An elongated COI fragment to discriminate botryllid species and as an improved ascidian DNA barcode

Abstract

Botryllids are colonial ascidians widely studied for their potential invasiveness and as model organisms, however the morphological description and discrimination of these species is very problematic, leading to frequent specimen misidentifications. To facilitate species discrimination and detection of cryptic/new species, we developed new barcoding primers for the amplification of a COI fragment of about 860 bp (860-COI), which is an extension of the common Folmer's barcode region. Our 860-COI was successfully amplified in 177 worldwide-sampled botryllid colonies. Combined with morphological analyses, 860-COI allowed not only discriminating known species, but also identifying undescribed and cryptic species, resurrecting old species currently in synonymy, and proposing the assignment of clade D of the model organism Botryllus schlosseri to Botryllus renierii. Importantly, within clade A of B. schlosseri, 860-COI recognized at least two candidate species against only one recognized by the Folmer's fragment, underlining the need of further genetic investigations on this clade. This result also suggests that the 860-COI could have a greater ability to diagnose cryptic/new species than the Folmer's fragment at very short evolutionary distances, such as those observed within clade A. Finally, our new primers simplify the amplification of 860-COI even in non-botryllid ascidians, suggesting their wider usefulness in ascidians.

Figure 1

Mapping and description of the primers for the 860-COI fragment. (a) Primer mapping on the COI of B. schlosseri. Red: primers for the first PCR; blue: nested primers for the second (reamplification) PCR; in brackets: position of the primer 3′ end on the complete COI sequence of B. schlosseri (1548 bp, Acc. number: FM177702). (b) Forward primers aligned to the ascidian COI. Reference2013: WebLogo plot of the dataset of 18 complete COI sequences of ascidians used for primer design; Total: WebLogo plot of the dataset of 33 complete COI sequences of ascidians currently available; Total-cons100: 100% amino acid consensus sequence from the Total dataset; yellow: positions with differences to the LCO1490 primer; red amino acids: positions with a 100% conservation; hnp: the hydrophobic non-polar amino acids Met, Leu, Phe, or Val; arrow head: positions differing between the two logos. Sequence datasets are reported in Supplementary Table S2. Primer references: Cox_URO_LCO and Pyura_LCO2012 are described in Rubinstein; COIu27Asc and COIu25Asc in Monniot; Tun_Forward in Stefaniak; didF in da Silva; jglco1490 in Geller. (c) Our reverse primers aligned to the ascidian COI. Rc: primer in reverse-complement orientation. All other symbols are as in b). Differences from the amino acid consensus sequence are as below: 1 = Asn in Lissoclinum patella (KJ596323), with A in the 1st codon position; 2 = Thr in Diplosoma listerianum (FN313539) and Ile in 6 species of Aplousobranchia and Styelidae; 3 = Ala in two Phallusia species; Thr in Lissoclinum patella (KJ596323), Diplosoma listerianum (FN313539) and Didemnum vexillum (KM259616); 4 = Ser in Lissoclinum patella (KJ596323); 5 = Ser in three Ciona species.

Figure 2

Bayesian majority rule consensus tree reconstructed from the "Elongated-856nt" alignment. Except for clade A, support values are reported close to the nodes as BPP/ML bootstrap percentage, only if BPP ≥ 0.90 or ML bootstrap ≥ 70%. Only for clade A, node supports are shown in Table 4, and the result of species delimitation analyses are reported as vertical bar (white for the ABGD initial partitions; black for bPTP). Yellow diamonds: OTUs identified outside of clade A by the initial partitions of all ABGD analyses (see “Methods”) and by bPTP; red dots: main sub-clades (A1, A2, A2b and A3) within clade A, described in Table 4; underlined names: genus Botryllus. The specimens of B. schlosseri clade A used as neotype or for which the entire mitochondrial or nuclear genomes were sequenced^, are also indicated.

Figure 3

Bayesian majority rule consensus tree reconstructed from the "Folmer-524nt" alignment. Except for clade A, support values are reported close to the nodes as BPP/ML bootstrap percentage, only if BPP ≥ 0.90 or ML bootstrap ≥ 70%. Only for clade A, node supports are shown in Table 4. Yellow diamonds: OTUs identified by the initial partitions of all ABGD analyses (see “Methods”) and by bPTP; red dots: main sub-clades (A1, A2, A2b and A3) within clade A, described in Table 4; underlined names: genus Botryllus. The specimens of B. schlosseri clade A as neotype or for which the entire mitochondrial or nuclear genomes were sequenced^, are also indicated.

Figure 4

Distribution of the former clades of B. schlosseri species complex in the Mediterranean Sea, and presence of the rare clades along the European Atlantic coasts. Red: this study, with localities and dates of our samplings detailed in Supplementary Table S1; blue: data from literature^,,,,. Notes: in Reem et al. clade E is reported as clade IV, while clade A includes the clades I, II, and III; localities Fornelos and Ferrol cited in Lopez-Legentil and Pérez-Portela, respectively, are here reported as "Ferrol Estuary", according to the data provided by the authors (Turon X, personal communication). Indeed, Fornelos is one of the marinas in the city of Ferrol, inside the Ría de Ferrol, while Ferrol is a natural environment at the entrance of the estuary. The map was created with ArcView GIS 3.2 (https://www.esri.com/).