Research Article |
Corresponding author: David H. Kavanaugh ( dkavanaugh@calacademy.org ) Academic editor: Thorsten Assmann
© 2021 David H. Kavanaugh, David R. Maddison, W. Brian Simison, Sean D. Schoville, Joachim Schmidt, Arnaud Faille, Wendy Moore, James M. Pflug, Sophie L. Archambeault, Tinya Hoang, Jei-Ying Chen.
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:
Kavanaugh DH, Maddison DR, Simison WB, Schoville SD, Schmidt J, Faille A, Moore W, Pflug JM, Archambeault SL, Hoang T, Chen J-Y (2021) Phylogeny of the supertribe Nebriitae (Coleoptera, Carabidae) based on analyses of DNA sequence data. In: Spence J, Casale A, Assmann T, Liebherr JК, Penev L (Eds) Systematic Zoology and Biodiversity Science: A tribute to Terry Erwin (1940-2020). ZooKeys 1044: 41-152. https://doi.org/10.3897/zookeys.1044.62245
|
The phylogeny of the carabid beetle supertribe Nebriitae is inferred from analyses of DNA sequence data from eight gene fragments including one nuclear ribosomal gene (28S), four nuclear-protein coding genes (CAD, topoisomerase 1, PEPCK, and wingless), and three mitochondrial gene fragments (16S + tRNA-Leu + ND1, COI (“barcode” region) and COI (“Pat/Jer” region)). Our taxon sample included 264 exemplars representing 241 species and subspecies (25% of the known nebriite fauna), 39 of 41 currently accepted genera and subgenera (all except Notiokasis and Archileistobrius), and eight outgroup taxa. Separate maximum likelihood (ML) analyses of individual genes, combined ML analyses of nuclear, nuclear protein-coding, and mitochondrial genes, and combined ML and Bayesian analyses of the eight-gene-fragment matrix resulted in a well-resolved phylogeny of the supertribe, with most nodes in the tree strongly supported. Within Nebriitae, 167 internal nodes of the tree (out of the maximum possible 255) are supported by maximum-likelihood bootstrap values of 90% or more. The tribes Notiophilini, Opisthiini, Pelophilini, and Nebriini are well supported as monophyletic but relationships among these are not well resolved. Nippononebria is a distinct genus more closely related to Leistus than Nebria. Archastes, Oreonebria, Spelaeonebria, and Eurynebria, previously treated as distinct genera by some authors, are all nested within a monophyletic genus Nebria. Within Nebria, four major clades are recognized: (1) the Oreonebria Series, including eight subgenera arrayed in two subgeneric complexes (the Eonebria and Oreonebria Complexes); (2) the Nebriola Series, including only subgenus Nebriola; (3) the Nebria Series, including ten subgenera arrayed in two subgeneric complexes, the Boreonebria and Nebria Complexes, with the latter further subdivided into three subgeneric subcomplexes (the Nebria, Epinebriola, and Eunebria Subcomplexes)); and (4) the Catonebria Series, including seven subgenera arrayed in two subgeneric complexes (the Reductonebria and Catonebria Complexes). A strong concordance of biogeography with the inferred phylogeny is noted and some evident vicariance patterns are highlighted. A revised classification, mainly within the Nebriini, is proposed to reflect the inferred phylogeny. Three genus-group taxa (Nippononebria, Vancouveria and Archastes) are given revised status and seven are recognized as new synonymies (Nebriorites Jeannel, 1941 and Marggia Huber, 2014 = Oreonebria Daniel, 1903; Pseudonebriola Ledoux & Roux, 1989 = Boreonebria Jeannel, 1937; Patrobonebria Bänninger, 1923, Paranebria Jeannel, 1937 and Barbonebriola Huber & Schmidt, 2017 = Epinebriola Daniel & Daniel, 1904; and Asionebria Shilenkov, 1982 = Psilonebria Andrewes, 1923). Six new subgenera are proposed and described for newly recognized clades: Parepinebriola Kavanaugh subgen. nov. (type species: Nebria delicata Huber & Schmidt, 2017), Insulanebria Kavanaugh subgen. nov. (type species: Nebria carbonaria Eschscholtz, 1829), Erwinebria Kavanaugh subgen. nov. (type species Nebria sahlbergii Fischer von Waldheim, 1828), Nivalonebria Kavanaugh subgen. nov. (type species: Nebria paradisi Darlington, 1931), Neaptenonebria Kavanaugh subgen. nov. (type species: Nebria ovipennis LeConte, 1878), and Palaptenonebria Kavanaugh subgen. nov. (type species: Nebria mellyi Gebler, 1847). Future efforts to better understand relationships within the supertribe should aim to expand the taxon sampling of DNA sequence data, particularly within subgenera Leistus and Evanoleistus of genus Leistus and the Nebria Complex of genus Nebria.
DNA, evolutionary tree, ground beetles, molecular phylogenetics, nomenclature, systematics, taxonomy
The carabid beetle supertribe Nebriitae is a moderately diverse group comprised of small to medium-sized beetles, varied in form (see Figs
Among these, Notiokasiini is the least diverse, represented by a single known species (Fig.
Habitus images of Nebriitae A Notiokasis chaudoiri Kavanaugh & Nègre B Pelophila borealis (Paykull) C Opisthius richardsoni Kirby D Paropisthius indicus chinensis Bousquet & Smetana E Notiophilus palustris Duftschmid F Archileistobrius hwangtienyuni Shilenkov & Kryzhanovskij. Scale bars: 1.0 mm. Photograph credits: A, C David Maddison; B, E Kiril Makarov; D, F Alexander Anischenko.
Pelophilini includes only two described species, both in genus Pelophila Dejean, 1821 (Fig.
Opisthiini is comprised of six described species-group taxa, with one species in the Nearctic genus Opisthius Kirby, 1837 (Fig.
Notiophilini includes ca. 60 described species and six additional subspecies, all in genus Notiophilus Duméril, 1806 (Fig.
Nebriini is the most diverse tribe of the supertribe with ca. 1,000 described species and subspecies, and additional new species are discovered, mainly in Asia, almost every year. The tribe is Holarctic in distribution, with more than 90% of the described species-group taxa and most of the genus-group taxa found in the Palearctic Region. The number of extant genera currently recognized as distinct ranges from four to seven, depending on the authors. At present, the widely accepted genera include Leistus Frölich, 1799, Nebria Latreille, 1802, Archastes Jedlička, 1935, and Archileistobrius Shilenkov & Kryzhanovskij, 1983.
Archileistobrius (Fig.
Genus Leistus (Fig.
Archastes (Fig.
Habitus images of Nebriini A Leistus (Nebrileistus) nubivagus Wollaston B L. (Leistus) ferruginosus Mannerheim C Archastes solitarius (Ledoux & Roux) D Nippononebria (Vancouveria) virescens (Horn) E Nebria (Oreonebria) castanea Bonelli F N. (Eurynebria) complanata (Linnaeus). Scale bars: 1.0 mm. Photograph credits: A, D–F David Maddison; B, C Alexander Anischenko.
Nebria (Figs
Genus-group names in supertribe Nebriitae. All published genus-group names are represented in this list in order of publication and include author, year of publication, type species and current taxonomic status (
Genus group name | Type species | Current status | Type | Div | Rep | % Rep | Proposed status | Result |
---|---|---|---|---|---|---|---|---|
Leistus Frölich, 1799 | Leistus testaceus Frölich = Carabus ferrugineus Linnaeus | valid genus and subgenus | Y | 52 | 11 | 21 | valid genus and subgenus | |
Pogonophorus Latreille, 1802 | Carabus spinibarbis Fabricius | valid subgenus of Leistus | N | 60 | 3 | 5 | valid subgenus of Leistus | |
Nebria Latreille, 1802 | Carabus brevicollis Fabricius | valid genus and subgenus | Y | 109 | 6 | 6 | valid genus and subgenus | |
Manticora Panzer, 1803 | Manticora pallipes Panzer = Leistus spinibarbis Fabricius | junior homonym of Manticora |
N | – | – | – | junior homonym of Manticora |
|
Notiophilus Duméril, 1805 | Cicindela aquatica Linnaeus | valid genus | N | 66 | 4 | 6 | valid genus | |
Alpaeus Bonelli, 1810 | Carabus hellwigii Panzer | subjective junior synonym of Nebria s. str. | N | – | – | – | subjective junior synonym of Nebria s. str. | |
Pelophila Dejean, 1821 | Carabus borealis Paykull | valid genus | Y | 2 | 2 | 100 | valid genus | |
Helobia Stephens, 1828 | Carabus brevicollis Fabricius | objective junior synonym of Nebria s. str. | Y | – | – | – | objective junior synonym of Nebria s. str. | |
Eonebria Semenov & Znojko, 1828 | Nebria komarovi Semenov & Znojko | valid subgenus of Nebria | Y | 77 | 6 | 7 | valid subgenus of Nebria | |
Opisthius Kirby, 1837 | Opisthius richardsoni Kirby | valid genus | Y | 1 | 1 | 100 | valid genus | |
Harpazobia Gistel, 1856 | Carabus brevicollis Fabricius | objective junior synonym of Nebria s. str. | Y | – | – | – | objective junior synonym of Nebria s. str. | |
Eurynebria Ganglbauer, 1891 | Carabus complanatus Linnaeus | valid subgenus of Nebria | Y | 1 | 1 | 100 | valid subgenus of Nebria | |
Leistidius Daniel, 1903 | Leistus piceus Frölich | subjective junior synonym of Leistus s. str. | N | – | – | – | subjective junior synonym of Leistus s. str. | |
Oreonebria Daniel, 1903 | Alpaeus castaneus Bonelli | valid subgenus of Nebria | Y | 26 | 9 | 35 | valid subgenus of Nebria | |
Oreobius Daniel, 1903 | Leistus gracilis Fuss | subjective junior synonym of Pogonophorus | N | – | – | – | subjective junior synonym of Pogonophorus | |
Nebriola Daniel, 1903 | Nebria laticollis Dejean | valid subgenus of Nebria | Y | 19 | 2 | 11 | valid subgenus of Nebria | |
Epinebriola Daniel & Daniel, 1904 | Nebria oxyptera Daniel & Daniel | valid subgenus of Nebria | Y | 28 | 6 | 21 | valid subgenus of Nebria | |
Chaetoleistus Semenov,1904 | Leistus relictus Semenov | subjective junior synonym of Pogonophorus | N | – | – | – | subjective junior synonym of Pogonophorus | |
Acroleistus Reitter, 1905 | Leistus denticollis Reitter | subjective junior synonym of Leistus s. str. | N | – | – | – | subjective junior synonym of Leistus s. str. | |
Euleistulus Reitter, 1905 | Leistus ellipticus Reitter = Leistus fulvus Chaudoir | subjective junior synonym of Leistus s. str. | Y | – | – | – | subjective junior synonym of Leistus s. str. | |
Leistophorus Reitter, 1905 | Leistus fulvibarbis Dejean | subjective junior synonym of Leistus s. str. | Y | – | – | – | subjective junior synonym of Leistus s. str. | |
Spelaeonebria Peyerimhoff, 1911 | Spelaeonebria nudicollis Peyerimhoff | valid subgenus of Nebria | Y | 2 | 1 | 50 | valid subgenus of Nebria | |
Eurinoleistus Breit, 1914 | Leistus depressus Breit | subjective junior synonym of Pogonophorus | N | – | – | – | subjective junior synonym of Pogonophorus | |
Paropisthius Casey, 1920 | Opisthius indicus Chaudoir | valid genus | Y | 5 | 4 | 80 | valid genus | |
Patrobonebria Bänninger, 1923 | Nebria desgodinsi Oberthür | valid subgenus of Nebria | Y | 11 | 4 | 36 | subjective junior synonym of Epinebriola | New Synonymy |
Psilonebria Andrewes, 1923 | Nebria superna Andrewes | valid subgenus of Nebria | Y | 4 | 2 | 50 | valid subgenus of Nebria | |
Nebrileistus Bänninger, 1925 | Leistus ellipticus Wollaston | valid subgenus of Leistus | N | 2 | 1 | 50 | valid subgenus of Leistus | |
Archastes Jedlička, 1935 | Archastes sterbai Jedlička | valid genus | N | 42 | 1 | 2 | valid subgenus of Nebria | New Status |
Boreonebria Jeannel, 1937 | Carabus rufescens Ström = Carabus gyllenhali Schönherr | valid subgenus of Nebria | Y | 42 | 19 | 45 | valid subgenus of Nebria | |
Eunebria Jeannel, 1937 | Carabus psammodes Rossi | valid subgenus of Nebria | Y | 54 | 15 | 28 | valid subgenus of Nebria | |
Paranebria Jeannel, 1937 | Carabus lividus Linnaeus | valid subgenus of Nebria | Y | 3 | 2 | 67 | subjective junior synonym of Epinebriola | New Synonymy |
Neonebria Hatch, 1939 | none designated | nomen nudum | – | – | – | – | nomen nudum | |
Nebriorites Jeannel, 1941 | Alpaeus gagates Bonelli | valid subgenus of Nebria | Y | 2 | 1 | 50 | valid subgenus of Nebria | |
Alpaeonebria Csiki, 1946 | Nebria fuscipes Fuss | valid subgenus of Nebria | N | 32 | 1 | 3 | subjective junior synonym of Nebria s. str. | New Synonymy |
Nippononebria Uéno, 1955 | Nebria pusilla Uéno | valid subgenus of Nebria | N | 7 | 2 | 29 | valid genus and subgenus | New Status |
Agonoamara Jedlička, 1962 | Agonoamara chujoi Jedlička = Nebria chalceola Bates | subjective junior synonym of Nippononebria | Y | – | – | – | subjective junior synonym of Nippononebria s. str. | |
Evanoleistus Jedlička, 1965 | Leistus nepalensis Jedlička | valid subgenus of Leistus | N | 154 | 9 | 5 | valid subgenus of Leistus | |
Neoleistus Erwin, 1970 | Leistus ferruginosus Mannerheim | valid subgenus of Leistus | Y | 3 | 3 | 100 | subjective junior synonym of Leistus s. str. | New Synonymy |
Germaria Jeanne, 1972 | Nebria bremii Germar | junior homonym of Germaria Robineau-Desvoidy, 1830 | Y | – | – | – | junior homonym of Germaria Robineau-Desvoidy, 1830 | |
Catonebria Shilenkov, 1975 | Carabus nitidulus Fabricius = Nebria banksii Crotch | valid subgenus of Nebria | Y | 42 | 39 | 93 | valid subgenus of Nebria | |
Orientonebria Shilenkov, 1975 | Nebria coreica Solsky | valid subgenus of Nebria | Y | 1 | 1 | 100 | valid subgenus of Nebria | |
Reductonebria Shilenkov, 1975 | Nebria ochotica Sahlberg | valid subgenus of Nebria | Y | 17 | 15 | 83 | valid subgenus of Nebria | |
Sardoleistus Perrault, 1980 | Leistus sardous Baudi di Selve | valid subgenus of Leistus | Y | 1 | 1 | 100 | valid subgenus of Leistus | |
Tetungonebria Shilenkov, 1982 | Nebria tetungi Shilenkov | subjective junior synonym of Eunebria | N | – | – | – | incertae sedis | |
Asionebria Shilenkov, 1982 | Nebria roborowskii Semenov | valid subgenus of Nebria | Y | 9 | 3 | 33 | subjective junior synonym of Psilonebria | New Synonymy |
Notiokasis Kavanaugh & Nègre, 1983 | Notiokasis chaudoiri Kavanaugh & Nčgre | valid genus | N | 1 | 0 | 0 | valid genus | |
Archileistobrius Shilenkov & Kryzhanovskij, 1983 | Archileistobrius hwangtienyuni Shilenkov & Kryzhanovskij | valid genus | N | 1 | 0 | 0 | incertae sedis | |
Germarina Jeanne, 1985 | Nebria bremii Germar | junior homonym of Germarina Mesnil 1963 | Y | – | – | – | junior homonym of Germarina Mesnil 1963 | |
Himalayonebria Ledoux, 1985 | Nebria nouristanensis Ledoux | subjective junior synonym of Epinebriola | N | – | – | – | incertae sedis | |
Pseudonebriola Ledoux & Roux, 1989 | Nebria saurica Shilenkov | valid subgenus of Nebria | N | 18 | 7 | 39 | subjective junior synonym of Boreonebria | New Synonymy |
Latviaphilus Barsevskis, 1994 | Elaphrus biguttatus Fabricius | subjective junior synonym of Notiophilus | N | – | – | – | subjective junior synonym of Notiophilus | |
Makarovius Barsevskis, 1994 | Notiophilus rufipes Curtis | subjective junior synonym of Notiophilus | N | – | – | – | subjective junior synonym of Notiophilus | |
Sphodronebria Sciaky & Pavesi, 1994 | Nebria paradoxa Sciaky & Pavesi = Nebria tetungi Shilenkov | objective junior synonym of Tetungonebria | N | – | – | – | objective junior synonym of Tetungonebria | |
Vancouveria Kavanaugh, 1995 | Nebria virescens Horn | valid subgenus of Nebria | Y | 3 | 3 | 100 | valid subgenus of Nippononebria | |
Ledouxnebria Deuve, 1998 | Ledouxnebria brisaci Deuve | valid genus (fossil only) | – | 1 | – | – | valid genus (fossil only) | |
Epispadias Ledoux & Roux, 1999 | Nebria janschneideri Ledoux & Roux | valid subgenus of Nebria | N | 2 | 1 | 50 | valid subgenus of Nebria | |
Falcinebria Ledoux & Roux, 2005 | Nebria reflexa Bates | valid subgenus of Nebria | Y | 16 | 2 | 13 | valid subgenus of Nebria | |
Nakanebria Ledoux & Roux, 2005 | Nebria kurosawai Nakane | valid subgenus of Nebria | N | 6 | 1 | 17 | valid subgenus of Nebria | |
Sadonebria Ledoux & Roux, 2005 | Nebria sadona Bates | valid subgenus of Nebria | N | 17 | 4 | 24 | valid subgenus of Nebria | |
Tyrrhenia Ledoux & Roux, 2005 | Carabus rubicundus Quesnel | valid subgenus of Nebria | Y | 18 | 2 | 11 | valid subgenus of Nebria | |
Marggia Huber, 2014 | Nebria bremii Germar | valid subgenus of Nebria | Y | 2 | 1 | 50 | subjective junior synonym of Oreonebria | New Synonymy |
Barbonebriola Huber & Schmidt, 2017 | Nebria barbata Andrewes | valid subgenus of Nebria | N | 6 | 1 | 17 | subjective junior synonym of Epinebriola | New Synonymy |
Archaeonebria Kavanaugh & Schmidt, 2019 | Archaeonebria inexspectata Schmidt & Kavanaugh | valid genus (fossil only) | – | 1 | – | – | valid genus (fossil only) | |
Parepinebriola subgen. nov. | Nebria delicata Huber & Schmidt | part of Epinebriola Daniel | Y | 5 | 5 | ? | new subgenus | New Subgenus |
Insulanebria subgen nov. | Nebria snowi Bates | part of Reductonebria Shilenkov | Y | 2 | 2 | 100 | new subgenus | New Subgenus |
Erwinebria subgen. nov. | Nebria sahlbergii Fisher von Waldheim | part of Reductonebria Shilenkov | Y | 21 | 20 | 95 | new subgenus | New Subgenus |
Nivalonebria subgen. nov. | Nebria paradisi Darlington | part of Nakanebria Ledoux & Roux | Y | 2 | 2 | 100 | new subgenus | New Subgenus |
Neaptenonebria subgen. nov. | Nebria ovipennis LeConte | part of Catonebria Shilenkov | Y | 6 | 6 | 100 | new subgenus | New Subgenus |
Palaptenonebria subgen. nov. | Nebria mellyi Gebler | part of Catonebria Shilenkov | Y | 14 | 11 | 79 | new subgenus | New Subgenus |
Habitus images of Nebria A N. (Eonebria) djakonovi Semenov & Znojko B N. (Orientonebria) coreica Solsky C N. (Spelaeonebria) nudicollis Peyerimhoff D N. (Psilonebria) superna Andrewes E N. (Reductonebria) ochotica Sahlberg F N. (Catonebria) banksii Crotch. Scale bars: 1.0 mm. Photograph credits: A, B, F Kiril Makarov; C, D David Maddison; E Alexander Anischenko.
Two additional nebriine genera have been described from fossil remains. Ledouxnebria
The taxonomic history of nebriite carabids began with Linnaeus, who, in the Tenth Edition of his Systema Naturae (
As the number of described species increased rapidly in the 19th and 20th centuries and it became clear that groups of more or less similar species could be recognized, the description of nebriite genera and subgenera began, first through extracting species from genus Carabus of Linnaeus and then adding new genus-group taxa to accommodate species unknown to Linnaeus as needed. The first of these described was genus Leistus Frölich, 1799, followed by Nebria Latreille, 1802 and Notiophilus Duméril, 1805, and each of these included Linnaean species. Subsequently, Pelophila Dejean, 1821, Opisthius Kirby, 1837, Eurynebria Ganglbauer, 1891b, Paropisthius Casey, 1920, Archastes Jedlička, 1935, Nippononebria Uéno, 1955, Notiokasis Kavanaugh & Nègre, 1983, and Archileistobrius Shilenkov & Kryzhanovskij, 1983 were described. For three of these genera, one or more additional subgenera have been proposed.
The suprageneric classification of nebriites has been relatively stable during the last century and a half compared with that of many other carabid groups, although the ranking and inclusiveness of various taxa has changed somewhat through time. The post-Linnaean era of carabid classification, which
Few studies have explored phylogenetic relationships among the nebriite taxa through formal analyses and most of them have used morphological data.
Until now, the use of molecular data for phylogenetic studies of nebriites has been very limited.
Working with insights gained from phylogenetic studies based on morphological data gathered during the past 50 years and guided by
In total, 264 exemplar specimens were used in this study. These included representatives of 241 different nebriite species-group taxa plus 14 replicates and eight outgroup taxa plus one replicate (Table
Sampling of Nebriitae and outgroup species. The ID column indicates how each voucher specimen was identified. A number indicates that the specimen was identified by the lead author (DK) using the corresponding numbered references as follows: (1)
All but nine of the specimens used in this study were collected directly into 95–100% ethanol, which was subsequently replaced with fresh ethanol at least once. Six specimens were collected directly into silica gel and allowed to fully desiccate. All specimens were stored as soon as possible in a freezer at -20 °C awaiting DNA extraction. The remaining three specimens were pinned museum specimens killed by unknown means. Depositories for all voucher specimens are listed in Appendix
Nearctic Nebria and Nippononebria specimens were identified by DHK, based on his revision (
DNA was extracted from specimens preserved in ethanol or desiccated in silica gel using Qiagen DNeasy (Qiagen) Blood & Tissue kits and the manufacturer’s recommended protocol. The first 25 extractions (in 2004) were made using thoracic muscle tissue; but all subsequent extractions were made from one or more legs, depending on the size of the specimen. The eight gene fragments used in this study and their abbreviations are: 28S: ca. 955 bases of the nuclear ribosomal DNA subunit 28S (D1–D3 regions); 16S-ND1: ca. 812 bases of the mitochondrial gene fragment including 16S ribosomal RNA gene (partial sequence), tRNA-leucine gene (complete sequence), and NADH dehydrogenase subunit 1 (ND1) (partial sequence); COI BC: 658 bases of the “barcode” region of mitochondrial protein-coding cytochrome c oxidase subunit I; COI PJ: 829 bases of the ”Pat/Jer” region of mitochondrial protein-coding cytochrome c oxidase subunit I (no overlap with COI BC); CAD2: 804 bases of the nuclear protein-coding carbamoyl-phosphate synthetase domain of the “rudimentary” gene, corresponding to region 2 of
Gene fragments were amplified successfully using polymerase chain reaction (PCR) on several different thermal cyclers, including an Eppendorf Mastercycler 5333, an MJ Research PTC-100 Thermal Cycler, or a BioRad MyCycler Thermal Cycler, and using different Taq polymerases, including Eppendorf Hotmaster Taq, Invitrogen Taq, or TaKaRa Ex Taq, each with the basic protocols recommended by the manufacturer. Cycling conditions and primers used in this study are presented in Appendix
Multiple chromatograms for each gene fragment were assembled using SEQUENCHER (several versions, concluding with v 5.2.2) (Gene Codes Corporation) or PHRED (
Three of the specimens from which DNA was extracted were pinned specimens whose killing and preservation history was unknown: Nebria coreica DRM4721, Nebria komarovi DRM5171, and Nebria oxyptera DRM5172. We prepared specimens following
The extracted DNA was quantified using a Qubit Fluorometer (Life Technologies) with a Quant-iT dsDNA HS Assay Kit. The fragment length distribution was measured using a 2100 Bioanalyzer (Agilent Technologies) using the High Sensitivity DNA Analysis Kit and 1 μl of sample. The DNA of Nebria oxyptera was treated with NEBNext FFPE DNA Repair Mix (New England BioLabs) to repair damaged bases prior to library preparation. Libraries were prepared using NEBNext DNA Ultra II (New England BioLabs) kits and dual-indexed using NEBNext Multiplex Oligos for Illumina. Sequencing (150 base paired-end) was performed at the Oregon State University Center for Genome Research and Biocomputing on an Illumina HiSeq 3000. The Nebria libraries were sequenced on two separate lanes with libraries of other carabids (including bembidiines, chlaeniines, oodines, licinines, and scaritines). Nebria coreica DRM4721 was the only Nebriitae on its lane, while Nebria komarovi DRM5171 and Nebria oxyptera DRM5172 were sequenced on the same lane. In the latter case, the two Nebria libraries were given unique multiplex indices to minimize the possibility of adapter cross-talk.
Demultiplexing was performed using bcl2fastq v2 Conversion Software (Illumina). Paired-end reads were imported into CLC GENOMICS WORKBENCH version 9.5.3 (QIAGEN Aarhus A/S), using default options except for the minimum and maximum paired-read distances, which we determined by analyzing a dilution of the enriched library on a Bioanalyzer 2100 (Agilent Technologies). Failed reads were removed during import. We used the “Trim Sequences” tool in CLC (with default parameters) to remove read ends with low quality, ambiguous base calls, and adapter contamination.
De novo assemblies were generated using CLC GENOMICS WORKBENCH from paired, trimmed reads using an automatic word and bubble size, with the minimum contig length set to 200. The assemblies were converted to local, BLASTable databases using NCBI’s makeblastdb tool, and these databases were queried using the local BLAST tool in MESQUITE. The query sequences were those from Nebria (Boreonebria) gouleti DHK0027, chosen because it is expected to be equally distant from all expected placements for the pinned specimens that were sequenced. For each BLAST conducted, up to ten top matches from each database with eValues less than 1 E -80 were imported into the matrix in MESQUITE for analysis.
Two reference-based assemblies were conducted for each of the three pinned specimens. The first reference-based assembly for each used sequences of Nebria (Catonebria) piperi DHK0965 for the reference; the second used sequences from Nippononebria virescens DHK0348. Reference based assembly was performed using the “Map Reads to Reference” tool in CLC GENOMICS WORKBENCH. Default settings were used except for the “Length Fraction” and Similarity Fraction” parameters, which were both increased to 0.8 to decrease the possibility of spurious read mappings.
Some contigs produced by de novo and reference-based assemblies were discarded. For the de novo assembly, several samples yielded multiple contigs; in these cases, one of the contigs was much larger than the remainder (e.g., two contigs were returned when 16S-ND1 was BLASTed to the Nebria komarovi de novo assembly, one of 4677 bases, the other of 528 bases). In such cases, the shorter contig always appeared to be a contaminant, with one exception. The apparent contaminants were identical to a non-nebriine that was included in the multiplex pool on the same Illumina lane [e.g., Clivina bipustulata (Fabricius), Chlaenius tomentosus (Say)], suggesting that cross-talk occurred (
In order to judge the consistency of the two or three sequences (de novo assembly, reference-based to Nippononebria, and reference-based to Catonebria) from each pinned specimen, preliminary analyses were conducted on each of the seven genes and on the concatenated data. After model choice with JMODELTEST version 2.1.10 (
The gene fragments containing ribosomal DNA were aligned differently from the protein-coding genes. An alignment of 28S was conducted in MAFFT version 7.130b (Katoh and Standley 2013), using the L-INS-i search option and otherwise default parameter values. Boundaries of 16S, tRNA leu and ND1 were determined by aligning our sequences to the annotated complete mitochondrial genome of Trachypachus inermis (GenBank Accession NC_011329;
The methods used to align the protein-coding genes varied from gene to gene. COI BC, COI PJ, and Topo showed no evident insertion or deletion events in the history of the taxa studied, and so they could easily be aligned by inspection. Variation in wg suggested several insertion or deletion events present in the history of that taxa. This gene was aligned by first translating the sequence to amino acids in MESQUITE, aligning amino acids in MAFFT, and then forcing that amino acid alignment onto the original nucleotide alignment using MESQUITE. CAD2 and PEPCK contained introns, but few other insertions and deletions, so they were most easily aligned by inspection within MESQUITE’s data editors. CAD2 contained one intron present in Trachypachus, as well as one three-base insertion in Metrius; there were no insertions or deletions evident within nebriites. PEPCK contained one variable-length intron present in most nebriites as well as Pterostichus and Scaphinotus. In addition, there were two three-base deletions in two lineages of Nebria.
Site exclusion also varied from gene to gene. For the protein-coding genes, introns were excluded as well as starting region of PEPCK and Topo, as that region was sequenced in only a few taxa. For 28S and 16S-ND1, sites were excluded from consideration using the modified GBLOCKS (
Models of nucleotide evolution for single genes (except for the 16S-ND1 fragment) were chosen with the aid of JMODELTEST. For concatenated analyses and for the 16S-ND1 fragment, PARTITIONFINDER was used to determine the optimal partitioning, using the greedy algorithm and Bayesian Information Criterion (BIC). The beginning partition for the PARTITIONFINDER analysis had each codon position in each protein-coding gene in a separate part, and the components of the 16S-ND1 fragment in separate parts; thus, for the full concatenated matrix, the initial partition had 25 parts (three parts for each of COI BC, COI PJ, CAD2, ND1, PEPCK, Topo, and wg, as well as one part for each of 28S, 16S, tRNA-Leu, and the non-coding section of the 16S-ND1 fragment). For the final, concatenated matrix the “--raxml” option was used in PARTITIONFINDER, as otherwise program failure occurred during optimization.
Maximum likelihood estimates of phylogenetic relationships were reconstructed using RAXML version 8.2.12 (
Bayesian analyses were conducted on the eight-fragment concatenated matrix (8G B) with MRBAYES v 3.2.7 (
Bootstrap support for and against various clades was calculated using MESQUITE’s “Clade Frequencies in Trees” feature and the bootstrap trees produced by RAXML. Alternative trees with specific contradictory nodes were also included among the bootstrap support trees, and support values for or against those specific nodes were calculated by the same method. Bayesian posterior probability (BPP) values for and against clades were calculated in similar fashion, but using the trees produced by the Bayesian sampling.
To find unique molecular synapotypies for clades, MESQUITE’s “With State Distinguishing Selected Taxa” feature was used. In brief, the “Select Taxa in Clade” tool was used in the tree window, and then the “With State Distinguishing Selected Taxa” feature selected all characters with a unique state in the selected taxa. The unique state was then determined by visually examining those characters in the data matrix. Unique amino acids characterizing clades were found using the same procedure after converting each protein-coding gene to the implied amino acids using MESQUITE’s “Translate DNA to Protein” feature. To find unique base insertions and deletions in 28S sequences, the aligned matrix, including well-aligned internal portions of sequences excluded from ML and Bayesian analyses, was examined visually. Unique amino acid insertions and deletions in protein-coding gene sequences were identified by visually examining the matrix for each fragment converted to a protein matrix.
New sequences generated in this study have been deposited in GenBank with accession numbers MW359101 through MW361062. Two files have been deposited in the Dryad Digital Repository, at https://doi.org/10.5061/dryad.6wwpzgmxn: (1) a NEXUS file containing aligned sequence data as well as inferred maximum likelihood trees, and (2) a NEXUS file containing data about PEPCK intron lengths and the ancestral state reconstruction shown in Suppl. material
We were able to acquire sequence data for 97% (2041) of the 2112 possible sequences for the eight gene fragments for our 264 exemplar specimens (see Table
28S | 16S-ND1 | COI PJ | COI BC | CAD | PEPCK | TOPO | wg | totals | |
---|---|---|---|---|---|---|---|---|---|
specimens sampled | 264 | 264 | 264 | 264 | 264 | 264 | 264 | 264 | 2112 |
sequences obtained | 262 | 260 | 264 | 264 | 259 | 228 | 251 | 253 | 2041 |
% success | 99 | 98 | 100 | 100 | 98 | 86 | 95 | 96 | 97 |
The 16S-ND1 sequences for Nebria piute DHK1098 and N. catenata DHK1067 each contain a stop codon in the ND1 gene, providing evidence that those particular sequences might be nuclear copies of the mitochondrial genes (“numts”,
The only gene fragment for which we obtained less than 95% of the possible number was PEPCK; we speculate that failures in this gene may be the result of the variable-length intron within the gene, which may be long enough in some specimens to prevent amplification under the protocols we used. Of the 228 taxa from which we obtained sequences of PEPCK, 208 shared an intron, although the length of the intron varied between 27 and 830 bases. The intron was lacking from all outgroups except for Pterostichus and Scaphinotus; within nebriites it was absent from Notiophilus, Nebria (Parepinebriola), and the Orientonebria-Archastes-Oreonebria clade, but otherwise present. The only large group in which PEPCK was sequenced in all species was the Catonebria Series, a group in which the intron is relatively short.
GTR+I+G was chosen by the BIC for all single-gene analyses. This includes the 16S-ND1 fragment, for which PARTITIONFINDER found that the optimal partition had all sites in one part.
For the concatenated matrices, the best model scheme chosen by PARTITIONFINDER for all parts was GTR+I+G. The optimal partition scheme varied among concatenated matrices. For the eight-gene concatenated matrices (with or without the merging of the pinned specimens’ sequences), third positions of the mitochondrial genes were in one part, and all of the remaining sites in a second part. For the concatenated nuclear protein-coding gene matrix and concatenated nuclear gene matrix, all sites were in one part. For the concatenated mtDNA matrix, all third positions were in one part, with all remaining sites in a second part.
From both ML and Bayesian analyses, we infer a well-resolved tree (Fig.
Support for or against various clades. All columns provide maximum likelihood bootstrap values for or against a particular clade, except for column “8G B,” which shows the Bayesian posterior probability estimates for the eight-gene matrix. “8G ML” shows the bootstrap values for the eight-gene concatenated matrix, “Nuc G” for the concatenated nuclear genes, “NPC G” for the concatenated nuclear protein-coding genes, and “Mito G” for the concatenated mitochondrial genes. The remaining eight columns provide values for the single gene analyses. All values are expressed as percentages, with positive numbers indicating support for a clade and negative numbers indicating support for a contradictory clade having the highest support. Specific contradictory clades from alternative trees are highlighted in medium grey. Cells with bootstrap values ≥ 90 are shown in black, with values between 75 and 89 in dark grey, and values from 50 to 74 in light grey. Cells in white indicate clades present in the ML tree, but with bootstrap values < 50. Cells in red have bootstrap values for a contradictory clade ≥ 50. Cells in pink have bootstrap values for or against a clade < 50, and the clade is not present in the ML tree. A “-“ in a cell indicates that taxon sampling for that gene was not sufficient to assess monophyly of that clade. “#g” shows the number of single-gene analyses (maximum of eight) that support a clade with bootstrap values of 50 or more.
ML trees and majority rule consensus trees of bootstrap trees resulting from analyses of concatenated matrices for Nuc G, NPC G, and Mito G and for each of the single-gene matrices are presented in Suppl. material
Unique bases found to support particular clades are listed in Suppl. material
The intron length of PEPCK varies across the phylogeny, with some clades having small or no introns, and with a few clades with longer introns (Suppl. material
Morphology has provided only weak support for the monophyly of a supertribe Nebriitae because most features that have been used to characterize it are both plesiomorphic within carabids and not unique to this group.
Based on results from analyses of our molecular data, we find very strong evidence for a monophyletic Nebriitae (Fig.
Monophyly of the tribes Notiophilini, Pelophilini, and Opisthiini are well supported morphologically (
In spite of the fact that the diverse Nebriini is not well characterized morphologically, being recognized mainly on the basis of symplesiomorphic features, monophyly of the tribe is strongly supported by our molecular data. Both ML and Bayesian analyses of the eight-gene matrix and ML analyses of the concatenated nuclear gene and nuclear protein-coding gene matrices produced support values of 100 for the tribe (Chart 1, line 11). Nebriini is also strongly supported in analyses for CAD2 and Topo, moderately supported in the wg analysis and weakly supported in the 28S and PEPCK analyses (Chart 1, line 11), as well as by one unique amino acid (Suppl. material
Phylogenetic relationships among the nebriite tribes remain somewhat less clearly defined than their monophyly. Our results show moderate to strong support for a clade including Opisthiini, Pelophilini and Nebriini as sister to Notiophilini (Fig.
Our results do not strongly speak to the treatment of Pelophila as belonging to a separate monogeneric tribe. Pelophila had been included among the Nebriini (see
As noted in the Introduction, there have been differences of opinion among taxonomists concerning the number of distinct genera that should be recognized among the Nebriini as defined here (i.e., excluding Pelophila). Nevertheless, Leistus has been universally accepted as a distinct genus, based on its wonderful suite of mouthpart co-adaptations for small prey capture. Nebria also has been universally regarded as a distinct genus, but its inclusiveness has differed among authors. In the following sections, we discuss how our results provide new evidence to support phylogenetic relationships among Leistus subgenera and species and better determine which groups are part of a monophyletic genus Nebria and which ones are not.
The monophyly of genus Leistus is strongly supported in all of our concatenated and single-gene analyses (see Fig.
We had only one of the two known species of subgenus Nebrileistus Bänninger, 1925 in our sample, so we cannot comment further on its monophyly. In our analyses, Nebrileistus is shown as sister to a clade including the other Leistus subgenera (Figs
Subgenus Sardoleistus Perrault, 1980 is monobasic, as it includes only Leistus sardous Baudi di Selve, 1883. Results from our analyses suggest that this subgenus is sister to a clade including Pogonophorus Latreille, 1802, Evanoleistus Jedlička, 1965 and Leistus s. str. (Fig.
The subgenus Pogonophorus is strongly supported as monophyletic in all concatenated gene analyses and all single-gene analyses except for 28S (Chart 1, line 20). In the tree from the 28S analysis (Suppl. material
A clade including Evanoleistus and Leistus s. str. is strongly supported as monophyletic in all concatenated gene analyses and strongly to moderately supported in all single-gene ML analyses except for PEPCK (Chart 1, line 21). It is also supported by one base unique among nebriites (Suppl. material
Evanoleistus is the most diverse subgenus of Leistus with more than 150 species described. Our sample included only nine species, two of which are undescribed, so our sampling was small (ca. 5%) relative to that diversity and our results must be judged in that light. Nonetheless, the species in our sample form a monophyletic group (Figs
Leistus s. str. currently includes ca. 50 species and our sample represents a little more than 20% of them. Our sample of this subgenus is strongly supported as monophyletic in all concatenated analyses except for the Mito G analysis (Chart 1, line 24), in which it is only weakly supported (MLB = 42). This clade is shown also in ML trees for all but three of the single-gene analyses (COI BC, COI PJ and PEPCK; Suppl. material
Results from our analyses of molecular data provide strong support for a clade including only Nippononebria and Leistus (Chart 1, line 12). Our 8G ML, 8G B, Nuc G, and NPC G concatenated analyses and CAD2 single-gene analyses show strong support for this clade, PEPCK shows moderate support, and Mito G and 16S-ND1 show weak support. Conversely, a clade including only Nippononebria and Nebria (Chart 1, line 13) is unsupported, except weakly so by the 28S single gene analysis (MLB = 41). Monophyly of Nippononebria (including both Nippononebria s. str. and Vancouveria) (Chart 1, line 14) is also strongly supported in all concatenated analyses except Mito G and in 28S, CAD2, PEPCK and wg analyses. Two unique amino acids also support this clade (Suppl. material
Unlike Nippononebria, Archastes (Fig.
All our concatenated analyses, as well as 28S and Topo single-gene analyses, very strongly support a clade including Archastes within Nebria (Chart 1, line 26). Conversely, a Nebria clade excluding Archastes is strongly contra-indicated. This is shown whether Archastes is excluded alone (Chart 1, line 29) or along with the other Nebria subgenera to which our results show close its relationship (see further discussion below and Chart 1, lines 27 and 28). Consequently, we conclude that Archastes does not warrant status as a separate genus, but should be classified within genus Nebria.
Just as for Archastes, our analyses strongly support a monophyletic Nebria including Oreonebria (Chart 1, line 26), and a Nebria clade excluding Oreonebria is strongly contra-indicated. This is shown whether Oreonebria is excluded alone (Chart 1, line 30) or along with the other Nebria subgenera found from our analyses to be closely related (see further discussion below and Chart 1, lines 27). We conclude that Oreonebria does not warrant status as a separate genus and should be classified within genus Nebria.
Eurynebria was described as a new genus by
Once again, our analyses strongly support genus Nebria including Eurynebria as monophyletic (Chart 1, line 26) and a Nebria clade excluding Eurynebria as strongly contra-indicated (see further discussion below and Chart 1, line 31). We conclude that this taxon does not warrant generic status and should be classified within genus Nebria.
As noted above, the evidence for the monophyly of Nebria, including Archastes, Oreonebria, and Eurynebria while excluding Nippononebria and Vancouveria, is strong. All concatenated analyses and the 28S analysis yielded MLB support values ≥ 97 (Chart 1, line 26). The Topo analysis also provides solid support (MLB value = 89) for this clade, and it appears in ML trees from five of the remaining six single-gene analyses (Suppl. material
Results of our analyses of molecular data also support several large clades within Nebria, most including multiple subgenera. Following the system of informal hierarchic ranks between genus and subgenus used by
The first of these is the Oreonebria Series (Fig.
The Oreonebria Series is an endemic Eurasian lineage that includes more than 180 described species-group taxa. This clade occupies a disjunct geographical distribution comprising the Alps mountain system of Europe in the west and central Asia to far eastern Asia, including Japan and Taiwan, in the east. It includes two groups of subgenera, the Eonebria and Oreonebria Complexes (Fig.
The remaining three series of Nebria are strongly supported as a monophyletic group in both 8G ML and 8G B analyses (Fig.
The Nebriola Series, which includes only subgenus Nebriola Daniel, 1903, is a small group comprised of only 19 described species-group taxa. Members of all described species are small and flightless due to atrophied hindwings. Their combined geographical distribution is confined to mountain systems of western Europe, from the Pyrenees Mountains of northern Spain and southern France northeast through the Alps to the Jura Mountains of eastern France and Switzerland and the Black Forest of southern Germany.
The Nebria Series is the most diverse and widespread group in the genus. More than 320 species-group taxa have been described, and the geographical range of the group encompasses most of the Holarctic Region and the northern edge of the Oriental Region in Asia. The group includes two main groups of subgenera, the Boreonebria and Nebria Complexes, and the latter includes three subgeneric subgroups, the Nebria, Epinebriola and Eunebria Subcomplexes (Fig.
The Catonebria Series is a moderately diverse group with 104 described species, representing a majority of Nearctic Nebria species. It has a Holarctic distribution with a geographical range extending from central Asia east to eastern North America. It includes two subgeneric groups, the Reductonebria and Catonebria Complexes (Fig.
Although our results provide evidence that the Nebriola, Nebria, and Catonebria Series together form a clade and that each individually is monophyletic, relationships among them are not yet as clear. Trees from both the 8G ML (Fig.
The group of subgenera that we refer to as the Eonebria Complex (Figs
The most diverse subgenus in this clade is Eonebria (Fig.
Diversity within subgenus Sadonebria has increased from the three species recognized by
One of the most unexpected results of our study was the discovery of the group of five species represented in our sample, including Nebria delicata Huber & Schmidt, 2017 and Nebria retingensis Huber & Schmidt, 2017 and three undescribed species (Figs
The monophyly of a group including Sadonebria and Eonebria as sister to Parepinebriola (Fig.
The Oreonebria Complex of subgenera (Figs
Within the complex, two main groups of subgenera are apparent. The first is a clade including Falcinebria Ledoux & Roux, 2005 and Epispadias Ledoux & Roux, 2005. It is well supported (MBL and BPP values ≥ 96) in all our analyses except for wg (Chart 1, line 39), as well as by three unique amino acids (Suppl. material
The other major group of subgenera in this complex includes Orientonebria Shilenkov, 1975, Archastes and Oreonebria (Fig.
The Boreonebria Complex is comprised of two subgenera, Boreonebria and Nakanebria. This group includes 66 species-group taxa with a combined geographical range that covers most of the Holarctic Region. It differs from
Subgenus Boreonebria includes the 60 remaining species of this complex. Their combined geographical range extends across the Holarctic region, from Greenland and Iceland eastward across Eurasia, Beringia and across North America from Alaska to Ellesmere Island and the Island of Newfoundland. In Eurasia, they range from above the Arctic Circle to southern Europe (but not northern Africa), the Tian Shan, Altai, and Qilian mountain systems of central Asia, northern China, North Korea, and northern Japan. In North America, they range south into the northern Sierra Nevada in California, the southern Rocky Mountains in northern New Mexico, and into the southern Appalachian Mountains of Tennessee and North Carolina.
Within Boreonebria, several well-supported clades can be recognized (Fig.
Our results suggest that sister to the nivalis group is a clade that includes all the species currently assigned to subgenus Pseudonebriola and the remaining Boreonebria species, which we call the gyllenhali group Fig.
Our gyllenhali group of Boreonebria is only partially equivalent to that of
We have not had the opportunity to include in our study several of the species of Boreonebria that Ledoux and Roux included in this group from southern and eastern Europe, China (including Tibet), Kazakhstan, or Afghanistan. However, those authors provided illustrations of the female bursa copulatrix and spermathecal duct and reservoir for several of these, all but one of which appear to have the spermathecal duct inserted ventrally or apicoventrally on the bursa. The lone exception is Nebria klapperichi Bänninger, 1956, which appears to have the duct inserted mid-dorsally on the bursa (
The group of taxa that we recognize as the Nebria Complex (Fig.
This group of subgenera is equivalent to the “Nebrides” of
Tyrrhenia includes 18 species-group taxa with a combined geographical range extending from Portugal and Morocco across southern Europe and northern Africa to southeastern Turkey. Although we had only two species of this subgenus represented in our sample, they form a well- supported clade in all our analyses (Chart 1, line 66). Two base insertions and two deletions in 28S are unique to the clade (Suppl. material