Open Access
How to translate text using browser tools
23 December 2021 Overall phylogenetic relationships of Scutellaria (Lamiaceae) shed light on the origin of the predominantly Caucasian and Irano-Turanian S. orientalis group
Rashad A. Salimov, Gerald Parolly, Thomas Borsch
Author Affiliations +
Abstract

Scutellaria is one of the largest genera in the Lamiaceae with an estimated 400–500 species with a nearly worldwide distribution. Most species occur in the N hemisphere, with the Caucasus and the wider Irano-Turanian region housing a large number of taxa, many of them considered endemic. We present an overall phylogeny of the monophyletic genus Scutellaria based on rapidly evolving plastid regions (matK-trnK, rpl16, trnL-F). Three well-supported clades are evident, which render the currently accepted S. subg. Scutellaria paraphyletic to S. subg. Apeltanthus, which appears nested in “clade A”, in which the African S. schweinfurthii is sister to all remaining taxa, followed by other lineages of S. subg. Scutellaria. Ancestral states of 12 morphological characters frequently used as diagnostic from subgenus to species level were reconstructed with BayesTraits. The S. orientalis group appears as a major radiation in the Caucasus area and the Irano-Turanian region that may comprise up to a quarter of the species in the genus. This radiation corresponds to a monophyletically defined S. sect. Lupulinaria, characterized by decussate inflorescences and specialized (e.g. cucullate) bracts. Our phylogenetic data present significant resolution at the species level within the S. orientalis group, indicating complex geographically centred patterns of speciation in adaptation to steppe and high mountain habitats, including multiple evolution of pinnate and tomentose leaves. The detailed infrageneric classification of Juzepczuk (1951, 1954) mostly does not reflect natural groups.

Citation: Salimov R. A., Parolly G. & Borsch T. 2021: Overall phylogenetic relationships of Scutellaria (Lamiaceae) shed light on the origin of the predominantly Caucasian and Irano-Turanian S. orientalis group. – Willdenowia 51: 395–427.

Version of record first published online on 23 December 2021 ahead of inclusion in December 2021 issue.

Introduction

Scutellaria L. is one of the largest genera of Lamiaceae with probably 400–500 species worldwide. These species numbers range from 360 species estimated by Paton (1990a, 1990b) and Harley & al. (2004) to 543 accepted names in World Flora Online (2019). Scutellaria is a sub-cosmopolitan genus centred mainly in temperate regions, with high species diversity in the E part of the Mediterranean region and the vast ranges of the Irano-Turanian region (see Manafzadeh & al. 2017 for current circumscription and history of the concept as a floristic region). For the area covered by the Euro+Med PlantBase (Europe, the Mediterranean, SW Asia and the Caucasus), 54 species are listed (Euro+Med 2006+). In the tropics most species occur in mountain regions of the Andes (e.g. 15 species reported from Peru; Zarucchi & Harley 1993), whereas tropical Africa has just two or three species (African Plant Database 2020–2021). Many species are highly variable morphologically, which has led to significantly deviating treatments of species in different Flora accounts (Juzepczuk 1954; Edmondson 1982; Rechinger 1982). According to such treatments, Scutellaria is represented by 15 species in Azerbaijan (Isayev 1957), 13 species in Georgia (Gagnidze & Ivanishvili 1987), three species and subspecies in Armenia (Menitsky 1987), eight species in Dagestan (Murtuzaliyev 2009), 27 species and 10 infraspecific taxa in Iran (Jamzad 2012, 2013) and 17 species and 22 infraspecific taxa in Turkey (Celep & Dirmenci 2017). The treatment by Ter-Chatschaturova (1967) in Flora Kavkaza recognizes 36 species for the Caucasus. Differences in taxon concepts at species level affect overall estimates of species diversity in Scutellaria. For example, S. orientalis L. alone encompasses currently 14 accepted subspecies in the Euro+Med area (Euro+Med 2006+) which eventually had been treated at different ranks. Also new species continue to be described, in particular from S. sect. Lupulinaria A. Ham., to which S. orientalis belongs, e.g. from Turkey [S. anatolica Çiçek & Ketenoğlu (2011) and S. yildirimlii Çiçek & Yaprak (2013)] and Iran [S. bakhtiarica Ranjbar & C. Mahmoudi (2018)], but also in other sections, e.g. from Mexico (González-Gallegos & Vázquez-García 2013).

The genus Scutellaria occurs in a wide range of habitats. Many species are chamaephytes and grow in rocky places, gravelly slopes, and steppes, to which they are well adapted as dwarf shrubs and subshrubs that sometimes even exhibit a cushion-plant-like branching pattern (Juzepczuk 1954; Paton 1990b; Guseva & Cheryomushkina 2015). Other species grow, mostly as hemicryptophytes, in forests, moist grassland habitats or marshes. Among neotropical species, only a few are from tropical forests (Zarucchi & Harley 1993).

When Linnaeus (1753) described the genus, he listed 12 species. The infrageneric classification of Scutellaria began with Hamilton (1832), who subdivided the genus into three sections (S. sect. Lupulinaria, S. sect. Stachymacris A. Ham. and S. sect. Galericularia A. Ham., see  Appendix S1 (wi.51.51307_Appendix_S1.pdf) in Supplemental content online), which were mainly based on the arrangement of flowers in the inflorescences and the morphology of bracts. Scutellaria orientalis was classified within S. sect. Lupulinaria. This classification was used by Ledebour (1949) in Flora rossica. Bentham (1834) added two new sections, S. sect. Heteranthesia Benth. (species with spirally arranged flowers) and S. sect. Maschalostachys Benth. (species with axillary inflorescences). He essentially maintained this sectional classification in his treatment for Candolle's Prodromus (Bentham 1848) while adding several subordinate groups. Later, Bentham (1876) merged S. sect. Galericularia and S. sect. Stachymacris into a new section that he superfluously and illegitimately named S. sect. Vulgares Benth. Boissier (1879) focused on species of the Old World and grouped them into the two sections, S. sect. Lupulinaria and S. sect. Vulgares, dividing the latter section into three unranked groups.

Briquet (1896) also recognized two sections, based mainly on the presence or absence of nutlet bands or wings rather than inflorescence characters. His first section, Scutellaria sect. “Euscutellaria” comprises plants with wingless nutlets (i.e. without a membranous margin) and includes Bentham's and Hamilton's S. sect. Heteranthesia, S. sect. Lupulinaria and S. sect. Vulgares as subordinate groups, also indicated as sections. In doing so, he subdivided S. sect. Vulgares into smaller taxonomic units – “Reihe” – for which he emphasized differences in floral, inflorescence and general habit as diagnostic characters. In addition to the three unranked groups already proposed by Boissier, he newly described S. [Reihe] Coccineae Briq., S. [Reihe] Cuneatae Briq. and adjusted the rank of S. [Reihe] Angustifoliae (Benth.) Briq. and S. [Reihe] Lateriflorae (Benth.) Briq. Scutellaria sect. “Euscutellaria” included almost all species of the genus, with the exception of the endemic North American species S. nervosa Pursh and the Japanese S. guilielmi A. Gray, which he distinguished in S. sect. Scutellariop-sis Briq., characterized by conspicuously membranous banded nutlets. Moreover, he distinguished subgroups, but these were not given any formal rank.

Later regional treatments of Scutellaria in the 20th century, e.g. those by Epling (1942), Juzepczuk (1954), Wu & Li (1977), Edmondson (1982) and Rechinger (1982), were mainly based on Briquet's classification. However, these treatments differed from earlier infrageneric classifications because the complete global variation of the genus was not considered (Paton 1990a). Epling (1942) proposed 17 sections of the American species of Scutellaria, mainly based on floral, inflorescence, rhizome and general habit characteristics except for the genus Salazaria Torr., with one species in desert regions of SW North America, which he considered strongly deviating with its distinctly inflated calyx at maturity [Salazaria mexicana Torr., now Scutellaria mexicana (Torr.) A. J. Paton]. Three of Epling's sections contained species from the Old World. Juzepczuk (1951) accepted five sections, S. sect. Anaspis (Rech. f.) Juz., S. sect. Apeltanthus Nevski ex Juz., S. sect. Cystaspis Juz., S. sect. Lupulinaria and S. sect. Vulgares. In two cases, he derived the name from the former genera Anaspis Rech. f. and “Apeltanthus Nevski” (the latter never validly published). But later, in his treatment for Flora URSS, Juzepczuk (1954) raised the rank of these sections and subdivided the genus into four subgenera: S. subg. “Euscutellaria”, S. subg. Anaspis (Rech. f.) Juz., S. subg. Apeltanthus (Nevski ex Juz.) Juz. and S. subg. Cystaspis (Juz.) Juz., which were mainly based upon the characteristics of upper lip of the calyx. In his classification, Juzepczuk (1954) placed S. sect. Lupulinaria (with S. subsect. Alpinae Juz., S. subsect. Fruticosae Juz., S. subsect. Orientales (Grossh.) Juz. and S. subsect. Ramosissimae Juz.) along with three other sections in a broadly defined S. subg. “Euscutellaria”. Wu & Li (1977) accepted Briquet's classification for the most part but followed the earlier treatment of Bentham (1834) by including five sections in S. subg. “Euscutellaria”. Rechinger (1982) adopted the classification of Juzepczuk (1954) for the taxa occurring in the Flora iranica area, whereas in Flora of Turkey, Edmondson (1982) followed Hamilton's circumscription of sections based on inflorescence characters and added S. sect. Salviifoliae (Boiss.) Edmondson. The most recent global taxonomic review and infrageneric classification of Scutellaria was proposed by Paton (1990b). He recognized the two subgenera, S. subg. Scutellaria and S. subg. Apeltanthus, with five and two sections, respectively, based on a phenetic analysis of a spectrum of morphological characters. The chronology of the infrageneric taxonomy of Scutellaria is shown in  Appendix S1 (wi.51.51307_Appendix_S1.pdf) in Supplemental content online.

The Caucasus ecoregion and the Irano-Turanian region are the areas with the highest species diversity of Scutellaria worldwide. The majority of the taxa occurring in these areas are considered to belong to the S. orientalis group This aggregate of taxa, which were supposed to be closely related to the Linnaean species S. orientalis, was mentioned by many authors from Popov (1916) to Paton (1990a, 1990b), who informally described it as the ‘S. orientalis species-group’ next to the ‘S. glechomoides species-group’, the ‘S. linearis species-group’ and the ‘S. colpodea species-group’ in S. subsect. Lupulinaria (A. Ham.) A. J. Paton. It is noteworthy that all infrageneric classification systems assigned S. orientalis and its allies to S. sect. Lupulinaria. However, the species-level taxonomy has long been a challenge, and species classification has not been satisfactorily solved due to high morphological polymorphism and often gradual variation. Characters used were the branching pattern of stems; the size, shape and degree of incision of leaf blades; the kind of the indumentum of leaves and bracts; and the size, shape, texture, and colour of bracts. Apart from its centre of diversity in the Caucasus area and the Irano-Turanian region, some species related to S. orientalis extend eastward to W China and westward with a few taxa to Spain and NW Africa.

Popov (1916) already stated that some scientists accepted Scutellaria orientalis as a “species aggregate” with some varieties. Using the observations made by Popov (1916), who had examined all the material available at his time from Crimea and the Caucasus, Juzepczuk (1939) came to the conclusion that a considerable number of taxa can be recognized based on morphology, which also deviates in their distribution and ecology. According to Juzepczuk (1939), the most useful differential characters were habit, indumentum, shape of cauline leaves and floral bracts as well as flower size. He placed the so-called “S. orientalis species cycle” in S. sect. Lupulinaria and further divided it into two series, S. ser. Oxystegiae Juz. and S. ser. Platystegiae Juz. Grossheim (1945) then described seven new species of S. sect. Lupulinaria from the Caucasus and also published a key for identification mostly based on the shape of cauline and floral leaves, as well as distribution maps of 19 species. Grossheim further noted that the “S. orientalis sensu lato” of earlier authors may not represent a natural group and suggested its division into two subsections, S. subsect. Oxystegiae (Juz.) Grossh. and S. subsect. Platystegiae (Juz.) Grossh., with S. ser. Orientales Grossh. placed under the latter subsection. Juzepczuk (1954) divided the species of S. sect. Lupulinaria into four subsections ( Appendix S1 (wi.51.51307_Appendix_S1.pdf) in Supplemental content online), one of which was S. subsect. Orientales with 56 species recognized in Flora URSS. Of these, 28 were originally described by him (Juzepczuk 1954), predominantly from Central Asia and Siberia. In his 1954 treatment, Juzepczuk further divided S. subsect. Orientales into two groups, the Oxystegiae group with five series and the Platystegiae group with two series. Later, Fedorov (1967) reduced most of the Crimean-Caucasian species within the S. orientalis group to subspecific rank. He argued that although these “species” possess more or less independent geographical areas, they would be connected by transitional forms. In his treatment for the Flora of Turkey, Edmondson (1982) followed Fedorov's concept and accepted subspecific ranks for the members of the group. In Flora of Iran, Jamzad (2012) essentially followed Rechinger (1982) and again recognized the members of the S. orientalis group at species level in S. sect. Lupulinaria. Paton (1990b) provided a global view on the genus Scutellaria based on a detailed analysis of morphological characters, including a phenetic tree and a revised infrageneric classification. Although the taxa from the Caucasus, SW and Central Asia were not fully represented, Paton (1990b) mentioned the ‘S. orientalis species-group’ and its allies as an entity within S. subsect. Lupulinaria of S. sect. Lupulinaria.

Several morphological characters have been studied in detail in Scutellaria, such as mericarp micromorphology (Lane 1983) nand surface sculptures of the nutlets (Kim & Lee 1995; Billie & Delprete 1996). The first cladistic analysis based on 85 morphological and anatomical characters by Cantino & al. (1992) found Scutellaria in a clade together with Salazaria, Perilomia Kunth (= S. sect. Perilomia (Kunth) Epling), and Harlanlewisia Epling with a unique synapomorphy in the gynoecium (character 55, state 1, ovary lobes wholly free from each other but laterally attached to a more or less elongate upward extension to the disc that terminates between them), all of which are now regarded as synonyms of the genus Scutellaria (Paton 1990b; Harley & al. 2004). The African genera Renschia Vatke and Tinnea Kotschy ex Hook. f. were further resolved in a tritomy with Scutellaria (Cantino & al. 1992).

Molecular phylogenetic analyses of Lamiaceae (Wagstaff & Olmstead 1997; Bendiksby & al. 2011; Li & al. 2012; Li & al. 2016) resulted in the genera Wenchengia C. Y. Wu & S. Chow, Holmskioldia Retz. and Tinnea as successive sisters to Scutellaria. However, the sampling of Scutellaria in these studies was not representative enough to allow judging the monophyly of the genus and to contribute to understanding evolutionary relationships within Scutellaria based on molecular phylogenetic trees. Several studies generated sequence data of plastid genomic regions (trnL-F, rpl16) of a few Scutellaria species with the purpose of discriminating samples of these medicinal plants (Hosokawa & al. 2005); or studies looked at relationships within species complexes in specific geographical regions such as North America (Olmstead 1989), Taiwan (Chiang & al. 2012) and Iran (Safikhani & al. 2017). Datasets with a more comprehensive taxon sampling were generated by Zhao & al. (2017), who sequenced nuclear ribosomal (nr)ITS and ETS regions for 34 Chinese species, and Safikhani & al. (2018), who focused on resolving relationships of 42 Iranian species and subspecies using the nrITS and trnL-F regions as markers. Presenting the so-far taxonomically most representative sampling of Scutellaria in a phylogenetic analysis, the authors provided evidence for the monophyly of S. sect. Lupulinaria subsect. Lupulinaria in their ITS tree, whereas the trnL-F tree depicted a polytomy with S. stocksii Boiss. (S. sect. Apeltanthus). Salmaki & Müller (2019) used ITS sequences to place the Iranian endemic S. xylorrhiza Bornm. Son & Park (2015, only ENA/GenBank entry but no publication) generated a complete plastid genome from S. insignis Nakai, Zhang & al. (2016, only ENA/ GenBank entry but no publication) from S. lateriflora L., Jiang & al. (2017) from S. baicalensis Georgi, and recently Zhao & al. (2020) published sequences of plastid genomes of eight further Scutellaria species from China as well as Holmskioldia, Tinnea and Wenchengia.

The present study aims at testing the monophyly of Scutellaria and at illuminating overall phylogenetic relationships in the genus based on a representative taxon sampling both at a species level in Scutellaria (see Paton 1990b) and for related lineages in Lamiaceae (Li & al. 2016). The specific objectives of this investigation were: (1) to clarify if the Scutellaria orientalis group represents a monophyletic lineage using a broad sampling of Eurasian, particularly Caucasian, SW Asian and Central Asian taxa; (2) to explore the phylogenetic utility of plastid markers to resolve relationships in this group at species level; (3) to reconstruct the evolution of morphological characters used as diagnostic for S. orientalis and relatives in previous classifications; (4) to test the hypothesis that decussate inflorescences with flowers subtended by cucullate bracts are a synapomorphy of a monophyletic S. subg. Apeltanthus; and (5) to evaluate the infrageneric classification of Scutellaria in light of phylogenetic relationships and morphology.

Material and methods

Field work and sampling

Plants were collected in the field as much as possible to obtain complete specimens and to document the life form and ecology. In these cases, corresponding fresh leave samples were dried in silica gel. Field trips were carried out between April and July in Azerbaijan from 2016–2018, with voucher specimens deposited in the herbarium BAK (Institute of Botany, Azerbaijan National Academy of Sciences), and also in Georgia and Russia between 2009 and 2017. Duplicates were deposited in B (Botanic Garden and Botanical Museum Berlin). Further material was obtained from the living collection of the Botanic Garden Berlin and the herbaria B, BAK and MO (Missouri Botanical Garden). The collected samples of the Scutellaria orientalis group were mostly determined according to treatments in Flora of Azerbaijan (Isayev 1957) and Flora URSS (Juzepczuk 1954). However, with respect to accepting ranks of the respective entities at species or subspecies level, we followed the treatment in Euro+Med (2006+). Additional samples from Germany were sequenced in the course of the German Barcode of Life (GBOL) Project (Geiger & al. 2016), appreciating that the GBOL markers for angiosperms correspond to the genomic regions used in this investigation.

Selection of taxa for phylogenetic analysis

Our dataset includes 76 taxa of Scutellaria, representing most infrageneric entities recognized by Paton (1990b). We sampled taxa from 15 of the 34 species groups of S. sect. Scutellaria as well as four of the five sections of S. subg. Scutellaria, and from all sections of S. subg. Apeltanthus, including subsections and informal groups. We further included the remaining lineages of Scutellarioideae, the monotypic genera Holmskioldia and Wenchengia, and Tinnea with two species. Furthermore, 17 outgroup species of 11 genera were added to represent other major lineages of Lamiaceae (Li & al. 2016) such as Lamioideae (Galeopsis L., Lamium L., Leonurus L., Stachys L.) and Nepetoideae (Mentha L., Prunella L., Salvia L.) and Ajugoideae (Ajuga L., Teucrium L.). Ajuga was used to root the trees. Detailed information about taxa and voucher specimens and the respective herbaria, the geographical origin of specimens and ENA/GenBank accession numbers of sequences obtained from these specimens are provided in Appendix 1.

DNA extraction, amplification and sequencing

Total genomic DNA was extracted from c. 20 mg of silica-dried leaf tissue with the Macherey-Nagel's Nucleo-Spin® Plant II kit and for older herbarium material using a modified CTAB method (Borsch & al. 2003). Dilutions of 1:10 in purified water were used for polymerase chain reactions (PCR).

Because two plastid genomic regions (trnK-matK and trnL-F) have been widely used to infer phylogenetic relationships in Lamiaceae (Schäferhoff & al. 2010; Li & al. 2016), as well as (trnL-F) in some studies of Scutellaria species (Safikhani & al. 2018; Seyedipour & al. 2020), we also employed these two markers in this investigation. We further added the group II intron in rpl16 in combination with the rps3-rpl16 spacer. All three regions have been shown to provide high levels of hierarchical phylogenetic signal across different taxonomic levels (see Borsch & Quandt 2009), and trnK-matK as well as trnL-F are increasingly used as a barcode for flowering plants (e.g. Geiger & al. 2016). Our goal was here to promote some standardization of markers to achieve additivity of datasets, which is crucial for surveying diversity in speciose genera such as Scutellaria. Yet, no comparable source of data exists for any nuclear DNA region for Scutellaria. Although data of nrITS were published (Safikhani & al. 2018; Salmaki & Müller 2019), we could not obtain single clear bands with PCR, and directly sequenced products usually showed overlying signal from multiple ribotypes, which can be due to a lack of ITS-concerted evolution (Nieto Feliner & Rossello 2007) as in several Lamiales as well as hybridization as reported from other Lamiaceae genera. In this investigation, we therefore focused on plastid data and morphology. More detailed analyses of nuclear loci of the S. orientalis group at the species level will be topics of follow-up work.

The PCR reactions for the three genomic regions of cpDNA were carried out in a total volume of 50 µl, containing 4 µl DNA template with a concentration of 10 ng/ µl, 5 µl Taq buffer S (PeqLab, Erlangen), 2 µl of each forward and reverse primers (20 pm/µl), 10 µl dNTPs (each 1.25 mM), 5 units/µl of 0.3µl Hot Taq DNA polymerase (PeqLab), 3µl of 25 mM MgCl2, and 23.7 µl ultrapure H2O was added to obtain a total volume of 50 µl.

The trnK intron including the matK gene was amplified in two overlapping fragments using the primer pairs trnKF (Wicke & Quandt 2009) plus ACmatK1400R (Müller & Borsch 2005) and ACmatK500F (Müller & Borsch 2005) plus psbA5R (Shaw & al. 2005), therefore effectively also covering the trnK-psbA spacer. For some samples obtained from herbarium specimens this region had to be amplified in two shorter overlapping quarters using the primer pairs trnKF + LAMtrnK724R (5′-AATTCGATTTGAATCAAAGG-3′), LAMtrnK690F (5′-CTTGTTTTGACTGTATCGC-3′) + ACmatK1400R, ACmatK500F + LAMmatK1450R (5′-TCGTTGATAGAAATCATATCC-3′) and LAMmatK1260F (5′-AAGAGTTTGTATCGAATAAAG-3′) and psbA5R. The Lamiaceae-specific primers were newly designed in this study. The PCR programme was as follows: initial denaturation of 1 min 30 sec at 95°C, 1 min primer annealing at 50°C and 45 sec extension at 72°C, followed by 34 cycles each consisting of 30 sec denaturation at 95°C, 1 min annealing at 50°C and 45 sec extension at 72°C, with a further final extension of 10 min at 72°C.

Amplification and sequencing of the trnL intron and trnL-trnF intergenic spacer (the “trnL-F region”) were performed with the primers trnTc, trnTf and trnTd (Taberlet & al. 1991) and trnL460F (Worberg & al. 2007), using an initial denaturation of 1 min 30 sec at 95°C, 1 min primer annealing at 52°C and 45 sec extension at 72°C, followed by 34 cycles each consisting of 30 sec denaturation at 95°C, 1 min annealing at 52°C and 45 sec extension at 72°C, with a further final extension of 10 min at 72°C.

The rpl16 region was amplified as a whole with the Lamiaceae specific primers LAMrps3F (5′-GAAATT-GCACGTGTCGAATGG-3′) and LAMrpl16R (5′-TCGTCCTCTATGTTGTTTACG-3′) designed in this study. For some in shorter fragments using primers LAMrps3F + LAMrpl16-290R (5′-ACTGTCTTTTAT-GAGATGAC-3′) and LAMrpl16-140F (5′-AGTCAT-ATCTTTATAGCAAC-3′) + LAMrpl16R. The internal primers were designed based on an initial alignment of rpl16 sequences of Scutellaria and other Lamiaceae. The temperature profile for the PCR amplification consisted of an initial denaturation of 1 min 30 sec at 95°C, 1 min primer annealing at 55°C and 1 min 30 sec extension at 72°C, followed by 34 cycles each consisting of 30 sec denaturation at 95°C, 1 min annealing at 55°C and 1 min 30 sec extension at 72°C, with a further final extension of 10 min at 72°C.

The DNA amplifications were performed in a peqSTAR 96X HPL thermocycler (Peqlab). All PCR products were stained with 100× SYBER-Gold nucleic acid stain (Life Technologies no. S11494, Carlsbad, California) and electrophoresed using a 1.5% agarose gels in TAE (Tris-Acetate-EDTA) buffer (pH 8.0) for 45 min. Bands were excised from the gel and purified using the Gel/ PCR DNA Fragment Extraction Kit (Geneaid Biotech, New Taipei) following the manufacturer's protocols and sequenced via standard Sanger Sequencing at Macrogen Europe (Amsterdam).

Sequence alignment and phylogenetic reconstruction

PhyDE v.0.9971 (Müller & al. 2010) was used to check pherograms for errors in called bases and to generate the contigs for individual genomic regions. Sequences were then aligned following a motif-alignment approach and the rules described in Kelchner (2000) and Löhne & Borsch (2005). Mutational hotspots (including poly-A/T stretches) were excluded from analysis from the final matrix due to ambiguous homology assessment. Insertions and deletions (indels) were coded as binary characters according to the Simple Indel Coding method (Simmons & Ochoterena 2000) as implemented in SeqState v.1.4.1 (Müller 2005a). Sequences were submitted to ENA/GenBank using the script of Gruenstaeudl (2020).

Phylogenetic trees were reconstructed using maximum parsimony (MP), maximum likelihood (ML) and Bayesian inference (BI) methods. All trees were rooted with Ajuga reptans L. for the phylogenetic analyses including the whole dataset of 93 samples. In addition, we generated Bayesian trees using a matrix reduced to the Scutellarioideae and Holmskioldia as root for the analysis of morphological character evolution.

Analyses via MP were performed with heuristic searches in PAUP* v.4.0b10 (Swofford 2002) in conjunction with the parsimony ratchet (Nixon 1999) as implemented in PRAP v.2.0b3 (Müller 2004). Ratchet settings were 200 iterations with 25% of the positions randomly upweighted (weight = 2) during each replicate and 10 random addition cycles. Jackknife support (JK) calculations were performed in PAUP by conducting a heuristic search within a total number of 10 000 JK replicates using the tree bisection-reconnection (TBR) branch-swapping, saving one tree and deleting 36.79% of characters in each replicate (Müller 2005b). Homoplasy indices [consistency (CI), retention (RI), rescaled consistency (RC) and homoplasy indices (HI)] were calculated to assess homoplasy. The best model of nucleotide sequence evolution for each marker was selected using the Akaike information criterion (AIC) as implemented in jModelTest v.2.1.6 (Darriba & al. 2012). The plastid dataset was partitioned as follows: (1) trnK/matK with four partitions: trnK 5′ intron, matK gene, trnK 3′ intron and trnK-psbA spacer; (2) rpl16 as one partition; and (3) trnL-F with two partitions: trnL and trnL-F spacer. The following models were found to best fit the sequence data: (1) TVM+G for matK, rpl16 and the trnK-psbA and trnL-F spacers; (2) GTR+G for the trnK 5′ and trnK 3′ intron parts and the trnL intron. For the indel partition, a restriction site (binary model) was applied. Bayesian inference was carried out using MrBayes v.3.2.6 (Ronquist & al. 2012). For the combined plastid dataset, four parallel runs each with four Markov Chain Monte Carlo (MCMC) chains (one cold, three heated) were performed with a random starting tree for five million generations. The first 25% of generations of each run were discarded and only the trees after the burn-in sampled at a frequency of 1000. Analyses were stopped when the standard deviation of split frequencies was well below 0.01. ML tree inference and bootstrapping (BS) were conducted with RAxML v.8.2.8 (Stamatakis 2014). The model was set to GTRGAMMA and bootstrap analyses were carried out with 1000 replicates. The partition of microstructural mutations was not analysed in ML. Trees were visualized in TreeGraph2 v.2.10.1-641 beta (Stöver & Müller 2010).

Definition and assessment of morphological characters

The selection of morphological characters for ancestral state evolution is based on their previous usage to describe infrageneric taxa of Scutellaria in particular by Juzepczuk (1954) and Paton (1990a). We then also considered the work of further authors to define characters and states. The definitions of characters and states were evaluated in light of the state of knowledge on homology in the respective available literature. The assessment of character states was done from the herbarium specimens used to extract genomic DNA or the respective vouchers when samples were preserved in silica gel. This was of particular importance in S. sect. Lupulinaria where species limits are unclear, in order to make sure that the morphological data correspond to the same individual represented in the tree (see Kilian & al. 2015). For the other species, this procedure was applied as far as possible and complemented with information form the literature such as Flora treatments [Flora of Bhutan (Grierson & Long (1983–2002); Flora zambesiaca (Flora Zambesica Managing Committee 1950+)].

Ancestral character state reconstruction

Ancestral states were inferred with BayesTraits v. 2.0 (Pagel & al. 2004) and commands generated in Tree-Graph2 (Stöver & Müller 2010). We used a population of 500 trees that were randomly selected with Mesquite v. 3.5 (Maddison & Maddison 2021) after the burn-in from the Bayesian analysis of the plastid dataset confined to Scutellarioideae (see above). Probabilities for ancestral states were visualized on the Maximum Clade Credibility tree from the Bayesian analysis with TreeGraph2. Since the diversity of the other Lamiaceae lineages outside the Scutellarioideae in the phylogenetic trees was only scarcely sampled, which would have resulted in an arbitrary representation of morphological variation, we confined the analysis of morphological character coding to the Scutellarioideae.

Results

Sequence and alignment characterization

The detailed sequence statistics of dataset A (105 taxa, three plastid genomic regions) can be found in Table 1. The complete matrix includes 5328 sequence positions and 290 coded indels. In all introns and spacers, we excluded mutational hotspots of unclear homology from the final matrix, most of which were poly A/T micro-satellites. The trnK-matK-psbA region with a total alignment length of 3202 characters has hotspots in the trnK intron in alignment positions 402–406 (HS1), 511–518 (HS2), 552–661 (HS3), 2411–2432 (H4) and the trnK-psbA spacer in positions 178–186 (HS5) and 263–268 (HS6). The first 37 positions corresponding to the trnK 5′ exon were trimmed, and the 36 downstream positions belong to the psbA gene. The matK CDS lacks hotspots but has a frameshift mutation involving positions 1473–1498 of the CDS (-CCTCTTCGACTTTGCGAGCAGTATT). The trnL-F region with a total alignment length of 1133 positions (including the trnL 5′ exon and the trnF gene) has hotspots in the trnL intron in positions 198–207 (HS 1), 449–457 (HS2) and in the trnL-F spacer in positions 672–673 (HS3) and 718–736 (HS4). The alignment of the rpl16 region has 1931 position in total, including 155 positions of the rps3 gene (the first 58 were trimmed because of incomplete sequences) and 400 positions of the rpl16 3′ exon. The alignment was trimmed for 41 downstream positions of the intron due to incomplete sequences. Hotspots are in the rps3-rpl16 spacer in positions 217–223 (HS1), 281–308 (HS2) and in the rpl16 intron in positions 395–409 (HS3), 424–425 (HS4), 654–655 (HS5), 802–817 (HS6), 1125–1142 (HS7), 1302–1305 (HS8) and 1344–1357 (HS9). Dataset B (157 taxa, trnL-F) includes additional sequences of Scutellaria species available from ENA/ GenBank added to the trnL-F partition from dataset A. It has the same alignment length and hotspots as dataset A. The highest number of informative substitution characters is provided by the matK gene, whereas it has only few indels (Table 1). The trnK-matK-psbA region overall contributes 54.6% of the informative characters. Annotated alignments are provided in the Supplemental content online.

Table 1.

Sequence statistics and phylogenetic utility of chloroplast regions used.

img-z7-2_395.gif

Phylogenetic analyses

All three inference methods of the combined plastid markers (dataset A) yielded well-resolved and largely congruent trees that deviated only in some inconsistently resolved nodes in shallow terminal clades (see supplement  S1 (wi.51.51307_Appendix_S1.pdf)-S3). The Bayesian majority rule topology is presented in Fig. 1. The monophyletic genus Scutellaria received maximum support, as did three major clades (A–C) within it. Tinnea was found as sister, also with maximum statistical support. The members of the S. orientalis group are all depicted in clade A, and in a core clade corresponding to S. subg. Apeltanthus in the circumscription of Paton (1990b). Whereas S. subg. Apeltanthus is monophyletic, its S. sect. Apeltanthus is polyphyletic because S. poecilantha Nevski appears in a lineage different from S. leptosiphon Nevski and S. stocksii and nested within S. sect. Lupulinaria (Fig. 1). Furthermore, S. subsect. Lupulinaria shows members of S. subsect. Cystaspis (Juz.) A. J. Paton nested within it in different positions (S. heydei Hook. f., S. physocalyx Regel & Schmalh.). This means that the S. orientalis group also does not form a natural group, as currently understood. Most remarkably, there are two principal subclades with members of the S. orientalis group, one of which includes individuals currently identified as S. orientalis from the Lesser and Greater Caucasus along with S. novorossica Juz. and S. poecilantha. The other, more diverse subclade comprises the Central Asian taxa S. altaica Ledeb. ex A. Spreng, S. glabrata Vved. and S. intermedia Popov along with some individuals of S. orientalis from the Lesser Caucasus, along with S. bornmuelleri Hausskn. ex Bornm., S. grossheimiana Juz., S. pinnatifida A. Ham. and S. prilipkoana Grossh. and also the European S. alpina L. as nested among Caucasian taxa. Scutellaria subg. Scutellaria is broadly paraphyletic to S. subg. Apeltanthus. The first branches of clade A are constituted by a grade of different members of S. sect. Scutellaria and S. sect. Salviifoliae (S. pontica K. Koch). The African S. schweinfurthii Briq. (representing the ‘S. violascens species-group’) is sister to the remainder of clade A.

Clade B contains E Mediterranean species of Scutellaria subg. Scutellaria but also S. scandens D. Don from the Himalaya, representing the ‘S. repens species-group’ and a S. altissima L.–S. tournefortii Benth. clade. Clade C depicts a grade of New World lineages: the South American lineage (S. costaricana H. Wendl., S. scutellarioides (Kunth) R. Harley, etc.), a North American lineage (S. incana Spreng. etc.) followed by S. mexicana (S. sect. Salazaria (Torr.) A. J. Paton). Two terminal sublineages comprise S. galericulata L. and relatives and S. indica L. and the other a number of species from China (e.g. S. tsinyunensis C. Y. Wu & S. Chow), respectively (terminal part of clade C in Fig. 1).

The extended trnL-F tree (Fig. 2) also shows the same major clades A and B, but all relevant nodes are less well supported. Clade C is not fully resolved, with the South American clade appearing in a tritomy. Within clade A, the additional samples from ENA/GenBank, which are mostly from Iran, appear mostly in two subclades: the Scutellaria przewalskii Juz. lineage that consists of only one sample in the combined tree (Fig. 1) is depicted in the wider-sampled trnL-F tree (Fig. 2) as a subclade including S. araxensis Grossh., S. bornmuelleri, several subspecies of S. pinnatifida and S. platystegia Juz. In addition, S. glechomoides Boiss., S. nepetifolia Benth. and others are resolved together with S. physocalyx.

Morphological characters

We developed a list of twelve characters with their respective states, which are presented in the following.

1. Life form — (1) Dwarf shrubs (fruticose chamaephyte) s.l., here defined as erect plants strongly branched from base, woody throughout and taller than 25–50 cm (nevertheless terminal branches can develop into herbaceous climbing branches); (2) suffruticose chamaephytes, with branches woody at base and for at least 10 cm in height and an erect stem; (3) caespitose creeping and cushion-forming chamaephytes, with woodiness restricted to very base, mostly comprising rhizome only, herbaceous branches and a creeping or cushion-like habit; (4) erect herbs; (5) herbaceous stems sprouting from a woody xylopodium with buds hiding under soil surface.

Fig. 1.

Phylogeny of Scutellaria based on a combined matrix of matK-trnK, rpl16 and trnL-F sequence data. Node support values are presented on the Bayesian majority rule consensus tree with Bayesian posterior probabilities shown above branches and Maximum Likelihood BS values (right, italics) and Maximum Parsimony JK values (left, bold) shown below branches. Infrageneric entities annotated on the tree show the classification of Paton (1990b). Taxa in bold are members of the S. orientalis group. An asterisk next to the name indicates taxa explicitly mentioned by Juzepczuk (1954) as members of S. subsect. Orientales or, if not mentioned there, listed as members of the ‘S. orientalis species-group’ by Paton (1990b; two asterisks). Country codes are ISO 3166-1 alpha-3. For voucher information, see Appendix 1.

img-z8-1_395.jpg

These definitions follow Ellenberg & Mueller-Dombois (1967), but combine all life forms into one category which can form transitions and are therefore sometimes hard to distinguish.

Character state 1 combines dwarf shrubs in a strict sense (fruticose chamaephytes; by far prevailing in Scutellaria) and low-growing shrubs (caespitose nanophanerophytes) only to be distinguished by the overall height of the woody vegetation body along an artificial 50(–100) cm limit (Raunkiaer 1934; Ellenberg & Mueller-Dombois 1967). Very few woody Scutellaria species grow taller than 50(–100) cm and hence do not fully fit into the conventional definition of a chamaephyte proper. Shoots rapidly thinning out at the tip and which may die back to the woody branch during untoward conditions occur only within chamaephytes s. str.

Character state 2 corresponds to what is commonly understood as a subshrub.

Plants exhibiting character state 3 are often described as being largely herbaceous plants with creeping stems or at least non-erect stems. The length of the creeping branches distinguishes creeping (reptant) chamaephytes, rooting at the nodes or not (so-called carpet shrubs, espalier shrubs or mat-forming subshrubs) from pulvinate, i.e. cushion-forming, sometimes almost herbaceous chamaephytes (Ellenberg & Mueller-Dombois 1967). Due to the many transitional stages, we accept only one wide character state.

Character state 4 unites all types of non-creeping hemicryptophytes without any subtypes. The term “scan-dent”, which is used by Paton (1990b) for the life form of Scutellaria scandens, is not considered as its own state here. Sometimes erect herbs develop into straggling shoots, and these are still accepted under the state erect herbs, as for example S. scandens in Flora of China (Li & Hedge 1994).

2. Leaf shape — (1) oblong-lanceolate to oblanceolate; (2) elliptic-ovate to deltoid; (3) broadly ovate to orbicular.

Leaf shape refers to the outline of the lamina.

3. Leaf margin — (1) entire; (2) toothed (to repand), all kinds of crenation to dentations less than 30%; (3) pinnatifid, incised or more than 50%.

4. Leaf base — (1) sagittate or hastate; (2) leaf lamina abruptly narrowing into petiole, including cordate, rounded and truncate at base; (3) leaf lamina gradually merging into petiole, including cuneate and attenuate at base.

5. Leaf indumentum of upper leaf surface — (1) glabrous (to subglabrous); (2) hairy.

6. Leaf texture — (1) leathery (xeromorphic); (2) thin-textured (mesomorphic).

7. Inflorescence architecture — Flowers in Scutellarioideae are basically arranged in indeterminate thyrses, but individual cymes are mostly reduced to a single flower. Therefore, inflorescences appear raceme-like. Our character “inflorescence architecture” is understood as a complex character describing the general appearance of inflorescences: (1) flowers opposite and decussate, in a conspicuous inflorescence; (2) flowers opposite and secund, therefore appearing raceme-like, in a more or less conspicuous inflorescence; (3) flowers helically arranged, in a more or less conspicuous inflorescence; (4) flowers in axils of opposite cauline leaves, through curved peduncles also more or less oriented to one side; (5) cymes with branches of first order present, so that several flowers appear from axil of one leaf or bract (“cymose” in Li & al. 2016).

The terminology of inflorescences is according to Weberling (1989), and the presence of inflorescence types in different genera follows Harley & al. (2004). Character state 4 is exemplified by Scutellaria galericulata as also described in Paton (1990b: fig. 2). It encompasses the same arrangement of flowers as state 2 but differs by unreduced cauline leaves supporting flowers instead of reduced cauline leaves or bracts. As a consequence, there is no appearance of a morphologically well-defined inflorescence in relation to the vegetative branching system.

8. Shape of flower-supporting organs — (1) cucullate, bract-like and smaller than cauline leaves; (2) flat, bract-like and smaller than cauline leaves; (3) similar to cauline leaves in size, texture and indumentum, i.e. specific bracts absent.

The character state cucullate (Paton 1990b) is called navicular in Juzepczuk (1954). Character state 3 occurs e.g. in Scutellaria galericulata, and some E Asian species such as S. amoena C. H. Wright (Fig. 3J), in which flowers sit directly in the axils of cauline leaves. Alternatively, character 8 could have been defined for bracts only, which would have resulted in coding it as “not applicable” in species possessing states 4 and 5 of character 7.

9. Texture of flower-supporting organs at fruiting stage — (1) membranous to scarious; (2) herbaceous.

10. Shape of upper lip of calyx — (1) testudinate, slight ridge midway along at anthesis, becoming testudinate in fruit; (2) scutellate, the upper lip is folded midway along its length forming a rigid, erect sail-like structure usually less than 6 mm tall and wide; (3) scutellate-inflated, similar to character state 2, expanding in fruit becoming membranous inflated, usually more than 6 mm tall and wide; (4) bladder-like, upper and lower lips identical, inflated and papery in fruit, forming a bladder; (5) patelliform, disk-shaped with a narrow rim, saucer-shaped (in Holmskioldia); (6) funnelform.

Definitions for character states 1–4 are in line with Paton (1990b), character state 5 is based on Harley & al. (2004) and character state 6 corresponds to the description in Li & al. (2012: fig. 2G). In the reconstruction with BayesTraits, states 5 and 6 are put together because of limited degrees of freedom in the program.

11. Upper lip of corolla — (1) not galeate; (2) galeate.

12. Flower colour — (1) bright red; (2) mainly yellow; (3) several colours (various combinations).

img-z10-1_395.jpg

Fig. 2

(A, B). Extended phylogeny of Scutellaria based on trnL-F sequence data including sequences from published sources. Node support values are presented on the Bayesian majority rule consensus tree with Bayesian posterior probabilities shown above branches and Maximum Likelihood BS values (right, italics) and Maximum Parsimony JK values (left, bold) shown below branches. Infrageneric entities annotated on the tree in clade A show the classification of Juzepczuk (1954) for S. sect. Lupulinaria with its four subsections and the series in S. subsect. Orientales. Taxa in bold are members of the S. orientalis group. An asterisk next to the name indicates taxa explicitly mentioned by Juzepczuk (1954) as members of S. subsect. Orientales or, if not mentioned there, listed as members of the ‘S. orientalis species-group’ by Paton (1990b; two asterisks). Blue text indicates the infrageneric classification of S. sect. Lupulinaria sensu Juzepczuk (1954); for species outside of this section, black text indicates the infrageneric classification of Paton (1990b). For sequences downloaded from ENA/GenBank, the corresponding accession number is indicated with the taxon name instead of the Scu DNA-isolation code. Country codes are ISO 3166-1 alpha-3.

img-z11-1_395.jpg

Fig. 3.

Morphological diversity of Scutellaria. A–G correspond to the S. orientalis group and clade A; H and I represent clade B; J represents clade C. – A, B: S. prilipkoana, Salimov & Safarov RSA0417, Scu030, from Lerik, Talysh; plants with deeply pinnatifid leaves that have a dense indumentum; C: S. alpina, living collection in Botanic Garden Berlin; D: S. orientalis from Quba, Azerbaijan; note almost glabrous and only slightly dentate cauline leaves; E: S. orientalis, Borsch & al. 5804, Scu025, from Stavropol, N Caucasus, Russia; inflorescence with cucullate bracts and leaves slightly pinnatifid and sparsely pubescent; F, G: S. orientalis, Parolly & al. 15096, Scu021, from Georgia close to Tbilisi; leaves slightly pinnatifid and densely tomentose abaxially; H: S. albida from Artvin, Turkey; raceme leafy, with leaves but distinctly smaller than cauline leaves; I: S. tsinyunensis possessing a distinct raceme-like inflorescence with flower-supporting organs strongly reduced to small, membranous bracts; J: S. amoena with solitary flowers arising from axils of regular-sized cauline leaves. – Photographs: R. Salimov (A, B, D), G. Parolly (C, F, G, H), T. Borsch (E), C.-L. Xiang (I, J).

img-z12-1_395.jpg

Fig. 4.

Evolution of inflorescence architecture and flower morphology shown on the Bayesian majority rule consensus tree of the combined plastid dataset. Pie charts illustrate the probabilities for ancestral character states reconstructed for inflorescence type (character 7), texture of bracts in fruit (character 9), shape of bracts (character 8), shape of upper calyx (character 10), flower colour (character 12) and shape of upper lip of corolla (character 11). A specific colour is assigned to each state (character and state definitions are found in the Results), and the position of the pie chart for each character is indicated at the top left of the figure. Boxes next to terminal branches indicate the respective state present in the specimen included in the tree.

img-z13-1_395.jpg

Fig. 5.

Evolution of leaf morphology shown on the Bayesian majority rule consensus tree of the combined plastid dataset. Pie charts illustrate the probabilities for ancestral character states reconstructed for leaf indumentum (character 5), leaf texture (character 6), leaf shape (character 2), leaf base (character 4) and leaf margin (character 3). A specific colour is assigned to each state (character and state definitions are found in the Results), and the position of the pie chart for each character is indicated at the top left of the figure. Boxes next to terminal branches indicate the respective state present in the specimen included in the tree.

img-z14-1_395.jpg

Fig. 6.

Evolution of life forms shown on the Bayesian majority rule consensus tree depicted as a phylogram of the combined plastid dataset. Pie charts illustrate the probabilities for ancestral character states (definitions are found in the Results). Boxes next to terminal branches indicate the respective state present in the specimen included in the tree. Countries in which the specimens were collected as well as major geographic areas are annotated on the tree. Country codes are ISO 3166-1 alpha-3.

img-z15-1_395.jpg

Ancestral character state reconstruction

The results from the analysis with BayesTraits are presented in Fig. 46 as pie charts showing the respective probabilities for the ancestral states on the respective nodes. Fig. 4 summarizes inflorescence and flower characters, whereas Fig. 5 presents leaf characters. Fig. 6 presents the evolution of the life form and also shows an annotation of geographical areas for the respective samples.

Discussion

Monophyly of Scutellaria and overall relationships within the genus

This study comprises the so-far most complete sampling of species, many of which are included in a molecular phylogenetic analysis for the first time. It represents most sections recognized by Paton (1990b) on morphological grounds and also a broad geographical coverage, therefore entailing putative hitherto unrecognized lineages. The genus Scutellaria as circumscribed by Paton (1990b) is found as a clade with maximum support in all three tree inference methods in the combined tree (Fig. 1) and also with trnL-F sequence data alone (Fig. 2, 1.0 PP, 99% ML-BS, 93% MP-JK). The formerly accepted genus Salazaria (type and only species is S. mexicana), which was merged with Scutellaria by Paton (1990b) on the basis of morphology, is found deeply nested in clade C of our phylogeny (Fig. 1), and so is S. scutellarioides, representing the former genus Perilomia. The latter was already included in Scutellaria by Epling (1942) with its two Andean species and is resolved in a South American lineage of clade C (Fig. 2).

Our analyses further show Tinnea as the sister group of Scutellaria and depict Holmskioldia and Wenchengia as successive branches also receiving maximum statistical support in the combined matK-trnK+rpl16+trnL-F trees (Fig. 1) and high support with trnL-F alone (Fig. 2). Our results therefore confirm the monophyly of Scutellarioideae and the generic relationships presented by Li & al. (2012) and more recently Zhao & al. (2020) based on complete plastome sequences.

Our phylogeny resolves three well-supported major clades within the monophyletic genus Scutellaria (Fig. 1). Clade A comprises S. subg. Apeltanthus sensu Paton (1990b), which also includes all members of the S. orientalis group and an assemblage of lineages from S. subg. Scutellaria with S. schweinfurthii depicted with maximum support as sister to all other species in clade A (Fig. 1).

Clades B and C are exclusively composed of members of Scutellaria subg. Scutellaria, which is therefore paraphyletic to S. subg. Apeltanthus. Paton (1990b), considering the results from his broad set of phenotypic characters (vegetative, inflorescence and floral morphology, nutlet morphology, embryo shape) analysed by UPGMA, and further considerations on the polarity of some character state transitions, assumed that S. subg. Apeltanthus could be monophyletic and derived. However, Paton (1990b) did not implement a consistent phylogenetic analysis due to the difficulty of establishing outgroups at a time when there was no phylogenetic hypothesis for the overall relationships of the Lamiaceae. Interestingly, Paton considered members of S. sect. Salviifoliae as possible closest ancestors of S. subg. Apeltanthus, which appear as one of the branches of our clade A and therefore among the closest relatives of S. subg. Apeltanthus in our molecular phylogeny.

Safikhani & al. (2018) provided a tree of mostly Iranian species based on trnL-F sequence data that shows Scutellaria galericulata (S. subg. Scutellaria) as sister to all remaining species, and further depicts two major clades of S. subg. Apeltanthus (incl. S. orientalis, S. pinnatifida, S. stocksii) and members of S. subg. Scutellaria (S. albida L., S. tournefortii and S. xylorrhiza), respectively. These clades correspond to our three clades, although Safikhani & al. (2018) depicted their “Clade II” as composed only of S. subg. Apeltanthus, due to the lack of sampling the African ‘S. violascens species-group’ and other early-branching lineages in this clade such as S. baicalensis or S. pauciflora Baker (Fig. 1, 2). The topology inferred by Zhao & al. (2020) includes 11 species of Scutellaria, which are resolved in two clades. The “subclade I” as annotated by the authors corresponds to clade A of our investigation (Fig. 1) and is further subdivided into a S. altaicaS. przewalskii lineage (S. subg. Apeltanthus) sister to a lineage composed of S. amoena, S. baicalensis and S. kingiana Prain (S. subg. Scutellaria). The “subclade II” of Zhao & al. (2020) corresponds to a terminal core of our clade C, whereas clade B is not represented in their study. Albeit their much more restricted taxon sampling, these two studies also provide some evidence for the paraphyly of S. subg. Scutellaria to S. subg. Apeltanthus in line with our results.

Evolution of selected morphological characters

Inflorescences with decussately arranged flowers evolved once in the core of clade A, in the common ancestor of Scutellaria sect. Lupulinaria and S. sect. Apeltanthus (S. leptosiphonS. stocksii subclade; Fig. 4; blue colour in pie charts). As an ancestral state in the genus Scutellaria, unspecialized inflorescences are inferred (character 7, state 4) where flowers arise from the axils of cauline leaves so that the inflorescence is not conspicuously separated from the vegetative branching system of the plants. This state is present in many species of clade C, such as in S. galericulata (Fig. 4) or S. amoena (Fig. 2, 3). Our ancestral character state reconstruction reveals that inflorescences with opposite flowers arranged to one side in a more or less well-defined raceme-like structure (state 2, red colour in the pie charts in Fig. 4) evolved several times in all three major clades. These inflorescences have flower-supporting organs that are flat and herbaceous in texture but differ considerably from cauline leaves by their much smaller size (e.g. S. albida; Fig. 3) to become further reduced and membranous to scarious (e.g. in S. tsinyunensis, Fig. 3). Inflorescence with helically arranged flowers are derived in S. costaricana and S. incarnata Vent. (Fig. 13), whereas cymes with more than one flower are found in Tinnea as the sister group of Scutellaria and appear to be also plesiomorphic in Scutellarioideae.

The evolution of the shape and texture of bracts (characters 8 and 9; Fig. 4) seems to be closely linked to the evolution of the inflorescence architecture. Cucullate bracts (character 8, state 1, blue colour in pie charts) as well as their membranous to scarious nature were present in the common ancestor of Scutellaria sect. Lupulinaria and S. sect. Apeltanthus, but appear to have already been present earlier, before the divergence of S. pontica and S. baicalensis in clade A. However, the precise reconstruction of this character state transformation requires a clarification of the phylogenetic position of S. pontica, because the topological uncertainty with respect to the second and third branch of clade A certainly has influenced our calculated state probabilities for the second node of clade A. In case further phylogenetic research leads to resolving an earlier-diverging position of S. poecilantha (see below, discussion of relationships within clade A), cucullate bracts may be revealed to mark a core of clade A as a synapomorphy.

The evolution of the shape of the upper lip of the calyx (character 10) is difficult to assess with equal probabilities for states 1 (testudinate), 2 (scutellate) and 3 (scutellate-inflated) at the deep nodes of Scutellaria (Fig. 4, colours blue, red and green, respectively). There seems to be a trend toward scutellate upper calyx lips in most of the ‘S. orientalis species-group’ members, including S. litwinowii Bornm. & Sint. and S. luteocoerulea Bornm. & Sint., whereas the subclade with S. heydei, S. kotkaiensis Rech. f. and S. linearis Benth. evolved scutellate-inflated lips (in fruit), and testudinate lips evolved in parallel in S. poecilantha and the S. leptosiphonS. stocksii subclades (Fig. 4). The bladder-like inflated upper and lower calyx lips in fruit in S. mexicana are derived from a scutellate calyx, and also the patelliform and funnelform ones in Holmskioldia and Wenchengia mark developments specific to these two genera. Homoplasy is present in the evolution of flower colours, although the bright red corollas are a synapomorphy for S. sect. Perilomia as currently defined (character 12, state 1, blue colour in pie chart, Fig. 4).

The evolution of leaf characters 2 to 6 as summarized in Fig. 5 shows significant homoplasy, and this is evident for example for the prominent, deeply pinnatifid leaf margins present in e.g. Scutellaria pinnatifida subsp. mucida (Stapf) Rech. f. and S. prilipkoana (Fig. 3, character 3, state 3, green colour in pie charts in Fig. 5), which appear as further derived from already toothed leaves in S. grossheimiana (character 3, state 2, red colour in pie charts). Most of the states are a simplification of the actual gradual variation in the extension of the incision of the leaf margin or the density of trichomes, which was illustrated by e.g. Edmondson (1982). Using a more densely sampled tree of S. sect. Lupulinaria in the future, a direct illustration of the actual morphological variation in the specific samples as well as numerical approaches will be useful in species delimitation. However, leaf characters do not support major clades as synapomorphies, which is also expected due to their highly adaptive nature (e.g. a dense indumentum and pinnatifid lamina in dry and sunny habitats).

Most of the species of Scutellaria are erect perennial herbs (character 1, state 4, yellow colour in pie charts in Fig. 6). Clade A evolved different life forms, being chamaephytes with a woody base. One of these is a caespitose life form with decumbent stems and branches and woodiness restricted to the very base (character 1, state 3, green colour in pie charts; e.g. in S. grossheimiana and S. orientalis), whereas in a similar life form stems are suffruticose and upright (state 3, green colour in pie charts, e.g. in S. litwinowii). Both states resemble adaptations to open habitats with a longer dry season, where the plants form part of open grassland, steppe or savanna communities. These chamaephytic life forms may therefore have contributed to the diversification of clade A. In the African S. schweinfurthii, herbaceous stems sprout from a woody xylopodium following the rains (A. Paton, pers. comm.). Such vegetative structures are common in plants adapted to fire in savannah ecosystems. On the other hand, dwarf shrubs with numerous erect, woody stems characterize a clade of Central and South American species (e.g. S. costaricana, S. scutellarioides; character 1, state 1, blue colour in pie charts in Fig. 6).

There are further morphological characters, such as surface sculpture, indumentum and colour of nutlets, which will be informative at the level of infrageneric lineages as well as species. Ripe nutlets are missing in most herbarium specimens or even field gatherings, so that nutlet characters cannot be linked to individual specimens. Since a direct link of molecular and morphological data, derived from the same individual or specimen is crucial for testing species limits in the Scutellaria orientalis group, we have not included nutlet characters in this investigation. Nevertheless, they should be a subject in future research, and fruiting stages should be collected during field work. This also underscores the importance of documented living collections, where the same individual plant can be sampled in various stages.

Clade A and the Scutellaria orientalis group

Clade A comprises all members of Scutellaria subg. Apeltanthus in the wider circumscription (Paton 1990b), which is found monophyletic with maximum support in the combined dataset (Fig. 1) and is also recovered with trnL-F alone (Fig. 2, 0.98 PP, 73 ML-BS, 63 MP-JK). Scutellaria subg. Apeltanthus therefore forms a core of clade A. One of the noteworthy new findings in this investigation is that the African ‘S. violascens species-group’ sensu Paton (1990b) of S. subg. Scutellaria sect. Scutellaria, here represented by S. schweinfurthii subsp. paucifolia (Baker) A. J. Paton, is sister to the remainder of clade A, followed by a lineage of Central to E Asian species of S. sect. Scutellaria (S. amoena, S. baicalensis, S. rehderiana Diels, S. viscidula Bunge; Fig. 2) and S. sect. Anaspis (S. kingiana) and the Caucasian S. pontica (S. sect. Salviifoliae) in a tritomy. Scutellaria sect. Salviifoliae was considered as morphologically intermediate between S. subg. Apeltanthus and S. subg. Scutellaria (Paton 1990a), which is in line with the now inferred phylogenetic position. However, further taxon sampling will be needed to test if the other species of the section form a lineage with S. pontica. Also, a phylogenomic approach will be required to resolve the phylogenetic position of S. pontica and test if it is sister to S. subg. Apeltanthus sensu Paton (1990b; Fig. 1). Whereas S. kingiana appears in clade A, the more extensive trnL-F tree (Fig. 2) depicts other species currently included in S. sect. Anaspis (e.g. S. ariana Hedge; Paton 1990b; Safikhani & al. 2018) in clade B, related to S. albida. It is possible that S. kingiana does not belong to S. sect. Anaspis as supported by Paton (1990b) and Zhao & al. (2020), but rather in an E Asian lineage that may correspond to the ‘S. strigillosa species-group’ (Paton 1990b) together with S. baicalensis and S. rehderiana. A detailed comparative study of nutlet morphology and anatomy will be relevant to test if the nutlets with a grey-black colour and their surface only partly covered with hairs that are considered diagnostic for S. sect. Apeltanthus (see Paton 1990b) are homoplastic or could be confined to the lineage corresponding to S. sect. Apeltanthus (Fig. 1). Salmaki & Müller (2019) also depicted S. baicalensis, S. kingiana and S. rehderiana in a clade with their ITS data, which is in line with our plastid tree.

Further, the molecular data provide a strong signal for an early branch in the core of clade A that corresponds to a core of Scutellaria sect. Apeltanthus (sensu Paton 1990b). The remainder of the core clade A then corresponds to S. sect. Lupulinaria, which at the rank of section appears largely monophyletic as circumscribed by Paton (1990b). Scutellaria sect. Apeltanthus is here represented by S. leptosiphon and S. stocksii. Although there is currently no sequence for the type species, S. orbicularis Bunge, the suite of morphological characters shared by all species currently included in the section in concert with a relatively limited range (from the Afghan and Pamir mountains to Central Asia) may indicate its monophyly. From the species included by Juzepczuk (1954) in his more narrowly defined fourth subgenus, S. subg. Apeltanthus (corresponding to S. sect. Apeltanthus sensu Paton 1990a), S. poecilantha is inferred in our molecular trees among the taxa of the S. orientalis group (Fig. 2) and therefore would have to be excluded from a monophyletic S. sect. Apeltanthus. However, the respective node is supported by 1.0 PP in the Bayesian tree but receives only 82JK and 84 BS support in the parsimony and likelihood analyses, respectively (Fig. 1). Future analyses should therefore test the signal for this node with more completely sampled plastid genomes and if S. sect. Apeltanthus is paraphyletic to S. sect. Lupulinaria (incl. S. sect. Cystaspis). Also, further species placed here by Juzepczuk (1954) and Rechinger (1982) require inclusion in molecular datasets in the future (e.g. S. andrachnoides Vved., S. baldshuanica Nevski, S. guttata Nevski, S. macrodonta Nevski and S. rubromaculata Juz. & Vved.). Scutellaria immaculata Nevski also belongs to these species and is included in the trnL-F tree (Fig. 2), but this marker alone does not yield the necessary resolution to judge on the circumscription of S. sect. Apeltanthus.

Looking at the classification system of Paton (1990b), Scutellaria subsect. Lupulinaria is paraphyletic to S. subsect. Cystaspis (Fig. 2), with S. physocalyx appearing in an isolated position between two main subclades comprising samples of S. orientalis and allies, whereas S. heydei is resolved in a terminal shallow tritomy together with S. kotkaiensis and S. linearis. As a consequence, a membranous-inflated scutellum, which was used as diagnostic character for the entity at differing ranks (section, Juzepczuk 1951; subgenus, Juzepczuk 1954; or subsection, Paton 1990b), evolved convergently. On the other hand, all three species share a similar geographical distribution in the Afghan and Pamir mountains to the Himalaya, which would be in line with relatively recent speciation from a common ancestor. The close relationship of S. litwinowii and S. luteocoerulea found in the combined tree (Fig. 2) was also indicated by Safikhani & al. (2018). Whereas Paton (1990b) listed S. litwinowii for the ‘S. orientalis species-group’, Juzepczuk (1954) grouped both species in S. subsect. Fruticosae (Fig. 2) and not in S. subsect. Orientales. Scutellaria subsect. Fruticosae, according to Juzepczuk (1954) characterized as being subshrubs with numerous, usually virgate stems, may represent a sublineage within the Irano-Turanian radiation of S. subsect. Lupulinaria also including S. glechomoides [in its own species group sensu Paton (1990b)], S. multicaulis Boiss. and S. nepetifolia based on our trnL-F tree (Fig. 2). A similar lineage was depicted by Safikhani & al. (2018), who also analysed trnL-F sequences. However, resolution of this genomic region alone is not enough to make any conclusions about further relationships in S. subsect. Lupulinaria.

The narrowest circumscription of a taxon relating to the Scutellaria orientalis group was made by Juzepczuk (1939) with his “cycle Orientales”. As is evident from our phylogenetic trees [samples with taxon names in bold and marked with an asterisk in Fig. 1 and 2, indicating taxa that were explicitly mentioned by by Juzepczuk (1954) as members of S. subsect. Orientales], these species are concentrated in S. sect. Lupulinaria (two subclades in Fig. 1 and three subclades in the more extensively sampled trnL-F tree), but these subclades are not exclusively composed of members of this subsection. Grossheim (1945) also argued that”cycle Orientales” as circumscribed by Juzepczuk (1939) is not a natural group, in particular as he assumed that the Oxystegiae and Platystegiae groups have different origins. Juzepczuk (1954) defined S. subsect. Orientales as plants with decussate, 4-angled, spike-like, elongated inflorescences, in contrast to interrupted inflorescences in S. subsect. Fruticosae and short, subcapitate inflorescences in S. subsect. Ramosissimae. These three subsections were further described as possessing leaves with a silvery, tomentose indumentum beneath, whereas leaves in S. subsect. Alpinae are green or mostly greyish green, and plants have also spike-like, elongated and uninterrupted inflorescences (see tree in Fig. 2 for annotation of subsections). It needs to be considered that not all species sharing the respective characters occur in the area covered by the Flora URSS and are therefore not annotated (e.g. S. tomentosa Bertol.). On the other hand, the characters used by Juzepczuk to delimit his subsections are prone to rapid change in the course of speciation and adaptation to habitats differing in elevation and humidity. We have therefore also indicated in bold, with two asterisks, those species that were mentioned by (Paton 1990b) in the ‘S. orientalis species-group’. Paton (1990b) circumscribed his ‘S. orientalis species-group’ to contain all species with cucullate bracts clasping the calyx at anthesis, a thick, woody rhizome (therefore excluding S. glechomoides) and ovate to narrowly ovate, entire, dentate or pinnatifid leaves (therefore excluding S. linearis). Most of the taxa resolved as part of S. subsect. Lupulinaria are therefore included in this broader S. orientalis group, except S. glechomoides (Paton’s ‘S. glechomoides species-group’), S. linearis (Paton’s ‘S. linearis species-group’), S. poecilantha and the members of S. subsect. Cystaspis appearing nested in S. subsect. Lupulinaria.

The Oxystegiae and Platystegiae groups as defined by Juzepczuk (1954) are not monophyletic (see especially Fig. 2). Within the Oxystegiae group, S. ser. Araxenses Grossh. on one hand and the Mesostegiae and Pulchellae series on the other are resolved in two different lineages. Members of S. ser. Platystegiae appear in all three lineages, and the only member of the Adenostegiae series (the second series of the Platystegiae group) currently sampled (S. adenostegia) is in the same lineage as the Mesostegiae and Pulchellae series. It is noteworthy that the members of Scutellaria subsect. Orientales in the strict sense of Juzepczuk (1954) are resolved in two different, well-supported, major subclades. One subclade contains most samples from the Greater Caucasus (including N Caucasus, Scu22 and Scu25; see also Fig. 3), the Talysh mountains (SW of the Caspian Sea and part of the Iranian plateau, e.g. S. grossheimiana Scu29, Scu46, Scu47 and S. prilipkoana Scu28, Scu30 from Lerik) and Central Asia (S. adenostegia Briq., S. tuvensis Juz.; Fig. 1). Scutellaria alpina, although with a different distribution (the sample sequenced here is from the W Alps), is also found here, along with further species of Juzepczuk's S. subsect. Alpinae (Fig. 2). The position of S. alpina within the S. orientalis group is also supported by morphology of the inflorescence and the cucullate bracteoles, although this species differs by its blue flowers (Fig. 3).

The other subclade incudes samples from the Lesser and Greater Caucasus and adjacent steppe areas, westward to the Krasnodar area in Russia (Scutellaria novorossica, Scu23; Fig. 1). Most of the Iranian samples are found in a third, albeit weakly supported lineage of the trnL-F tree (Fig. 2), including S. bornmuelleri, S. multicaulis, S. platystegia and several subspecies of S. pinnatifida. Moreover, Pichugin (2015) suggested that the Crimean endemic S. taurica Juz. has affinities with S. stevenii Juz. (both from S. ser. Tauricae Grossh. of the Oxystegiae group), being distant from S. orientalis. Because their sampling was scarce and trnL-F sequences are not available, it remains unclear if there are further lineages in addition to those presented in Fig. 2. Future analyses of the species-level relationships and species limits within the S. orientalis group also need to extend the sampling to further geographical areas (especially the Balkan peninsula, W Mediterranean subregion and Anatolian highlands) and improve the sampling in Central Asia and Siberia.

Clade B

This major clade is constituted by the ‘Scutellaria albida species-group’ as defined by Paton (1990b) including S. altissima and S. tournefortii (S. subg. Scutellaria sect. Scutellaria) and a core of species from S. subg. Scutellaria sect. Anaspis (S. ariana, S. fragillima Rech. f., S. velutina Juz. & Vved., S. xylorrhiza; Paton 1990b; Safikhani & al. 2018). Scutellaria scandens, which was informally assigned to the ‘S. repens species-group’ in S. sect. Scutellaria by Paton (1990a), appears in the same clade as the other members of S. sect. Anaspis in clade B (Fig. 2). Scutellaria sect. Anaspis therefore appears nested among the lineages of clade B (Fig. 2). However, further increased taxon sampling has to test if S. altissima and S. tournefortii mark a specific lineage that is sister to the remainder of clade B. Also Safikhani & al. (2018) presented an ITS tree in which their “Clade I” depicts members of S. sect. Anaspis as sister to S. albida and S. megalaspis Rech. f., and S. tournefortii as sister to both. These nodes have gained statistical support, whereas their inferred position of S. galericulata as sister to the aforementioned taxa has no statistical confidence. Earlier authors (Paton 1990b) indicated the similarity in nutlet morphology of the ‘S. albida species-group’ with S. sect. Anaspis. The current results of our molecular phylogenetic analysis therefore appear to be consistent with morphology. Scutellaria diffusa Benth., which is a member of S. sect. Salviifoliae according to Edmondson (1982), is inferred as a sole species within the clade representing the ‘S. albida species-group’ (Fig. 2). This was recently also observed by Seyedipour & al. (2020), who presented trnL-F and ITS trees of Scutellaria species mostly from Iran and sequenced the same material as in our investigation (Ern 6923), and it is confirmed here with increased node support as inferred from our larger molecular character set. Considering the type species of the section, S. salviifolia Benth. (Paton 1990a), and the morphological similarity of this species to S. diffusa (Edmondson 1982; Ezer & Renda 2012), it is very possible that S. pontica just exhibits a convergent evolution of a prostrate habit [in S. diffusa and S. pontica, whereas S. salviifolia has upright stems; Ezer & Renda (2012)], with small and entire leaves, otherwise also characteristic of the core of S. sect. Salviifoliae. Scutellaria diffusa differs from S. salviifolia by a purple, versus yellow, flower colour, while purple flowers are also present in S. pontica. However, shifts in flower colour are homoplastic in Scutellaria, as is also evident in S. alpina with purple flowers but yellow-flowered ancestors (Fig. 4).

Rechinger (1982) accepted Scutellaria sect. Stachymacris including S. fragillima, S. megalaspis and S. tournefortii, all of which are resolved in clade B (Fig. 1, 2). As such, S. sect. Stachymacris could be a taxon concept to formally classify clade B, although it is currently at sectional rank (see below). Scutellaria fragillima was alternatively classified in S. sect. Anaspis (Juzepczuk 1954; Fig. 2), which represents one of the three sublineages of clade B.

Clade C

The phylogenetic position of Scutellaria sect. Perilomia, represented here by S. scutellarioides (type species of S. sect. Perilomia; Paton 1990a) and S. volubilis Kunth, is resolved among other neotropical members of Scutellaria representing the ‘S. costaricana species-group' (S. costaricana) and the ‘S. speciosa species-group' (S. incarnata), all of which are well supported with an early divergence in clade C as sister to all its other species. This lineage is also morphologically characterized by red or scarlet flowers, distally bent corolla tube and other characters that had resulted in a very distinct grouping in the UPGMA analysis of Paton (1990b). Our molecular tree suggests that the emended circumscription of S. sect. Perilomia by Paton (1990a) to include S. sect. Perilomia, S. sect. Perilomioideae (Benth.) Epling and S. sect. Theresa (Clos) Epling (Epling 1942) should be further extended to include other neotropical species. However, this has to await a more complete inclusion of those taxa in the molecular phylogeny, considering that there are several species groups of S. sect. Scutellaria present in Central and South America, with potential relatives. Most interestingly, there are several lineages within clade C that extend into the Neotropics, such as S. mexicana (S. sect. Salazaria) as the third branch (Fig. 1) and a lineage comprising S. ovata Hill (representing the ‘S. ovata species-group’ occurring in the U.S.A. and Mexico) sister to S. nummulariifolia Hook. f. from Argentina [the only species of the ‘S. nummulariifolia species-group’ sensu Paton (1990b)]. Our molecular phylogenetic results indicate that at least two lineages of Scutellaria may have migrated southward from North to South America (see below), underscoring the need of a very dense taxon sampling to better understand relationships among New World Scutellaria in future studies. Morphological similarities of the ‘S. ovata species-group’ to species in clade B as indicated by Paton (1990b), such as a pericarp (also present in S. schweinfurthii), and overall similarity to S. altissima of the ‘S. albida species-group’ must therefore be explained by convergence. Further, this investigation provides some evidence for the monophyly of the ‘S. galericulata species-group’ as circumscribed by Paton (1990b) including S. galericulata, S. minor Huds. and the E Asian S. dependens Maxim. as a core, but also including S. lateriflora and S. hastifolia L. (Fig. 2), albeit node support is yet weak. The morphological similarity of S. galericulata to S. lateriflora was pointed out by Paton (1990b), although he assigned the latter to its own species group (‘S. lateriflora species-group’). Another well-supported terminal subclade comprises species from China (e.g. S. franchetiana H. Lév., S. tsinyunensis; for the latter see also Fig. 3) and appears to correspond to a more species-rich radiation in subtropical parts of China identified by Zhao & al. (2017) based on ITS sequence data.

Species limits, species identification and species diversity in the Scutellaria orientalis group

The employed molecular markers (matK-trnK, rpl16, trnL-F) exhibit substantial phylogenetic signal within the Scutellaria orientalis group and indicate complex speciation patterns, which happened and appear to be ongoing more or less isolated from each other in different parts of the Caucasus and the Irano-Turanian region. Molecular phylogenetic results (Fig. 1, 2) contradict a treatment of S. orientalis as a single species with many subspecies as implemented by various authors such as Fedorov (1967), Edmondson (1982), Menitsky (1992), Celep & Dirmenci (2017) and reflected by the World Checklist of Selected Plant Families that is largely implemented also in the Euro+Med PlantBase (Euro+Med 2006+). Several of the entities treated at subspecific rank (but identified in our investigation at the level of species) appear in different subclades of the molecular trees. Assuming a concept of vicariance as a result of ongoing allopatric speciation, all subspecies of a species would have to be found in the same terminal clade as the nominate subspecies. The lectotype of S. orientalis (Herb. Linnaeus no. 751.1, LINN [image at  http://linnean-online.org/6677/]; Edmondson 1982) has long and slender stems that must have been procumbent with many short side branches, leaves with a pinnate margin and a short inflorescence with few flowers, the lower of which have just opened. Plants similar to this morphotype were included in this investigation under the name S. orientalis, but appear in two also geographically different subclades of S. sect. Lupulinaria (Fig. 2). The lectotype does not convey any locality information, whereas other historical material of S. orientalis in LINN comes from “Caucasus Mt.” or “Taurus” and apparently comes from morphologically deviant populations. Using morphology alone, a clarification of the proper identity of the type specimen may therefore be limited.

On the other hand, the phylogeny depicts several individuals of the morphotype and current species identification in the same terminal lineage, such as Scutellaria grossheimiana, which was described from Lerik (Juzepczuk 1951), and S. prilipkoana (see also Fig. 3), which was also described from Talysh (Grossheim 1945). Both taxa have very close affinities morphologically, phylogenetically and also geographically, and stand for the other species that may be range restricted or much narrower endemics that can be resolved by their genotypes found in a certain geographical area. Also Charadze (1961) argued that S. sect. Lupulinaria in the Caucasus contains several range-restricted oreophytes, indicating that speciation by adaptation to high mountain environments and probably geographic isolation plays a significant role. A much denser taxon sampling and a more representative geographical coverage will be needed throughout the S. orientalis group for species delimitation that samples type localities if possible and at best also type specimens directly. Our experience with herbarium material was quite good (Scu083 from 1933, Scu76 from 1962 and Scu075 from 1965), so this perspective exists for at least a significant number of type specimens.

We have also implemented an extended analysis by adding the available trnL-F sequences from other published studies to cover additional species and to ideally provide some insights into infraspecific variability and species limits through sequences from individuals from further geographic localities. The trnL-F region appeared as the most widely used marker, and in several cases the only plastid region employed (Safikhani & al. 2018; Seyedipour & al. 2020). The trnL-F tree in Fig. 2 therefore gives the currently most representative approximation on relationships within Scutellaria in terms of taxonomic coverage. As expected, the resolution by trnL-F alone is considerably less than through the combined plastid data set, so individuals of several closely related species appear in polytomies (e.g. S. altissima and S. tournefortii; compare Fig. 1 and 2). Nevertheless, the placement of some of the currently accepted species is in line with similarities in morphology in our trees, e.g. S. farsistanica Rech. f. and S. nepetifolia as sisters; see also Rechinger (1982), Ranjbar & Mahmoudi (2018), albeit the trnL-F topology presented by Safikhani & al. (2018: fig. 2) differs significantly. A possible explanation is differences in the matrix subjected to tree inference, by not excluding mutational hotspots of uncertain homology. Correct alignment matters greatly in plastid data sets, and in particular when genetic distances are low, as among closely related species (Escobari & al. 2021). Noteworthy is also the placement of all individuals of S. albida from Iran in a clade to which the individual from Greece appears close, but unresolved (Fig. 2), which points to phylogeographic patterns within this species. On the other hand, our Fig. 2 reveals several taxa in significantly different parts of the tree, such as S. litwinowii or S. platystegia. In the latter case, S. platystegia (LC384602) is from a specimen collected in the Kiamaghi Dagh mountains in N Iran. Considering the specimen (Safikhani 105466, TARI image!), this plant appears with deeply pinnatifid, tomentose cauline leaves and very densely arranged, shortly tomentose, cucullate bracts, whereas the type specimen (LE 01010199 image!) is a mesophytic plant with moderately pinnatifid leaves from a different geographic area in the Caucasus. As a consequence, the sequence LC384602 could also be misidentified. This example shows that a direct comparison with the respective type specimens is essential in a situation with so-far prevailing regional treatments and identification keys as well as species diagnoses based largely on gradually varying leaf, inflorescence and flower characters.

The employed combined plastid regions (matK-trnKpsbA+rpl16+trnL-F) form a good basis for an overall species-level phylogeny within Scutellaria sect. Lupulinaria, which confirms the high levels of hierarchical phylogenetic signal observed in these genomic regions in other angiosperm lineages (Barniske & al. 2012; Korotkova & al. 2011). Compared to recent radiations within angiosperm genera such as Dianthus L. (Caryophyllaceae; G. Fassou pers. comm. 2021), Gynoxys Cass. (Asteraceae; Escobari & al. 2021) or Pyrus L. in the Caucasus (Rosaceae; Korotkova & al. 2018), the genetic distances of the same plastid regions found here in S. orientalis and relatives (Table 1) are much greater. This not only suggests that the S. orientalis group is composed of many phylogenetically defined entities, but also holds promise in further improving the resolution of phylogenetic trees in this group by complementing with additional highly variable spacer regions (Zhao & al. 2020) and the application of numerous loci through a phylogenomic approach. Nuclear loci will be required to identify possible reticulate evolution and incomplete lineage sorting.

Despite the resolution obtained with markers also used in other DNA barcoding projects (e.g. the German Barcode of Life, GBOL; Geiger & al. 2016), the still largely unclear species limits, with the need to link samples represented in molecular datasets to formally described taxa (including type specimens), hamper attempts of molecular species identification in the Scutellaria orientalis group at the moment. What is needed is an integrative taxonomic approach, using structured character data, both molecular and morphological (Kilian & al. 2015), and also testing geographical and ecological scenarios relevant for speciation. In the example of S. galericulata, which is a morphologically well-characterized species, the matK-trnK and rpl16 plastid genomic regions reveal several substitutions that resolve the species as monophyletic but show the individual from the German state of Bavaria (Scu001) to possess a haplotype closer related to that in the individual from the Caucasus (Scu024) than to the sampled individual from the German state of Saxony-Anhalt (Fig. 1). This underscores the complexity in the evolution of the flora of the N hemisphere and the need to implement range-wide sampling in the study of widespread taxa.

As a consequence, none of the currently available Flora treatments or checklists at continental to global level will describe the species diversity of the Scutellaria orientalis group in a way reflecting only biologically meaningful entities as taxa at species level. An interim strategy for a species-level checklist will therefore be to accept species, rather than infraspecific taxa with doubtful assignments, and to evaluate the species to be accepted on the basis of all existing data, even if this results in a mosaic-like pattern of different stages of knowledge for different taxa. The level of taxonomic evaluation should be indicated and the authority of the accepted respective taxon concept cited to facilitate the stepwise improvement of the taxonomic backbone. Such an approach has been taken by the Taxonomic Expert Networks (TENs) of the World Flora Online consortium (Borsch & al. 2020), and would ideally also be followed up in the case of Scutellaria.

Implications for biogeography

The core of clade A entails the members of Scutellaria subg. Apeltanthus sensu Paton, which are mainly distributed from the Anatolian plateau, the Caucasus, to the Altay, Tien Shan and Pamir Mountains in the east to the Iranian plateau in the south. The majority of taxa is therefore concentrated in the W Irano-Turanian region (see Manafzadeh & al. 2017 for review and circumscription), which corresponds to the earlier idea that the Irano-Turanian region is a major centre of diversification of S. subg. Apeltanthus (Paton 1990b). The phylogenetic position of S. alpina within the subclade of the S. orientalis group, which is mostly composed of samples from the Greater Caucasus and Central Asia, suggests a migration from this area westward to the Alps and Pyrenees. Considering that the species ranges from the W coast of the Black Sea (Bulgaria, Romania) over the Balkan peninsula and Carpathians to the Alps (Euro+Med 2006+), it would be interesting to test if the western genotypes are more derived. Plant migrations out of the Caucasus to Central Europe have also been documented in Campanula L. (Campanulaceae; Jones & al. 2017) and Primula L. (Primulaceae; Volkova & al. 2020).

The subdivision of the Scutellaria orientalis group into two major (Fig. 1) or even a third (Fig. 2) subclade, each of which corresponds to more restricted geographical areas, contradicts the subdivision of the S. orientalis group by Juzepczuk (1954) into the Platystegiae group, which was considered to contain predominantly mesophytic species, versus xerophilic species in the Oxystegiae group. It is very possible that the S. sect. Lupulinaria radiation encompasses several subclades that diversified predominantly in particular mountain ranges such as the Greater or Lesser Caucasus but also the Iranian plateau. Silakadze & al. (1990b) found similar patterns in the diversification of the Caucasian high-mountain bellflowers (Campanula sect. Scapiflorae (Boiss.) Kharadze), which evolved from a common ancestor. On the other hand, Szukala & al. (1990b) showed for the genus Jurinea Cass. (Asteraceae), which has a largely similar distribution to S. sect. Lupulinaria, that the Caucasus was colonized several times during the evolutionary history of the genus, followed by subsequent speciation, especially in the N Caucasus. Also, different subclades were found including species in Central Asia and SW Asia, suggesting several independent species radiations in the wider Irano-Turanian region. So far, our phylogenetic trees of the S. orientalis group are not yet sufficiently conclusive with respect to species-level relationships of the Central Asian taxa such as S. altaica and S. supina L. (Fig. 2, 6), but this is likely to be resolved with a more extensive molecular character set.

The genus Scutellaria appears to be a N hemisphere group, from which several lineages migrated south, once to Africa (S. schweinfurthii and relatives) and at least two times to South America (Fig. 6) within clade C (S. subg. Scutellaria as defined in this study). The patterns are very similar to other temperate genera (e.g. Pinguicula L., Lentibulariaceae; Cieslack & al. 2005), where a southward migration also occurred from ancestors in Central America and SE North America. On the other hand, there are E Asian as well as North American species groups that appear rather unrelated to the lineages that gave rise to the Central and South American taxa, and these may be explained by recent migration in the N hemisphere via Beringia. A dense taxon sampling of S. subg. Scutellaria is needed to reveal the biogeographic history and test if there are other hitherto unrecognized migration events to South America. Scutellaria subg. Anaspis as defined here represents a Eurasian lineage that is centred in the E Mediterranean, SW Asia and the Caucasus and may have diversified to Central Europe in the west and the Himalaya in the east.

Implications for a phylogeny-based infrageneric classification

Based on phylogenetic evidence and the evolution of morphological characters, an infrageneric classification system of the genus Scutellaria with more than two subgenera will be necessary in order to make these monophyletic. The current taxon sampling entails three major clades, A, B and C, with S. subg. Scutellaria largely paraphyletic to S. subg. Apeltanthus (Fig. 1). The core of clade A is formed by S. subg. Apeltanthus, and one option could be to extend the circumscription of this subgenus to clade A. Scutellaria orbicularis was designated as the type species of S. sect. (and subg.) Apeltanthus by Paton (1990a), but its position in the molecular phylogenetic tree is yet to be determined. The disadvantage of such a treatment based on the current phylogenetic results would be that there would be no synapomorphic morphological character state. The situation is not only complicated by the hitherto unresolved phylogenetic position of S. pontica but also by the position of a lineage of S. baicalensis and relatives (S. subg. Scutellaria sect. Scutellaria) among the earliest branches of clade A, which contains species that are more similar to the E Asian taxa in clade C in terms of the morphological characters investigated here (Fig. 46). Any formal infrageneric treatment therefore needs to await a much more comprehensive phylogenetic analysis that can test which other species potentially belong to this E Asian lineage within clade A and examine the evolution of further morphological characters (e.g. of nutlets). Based on the current results, the African S. schweinfurthii lineage, which deviates morphologically from the core of clade A, could also be classified at subgeneric rank. However, the specific life form (Fig. 6) may not suffice as a synapomorphy, because there seems to be some variability in potential close allies of the ‘S. violascens species-group’ not yet sampled here.

The core of Scutellaria sect. Anaspis falls within clade B, so that there is a name at subgeneric rank for this clade. Scutellaria subg. Anaspis was accepted by Wu & Li (1977) and Rechinger (1982), and the type species is S. fedtschenkoi Bornm. (see Paton 1990a). Although previous authors argued for the consistency of the section based on morphology, and several taxa are included at least in our trnL-F tree (Fig. 2), there are yet no sequence data available for S. fedtschenkoi. Additional sequencing needs to confirm that this species falls within clade B to support eventual acceptance of the name S. subg. Anaspis in a wider circumscription of this clade. Paton (1990b) also mentioned that nutlet anatomy links the ‘S. albida species-group’ to S. sect. Anaspis. Altough S. sect. Stachymacris. would match clade B, there is no validation of this name at subgeneric rank. Therefore, the name S. subg. Anaspis will have priority when formally classifying clade B at subgeneric level in the future.

Scutellaria subg. Scutellaria would have to be limited to clade C through a narrower circumscription, and S. galericulata as the type species is included in our phylogeny. A majority of species in this clade seems to possess unspecialized inflorescences (character 7 state 4), which are therefore clearly resolved as an ancestral state of clade C (Fig. 4). However, because raceme-like inflorescences with opposite and secund flowers, which are consistently present in clade B, are reconstructed to have also evolved in parallel in a terminal lineage within clade A (e.g. S. indica, S. ovata), there is no actual synapomorphy among the currently investigated morphological characters.

Author contributions

The study was designed by R.S. and T.B. Field and herbarium work as well as the identification of plant material and the assessment of morphological character states was mostly done by R.S. with smaller contributions by G.P. and T.B. The molecular dataset was generated and analysed by R.S. with help from T.B. and with technical support in the lab. T.B. processed the GBOL data. All authors contributed to the text and approved the final version of the manuscript.

Acknowledgements

We gratefully acknowledge Dr. James C. Solomon and Dr. James S. Miller, Missouri Botanical Garden, for providing samples from many relevant specimens and for their help during a visit by R.S. Dr. Chun-Lei Xiang, Kunming Institute of Botany, Chinese Academy of Sciences, sent us plastid sequences of Holmskioldia and Wenchengia even before they were published as well as photos of Chinese Scutellaria species. We thank Prof. Dr. Valida Alizade and Dr. Esmira Alirzayeva for support as focal and local coordinators of the project “Developing tools for conserving the plant diversity of the South Caucasus” funded by the VolkswagenStiftung (AZ85021 and AZ89950). The sequence data for the Scutellaria samples from Germany were generated at the Botanic Garden and Botanical Museum Berlin (BGBM) in the context of the German Barcode of Life GBOL5 Project (Botany) funded by the German Federal Ministry of Education and Research (BMBF) under FKZ01LI1101E and FKZ01LI1501E. R.S. appreciates the many years of support and mentoring from Prof. Dr. Alizade, which was very important for the successful implementation of this project. Thanks are due to the herbarium and lab staff at the BGBM, in particular Bettina Giesicke, Kim Govers and Peter Hein, and to Nicholas Turland for insightful discussions regarding various aspects of nomenclature, classification and historical literature. We acknowledge various kinds of help and scientific discussion during the implementation of this project by Dr. Ramazan Murtazaliyev (Makhachkala), Prof. Dr. Ahmed Ouhammou (Marrakech), Prof. Dr. Mostafa Assadi (Tehran), Prof. Dr. Komiljon Tojibaev (Tashkent) and Dr. Valery Tikhomirov (Minsk). The insightful comments by Dr. Dmitry Geltman (Saint Petersburg) and Dr. Alan Paton (Kew), who reviewed an earlier version of this paper, are greatly appreciated.

Supplemental content online

See  https://doi.org/10.3372/wi.51.51307

 Appendix S1 (wi.51.51307_Appendix_S1.pdf). Chronology and comparison of infrageneric taxonomy of Scutellaria proposed by different authors. Subgenera are bold and underlined, sections are bold and italic, lower ranks are only italic.

 Appendix S2 (wi.51.51307_Appendix_S2.xlsx). Matrix of morphological characters.

Annotated alignments in Nexus format for matK-trnKpsbA (dataset A), rpl16 (dataset A), trnL-F (dataset A) and trnL-F (dataset B).

  (wi.51.51307_suppl.zip)

References

1.

African Plant Database 2020–2021: African Plant Database, version 3.4.0. – Geneva: Conservatoire et Jardin botaniques de la Ville de Genève; Pretoria: South African National Biodiveristy Institute. – Published at  https://www.ville-ge.ch/musinfo/bd/cjb/africa/index.php [accessed 4 Oct 2020 and 12 Aug 2021]. Google Scholar

2.

Barniske A., Borsch T., Müller K., Krug M., Neinhuis C. & Quandt D. 2012: Phylogenetics of early branching eudicots: comparing phylogenetic signal across chloroplast introns, spacers, and genes. – J. Syst. Evol. 50: 85–108.  CrossrefGoogle Scholar

3.

Bendiksby M., Thorbek L., Scheen A. C., Lindqvist, C. & Ryding O. 2011: An updated phylogeny and classification of Lamiaceae subfamily Lamioideae. – Taxon 60: 471–484.  CrossrefGoogle Scholar

4.

Bentham G. 1834: Scutellaria and Perilomia. – Pp. 419–446 in Labiatarum genera et species. – London: Ridgeway. Google Scholar

5.

Bentham G. 1848: Labiatae. – Pp. 412–431 in: Candolle A. P. de (ed.),  Prodromus systematis naturalis regni vegetabilis 12. – Paris: Treuttel et Würtz. Google Scholar

6.

Bentham G. 1876: Scutellaria, Salazaria and Perilomia. – Pp. 1201–1203 in: Bentham G. & Hooker J. D.,  Genera plantarum 2. – London: Lovell Reeve & Co. Google Scholar

7.

Billie L. T. & Delprete P. G. 1996: Nutlet sculpturing in Scutellaria sect. Resinosa (Lamiaceae) and its taxonomic utility. – Pl. Syst. Evol. 199: 109–120.  CrossrefGoogle Scholar

8.

Boissier P. E. 1879: Scutellaria . – Pp. 681–691 in:  Flora orientalis 4. – Genevae et Basileae: H. Georg. Google Scholar

9.

Borsch, T., Berendsohn W. G., Dalcin E., Delmas M., Demissew S., Elliott A., Fritsch P., Fuchs A., Geltman D., Güner A., Haevermans T., Knapp S., le Roux M. M., Loizeau P.-A., Miller C., Miller J., Miller J T., Palese R., Paton A., Parnell J., Pendry C., Qin H.-N., Sosa V., Sosef M., von Raab-Straube E., Ranwashe F., Raz L., Salimov R., Smets E., Thiers B., Thomas W., Tulig M., Ulate W., Ung V., Watson M., Wyse Jackson P. & Zamora N. 2020: World Flora Online: placing taxonomists at the heart of a definitive and comprehensive global resource on the world's plants. – Taxon 69: 1311–1341.  CrossrefGoogle Scholar

10.

Borsch T., Hilu K. W., Quandt D., Wilde V., Neinhuis C. & Barthlott W. 2003: Noncoding plastid trnT-trnF sequences reveal a well resolved phylogeny of basal angiosperms. – J. Evol. Biol. 16: 558–576.  CrossrefGoogle Scholar

11.

Borsch T. & Quandt D. 2009: Mutational dynamics and phylogenetic utility of noncoding chloroplast DNA. – Pl. Syst. Evol. 282: 169–199.  CrossrefGoogle Scholar

12.

Briquet J. 1896: Scutellaria, Salazaria and Perilomia. – Pp. 224–227, 232–233 in: Engler A. (ed.),  Die natürlichen Pflanzenfamilien IV(3a). – Leipzig: Wilhelm Engelmann. Google Scholar

13.

Cantino P. D., Harley R. M. & Wagstaff S. J. 1992: Genera of Labiatae: status and classification. – Pp. 511–522 in: Harley R. M. & Reynolds T. (ed.), Advances in Labiate science. – Richmond: Royal Botanic Gardens, Kew. Google Scholar

14.

Celep F. & Dirmenci T. 2017: Systematic and biogeographic overview of Lamiaceae in Turkey. – Natural Volatiles & Essential Oils 4( 4 ): 14–27. Google Scholar

15.

Charadze A. 1961: Materials for the study of Scutellaria species of the Greater Caucasus from the section Lupulinaria A. Hamilt. (“Ad cognitionem specierum generis Scutellaria L. sectionis Lupulinaria A. Hamilt. E Caucaso magno”). – Zametki Sist. Geogr. Rast. 22: 44–57. Google Scholar

16.

Chen Q. & Zhang D.-Q. 2019: Characterization of the complete chloroplast genome of Scutellaria amoena C. H. Wright (Lamiaceae), a medicinal plant in southwest China. – Mitochondrial DNA Part B 4: 3057–3059.  CrossrefGoogle Scholar

17.

Chiang Y.-C., Huang B.-H. & Liao P.-C. 2012: Diversification, biogeographic pattern, and demographic history of Taiwanese Scutellaria species inferred from nuclear and chloroplast DNA. – PloS ONE 7( 11 ): e50844.  CrossrefGoogle Scholar

18.

Çiçek M. & Ketenoğlu O. 2011: Scutellaria anatolica (Lamiaceae), a new species from Turkey. – Ann. Bot. Fenn. 48: 276–279.  CrossrefGoogle Scholar

19.

Çiçek M. & Yaprak A. E. 2013: Scutellaria yildirimlii (Lamiaceae), a new species from Turkey. – Phytotaxa 132: 53–58.  CrossrefGoogle Scholar

20.

Cieslack T., Polepalli J. S., White A., Müller K., Borsch T., Barthlott W., Steiger J., Marchant A. & Legendre L. 2005: Phylogenetic analysis of Pinguicula (Lentibulariaceae): chloroplast DNA sequences and morphology support several geographically distinct radiations. – Amer. J. Bot. 92: 1723–1736.  CrossrefGoogle Scholar

21.

Darriba D., Taboada G. L., Doallo R. & Posada D. 2012: jModelTest2: more models, new heuristics and parallel computing. – Nature, Meth. 9: 772–772.  CrossrefGoogle Scholar

22.

Edmondson J. 1982: Scutellaria. – Pp. 78–100 in: Davis P. H. (ed.), Flora of Turkey and the East Aegean Islands 7. – Edinburgh: Edinburgh University Press. Google Scholar

23.

Ellenberg H. & Mueller-Dombois D. 1967: A key to Raunkiaer plant life forms with revised subdivisions. – Ber. Geobot. Inst. E.T.H. Stiftung Rübel. 37: 56–73. Google Scholar

24.

Epling C. 1942: American species of Scutellaria. – Univ. Calif. Publ. Bot. 20: 1–146. – Berkeley & Los Angeles: University of California Press. Google Scholar

25.

Escobari B., Borsch T., Quedensley T. S. & Gruenstaeudl M. 2021: Plastid phylogenomics of the Gynoxoid group (Senecioneae, Asteraceae) highlights the importance of motif-based sequence alignment amid low genetic distances. – Amer. J. Bot. 108: 2235–2256.  CrossrefGoogle Scholar

26.

Euro+Med 2006+ [continuously updated]: Euro+Med PlantBase – the information resource for Euro-Mediterranean plant diversity. – Published at  http://ww2.bgbm.org/EuroPlusMed/ [accessed 4 Oct 2020]. Google Scholar

27.

Ezer N. & Renda G. 2012: Comparative morphological studies on Scutellaria salviifolia Benth. and Scutellaria diffusa Benth. (Lamiaceae) growing in Turkey. – FABAD J. Pharm. Sci. 37: 197–203. Google Scholar

28.

Fedorov A. 1967: 4838. Scutellaria orientalis L. subsp. platystegia (Juz.) Fed. comb. nova. – Pp. 73–74 in: Schedae ad herbarium florae URSS 17. – Leningrad: Nauka. Google Scholar

29.

Flora Zambesiaca Managing Committee 1950+: Flora zambesiaca 1–45. – Richmond: Royal Botanic Gardens, Kew. Google Scholar

30.

Gagnidze R. I. & Ivanishvili M. A. 1987: Scutellaria L. – Pp. 35–51 in: Gagnidze R. I. (ed.), Flora Georgia, ed. 2, 11. – Tbilisi: Metsniereba. Google Scholar

31.

Geiger M. F., Astrin J. J., Borsch T., Burkhardt U., Grobe P., Hand R., Hausmann A., Hohberg K., Krogmann L., Lutz M., Monje C., Misof B., Morinière J., Müller K., Pietsch S., Quandt D., Rulik B., Scholler M., Traunspurger W., Haszprunar G. & Wägele W. 2016: How to tackle the molecular species inventory for an industrialized nation—lessons from the first phase of the German Barcode of Life initiative GBOL (2012–2015). – Genome 59: 1–10.  CrossrefGoogle Scholar

32.

González-Gallegos J. G. & Vázquez-García J. A. 2013: Scutellaria cuevasiana and Scutellaria sublitoralis (Lamiaceae), two new species from Jalisco and Nayarit, Mexico. – Revista Mex. Biodivers. 84: 20–29.  CrossrefGoogle Scholar

33.

Grierson A. J. C. & Long D. G. (ed.) 1983–2002: Flora of Bhutan, 3 vols. – Edinburgh: Royal Botanic Garden Edinburgh. Google Scholar

34.

Grossheim A. A. 1945: Кaвкaзcкзe пpeпcтaвзтeлз ceкцзз Lupulinaria Ham. poпa Scutellaria L. [Caucasian representatives of the section Lupulinaria Ham. of the genus Scutellaria L.]. – Izv. Azerbaidzhansk. Fil. 3: 79–90. Google Scholar

35.

Gruenstaeudl M. 2020: annonex2embl: automatic preparation of annotated DNA sequences for bulk submission to ENA. – Bioinformatics 36: 3841–3848.  CrossrefGoogle Scholar

36.

Guseva A. & Cheryomushkina V. 2015: Polyvariation of the development Scutellaria L. species and the structure of their populations in Siberia. BIO Web of Conferences 16, Results and Prospects of Geobotanical Research in Siberia.  CrossrefGoogle Scholar

37.

Hamilton A. 1832:  Esquisse d'une monographie du genre Scutellaria ou toque. – Lyon: Louis Perrin. Google Scholar

38.

Harley R. M., Atkins S., Budantsev A. L., Cantino P. D., Conn B. J., Grayer R., Harley M. M., Kok R. de, Krestovskaja T., Morales R., Paton A. J., Ryding O. & Upson T. 2004: Labiatae. – Pp. 167–275 in: Kubitzki K. (ed.), The families and genera of vascular plants 7. – Berlin: Springer.  CrossrefGoogle Scholar

39.

Hosokawa K., Minami M., Nakamura I., Hishida A. & Shibata T. 2005: The sequences of the plastid gene rpl16 and the rpl16-rpl14 spacer region allow discrimination among six species of Scutellaria. – J. Ethnopharmacol. 99: 105–108.  CrossrefGoogle Scholar

40.

Isayev Y. 1957: Scutellaria L. – Pp. 234–245 in: Karyagin I. I. (ed.), Flora Azerbaijana 7. – Baku: Akademia Nauk Azerbaijan SSR. Google Scholar

41.

Jamzad Z. 2012: Scutellaria. – Pp. 85–139 in: Assadi M. M. (ed.), Flora of Iran 76. – Tehran: Research Institute of Forests & Rangelands. Google Scholar

42.

Jamzad Z. 2013: A survey of Lamiaceae in the flora of Iran. – Rostaniha 14: 59–67. Google Scholar

43.

Jiang D., Zhao Z.-Y, Zhang T., Zhong W.-H., Liu C.-S., Yuan Q.-J. & Huang L.-Q. 2017: The chloroplast genome sequence of Scutellaria baicalensis provides insight into intraspecific and interspecific chloroplast genome diversity in Scutellaria. – Genes 8: 227.  CrossrefGoogle Scholar

44.

Jones K. E., Korotkova N., Petersen J., Henning T., Borsch T. & Kilian N. 2017: Dynamic diversification history with rate upshifts in Holarctic bell-flowers (Campanula and allies). – Cladistics: 33: 637–666.  CrossrefGoogle Scholar

45.

Juzepczuk S. V. 1939: O «Scutellaria orientalis» aвтopoв кpымcкoй флopы [About “Scutellaria orientalis” by the authors of the Crimean flora]. – Bot. Zhurn. S.S.S.R. 24: 430–436. Google Scholar

46.

Juzepczuk S. V. 1951: Scutellariarum novarum decades I–VI. – Bot. Mater. Gerb. Bot. Inst. Komarova Akad. Nauk S.S.S.R. 14: 356–435. Google Scholar

47.

Juzepczuk S. V. 1954: Scutellaria. – Pp. 72–225 in: Shishkin B. & Juzepczuk S. V. (ed.), Flora URSS. 20. – Moscow & Leningrad: Editio Academiae Scientarum URSS. Google Scholar

48.

Kelchner S. A. 2000: The evolution of non-coding chloroplast DNA and its application in plant systematics. – Ann. Missouri Bot. Gard. 87: 482–498.  CrossrefGoogle Scholar

49.

Kilian N., Henning T., Plitzner P., Müller A., Güntsch A., Stöver B. C., Müller K. F., Berendsohn W. G. & Borsch T. 2015: Sample data processing in an additive and reproducible taxonomic workflow by using character data persistently linked to preserved individual specimens. – Database 2015: bav094.  CrossrefGoogle Scholar

50.

Kim S. T. & Lee S. T. 1995: Fruit surface morphology of Scutellaria (Lamiaceae ) in Korea and its taxonomic implication. – Korean J. Pl. Taxon. 25: 165–176.  CrossrefGoogle Scholar

51.

Korotkova N., Borsch T., Quandt D., Taylor N. P., Müller K. F. & Barthlott W. 2011: What does it take to resolve relationships and to identify species with molecular markers? An example from the epiphytic Rhipsalideae (Cactaceae). – Amer. J. Bot. 98: 1549–1572.  CrossrefGoogle Scholar

52.

Korotkova N., Parolly G., Khachatryan A., Ghulikyan L., Akopian J. A., Borsch T. & Gruenstaeudl M. 2018: Towards resolving the evolutionary history of Caucasian pears (Pyrus, Rosaceae)—phylogenetic relationships, divergence times and leaf trait evolution. – J. Syst. Evol. 56: 35–47.  CrossrefGoogle Scholar

53.

Lane T. 1983: Mericarp micromorphology of Great Plains Scutellaria (Labiatae). – S. W. Naturalist 28: 71–79.  CrossrefGoogle Scholar

54.

Ledebour C. F. 1849: Scutellaria. – Pp. 394–400 in:  Flora rossica 3. – Stuttgartiae: E. Schweizerbart. Google Scholar

55.

Li B., Cantino P. D., Olmstead R. G., Bramley G. L. C., Xiang C.-L., Ma Z.-H., Tan Y.-H. & Zhang D.-X. 2016: A large-scale chloroplast phylogeny of the Lamiaceae sheds new light on its subfamilial classification. – Sci. Rep. 6: 34343.  CrossrefGoogle Scholar

56.

Li B., Xu W.-X., Tu T.-Y., Wang Z.-S., Olmstead R. G., Peng H., Francisco-Ortega J., Cantino P. D. & Zhang D.-X. 2012: Phylogenetic position of Wenchengia (Lamiaceae): a taxonomically enigmatic and critically endangered genus. – Taxon 61: 392–401.  CrossrefGoogle Scholar

57.

Li X. W. & Hedge I. 1994: Labiatae. – Pp. 50–299 in: Wu Z. Y. & Raven P. H. (ed.), Flora of China 17 (Verbenaceae through Solanaceae). – Beijing: Science Press; St. Louis: Missouri Botanical Garden Press. Google Scholar

58.

Linnaeus C. 1753: Species plantarum. – Holmiae: Laurentii Salvii.  CrossrefGoogle Scholar

59.

Löhne C. & Borsch T. 2005: Molecular evolution and phylogenetic utility of the petD group II Intron: a case study in basal angiosperms. – Molec. Biol. Evol. 22: 317–332.  CrossrefGoogle Scholar

60.

Maddison W. P. & Maddison D. R. 2021: Mesquite: a modular system for evolutionary analysis. Version 3.70. – Published at  http://www.mesquiteproject.org/  Google Scholar

61.

Manafzadeh S., Staedler Y. M. & Conti E. 2017: Visions of the past and dreams of the future in the Orient: the Irano-Turanian region from classical botany to evolutionary studies. – Biol. Rev. 92: 1365–1388.  CrossrefGoogle Scholar

62.

Menitsky Y. L. 1987. Scutellaria L. – Pp. 23–30 in: Tahtadzjan A. (ed.), Flora of Armenia 8. – Yerevan: Akademii nauk Armianskoi SSR. Google Scholar

63.

Menitsky Y. L. 1992: Synopsis of species of the family Lamiaceae (Labiatae) from the Caucasus. – Bot. Zhurn. (Moscow & Leningrad) 77( 6 ): 63–78. Google Scholar

64.

Müller K. F. 2004: PRAP—computation of Bremer support for large data sets. – Molec. Phylogen. Evol. 31: 780–782.  CrossrefGoogle Scholar

65.

Müller K. F. 2005a: SeqState. Primer design and sequence statistics for phylogenetic DNA datasets. – Appl. Bioinf. 4: 65–69.  CrossrefGoogle Scholar

66.

Müller K. F. 2005b: The efficiency of different search strategies in estimating parsimony jackknife, bootstrap, and Bremer support. – B.M.C. Evol. Biol. 5: 58.  CrossrefGoogle Scholar

67.

Müller K. & Borsch T. 2005: Phylogenetics of Amaranthaceae based on matK/trnK sequence data—evidence from parsimony, likelihood, and Bayesian analyses. – Ann. Missouri Bot. Gard. 92: 66–102. Google Scholar

68.

Müller K., Quandt D., Müller J. & Neinhuis C. 2010: PhyDE version 0.9971: phylogenetic data editor. – Published at  http://www.phyde.de/  Google Scholar

69.

Murtuzaliyev R. 2009. Scutellaria L. – Pp. 203–205 in: Murtuzaliyev R. (ed.), The conspect of the flora of Dagestan. – Makhachkala: Epocha. Google Scholar

70.

Nieto Feliner G. & Rossello J. A. 2007: Better the devil you know? Guidelines for insightful utilization of nrDNA ITS in species-level evolutionary studies in plants. – Molec. Phylogen. Evol. 44: 911–919.  CrossrefGoogle Scholar

71.

Nixon K. C. 1999: The parsimony ratchet, a new method for rapid parsimony analysis. – Cladistics 15: 407–414.  CrossrefGoogle Scholar

72.

Olmstead R. W. 1989: Phylogeny, phenotypic evolution, and biogeography of the Scutellaria angustifolia complex (Lamiaceae): inference from morphological and molecular data. – Syst. Bot. 14: 320–338.  CrossrefGoogle Scholar

73.

Pagel M., Meade A. & Barker D. 2004: Bayesian estimation of ancestral character states on phylogenies. – Syst. Biol. 53: 673–684.  CrossrefGoogle Scholar

74.

Paton A. 1990a: The phytogeography of Scutellaria L. – Notes Roy. Bot. Gard. Edinburgh 46: 345–359. Google Scholar

75.

Paton A. 1990b: A global taxonomic investigation of Scutellaria (Labiatae). – Kew Bull. 45: 399–450.  CrossrefGoogle Scholar

76.

Pichugin V. S. 2015: The characteristic of species' taxonomy of the genus Scutellaria L. sprouting in the Crimea. – Bot. Vestn. Severn. Kavkaza [Botanicheskii Vestnik Severnogo Kavkaza] 2: 40–45. Google Scholar

77.

Popov N. P. 1916: Labiatae [AjugaMarrubium]. – In: Kuznetsov N., Busch N. & Fomin A. (ed.), Flora caucasica critica. Materialy dlya flory Kavkaza. Kriticheskoe sistematicheskoe-geograficheskoe izsledovanie 4(3) (No. 43, 45). – Yur'ev: Tipographiya K. Mattisena. Google Scholar

78.

Ranjbar M. & Mahmoudi, C. 2018; A taxonomic revision of Scutellaria sect. Lupulinaria subsect. Lupulinaria (Lamiaceae) in Iran. – Feddes Repert. 128: 63–101.  CrossrefGoogle Scholar

79.

Raunkiaer C. 1934: The life forms of plants and statistical plant geography: being the collected papers of C. Raunkiaer. – Oxford: The Clarendon Press. Google Scholar

80.

Rechinger K. H. 1982: Scutellaria L. – Pp. 48–84 in: Rechinger K. H. (ed.), Flora iranica 150. – Graz: Akademische Druck- und Verlagsanstalt. Google Scholar

81.

Ronquist F., Teslenko M., Mark P. van der, Ayres D. L., Darling A., Höhna S., Larget B., Liu L., Suchard M. A. & Huelsenbeck J. P. 2012: MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. – Syst. Biol. 61: 539–542.  CrossrefGoogle Scholar

82.

Safikhani K., Jamzad Z. & Saeidi H. 2017: A taxonomic revision of Scutellaria multicaulis (Lamiaceae) species complex in Iran. – Iran. J. Bot. 23: 10–24. Google Scholar

83.

Safikhani K., Jamzad Z. & Saeidi H. 2018: Phylogenetic relationships in Iranian Scutellaria (Lamiaceae) based on nuclear ribosomal ITS and chloroplast trnL-F DNA data. – Pl. Syst. Evol. 304: 1077–1089.  CrossrefGoogle Scholar

84.

Salmaki Y. & Müller J. 2019: Rediscovery of the enigmatic Scutellaria xylorrhiza (Scutellarioideae; Lamiaceae)—a rare endemic species from Iran. – Phytotaxa 394: 267–275.  CrossrefGoogle Scholar

85.

Schäferhoff B., Fleischmann A., Fischer E., Albach D. C., Borsch T., Heubl G. & Müller K. F. 2010: Towards resolving Lamiales relationships: insights from rapidly evolving chloroplast sequences. – B.M.C. Evol. Biol. 10: 352.  CrossrefGoogle Scholar

86.

Seyedipour S., Salmaki Y. & Xiang C.-L. 2020: Molecular phylogeny of Scutellaria (Lamiaceae; Scutellarioideae) in Iranian highlands inferred from nrITS and trnL-F sequences. – Progr. Biol. Sci. 7: 169–181.  CrossrefGoogle Scholar

87.

Shan Y.-Y., Pei X.-Y., Yong S.-Y., Li J.-L., Qin Q.-L-, Zeng S.-Y. & Yu J. 2021: Analysis of the complete chloroplast genomes of Scutellaria tsinyunensis and Scutellaria tubifera (Lamiaceae). – Mitochondrial DNA Part B 6: 2672–2680.  CrossrefGoogle Scholar

88.

Shaw J., Lickey E. B., Beck J. T., Farmer S. B., Liu W.-S., Miller J., Siripun K. C., Winder C. T., Schilling E. E. & Small R. L. 2005: The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. – Amer. J. Bot. 92: 142–166.  CrossrefGoogle Scholar

89.

Silakadze N., Kilian N., Korotkova N., Mosulishvili M. & Borsch T. 2019: Multiple evolutionary origins of high mountain bellflowers with solitary flowers and calyx scales render a core Caucasian clade of the Scapiflorae group (Campanulaceae). – Syst. Biodivers. 17: 690–711.  CrossrefGoogle Scholar

90.

Simmons M. P. & Ochoterena H. 2000: Gaps as characters in sequence-based phylogenetic analyses. – Syst. Biol. 49: 369–381.  CrossrefGoogle Scholar

91.

Stamatakis A. 2014: RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. – Bioinformatics 30: 1312–1313.  CrossrefGoogle Scholar

92.

Stöver B. C. & Müller K. F. 2010: TreeGraph 2: combining and visualizing evidence from different phylogenetic analyses. – B.M.C. Bioinf. 11: 7.  CrossrefGoogle Scholar

93.

Swofford D. L. 2002: PAUP* Phylogenetic Analysis Using Parsimony (*and other methods). Version 4.0b10. – Sunderland: Sinauer Associates. Google Scholar

94.

Szukala A., Korotkova N., Gruenstaeudl M., Sennikov A. N., Lazkov G. A., Litvinskaya S. A., Gabrielian E., Borsch T. & Raab-Straube E. von 2019: Phylogeny of the Eurasian genus Jurinea (Asteraceae: Cardueae): support for a monophyletic genus concept and a first hypothesis on overall species relationships. – Taxon 68: 112–131.  CrossrefGoogle Scholar

95.

Taberlet P., Gielly L., Pautou G. & Bouvet J. 1991: Universal primers for amplification of three non-coding regions of chloroplast DNA. – Pl. Molec. Biol. 17: 1105–1109.  CrossrefGoogle Scholar

96.

Ter-Chatschaturova S. Ya. 1967: Scutellaria L. – Pp. 308–322 in: Grossheim A. A., Flora Kavkaza, ed. 2, 7. – Leningrad: Nauka. Google Scholar

97.

Volkova P., Laczkó L., Demina O., Schanzer I. & Sramkó G. 2020: Out of Colchis: the colonization of Europe by Primula vulgaris Huds. (Primulaceae). – Acta Soc. Bot. Poloniae 89( 3 ): 89313.  CrossrefGoogle Scholar

98.

Wagstaff S. J. & Olmstead R. G. 1997: Phylogeny of Labiatae and Verbenaceae inferred from rbcL sequences. – Syst. Bot. 22: 165–179.  CrossrefGoogle Scholar

99.

Wang X.-G., Hai M.-R. & Liu T. 2012: Analysis of trnL-F sequences of the medicinal plants Scutellaria baicalensis and its adulterants. – Chin. Agric. Sci. Bull. 28: 173–177. Google Scholar

100.

Weberling F. 1989: Morphology of flowers and inflorescences. – Cambridge: Cambridge University Press. Google Scholar

101.

Wicke S. & Quandt D. 2009: Universal primers for the amplification of the plastid trnK/matK region in land plants. – Anales Jard. Bot. Madrid 66: 285–288.  CrossrefGoogle Scholar

102.

Worberg A., Quandt D., Barniske A.-M., Löhne C., Hilu K. W. & Borsch T. 2007: Phylogeny of basal eudicots: insights from non-coding and rapidly evolving DNA. – Org. Divers. Evol. 7: 55–77. Crossref. Google Scholar

103.

World Flora Online 2019: World Flora Online. – Published at  http://www.worldfloraonline.org/ [accessed 10 Nov 2019]. Google Scholar

104.

Wu C. Y. & Li H. W. 1977: Scutellaria. – Pp. 124–248 in: Flora Reipublicae Popularis Sinicae 65(2). – Beijing: Science Press. Google Scholar

105.

Zarucchi J. L. & Harley R. M. 1993: Lamiaceae. – Pp. 579–590 in: Zarucchi J. L. & Brako L. (ed.), Catalogue of the flowering plants and gymnosperms of Peru. – Monogr. Syst. Bot. Missouri Bot. Gard. 45. – St. Louis: Missouri Botanical Garden Press. Google Scholar

106.

Zhao F., Li B., Drew B. T., Chen Y.-P., Wang Q., Yu W.-B., Liu E.-D., Salmaki Y., Peng H. & Xiang C.-L. 2020: Leveraging plastomes for comparative analysis and phylogenomic inference within Scutellarioideae (Lamiaceae). – PLoS ONE 15( 5 ): e0232602.  CrossrefGoogle Scholar

107.

Zhao F., Liu E.-D., Peng H. & Xiang C.-L. 2017: A new species of Scutellaria (Scutellarioideae, Lamiaceae) from Sichuan Province in southwest China. – Peer J. 5: e3624.  CrossrefGoogle Scholar

Appendices

Appendix 1.

Voucher information for phylogenetic analyses and ENA/GenBank accession numbers. The first bloc lists specimens from which sequences were newly generated in this investigation and incudes the respective DNA bank (DB) numbers as well as the DNA lab code (Scu). The second bloc includes sequences of Scutellaria that were downloaded from ENA/GenBank. In most cases a voucher specimen could be found in the respective publication and is indicated in the voucher column. The corresponding publications are 1Wang & al. (2012), 2Jiang & al. (2017), 3Safikhani & al. (2017), 4Safikhani & al. (2018), 5Chen & Zhang (2019), 6Seyedipour & al. (2020), 7Zhao & al. (2020) and 8Shan & al. (2021). In some cases, mostly of complete plastid genomes, only an entry in ENA/GenBank could be found without actual publication and voucher citation. These were aSong & Park (2015), bZhang & Handy (2016), cCai (2020), dHuang (2020), eLi & al. (2020) and fZhang (2020).

img-z28-1_395.gif

Continued

img-z29-4_395.gif

Continued

img-z30-1_395.gif

Continued

img-z31-1_395.gif

Continued

img-z32-1_395.gif

Continued

img-z33-1_395.gif
© 2021 The Authors ·
Rashad A. Salimov, Gerald Parolly, and Thomas Borsch "Overall phylogenetic relationships of Scutellaria (Lamiaceae) shed light on the origin of the predominantly Caucasian and Irano-Turanian S. orientalis group," Willdenowia 51(3), 395-427, (23 December 2021). https://doi.org/10.3372/wi.51.51307
Received: 30 October 2021; Accepted: 26 November 2021; Published: 23 December 2021
KEYWORDS
Caucasus ecoregion
character evolution
Infrageneric classification
Irano-Turanian floristic region
Labiatae
Lamiaceae
molecular phylogeny
Back to Top