Research Article |
Corresponding author: Antonio S. Ortiz ( aortiz@um.es ) Academic editor: Reza Zahiri
© 2023 Antonio S. Ortiz, Rosa M. Rubio, Josef J. de Freina, Juan J. Guerrero, Manuel Garre, José Luis Yela.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Ortiz AS, Rubio RM, de Freina JJ, Guerrero JJ, Garre M, Yela JL (2023) DNA barcoding and morphology reveal European and western Asian Arctia villica (Linnaeus, 1758) as a complex of species (Lepidoptera, Erebidae, Arctiinae). ZooKeys 1159: 69-86. https://doi.org/10.3897/zookeys.1159.95225
|
Currently, the genus Arctia Schrank, 1802 includes approximately 16 species in the Palaearctic region, depending on the taxonomic interpretation. Here, populations of the Arctia villica (Linnaeus, 1758) morphospecies complex were studied from Europe to the Middle East (Turkey, northern Iran) by molecular methods. Morphological treatment has traditionally revealed the presence of five nominal taxa: A. villica (Linnaeus, 1758), A. angelica (Boisduval, 1829), A. konewkaii (Freyer, 1831), A. marchandi de Freina, 1983, and A. confluens Romanoff, 1884. The molecular approach tests whether they represent well-delimited species. Subsequently, this study corroborates the suitability of the mitochondrial cytochrome c oxidase subunit 1 (COI) marker sequence for species delimitation. In total, 55 barcodes of the Arctia villica complex were compared, and two molecular species delimitation algorithms were applied to reveal the potential Molecular Operational Taxonomic Units (MOTUs), namely the distance-based Barcode Index Number (BIN) System, and the hierarchical clustering algorithm based on a pairwise genetic distances approach using the Assemble Species by Automatic Partitioning (ASAP). The applied ASAP distance-based species delimitation method for the analysed dataset revealed an interspecific threshold of 2.0–3.5% K2P distance as suitable for species identification purposes of the Iberian A. angelica and the Sicilian A. konewkaii and less than 2% for the three taxa of the A. villica clade: A. villica, A. confluens, and A. marchandi. This study contributes to a better understanding of the taxonomy of the genus Arctia and challenges future revision of this genus in Turkey, the Caucasus, Transcaucasia as well as northern Iran using standard molecular markers.
Arctia, COI, DNA barcoding, Europe, species delimitation, western Asia
The European fauna of Arctiinae moths (Erebidae) is particularly well-known comprising 113 species (
More recently, the use of eight molecular markers coupled with proper analytical algorithms (Maximum Likelihood and Bayesian Inference) has postulated a much more comprehensive view of the genus Arctia and closely related genera, including Acerbia Sotavalta, 1863, Ammobiota Wallengren, 1885, Atlantarctia Dubatolov, 1990, Borearctia Dubatolov, 1884, Epicallia, Eucharia, Hyphoraia Hübner, [1820], Pararctia Sotavalta, 1965, Parasemia Hübner, [1820], and Pericallia as synonyms or potential subgenera (
Following the current view, Arctia villica is an extremely variable species in body size, coloration, wing pattern, and size of spots including the so far considered subspecies A. v. villica s.str., A. v. angelica (Boisduval, 1829), A. v. konewkaii (Freyer, 1831), A. v. confluens Romanoff, 1884 and A. v. marchandi de Freina, 1983 (
In the revisions of the Arctia villica complex,
The study of the morphological features and genitalia structures provides valuable criteria for species recognition according to
In this paper, we present new insights from DNA barcodes of material collected along the geographical distribution of the different taxa of Arctia villica complex to be added to the previous morphological studies in order to substantiate their taxonomic range.
MWM Museum WITT, München (in
This study is based on the results of the morphological study of a large amount of collected material deposited in the
Distribution of Arctia samples sequenced. Note that each point may represent more than one specimen. The map was created using www.simplemappr.net.
Thirty-four adult specimens of the Arctia villica complex were processed and sequenced at the Canadian Centre for DNA Barcoding (CCDB, Guelph) to obtain DNA barcodes using the standard high-throughput protocol described by
Taxon names, old (BIN1) and present BINs (BIN2), BOLD accession numbers for the specimens used in distance estimations (Process ID), haplotype, and locality information (Country and Territory).
Taxon | BIN1 | BIN2 | Process ID | Haplotype | Country | Territory |
---|---|---|---|---|---|---|
Arctia villica | ACP7520 | AAC8627 | ABOLA146-14 | 1 | Austria | Kaernten |
ACP7520 | AAC8627 | ABOLC051-16 | 1 | Austria | Niederöesterreich | |
ACP7520 | AAC8627 | ABOLD441-16 | 1 | Austria | Niederöesterreich | |
ACP7520 | AAC8627 | FBLMZ558-12 | 1 | Germany | Bavaria | |
ACP7520 | AAC8627 | GWOSI563-10 | 1 | Germany | Bavaria | |
ACP7520 | AAC8627 | IBLAO1063-14 | 1 | Hungary | Bacs-Kiskun | |
ACP7520 | AAC8627 | NOENO403-17 | 1 | Austria | Niederöesterreich | |
ACP7520 | AAC8627 | LEASW136-19 | 2 | Greece | Peloponnese | |
Arctia marchandi | ACP7520 | AAC8627 | VNMB685-08 | 3 | Syria | Aleppo |
AAC8627 | AAC8627 | IBLAO1080-14 | 4 | Turkey | Hakkari | |
AAC8627 | AAC8627 | VNMB690-08 | 4 | Turkey | Hakkari | |
AAC8627 | AAC8627 | VNMB689-08 | 5 | Turkey | Hakkari | |
Arctia villica | ACP7477 | AAC8627 | CGUKD205-09 | 6 | U. Kingdom | Devon |
ACP7477 | AAC8627 | CGUKD277-09 | 6 | U. Kingdom | Norfolk | |
Arctia angelica | ACP7477 | AAC8627 | IBLAO1060-14 | 6 | Spain | Castilla-León |
ACP7477 | AAC8627 | IBLAO1061-14 | 6 | Spain | Castilla-León | |
Arctia villica | ACP7477 | AAC8627 | IBLAO1073-14 | 6 | Spain | Cataluña |
ACP7477 | AAC8627 | IBLAO1124-14 | 6 | Spain | Aragon | |
ACP7477 | AAC8627 | IBLAO1125-14 | 6 | Spain | Asturias | |
ACP7477 | AAC8627 | IBLAO1126-14 | 6 | Spain | Asturias | |
ACP7477 | AAC8627 | IBLAO951-14 | 6 | Spain | Aragón | |
ACP7477 | AAC8627 | IBLAO952-14 | 6 | Spain | Aragón | |
ACP7477 | AAC8627 | LEATC649-13 | 6 | Italy | Südtirol | |
ACP7477 | AAC8627 | GWORZ116-10 | 7 | Italy | Basilicata | |
ACP7477 | AAC8627 | IBLAO553-12 | 8 | Spain | Cataluña | |
ACP7477 | AAC8627 | PHLAC475-10 | 9 | Italy | Südtirol | |
ACP7477 | AAC8627 | VNMB688-08 | 10 | Turkey | Kars | |
Arctia confluens | ACP7477 | AAC8627 | IBLAO1077-14 | 11 | Azerbaijan | Masally |
ACP7477 | AAC8627 | IBLAO1078-14 | 11 | Azerbaijan | Lerik | |
ACP7477 | AAC8627 | VNMB686-08 | 11 | Russia | Dagestan | |
ACP7477 | AAC8627 | IBLAO1082-14 | 12 | Russia | Dagestan | |
ACP7477 | AAC8627 | IBLAO1081-14 | 13 | Russia | Dagestan | |
ACP7477 | AAC8627 | IBLAO1083-14 | 14 | Russia | Dagestan | |
ACP7477 | AAC8627 | VNMB687-08 | 15 | Kyrgyzstan | Namagan | |
ACP8428 | AAC8627 | IBLAO1064-14 | 16 | Iran | Zanjan | |
Arctia angelica | ABY6789 | ABY6789 | GBMIN80080-17 | 17 | n/a | n/a |
ABY6789 | ABY6789 | IBLAO1068-14 | 17 | Spain | Andalusia | |
ABY6789 | ABY6789 | IBLAO1069-14 | 17 | Spain | Andalusia | |
ABY6789 | ABY6789 | IBLAO1070-14 | 17 | Spain | Andalusia | |
ABY6789 | ABY6789 | IBLAO1246-20 | 17 | Spain | Andalusia | |
ABY6789 | ABY6789 | IBLAO733-12 | 17 | Spain | Andalusia | |
ABY6789 | ABY6789 | IBLAO734-12 | 17 | Spain | Andalusia | |
ABY6789 | ABY6789 | IBLAO1067-14 | 18 | Spain | Andalusia | |
ABY6789 | ABY6789 | IBLAO1071-14 | 18 | Spain | Castilla-La Mancha | |
ABY6789 | ABY6789 | IBLAO1107-14 | 18 | Spain | Castilla-La Mancha | |
ABY6789 | ABY6789 | IBLAO1127-14 | 18 | Spain | Castilla-La Mancha | |
ABY6789 | ABY6789 | IBLAO1128-14 | 18 | Spain | Castilla-La Mancha | |
ABY6789 | ABY6789 | IBLAO1058-14 | 19 | Morocco | Marrakech-Tensift-Al Hauz | |
ABY6789 | ABY6789 | IBLAO1059-14 | 19 | Morocco | Marrakech-Tensift-Al Hauz | |
Arctia konewkaii | ACL5457 | ACL5457 | GBMIN80081-17 | 20 | Italy | Sicily |
ACL5457 | ACL5457 | IBLAO1074-14 | 21 | Italy | Sicily | |
Arctia konewkaii | ACL5457 | ACL5457 | IBLAO1076-14 | 21 | Italy | Sicily |
ACL5457 | ACL5457 | IBLAO1188-19 | 21 | Italy | Sicily | |
ACL5457 | ACL5457 | IBLAO1075-14 | 22 | Italy | Sicily | |
Arctia confluens | AAC8628 | AAC8628 | VNMB693-08 | 23 | Iran | Mazandaran |
Arctia caja | AAA8530 | IBLAO551-12 | Spain | Catalonia | ||
Arctia festiva | ABW9262 | IBLAO1091-14 | Spain | La Rioja | ||
Arctia flavia | AAV9830 | ABOLA435-14 | Austria | NordTirol | ||
Arctia lapponica | ACF2201 | LON720-09 | Norway | Sor-Varanger |
Sequence divergences for the barcode region were calculated using the Kimura 2-parameter (K2P) model (
To elucidate the taxonomic status of some of the A. villica species complex studied, two molecular species delimitation methods were applied to reveal the potential Molecular Operational Taxonomic Units (MOTUs) that could represent putative cryptic species. The two methods were distance-based: Barcode Index Number (BIN) System (
The BIN method (
The ASAP method (
In the dataset composed of 59 sequences, 55 specimens of the A. villica complex were sequenced or their conspecific sequences were acquired from the databases (BOLD) to analyse taxonomic identity and geographical species grouping, obtaining more than 572 bp for the barcode region (48 of them with 658 bp). In total, 23 different haplotypes were found in the 55 barcode sequences analysed of the five lineages of A. villica species complex ranging from eleven haplotypes in A. villica to three in A. marchandi (Table
Neighbour-Joining (NJ) and Maximum Likelihood (ML) trees of COI barcode region generated using MEGA software recovered the same topology and were practically identical, and all haplotypes could be unequivocally assigned to one of the eight major clades (Table
Three well-supported clades (bootstrap values higher than 70%) were found and were thereafter treated as three MOTUs considered as species, namely A. villica, A. angelica, and A. konewkaii. The divergence between these three groups varies between 2.4% and 3.5% (mean 2.8%; Table
Interspecific mean K2P (Kimura 2-Parameter) divergences (mean pairwise distances) based on the analysis of COI fragments (> 500 bp) among Arctia species.
A. angelica | A. confluens | A. konewkaii | A. marchandi | A. villica01 | |
---|---|---|---|---|---|
A. confluens | 2.7 | ||||
A. konewkaii | 3.2 | 3.5 | |||
A. marchandi | 2.2 | 1.8 | 2.7 | ||
A. villica01 | 2.4 | 1.9 | 3.3 | 1.3 | |
A. villica02 | 2.4 | 1.5 | 3.0 | 1.4 | 1.5 |
The A. villica clade shows five different sub-clades with genetic differences of less than 2% among them including the clade 1 sequences from Northern Spain, Italy, and the United Kingdom (15 samples with 6 haplotypes); the clade 2, with samples from Germany, Austria, and Hungary (7 samples with 1 haplotype); another clade with samples from Russia, Kyrgyzstan, and Azerbaijan (7 samples with 6 haplotypes); the Turkey clade (3 samples with 2 haplotypes); and the one from Syria (1 sample with one haplotype). West and Central European specimens initially identified as A. villica clades 1 and 2 differed up to 1.5% while A. villica clade 2 from Central Europe differed at 1.5% and 1.4% from A. confluens and A. marchandi, respectively (Table
The mean K2P values between the morphologically determined species were used to study species delimitation using two different methods. The BIN System is an online framework that clusters barcode sequences algorithmically and is recalculated from time to time as the number of sequences of each species increases.
All the COI sequences from the five lineages were uploaded and examined into the Barcode of Life Data System (BOLD), resulting in four BINs (AAC8627, ABY6789, ACL5457, and AAC8628) (n = 55 seqs; 23 COI haplotypes). The BIN AAC8627 was attributed to 35 sequences (15 COI haplotypes: 577–658 bp) from different continuous localities along Europe, from Spain and the United Kingdom, to Asia Minor, Caucasus, and Iran; ABY6789 was associated with 14 sequences, and three COI haplotypes (658 bp) located in the southern half of the Iberian Peninsula and Morocco; ACL5457 is unique from Sicily, with five sequences and three haplotypes (658 bp); and AAC8628 for one specimen identified as A. confluens (Table
ASAP method was performed on the data set of 54 sequences, representing all specimens sequenced. For the ASAP analysis, the sequence of a specimen identified as A. confluens from Iran (VNMB693-08; BIN: AAC8628) was excluded from further analysis because of the doubtful sequence that separates it from the rest of the groups. The analysis resulted in partitioning all COI sequences into eight MOTUs (hypothetical species) corresponding to five main lineages: A. angelica, A. konewkaii, A. villica (2 groups), A. marchandi (2 groups, one of them with only one barcode), and A. confluens (2 groups, one of them with only one barcode). ASAP score 2.5 was the smallest within the range of genetic distances and was calculated as the average among two values: the fourth largest p-value (0.255) and the smallest rank of relative barcode gap width (1.22e-04). The value of probability (p-value) quantifies the chances that each of its new groups represents a single species, and the rank calculates the width of the barcode gap between the previous and this new partition. Both metrics are combined into a single ASAP-score that is used to rank the partitions (see
The two methods used for species delimitation are congruent in recognizing the five main lineages as distinct from any other species studied. The exception are the specimens identified as A. confluens, since they are related to two different BINs: specimens collected from Iran with BIN AAC8628 and the general A. villica BIN AAC8627 including specimens from Caucasus and Syria. This high genetic variation needs to be further investigated with additional samples, particularly from Turkey, Caucasus, and Transcaucasia.
The contemporary species definitions and the properties upon which they are based were presented and discussed by e.g.,
The Arctia villica complex is a species group with high morphological and colour variability, characterized by the extensive black forewing background with small, well-defined, and rather distinct white patches partly or fully fused, but do not show clearly distinctive differences in the genitalia except in the slight variation of the aedeagus angle, although within each of the species, there is a wide range of variation of the characters with singular overlaps between the peripheric populations where introgression processes probably appear.
In this group of species, wing morphology, zoogeographical distribution, and maternal mitochondrial DNA (barcoding) are properties of lineage separation. Analysis of extensive samples of European and Asian populations of the Arctia villica species complex based on morphological features and zoogeographical and biological information revealed the presence of five species, including different forms and varieties, named A. villica, A. angelica, A. konewkaii, A. marchandi, and A. confluens (
Nevertheless,
Although the main criterion to separate Lepidoptera species is genital morphology, other pre- and post-reproductive isolation mechanisms must be considered that prevent the forthcoming of fertile offspring, even though mating may occur (e.g.,
Detailed morphological, ecological, and genetic analysis can discriminate closely related species that show slight sequence divergence from their nearest neighbour. Molecular analyses enable initial biodiversity evaluation in such taxa, but there is no objective way to select the algorithm or input parameters that best recover actual species boundaries. In different groups of invertebrate taxa, a sequence divergence in the barcode region lower than 2% often corresponds to intraspecific differences, while higher values are typical of interspecific variation and recognized as distinct MOTUs (
The discrimination of divergences involving these young species requires more algorithmic finesse, and the selection of an effective algorithm for MOTU recognition is necessary. In our study, BIN and ASAP methods were selected to analyse the Arctia villica complex sequences.
The Barcode Index Number (BIN) system is a persistent registry for animal MOTUs recognized through sequence variation in the barcode region. The BIN pipeline analyzes new sequence data for the barcode region as they are uploaded to BOLD, and BIN metadata are dynamic because key elements of specimen records on BOLD, especially taxonomic assignments, are frequently revised by data providers and because of the high flow of new records (
In 2017, BOLD calculated seven different BINS for the five species with singleton BINs for A. angelica, A. konewkaii, and A. marchandi, two for A. villica and three for A. confluens. Currently, four BINs have been calculated clearly differentiating A. konewkaii and A. angelica, while A. villica formed a group that includes A. villica, A. marchandi, and A. confluens. This variation in the BIN values suggests the presence of different species with unique and specific identifiers within the studied Arctia villica complex.
This case of discordance between BIN assignments and the Arctia villica species taxonomy proposed can reflect the inability of sequence variation at COI to diagnose species because of introgression or their young age. These “merged species” have diagnostic nucleotide substitutions in the barcode region that can be correlated with the morphological or ecological traits used in species diagnosis. BIN sharing can be made when algorithms used as ABGD (Automatic Barcode Gap Discovery,
Concerning the species delimitation analyses,
The ASAP procedure of the barcoding data obtained in our study indicates the existence of five Arctia species within the formerly known Arctia villica subspecies distributed from the Iberia Peninsula to the areas around Transcaucasia and Iran. Two well-supported and reciprocally monophyletic groups were found in the Iberian Peninsula, A. angelica, and Sicily, A. konewkaii. The other group was made up by sequences of the polymorphic A. villica within two clusters which show notable sequence variability as expected from its wide geographic distribution across many European and Asian localities. These two groups represent two A. villica populations related to both species A. marchandi and A. confluens, respectively. The A. villica populations (clade 1) from Austria, Germany, and Hungary were related to the A. marchandi population from Turkey while one specimen from Syria identified as A. villica is equally related to both groups. The other A. villica population (clade 2) from Great Britain, Italy, and half of the Iberian Peninsula was related to the polymorphic A. confluens (Fig.
Hence, with the combined evidence from comparative morphological studies and the DNA barcode results presented above, Arctia species belonging to the villica group could be considered as metapopulations lineages separately evolving extended through time and including connected subpopulations according to the Unified Species Concept and Species Delimitation (de Queiroz, 2007). In our study, these Arctia species must have evolved separately from other lineages with separated historical biogeographic processes (Fig.
The current distributions of A. villica species complex is likely to be the result of multiple range shifts driven by past climatic changes. Cyclic climatic change during the Pleistocene has caused repeated range shifts in most European taxa, profoundly influencing the biodiversity of Europe (for reviews, see
Three peninsular refugia in the south of the continent, Iberia, Italy, and Balkan harbour high biodiversity and endemism rates due in large part to their long-term environmental stability, which enables the persistence of palaeoendemic taxa (e.g.,
The major lineage of A. villica may have diversified in four complexes represented by a widespread group of ancient A. villica from Spain to Russia and three species groups restricted to known areas in the Iberian Peninsula (A. angelica), Sicily (A. konewkaii), and around the Caucasus (A. confluens and A. marchandi), showing a pattern that could be considered evidence for similar ecological preferences or parallel histories for these species during the Quaternary. The fact that these four taxa with more restricted distribution make up separated clades to the broadly distributed A. villica suggests that all clades share it as their recent common ancestor. This phylogeographic pattern has been suggested for other Lepidoptera (e.g. those of the genera Polyommatus Latreille, 1804, Erebia Dalman, 1816, Melitaea Fabricius, 1807, Parnassius Latreille, 1804, Chelis Rambur, [1866]), in which it is postulated a marked role of climatic oscillations during the Pleistocene on population isolation and differentiation (
The apparent absence of a clear lineage sorting between A. villica and A. confluens-A. marchandi may be due to relatively recent radiation. Preliminary information in
However, an increasing number of studies indicate that many endemic taxa inhabiting refugial regions are of Pleistocene origin and formed by allopatric fragmentation. In some cases, they are described as distinct species and, in other cases, these taxa are considered to be subspecies or lineages within species. The present occurrence of A. angelica, A. confluens and A. marchandi in the Iberian Peninsula and around the Caucasus, respectively, showing a sympatric distribution with A. villica, is thought to be the result of a range expansion of A. villica since the last glaciation events, due to the high ecological plasticity of this species and the finding of many suitable habitats, including A. konewkaii in Sicily.
Additional information from nuclear markers and a greater number of samples from the Caucasus will be crucial for the resolution of the different questions that are currently unresolved, such as A. confluens and A. marchandi intraspecific variability and whether they can be delimited and recognized as species utilizing integrated taxonomy.
We are very grateful to the staff at the Canadian Centre for DNA Barcoding for sequence analysis. Paul D.N. Hebert and many other colleagues of the Barcode of Life project (Biodiversity Institute of Ontario, Guelph, Canada) contributed to the success of this study. We are particularly grateful to Ramón Macià, Carlos Antonietty and Pablo Valero for loaning specimens and to Peter Huemer from Tiroler Landesmuseum in Innsbruck (Austria) and Axel Hausmann from Bavarian State Collection of Zoology, of Munich (Germany), for granting access to their projects on BOLD. Thanks are also due to Alejandro López, Ana Isabel Asensio, and Manuel González Veiga for their comments, suggestions and technical support. John Girdley and Claire Ward also helped with comments and linguistic advice.
Environmental Authorities in the National Park of Sierra Nevada (Andalusia), National Park of Picos de Europa (Asturias), Aragón Region, Castilla-La Mancha Region, Natural Park of Alt Pirineu, and Aran Valley (Catalonia) permitted collection and access to field sites. We also thank to Reza Zahiri and various anonymous referees for their comments. We are very grateful for this collegial and kind support. This study has been supported by the project Fauna Ibérica XII – Lepidoptera: Noctuoidea I (PGC2018-095851-B-C63) of the Spanish Ministry of Research and Science and part of DNA sequencing was supported by Genome Canada (Ontario Genomics Institute) in the framework of the iBOL program, WP 1.9. The Instituto de Ciencias Ambientales of Toledo (ICAM-UCLM) provided infrastructure and collaborated in logistic tasks; we thank its director, Prof. Federico Fernández González, for his kind support.
Complete similarity matrix
Data type: Similarity index matrix
Explanation note: Complete similarity index matrix among Arctia villica species.