The Genetic and Cytogenetic Relationships among Subgenera of Chrysolina Motschulsky, 1860 and Oreina Chevrolat, 1837 (Coleoptera: Chrysomelidae: Chrysomelinae)

Introduction

A recent insightful and detailed taxonomic review of the genus Chrysolina Motschulsky, 1860, based on morphological and male genitalia adult characters [1], has prompted us to set up the possible genetic interrelationships among many of the described subgenera, taking into account mitochondrial gene sequences [2, 3, 4, 5], and both mitochondrial and nuclear DNA [6], together with chromosomal studies [7, 8, 9, 10, 11, 12, 13, 14, 15], and host-plant affiliations because most species of Chrysolina and its near genus Oreina, are highly specific feeders on one or less often two botanic families [16, 17, 18].

Material and Methods

A total of 84 species of Chrysolina of 43 subgenera and 19 species of Oreina of the seven described subgenera have been genetically analyzed, either by DNA sequences and/ or by karyotypes. Three different molecular markers were selected for building a phylogenetic tree: a partial sequence of the mitochondrial 16S rDNA (rrnL), a partial sequence of the mitochondrial cytochrome c oxidase subunit 1 gene (cox1), and a fragment from the nuclear histone 3 gene (nuH3) [6]. We have also used the published mitochondrial phylogenetic trees based on the small subunit ribosomal RNA (12S) and the large subunit ribosomal RNA (16S) [5], that on the large subunit ribosomal RNA (16S) and the cytochrome c oxidase subunit 1 gene (cox1) [3], and that on the large subunit ribosomal RNA (16S) only [4]. The chromosomal analyses in meiotic and mitotic testis cells, by classical methods, have given chromosome numbers, male sex-chromosome systems, and in some cases detailed karyotypes of Chrysolina and Oreina species, reported in multiple publications [6, 7, 10]. In addition, host-plant affiliations to different botanic families are also considered as worth data for phylogenetic trees. Bayesian, Neighbour-joining (NJ), Maximum Likelihood (ML) and/or Maximum Parsimony (PAUP) trees have been implemented in the different published cladograms [2, 3, 4, 5, 6].

Results

The genetic findings, DNA sequences and/or karyotypes, on 84 species of Chrysolina among a total number of almost 500, belonging to 43 subgenera, 61% of the whole 70 described ones [1], have been used for our review. The DNA data of Chrysolina are compiled in 63 species and the chromosomes in 69 species, both corresponding to 35 subgenera, and in the closely related genus Oreina, 19 species of the seven described subgenera, have been analyzed by DNA sequences [3, 4, 5, 6] and/or chromosomes [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25]. Moreover, the findings on plant affiliations in these species of Chrysolina and Oreina are also taken from several authors [26, 27, 28, 29, 30, 31]. Therefore, we believe that this sample is enough representative for getting reliable genetic interrelationships among these subgenera of Chrysolina and Oreina.

Discussion and Conclusions

The Bayesian and Maximum Likelihood phylogenetic trees obtained from the combined analysis of mtDNA cox1 and rrnl genes and nuDNA H3 gene, have shown three distinctive most plesiomorphous clades of species feeding on Lamiaceae plants (P=0.98 [6]: 1) that including Ch. (Taeniochrysa) americana and Ch. (Chrysolinopsis) gemina, 2) another with three species of the subgenus Synerga Weise, 1900, Ch. coerulans, according with Hsiao and Pasteels [5], Ch. herbacea (Duftschmid, 1825) and Ch. viridana (Küster, 1844), two species of the subgenus Melasomoptera Bechyné, et al. [32], Ch. grossa (Fabricius, 1792), according with Hsiao and Pasteels [5], and Ch. lucida (Olivier, 1807), one of the subgenus Chrysomorpha Motschulsky, et al. [33], Ch. cerealis cyaneoaurata (Motschulsky, 1860) and another of the subgenus Euchrysolina Bechyné [32], Ch. graminis (Linnaeus, 1758), this latter species according with Hsiao and Pasteels [5], 3) a third large clade made of three taxa of the subgenus Maenadochrysa Bechyné [32] Ch. affinis baetica (Suffrian, 1851), Ch. affinis mesatlantica (Kocher, 1958) and Ch. femoralis (Olivier, 1790), close to them the monotypic subgenus Erythrochrysa Bechyné [32], with Ch. polita (Linnaeus, 1758), the Asian subgenus Chrysocrosita Bechyné [32], with Ch. jakowlewi (Weise, 1894), and also in the same clade but in an intermediate position between the subgenus Maenadochrysa and those of these two latter taxa, Ch. (Colaphosoma) sturmi (Westhoff, 1882), the unique species of this last subgenera, plus a doubtful and misleading subclade composed of Ch. (Centoptera) bicolor (Fabricius, 1775), Ch. (Fasta) fastuosa (Scopoli, 1763). All species of these three clades, except Ch. sturmi which has not been cytogenetically studied yet, display 2n=24(Xyp) chromosomes and similar karyotypes [11, 19, 20, 23], and the nine karyotypes which have been carefully analyzed, show meta- or submetacentric chromosomes and among them a satellited first autosome [7, 23], that could be a marker chromosome for all these group of species, and bearing the ribosomal DNA locus, very likely. Two further species feeding also on Lamiaceae, that were only chromosomally analyzed but not by DNAs, and sharing also 2n=24(Xyp), Ch. (Bittotaenia) salviae (Germar, 1824) and Ch. (Rhyssoloma) fragariae (Wollaston, 1824) (7,43,44), should probably join to any one of these three previous clades of Chrysolina.

In the Maximum Likehood phylogenetic tree obtained by a combined analysis of the three sequenced genes [6], and close to the latter subclade of Ch. bicolor and Ch. fastuosa, there is another one (P=0.527 [6], made of five species of Oreina, of two different subgenera, Chrysochloa Hope, 1840 and Oreina s.str., all of them but one sharing karyotypes with 2n=24(Xyp) chromosomes [8, 21], but with an additional species of Chrysolina, Ch. (Timarchoptera) haemochlora (Gebler, 1823). This last species displays a highly distinct karyotype of 2n(♂)=27(XO) chromosomes, strikingly diverse in size, with two autosome pairs threefold larger than any of the remainders and lacking the Y-chromosome [11]. The feeding affiliations of the species in this subclade are very heterogeneous, Ch. haemochlora and two species of Oreina on Apiaceae, and the other three species of Oreina on Asteraceae, contrary to all the above mentioned species on Lamiaceae plants [5, 6, 29]. Therefore, the presence in this subclade of at least one species of Chrysolina with five ones of Oreina has induced us to apply an Approximately Unbiased test [34] to see whether Oreina is a different genus from Chrysolina, and the AU test has given a statistically significant value of 0.016, P < 0.05, rejecting this hypothesis [6], so they might be joined, although most recent authors still keep them separated [1, 18, 35, 36].

Another new and very heterogeneous main clade, the fourth one, includes subgenera affiliated to all the nine plant families so far recorded in Chrysolina. Those on Apiaceae plants encloses the three analyzed species in the subgenus Threnosoma Motschulsky, 1860, Ch. helopioides (Suffrian, 1851), Ch. obscurella (Suffrian, 1851) and Ch. timarchoides, one of the subgenus Crositops Marseul, 1883, Ch. pedestris, one also in the subgenus Timarchoptera Motschulsky, 1860, Ch. soiota [6], and three in the subgenus Minckia Strand, 1835 (= Sulcicollis J. Sahlberg, 1913), Ch. oricalcia (O.F. Müller, 1776), Ch. peregrina (Herrich-Schäffer, 1831) and Ch. rufoaenea (Suffrian, 1851) [5, 6], all of them, out of maybe Ch. soiota, feeding on this plant family [6, 16, 17, 29]. Their chromosome numbers and sex-chromosome systems are different but all showing high values, thereby the three quoted species of Threnosoma share 2n(♂ )= 47(XO), those of Minckia, Ch. oricalcia 2n=40(Xyp), Ch. peregrina 2n=46(Xyp) and Ch. rufoanenea 2n=50(Xyp), the highest updated number in the Chrysolina and the subfamily Chrysomelinae, plus the unique species of the subgenus Cyrtochrysolina Kippenberg (2012), Ch. marcasitica (Germar, 1824) 2n=42(Xyp) [8, 11, 21, 23], and Ch. pedestris with 2n(♂)= 47(XO), as those of the subgenus Threnosoma, though having two large ring-shaped autosome bivalents, very remarkable among the remaining medium or small ones, and a tiny X-chromosome. Mikhailov [38], and recently also Bienkowski [1], synonymyzed the subgenera Timarchoptera Motschulsky (1860) and Paraheliostola L. Medvedev, et al. [38], but the molecular phylogeny of their respective typical species, Ch. haemochlora and Ch. soiota, does not support this view, because when we applied the AU Shimodaira test, the hypothesis of monophyly for both species was statistically rejected [6]. Therefore, Paraheliostola L. Medvedev (1992) should be stat. ressur. as a good new subgenus.

Two further subgenera, Hypericia Bedel, 1899 and the monotypic Sphaeromela Bedel, 1899, the first with four DNA sampled species, Ch. brunsvicensis (Gravenhorst, 1807), Ch. geminata (Paykull, 1799), Ch. hyperici (Forster, 1771) and Ch. quadrigemina (Suffrian, 1851), and the second with Ch. varians (Schaller, 1783) [3, 4, 5, 6, 16, 17], constitute a well-defined subclade. All these species are feeding on Hypericaceae plants, a botanic affiliation which has been found in no other species of the remaining subgenera [16, 17], out of the Moroccan Ch. seriatopora, of the monotypic subgenus Atlasiana Bourdonné et Doguet, 1991 [17], which has not been genetically studied yet. The chromosomes of the three checked species of Hypericia share 2n=38(Xyp) elements, mostly metacentrics of medium and small size [8, 21], while Ch. varians displays 2n=32(Xyp) [8, 24].

The five updated DNA and/or cytogenetically sampled species of the subgenus Chalcoidea Motschulsky, 1860, Ch. analis (Linnaeus, 1767), Ch. carnifex (Fabricius, 1792), Ch. interstincta (Suffrian, 1851), Ch. janbechynei Cobos, 1953 (= Ch. curvilinea Weise, 1884) and Ch. marginata (Linnaeus, 1758) belong to a clearly supported subclade [3, 5, 6], with three more species of two subgenera, Ch. (Pezocrosita) convexicollis (Jacobson, 1901), very close to Ch. carnifex, and two a bit more distant to the Chalcoidea species, Ch. (Anopachys) aurichalcea (Gebler, 1825) and Ch. (Anopachys) eurina (Frivaldszky, 1883). All these species feed on Asteraceae plantsn [6, 16, 17], and at least the five species of Chalcoidea share 2n=40(Xyp) chromosomes [0, 12, 14, 19, 21], whereas the cytogenetic findings on Ch. aurichalcea, 2n(♂)=31(XO) to 2n(♂)=45(XO) chromosomes, have proved the existence of several cryptic species [21, 40]. Although none of the three chromosomally studied species of the subgenus Apterosoma Motschulsky, 1860 (= Caudatochrysa Bechyné, et al. [32] have been surveyed at the DNA level, their similar cytogenetic results of 2n(♂)=39(XO) and 2n=42(XY) chromosomes [13], and their trophic affiliation on Asteraceae (Jolivet and Petitpierre, 1976), point out to a probable close genetic resemblance, in agreement with the proposal by Bourdonné and Doguet [ 17] who joined the species of Chalcoidea, Anopachys, Apterosoma (sub Caudatochrysa) and Allochrysolina Bechyné, 1950, in his group 9. In relation with this last subgenus Allochrysolina, the molecular phylogeny of the two analyzed species, Ch. fuliginosa (Olivier, 1807) and Ch. lepida (Olivier, 1807) join them in a minor subclade within the large one of Anopachys and Chalcoidea subgenera [6], and also in accordance with an identical chromosome number of 2n=42(Xyp), although Ch. fuliginosa and Ch. lepida have distinguishable karyotypes [8]. In relation to the subgenus Pezocrosita, the DNA analysis of only one species, Ch. convexicollis, may not assume that the subgenus should be included in this latter clade, because it is a highly complex taxon made of 34 species in seven different groups [1], and three or four more species of Pezocrosita at least, should be genetically analyzed before coming to this conclusion.

Pasteels et al. [42] suggested a possible relationship between the subgenera Chalcoidea and Hypericia, based on their defensive toxins, and Hsiao and Pasteels [5] another between the subgenera Allochrysolina and Anopachys based on mtDNA phylogeny, host-plant affiliation and chemical defense, and both proposals have not been rejected by our AU tests (P= 0.066 and P=0.215, respectively [6]. Likewise, Bourdonné [46] proposed a monophyly for the subgenera Allochrysolina, Chalcoidea and Pezocrosita, based on adult morphology and host-plant affiliations, an hypothesis that could not be statistically rejected by our AU test too (P=0.205) [6]. Therefore, these three phyletic assertions can be possible.

The five checked species of the subgenus Chrysolina s. str. Motschulsky, 1860, Ch. bankii (Fabricius, 1775), Ch. costalis (Olivier, 1807, = Ch. obsoleta Brullé, 1838), Ch. rufa (Duftschmid, 1825), Ch.staphylaea (Linnaeus, 1758) and Ch. wollastoni Bechyné, 1957 (= Ch. rutilans Wollaston, 1864 nec Gravenhorst, 1807), constitute a highly supported subclade either by DNA molecular systematics [3, 4, 6] or by their 2n(♂)= 23(XO) chromosomes (19,21,23; Petitpierre et al., 1988; Petitpierre et al., 2004). Nevertheless, in another species, Ch. (Lithopteroides) exanthematica, which shares also 2n(♂)= 23(XO) chromosomes, we do not know whether it is related to this previous clade or not, because we should await to DNA analyses. Moreover, Bienkowski [1] placed the Canarian endemic species Ch. costalis and Ch. wollastoni within the subgenus Rhyssoloma Wollaston, 1854, whose typical species is the Madeira island endemic Ch. fragariae a unique species for this subgenus, according with Bechyné [32], Bienkowski [33] and Kippenberg [35], a viewpoint that we follow too because Ch. fragariae, apart from being a Madeiran endemic, has a very unusual longitudinal relief in elytra and it shows 2n=24(Xyp) chromosomes [7] instead of 2n(♂)= 23(XO) chromosomes of the above mentioned species, despite of the absence of DNA findings in Ch. fragariae that may give an additional stronger support to this view.

The subgenera Arctolina Kontkanen, 1959, with two genetically sampled species, Ch. oirota Lopatin, 1990, and Ch. poretzky, and Pleurosticha Motschulsky, [33], with three other ones, Ch. gebleri L. Medvedev, et al. [38] Ch. lagunovi Mikhailov, et al. [31] and Ch. sylvatica (Gebler, 1823), display a remarkable DNA resemblance, with 2n=26(Xyp) or 2n=26 (XYp) chromosomes, that may allow to a possible synonymisation of these two subgenera. However, Ch. oirota and Ch. poretzky seem to have a double feeding affiliation on Lamiaceae and Scrophulariaceae, whereas Ch. lagunovi Mikhailov, 2006 and Ch. sylvatica have only one on Lamiaceae and Ranunculaceae, respectively.

Two species belonging to the subgenus Allohypericia Bechyné, et al. [32], Ch aeruginosa (Faldermann, 1835) and Ch. perforata (Gebler, 1830), plus probably three more of the subgenus Colaphoptera Motschulsky, 1860, Ch. bigorrensis (Fairmaire, 1865), Ch. globosa (Panzer, 1805), and Ch. purpurascens (Germar, 1822), join to a sister subclade of the previous one of Arctolina and Pleurosticha subgenera. The chromosome number of Ch. bigorrensis (Fairmaire, 1865) and Ch. globosa (Panzer, 1805), 2n=26(Xyp), agrees with those of Ch. (Arctolina) poretzkyi (Jacobson, 1897), Ch. (Pleurosticha) gebleri L. Medvedev, 1979, Ch. (Pleurosticha) lagunovi and Ch. (Pleurosticha) sylvatica whereas that of Ch. purpurascens crassimargo (Germar, 1824), 2n=24(Xyp), is quite similar. Moreover, the host-plants of Ch. bigorrensis and Ch. globosa are Lamiaceae, but those of Ch. aeruginosa and Ch. perforata are Lamiaceae and Asteraceae, as well as those of Ch. oirota and Ch. Poretzky.

Another well-supported subclade (P=0.91, [6]), include sister minor subclades of species mostly feeding on Plantaginaceae plants, such as those of the subgenera Taeniosticha Motschulsky, 1860, Ch. petitpierrei Kippenberg, 2004 and Ch. reitteri (Weise, 1884), whose feedings are not yet known, Palaeosticta Bechyné, 1952, Ch. diluta (Germar, 1824) and Ch. kocheri (Codina Padilla, 1961), Colaphodes Motschulsky, 1860, Ch. haemoptera (Linnaeus, 1758), and Ovosoma Motschulsky, 1860, Ch. vernalis pyrenaica (Dufour, 1843) and probably Ch. susterai Bechyné, 1950, though this last species has only been studied chromosomally but not by DNA. The chromosomes of the species included in this large subclade are known for Ch. diluta, 2n=36(Xyp) [21] (Petitpierre, 1981), Ch. haemoptera, Ch. vernalis pyrenaica and Ch susterai, all of them with 2n=40(Xyp) [11, 20, 21, 44]. Another species that would be possibly included in this subclade, when the DNA sequences were become available, is Ch. (Zeugotaenia) limbata (Fabricius, 1775), because it feeds on Plantaginaceae [16, 17], it shows a similar chromosome number of 2n=42(Xyp) [8], and shares some morphological features with the subgenus Taeniosticha [37], sub genus Craspeda Motschulsky, 1860), and the subgenus Palaeosticta [37].

An additional highly evolved subclade is that of the subgenus Stichoptera Motschulsky, 1860, where eight species have been analyzed, either by their DNA sequences and/or by their chromosomes [8, 9, 12, 21, 23, 39], with a highly variable numbers from 2n=22(Xyp) of Ch. kuesteri (Helliessen, 1912) to 2n=34(Xyp) of Ch. sanguinolenta (Linnaeus, 1758), a modal value of 2n=24(Xyp) of Ch. colasi (Cobos, 1952), Ch. latecincta (Demaison, 1896) and Ch. oceanoripensis Bourdonné, Doguet and Petitpierre, [39] (= Ch. ripoceanesis auct.), and three species with high chromosome numbers, Ch. lucidicollis grossepunctata (Lindberg, 1950) with 2n=30(Xyp), and Ch. gypsophilae (Küster, 1845) and Ch. rossia (Illiger, 1802)

with 2n=32(Xyp). A distinctive cytogenetic trait of these eight karyotypes is their asymmetry of autosome sizes, with a few large autosome pairs and the remaining small pairs, being the number of these large autosome pairs inversely correlated with the chromosome numbers [9]. The trophic affiliation of these species of Stichoptera is quite eclectic, mainly on Scrophulariaceae, but at least two species only on Lamiaceae, and another one with a secondary choice on Plantaginaceae in adults [6, 39].

Finally, the subgenus Sibiriella L. Medvedev, 1999, has been cytogenetically studied by two species, Ch. montana (Gebler, 1848) and Ch. schewyrewi (Jacobson, 1895), with 2n=40(XYp) and 2n=40(Xyp) chromosomes, respectively [8, 11], This number of 2n=40 chromosomes and their feeding on Asteraceae plants are also found in species of the subgenus Chalcoidea which might place Sibiriella phyletically close to them. Likewise, the monotypic subgenus Sphaerochrysolina Kippenberg [35], with Ch. umbratilis (Weise, 1887), having 2n=30(XY) chromosomes [14], but without DNA findings, as in the previous subgenus, we could not place it in a molecular phylogenetic tree. However, Sphaerochrysolina resembles morphologically, among others, to the subgenera Chrysolina s.str., Colaphoptera and Cyrtochrysolina [1], but it is strikingly different by its karyotype of 2n=30 chromosomes and the chiasmatic XY sex-chromosome system, whereas Chrysolina s.str. Displays 2n (♂)=23(XO), Colaphoptera 2n=26(Xyp) and 2n=24(Xyp) y Cyrtochrysolina 2n=42(Xyp) chromosomes. Furthermore, the trophic affiliation of Sphaerochrysolina on Asteraceae plants [29], is clearly different from those of the above subgenera on Lamiaceae, Plantaginaceae, Ranunculaceae and Apiaceae [6].

Although Lamiaceae are the first plant-affiliation for the genus (P=0.98) [6] the species of Chrysolina-Oreina have extended their trophism to eight other botanic families [6, 15, 16, 17]. The different shifts from Lamiaceae to new host-plants families have not been unique events for each of them, except for Hypericaceae selected by the subgenera Hypericia and Sphaeromela [3, 6] namely, that from Lamiaceae to Apiaceae, has taken place in a clade grouping Oreina s.str. and Ch. (Timarchoptera) haemochlora (P=0.37) [6], in a subclade with the subgenera Minckia and Threnosoma, and Ch. (Timarchoptera) soiota (P=0.62) [6]. Also, that from Lamiaceae to Asteraceae in a subclade of Oreina subgenus Chrysochloa (P=0.54) [6] and in another with the subgenera Allochrysolina, Anopachys, Apterosoma, Chalcoidea and Pezocrosita (Ch. convexicollis) (P=0.94) [46]. In agreement with this last subclade, Bourdonné [46] has put within his “genus” Chalcoidea Motschulsky, the subgenera Allochrysolina, Chalcoidea s.str. and Pezocrosita. It is interesting to note that some subgenera subclades, such as that of Stichoptera, shows a double trophic shift from possibly Plantaginaceae to Scrophulariaceae and Lamiaceae (P=0.66) [6], while others such as Allohypericia, Arctolina and Pleurosticha display feedings in three plant families, on Asteraceae, Lamiaceae and Ranunculaceae [15, 30, 31], probably also from a Plantaginaceae ancestor, in reversal shifts (P=0.5 in these two cases) [6]. The species of subgenus Chrysolina s. str. are paradigmatic in this sense because some of them, as Ch. bankii, can feed on three plant families at least, Asteraceae, Lamiaceae and Plantaginaceae, and Ch. staphylaea on two families, Lamiaceae and Ranunculaceae [16]. In addition, Ch. (Euchrysolina) graminis has a double feeding too, on Asteraceae and Lamiaceae [16, 17], as well as the two sampled species of the subgenus Allohypericia, Ch. aeruginosa and Ch. perforata, quoted before. In all these examples of a relative polyphagy, there are no new plant-affiliations, because these selected botanic families are always chosen by other congeneric species. Thus, most species of Chrysolina-Oreina have possibly maintained a common potential genetic basis in their genomes for plant affiliations in these botanic families.