-100 or if the keyword ‘opsin’ was present in their blast description. Candidate sequences were aligned using MAFFT42 (L-INSI algorithm). Potential opsin sequences with E-values worse than 1E-15 included in these initial alignments were critically inspected: such sequences were only kept for further processing if the best hit of a search against the NCBI nr (non redundant) database was annotated as an opsin and if key structural motifs of opsins were present (for example a conserved lysine needed to form a Schiff base linkage to the chromophore). Sequences with the longest coding regions were manually identified and highly similar sequences were designated as isoforms (maximum of 5 different amino acids within the coding region, often located at the more error prone contig ends). Isoforms were excluded from further analysis (but are included in Supplementary Table S2 and in Supplementary Data S3 and S4 for full transparency). Three sequences included in the further analysis (Table S2) had frameshifts that were corrected using the program HMM-FRAME43 with default options and the same HMM profile that was used for the initial HMMER search. Out of these three, only the blue sensitive opsin transcript in Apteropanorpa tasmanica did not exhibit an ambiguity code at the site of the frameshift, so the underlying cause could either be sequencing error or an actual mutation. All identified opsin sequences are supplied as Supplementary material (S25).

Salmon 0.7.244 was used to quantify relative transcript abundance (quasi-mapping mode, k-mer length 31). A subset of eight mecopteran species was selected for this analysis based on the presence of all four visual opsins.

A number of opsin sequences from selected additional taxa (Table S1), belonging to different known opsin classes8,12,38, was added to the final set of opsin amino acid sequences. After alignment with MAFFT (L-INSI) the 5′ and 3′ untranslated regions were trimmed. To infer phylogenetic trees three different programs were used: IQ-TREE 1.645, RAxML 8.2.046 and ExaBayes 1.4.147. To test for model fit IQ-TREE was run with the -mfp (Model Finder Plus48) option, explicitly including a general time reversible (GTR) substitution model using the option -mset GTR20. An LG matrix with empirical base frequencies and 6 free rate categories (LG + F + R6) fits the data best according to the Bayesian information criterion. Under the Akaike information criterion a GTR model with empirical base frequencies and six rate categories is preferred (GTR20 + F + R6). The LG model was used for the subsequent IQ-TREE maximum likelihood (ML) analyses together with options to perform a standard bootstrap analysis (500 replicates) and an ultrafast bootstrap analysis (UFBoot49, 2000 replicates). Since trials indicated that setting the -numstop option to 500 and the -pers option to 0.3 yields better trees than with the defaults, we used these settings for the final analysis.

RAxML was utilised to reconstruct a ML gene tree of the dataset using the rapid hill climbing algorithm under the GAMMA model with the GTR substitution matrix. Using this model the program found trees with better ML than IQ-TREE with the GTR20 model. Bootstrap support was inferred using the implemented rapid bootstrap algorithm50. The number of bootstrapping replicates (549) was automatically determined by RAxML, using the “bootstop” criterion that checks for convergence of bootstrap replicates on the fly51, using the extended majority rule consensus convergence criterion (autoMRE) and the default value of −B 0.03.

For Bayesian tree inference with ExaBayes we ran 1030000 generations in four runs with two Metropolis-coupled chains each. The GTR model and random starting trees were used. An average deviation of split frequencies of 5% was set as a criterion for topological convergence and the final average potential scale reduction factor (PSRF) was 1.02 ± 0.03 and the average effective sampling size (ESS) 180 ± 98. An extended majority rule consensus tree was built discarding the first 25% of samples.

Together with the added previously characterized opsins (e.g. from Drosophila), the opsin gene trees were used to assign each siphonapteran and mecopteran opsin to a peak absorbance class (also referred to as predicted spectral sensitivity). The distinction between visual and non-visual opsins in Mecoptera and Siphonaptera as well is based solely on sequence similarity with known physiologically characterized opsins.


Results from screening the transcriptomes of 15 Mecoptera and six Siphonaptera species are shown in Table 1. In the majority of investigated mecopteran transcriptomes four visual opsins are present (Table 1, Fig. 1, Supplementary Material S2, S3, S4, S5 and Supplementary Fig. S6): LwA = long wavelength sensitive A, LwB = long wavelength sensitive B, Blue = blue sensitive and UV = ultraviolet sensitive. The ocelli-less species Caurinus tlagu potentially lacks LwA while in the likewise ocelli-less Apteropanorpa species no Blue and only a very short potential LwA candidate were recovered. LwA could not be found in the sequenced Nannochorista larvae. Neither blue or UV sensitive opsins were detected in all examined Siphonaptera. Apart from the mentioned visual opsins, further opsin classes, with less well understood function, occur: peropsins in Chorista australis and Panorpa communis, Rh7-like opsins in Nannochorista dipteroides. In the studied Siphonaptera we found a Rh7-like opsin in Ctenocephalides felis and Xenopsylla cheopis. Pteropsin transcripts (also known as encephalopsin or OPN3-like opsins) are present in the Xenopsylla cheopis and Archaeopsylla erinacei data.

Figure 1
Figure 1

Opsin occurrence in Mecoptera. Best maximum likelihood tree annotated with ultrafast bootstrap support (first number; 2000 replicates) and non-parametric slow bootstrap support (second number; 500 replicates) inferred by IQ-TREE 1.6. A third support value when inferred with the Bayesian approach (ExaBayes, see Supplementary Fig. S6) represents posterior probabilities. The tree was rooted with a Danio rerio melanopsin sequence. Visual opsins are highlighted in colours indicating their predicted peak absorbance. Sequences shorter than 100 amino acids included to calculate this tree, are marked by asterisks. Mecopteran and siphonapteran sequence names are highlighted in red and brown, respectively. Appended codes represent NCBI accession numbers (also available in Table S1) or contig and scaffold numbers of the assemblies (Table S2). Abbreviations: Blue = blue sensitive, c = ciliary, LwA = long wavelength sensitive group A, LwB = long wavelength sensitive group B, MW = arthropod medium wavelength sensitive, Rh7 = rhodopsin 7-like opsin. Full sequences and the alignment can be found in the Supplementary Material (S3, S4, S5).

Apart from terminal regions with low signal-to-noise ratio, the Bayesian tree (ExaBayes, Supplementary Fig. S6) was topologically consistent in deep major nodes with the topology inferred by the ML approach (RAxML, Supplementary Fig. S6; IQ-TREE, Fig. 1). A sister group relationship between the LwA and LwB clades is recovered by all three tree reconstruction programs, albeit with weak support.

A comparison of the visual opsin transcript abundance (Fig. 2) shows that LwA is expressed less (up to 100-fold) than LwB – latter generally shows the highest relative expression. Only in the female Panorpa communis specimen, Blue shows the highest expression and LwA expression is approximately 30 times higher than LwB expression. In the Panorpa vulgaris data, which is a 1:1 mixture of male and female tissue29, Blue expression is elevated as well. Due to the influence of diurnal variation of expression, sexual dimorphism and technical biases the present results cannot be considered well supported and an in-depth analysis of opsin expression levels would necessitate a more sophisticated experimental design since sampled data from other sources (NCBI & 1KITE) were originally not designed for such purposes. Nevertheless, these values can give a first impression of the relative opsin expression and may serve as a starting point for further investigations.

Figure 2
Figure 2

Relative abundance of opsin transcripts in Mecoptera. This plot shows TPM (transcripts per million, logarithmic scale) values calculated by Salmon 0.7.2. In the majority of samples LwB is the most expressed opsin, followed by Blue, UV and LwA. Only in Panorpa communis and Panorpa vulgaris, the most abundant transcript is Blue. In the only female sequenced, P. communis, LwA expression is much stronger than LwB expression. The sex of the investigated specimens is indicated by the pictograms below the figure. Abbreviations: Blue = blue sensitive, LwA = long wavelength sensitive group A, LwB = long wavelength sensitive group B, MW = arthropod medium wavelength sensitive, n.d. = not determined, UV = ultraviolet.


With around 500 extant species classified into 9 families52, Mecoptera are a species-poor group compared to other mega-diverse holometabolous insect orders. Some Mecoptera, specifically the genus Panorpa and Bittacidae (represented in our study by Harpobittacus), show complex courtship and mating behaviours:52,53 Males secrete pheromones for long-range attraction of females and go on to present nuptial gifts (salivary secretions or dead arthropods) while slowly flapping their wings which exhibit a species specific pattern of black spots and/or stripes54. It has been stated54 that the vision of Panorpa is very coarse based on an assumed interommatidial angle Δϕ of 1°. However, an interommatidial angle of 6° has been found in the central region of the eye of Panorpa dubia26. Unless Δϕ is lower in an acute zone of the eye this means that two individuals of this species can only be 19 mm apart for being able to discern stripes of 2 mm width (under the assumption of equally sized gaps between the stripes). That would limit the use of visual communication in Panorpa to short ranges. Whether the predaceous Bittacidae, which can catch flying insects while hanging from vegetation or in mid-air52, have eyes with better acuity remains to be investigated.

During courtship many Panorpidae also produce a multimodal vibrational and visual signal by moving their abdomen up and down54. The abdominal tip of male Panorpa bears the clamp-like notal organ that is reddish in most species and the abdomen itself can be brightly coloured as well. While it seems likely that their visual system enables Mecoptera to detect such signals, colour vision in this group has not been investigated in detail so far. Preliminary electroretinogram recordings55 in Panorpa showed a peak in the green and a shoulder in the UV region. Apart from that, no electrophysiological data on mecopteran spectral sensitivity or conclusive behavioural studies on colour vision are available.

Opsins in Mecoptera

The first long wavelength sensitive opsin sequences reported for Mecoptera28 represent a mix of LwA (Boreus coloradensis) and LwB (Panorpa spp.). Recently, the additional presence of blue and UV sensitive opsins has been shown for Boreus coloradensis21. However, in the same study21 only a single long wavelength sensitive opsin is documented in Boreus coloradensis, Panorpa acuminata and Panorpa nebulosa. Our results imply that it is most likely that Manwaring et al.21 failed to retrieve all opsins due to insufficient sequencing depth or other methodological problems. The latter cannot be ruled out since the data is not publicly available. Furthermore, we can refuse their hypothesis21 that only female Panorpa express a single green sensitive opsin. In our female Panorpa communis we found the same set of visual opsins as in male individuals of Panorpa trizonata and Panorpa pryeri. The claim that electrophysiological data55 would support the presence of a single opsin in panorpids is not valid since the summative nature of electroretinograms is not suitable to detect a minority of blue or UV-sensitive receptor cells against a background of predominantly green sensitive receptor cells. Furthermore, these electroretinograms were restricted to the compound eyes55.

Only in mecopteran species lacking ocelli (Caurinus56,57, Apteropanorpidae58, Nannochorista larvae23,59) LwA and Blue (in Apteropanorpa) could not be detected. This could indicate that LwA expression is either very low or confined to the ocelli. In Boreidae, the ocelli are reduced in size58, or lack the median ocellus57, and a peculiar case of missing ocelli in an individual of a species that normally has ocelli (Boreus westwoodi) has been reported60. While this could mean that at least some Boreidae are on the verge of losing ocelli, the full set of visual opsins is still expressed in the species examined by us.

Since only one female Panorpa was sequenced for this study, we cannot provide a definite conclusion on a possible sexual dimorphism of the relative expression of opsins. However, assuming female Panorpa vulgaris show an expression pattern similar to the examined female Panorpa communis, the observed pattern in the Panorpa vulgaris data (three females and one male pooled, see Fig. 2) could probably be explained by a linear combination of male and female values. Of course, only further experimental work (e.g. qPCR) will be able to definitely resolve this question.

Evolution of opsins among Holometabola

While the monophyly of the Mecoptera and their internal relationships are not yet unambiguously settled61, Diptera and Siphonaptera have been regarded as their closest relatives in almost all traditional and recent analyses62. These three orders represent a monophyletic taxon designated Antliophora by Hennig63 (Fig. 3). With respect to mecopteran relationships, one particular problem is the phylogenetic position of the peculiarly relictual family Nannochoristidae.

Figure 3
Figure 3

Hypotheses about the phylogenetic relationships within Antliophora and occurrence of opsins within this clade. (A) Cladogram based on transcriptomic data29. (B) Cladogram based on morphological characters64 (modified). The pictograms represent the opsins which are reported in this study or are known to be present in Diptera; colours correspond to Fig. 1. The observed distribution suggests that the ancestor of Antliophora possessed at least 7 opsins, as shown by the pictograms at the root of the cladograms.

A comprehensive multigene approach61 revealed paraphyletic Mecoptera with two major lineages: (Nannochoristidae + (Siphonaptera + Boreidae)) and (Meropidae + ((Choristidae + Apteropanorpidae) + (Panorpidae + (Panorpodidae + Bittacidae)))). The morphological evidence cited supports a sister group relationship between Boreidae and Siphonaptera. More recently, the monophyly of Mecoptera was likewise not supported by a detailed morphological study64, and Nannochoristidae was the sister taxon of a clade comprising Diptera and Siphonaptera. Nevertheless, when considering the phylogenetic value of certain morphological characters in combination with molecular evidence, the authors concluded that the paraphyly of Mecoptera is probably artificial. Despite these uncertainties, all the above-mentioned studies recovered Nannochoristidae and Boreidae as more ancestral than Panorpidae. In the only transcriptomic molecular analysis29, Mecoptera appeared monophyletic when the amino acid data was analysed, yet support was moderate for the placement of Nannochorista within the clade: (Boreus + (Nannochorista + (Bittacus + (Panorpa)))).

What can we learn about the evolution of opsins in light of this phylogenetic background? Mapping the results of our study onto a cladogram (Fig. 3) shows that all four visual opsins known for Mecoptera must have been present in the common ancestor of Antliophora. However, this condition is probably much ancestral since it is also found in non-holometabolan insects, such as Orthoptera38. Apart from opsins with a known visual function, various other functionally less well-understood opsins (e.g. Rh7-like, pteropsin, arthropsins and peropsins) are present in Pterygota12,65. Holometabola initially possessed a pteropsin (reported in Diptera, Lepidoptera, Coleoptera and Hymenoptera65,66,67) which was not found in Mecoptera. So far, the only holometabolans for which peropsins have been reported are Lepidoptera12 and Coleoptera (NCBI NW_017259703). Due to the large amount of genomic and transcriptomic data available for Diptera and Hymenoptera, the absence of peropsin in these taxa is fairly well established. Since peropsins are weakly expressed in Panorpa communis and Chorista australis, it is conceivable that these proteins can be found in other (if not all) Mecoptera, provided the sequencing included both larvae and adults collected at different times of the day. Likewise, other opsins, such as Rh7-like and pteropsins, may be more widespread than the transcriptome data suggest.

To summarize, the common ancestor of Antliophora not only possessed a full set of presumably visual opsins but also pteropsin, peropsin and a Rh7-like opsin. In the course of evolution of this clade, individual opsin genes were either lost in several lineages (Fig. 3) or were not yet detected given the presently available datasets.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1.

    Kelber, A. Colour in the eye of the beholder: receptor sensitivities and neural circuits underlying colour opponency and colour perception. Curr. Opin. Neurobiol. 41, 106–112 (2016).

  2. 2.

    Menzel, R. Spectral sensitivity and colour vision in invertebrates. In Autrum, H. (ed.) Comparative Physiology and Evolution of Vision in Invertebrates. A: Invertebrate Photoreceptors, vol. VII/6A of Handbook of Sensory Physiology (1979).

  3. 3.

    Shichida, Y. & Matsuyama, T. Evolution of opsins and phototransduction. Philos. Trans. R. Soc. B 364, 2881–95 (2009).

  4. 4.

    Briscoe, A. D. & Chittka, L. The evolution of color vision in insects. Annu. Rev. Entomol. 46, 471–510 (2001).

  5. 5.

    Porter, M. L., Cronin, T. W., McClellan, D. A. & Crandall, K. A. Molecular characterization of crustacean visual pigments and the evolution of pancrustacean opsins. Mol. Biol. Evol. 24, 253–68 (2007).

  6. 6.

    Davies, W. I. L., Wilkie, S. E., Cowing, J. A., Hankins, M. W. & Hunt, D. M. Anion sensitivity and spectral tuning of middle- and long-wavelength-sensitive (MWS/LWS) visual pigments. Cell. Mol. Life Sci. 69, 2455–64 (2012).

  7. 7.

    Lebhardt, F. & Desplan, C. Retinal perception and ecological significance of color vision in insects. Curr. Opin. Insect Sci. 24, 75–83 (2017).

  8. 8.

    Porter, M. L. et al. Shedding new light on opsin evolution. Proc. R. Soc. B 279, 3–14 (2012).

  9. 9.

    Kingston, A. C. N. & Cronin, T. W. Diverse distributions of extraocular opsins in crustaceans, cephalopods, and fish. Integr. Comp. Biol. 56, 820–833 (2016).

  10. 10.

    Ramirez, M. et al. The last common ancestor of most bilaterian animals possessed at least 9 opsins. Genome Biol. Evol. 8, 3640–3652 (2016).

  11. 11.

    Hering, L. & Mayer, G. Analysis of the opsin repertoire in the tardigrade Hypsibius dujardini provides insights into the evolution of opsin genes in Panarthropoda. Genome Biol. Evol. 6, 2380–91 (2014).

  12. 12.

    Henze, M. J. & Oakley, T. H. The dynamic evolutionary history of pancrustacean eyes and opsins. Integr. Comp. Biol. (2015).

  13. 13.

    Marshall, J., Carleton, K. & Cronin, T. Colour vision in marine organisms. Curr. Opin. Neurobiol. 34, 86–94 (2015).

  14. 14.

    Wernet, M., Perry, M. & Desplan, C. The evolutionary diversity of insect retinal mosaics: common design principles and emerging molecular logic. Trends Genet. 31, 316–328 (2015).

  15. 15.

    Bybee, S. et al. Odonata (dragonflies and damselflies) as a bridge between ecology and evolutionary genomics. Front. Zool. 13, 46 (2016).

  16. 16.

    Futahashi, R. Color vision and color formation in dragonflies. Curr. Opin. Insect Sci. 17, 32–39 (2016).

  17. 17.

    Futahashi, R. Molecular mechanisms underlying color vision and color formation in dragonflies. In: T. Sekimura, H. Nijhout (eds.). Diversity and Evolution of Butterfly Wing Patterns, chap. 17, (303–321 (2017).

  18. 18.

    Tierney, S. et al. Opsin transcripts of predatory diving beetles: a comparison of surface and subterranean photic niches. R. Soc. Open Sci. 2, 140386 (2015).

  19. 19.

    Tierney, S. et al. Consequences of evolutionary transitions in changing photic environments. Austral Entomology 56, 23–46 (2017).

  20. 20.

    Friedrich, M. et al. Phototransduction and clock gene expression in the troglobiont beetle Ptomaphagus hirtus of Mammoth cave. J. Exp. Biol. 214, 3532–3541 (2011).

  21. 21.

    Manwaring, K., Whiting, M. F., Wilcox, E. & Bybee, S. M. A study of common scorpionfly (Mecoptera: Panorpidae) visual systems reveals the expression of a single opsin. Org. Div. Evol. 16, 201–209 (2016).

  22. 22.

    Paulus, H. F. Eye structure and the monophyly of the Arthropoda. In Gupta, A. P. (ed.) Arthropod phylogeny, 299–383 (Van Nostrand Reinhold Corporation, New York, 1979).

  23. 23.

    Melzer, R. R., Paulus, H. F. & Kristensen, N. P. The larval eye of nannochoristid scorpionflies (Insecta, Mecoptera). Acta Zool. 75, 201–208 (1994).

  24. 24.

    Chen, Q., Li, T. & Hua, B. Ultrastructure of the larval eye of the scorpionfly Panorpa dubia (Mecoptera: Panorpidae) with implications for the evolutionary origin of holometabolous larvae. J. Morphol. 273, 561–571 (2012).

  25. 25.

    Chen, Q., Wei, Y. & Hua, B. Fine structure of the ommatidia of the short-faced scorpionfly Panorpodes kuandianensis (Mecoptera: Panorpodidae). Microsc. Res. Tech. 76, 862–869 (2013).

  26. 26.

    Chen, Q.-X. & Hua, B.-Z. Ultrastructure and morphology of compound eyes of the scorpionfly Panorpa dubia (Insecta: Mecoptera: Panorpidae). PLoS One 11, e0156970 (2016).

  27. 27.

    Wei, Y. & Hua, B. Ultrastructural comparison of the ocelli of Sinopanorpa tincta and Bittacus planus (Mecoptera). Microsc. Res. Tech. 74, 502–511 (2011).

  28. 28.

    Taylor, S., D de la Cruz, K., Porter, M. L. & Whiting, M. Characterization of the long-wavelength opsin from Mecoptera and Siphonaptera: Does a flea see? Mol. Biol. Evol. 22, 1165–1174 (2005).

  29. 29.

    Misof, B. et al. Phylogenomics resolves the timing and pattern of insect evolution. Science 346, 763–767 (2014).

  30. 30.

    Peters, R. S. et al. The evolutionary history of holometabolous insects inferred from transcriptome-based phylogeny and comprehensive morphological data. BMC Evolutionary Biology 14, 52 (2014).

  31. 31.

    Hu, G.-L., Yan, G., Xu, H. & Hua, B.-Z. Molecular phylogeny of Panorpidae (Insecta: Mecoptera) based on mitochondrial and nuclear genes. Mol. Phylogenet. Evol. 85, 22–31 (2015).

  32. 32.

    Misof, B., Erpenbeck, D. & Sauer, K. Mitochondrial gene fragments suggest paraphyly of the genus Panorpa (Mecoptera, Panorpidae). Mol. Phylogenet. Evol. 17, 76–84 (2000).

  33. 33.

    Sauer, K. P. & Hensle, R. Reproduktive Isolation, ökologische Sonderung und morphologische Differenz der Zwillingsarten Panorpa communis L. und P. vulgaris Imhoff und Labram (Insecta, Mecoptera). J. Zoolog. Syst. Evol. Res. 15, 169–207 (1977).

  34. 34.

    Tillier, P. Contribution à l’étude des Mécoptères de France. 2ème partie: Clé d’identification des Panorpa de France (Mecoptera Panorpidae). L’Entomologiste 64, 21–30 (2008).

  35. 35.

    Peters, R. S. et al. Evolutionary history of the Hymenoptera. Curr. Biol. 27, 1013–1018 (2017).

  36. 36.

    Aronesty, E. Comparison of sequencing utility programs. Open Bioinformatics J. 7, 1–8 (2013).

  37. 37.

    Peng, Y. et al. IDBA-tran: a more robust de novo de Bruijn graph assembler for transcriptomes with uneven expression levels. Bioinformatics 29, i326–i334 (2013).

  38. 38.

    Henze, M. J., Dannenhauer, K., Kohler, M., Labhart, T. & Gesemann, M. Opsin evolution and expression in arthropod compound eyes and ocelli: Insights from the cricket Gryllus bimaculatus. BMC Evol. Biol. 163 (2012).

  39. 39.

    Eddy, S. R. Accelerated profile HMM searches. PLoS Comput. Biol. 7, 1–16 (2011).

  40. 40.

    Camacho, C. et al. BLAST+: architecture and applications. BMC bioinformatics 10, 421 (2009).

  41. 41.

    The UniProt Consortium. UniProt: a hub for protein information. Nucleic Acids Res. 43, D204–D212 (2015).

  42. 42.

    Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Mol. Biol. Evol. 30, 772–80 (2013).

  43. 43.

    Zhang, Y. & Sun, Y. HMM-FRAME: accurate protein domain classification for metagenomic sequences containing frameshift errors. BMC Bioinformatics 12, 198 (2011).

  44. 44.

    Patro, R., Duggal, G., Love, M. I., Irizarry, R. A. & Kingsford, C. Salmon provides fast and bias-aware quantification of transcript expression. Nat. Methods 14, 417–419 (2017).

  45. 45.

    Nguyen, L.-T., Schmidt, H., von Haeseler, A. & Minh, B. Q. IQ-TREE: A fast and effective stochastic algorithm for estimating maximum likelihood phylogenies. Mol. Biol. Evol. 32, 268–274 (2015).

  46. 46.

    Stamatakis, A. RAxML version 8: A tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313 (2014).

  47. 47.

    Aberer, A. J., Kobert, K. & Stamatakis, A. ExaBayes: Massively parallel Bayesian tree inference for the whole-genome era. Mol. Biol. Evol. 31, 2553–2556 (2014).

  48. 48.

    Kalyaanamoorthy, S., Minh, B. Q., Wong, T. K. F., von Haeseler, A. & Jermiin, L. S. Modelfinder: fast model selection for accurate phylogenetic estimates. Nat. Methods 14, 587–589 (2017).

  49. 49.

    Hoang, D., Chernomor, O., von Haeseler, A., Minh, B. Q. & Vinh, L. UFBoot2: Improving the ultrafast bootstrap approximation. Mol. Biol. Evol. 35, 518–522 (2018).

  50. 50.

    Stamatakis, A., Hoover, P. & Rougemont, J. A rapid bootstrap algorithm for the RAxML web servers. Syst. Biol. 57, 758–771 (2008).

  51. 51.

    Pattengale, N. D., Alipour, M., Bininda-Emonds, O. R. P., Moret, B. M. E. & Stamatakis, A. How many bootstrap replicates are necessary? J. Comput. Biol. 17, 337–354 (2010).

  52. 52.

    Byers, G. & Thornhill, R. Biology of the Mecoptera. Annu. Rev. Entomol. 28, 203–228 (1983).

  53. 53.

    Thornhill, R. Panorpa (Mecoptera: Panorpidae) scorpionflies: systems for understanding resource-defense polygyny and alternative male reproductive efforts. Ann. Rev. Ecol. Syst. 12, 355–86 (1981).

  54. 54.

    Hartbauer, M., Gepp, J., Hinteregger, K. & Koblmüller, S. Diversity of wing patterns and abdomen-generated substrate sounds in 3 European scorpionfly species. Insect Sci. 22, 521–31 (2015).

  55. 55.

    Burkhardt, D. & de la Motte, I. Electrophysiological studies on the eyes of Diptera, Mecoptera and Hymenoptera. In Wehner, R. (ed.) Information Processing in the Visual Systems of Anthropods, 147–153 (Springer, Berlin Heidelberg, 1972).

  56. 56.

    Sikes, D. S. & Stockbridge, J. Description of Caurinus tlagu, new species, from Prince of Wales island, Alaska (Mecoptera, Boreidae, Caurininae). Zookeys 35–53 (2013).

  57. 57.

    Beutel, R. G., Friedrich, F. & Whiting, M. F. Head morphology of Caurinus (Boreidae, Mecoptera) and its phylogenetic implications. Arthropod Struct. Dev. 37, 418–33 (2008).

  58. 58.

    Hepburn, H. The skeleto-muscular system of Mecoptera: the head. Univ. Kansas Sci. Bull. 48, 721–765 (1969).

  59. 59.

    Beutel, R. G., Kristensen, N. P. & Pohl, H. Resolving insect phylogeny: The significance of cephalic structures of the Nannomecoptera in understanding endopterygote relationships. Arthropod Struct. Dev. 38, 427–60 (2009).

  60. 60.

    Kreithner, A. Ü. Boreus - Arten aus den Alpen: Taxonomische Charakterisierung und Verbreitung (Insecta, Mecoptera, Boreidae). Ber. nat.-med. Verein Innsbruck 88, 213–236 (2001).

  61. 61.

    Whiting, M. F. Mecoptera is paraphyletic: multiple genes and phylogeny of Mecoptera and Siphonaptera. Zool. Script. 31, 93–104 (2002).

  62. 62.

    Willmann, R. The phylogenetic system of the Mecoptera. Syst. Entomol. 12, 519–524 (1987).

  63. 63.

    Hennig, W. Die Stammesgeschichte der Insekten (Kramer, Frankfurt am Main, 1969).

  64. 64.

    Friedrich, F., Pohl, H., Beckmann, F. & Beutel, R. G. The head of Merope tuber (Meropeidae) and the phylogeny of Mecoptera (Hexapoda). Arthropod Struct. Dev. 42, 69–88 (2013).

  65. 65.

    Futahashi, R. et al. Extraordinary diversity of visual opsin genes in dragonflies. Proc. Natl. Acad. Sci. USA 112, E1247–1256 (2015).

  66. 66.

    Koyanagi, M., Takada, E., Nagata, T., Tsukamoto, H. & Terakita, A. Homologs of vertebrate Opn3 potentially serve as a light sensor in nonphotoreceptive tissue. Proc. Natl. Acad. Sci. USA 110, 4998–5003 (2013).

  67. 67.

    Feuda, R., Marlétaz, F., Bentley, M. A. & Holland, P. W. H. Conservation, duplication, and divergence of five opsin genes in insect evolution. Genome Biol. Evol. 8, 579–587 (2016).

Download references


We thank the Austrian Science Fund FWF for financial support (P 23251-B17). The computational results presented have been partially achieved using the Vienna Scientific Cluster (VSC2). Sequencing was done by the Vienna Biocenter Core Facilities (VBCF). We thank the 1KITE Antliophora group ( for granting us access to unpublished NCBI accepted transcriptome data of mecopteran and siphonapteran species. KM thanks David Yeates (CSIRO National Research Collections Australia).

Author information


  1. Department of Integrative Zoology, University of Vienna, Althanstraβe 14, 1090, Vienna, Austria

    • Alexander Böhm
    •  & Günther Pass
  2. Department of Evolutionary Biology and Ecology, Institute for Biology I (Zoology), University of Freiburg, Hauptstraβe 1, 79104, Freiburg, Germany

    • Karen Meusemann
  3. Center for Molecular Biodiversity Research, Zoological Research Museum Alexander Koenig, Adenauerallee 160, 53113, Bonn, Germany

    • Karen Meusemann
    •  & Bernhard Misof
  4. Australian National Insect Collection, CSIRO National Research Collections Australia (NRCA), Acton, ACT, 2601, Australia

    • Karen Meusemann


  1. Search for Alexander Böhm in:

  2. Search for Karen Meusemann in:

  3. Search for Bernhard Misof in:

  4. Search for Günther Pass in:


A.B. conceived the study, performed the analyses and drafted the manuscript. K.M., B.M. and G.P. contributed to the manuscript.

Competing Interests

The authors declare no competing interests.

Corresponding author

Correspondence to Alexander Böhm.

Electronic supplementary material

About this article

Publication history






By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.