4.2. Comparing protocols and paving the road for comprehensive single-specimen systematics
Besides providing the first multi-locus, vouchered phylogeny estimate for Facetotecta, our preliminary phylogeny of this group (Fig. 1) demonstrates the advantages and disadvantages of the three protocols that we outlined above. Protocol 1 outperformed all previous attempts to study the molecular phylogenetics and evolution of Facetotecta. First of all, it retained the information obtained from living specimens (y-nauplii, y-cyprid) throughout their larval development and also retained voucher exuviae of single specimens after their DNA extraction. Secondly (Fig. 1) it allowed for consistent amplification of long (>1500bp) and short (<400bp) DNA templates of protein-coding and ribosomal loci of both the mitochondrial and nuclear genomes (Table 1, 2, 3, S2). Protocol 1 thus ensured maximal gathering of morphological and molecular information from individual specimens and allowed the scoring both y-naupliar and y-cyprid characters. Given the remarkably high diversity at local scales such Sesoko Island (Glenner et al., 2008; Olesen et al., 2022; Dreyer et al., in press), this protocol will surely become an important asset to understanding Facetotecta diversity. It has already proven invaluable for species descriptions when dealing with sympatric distributions of multiple species (Olesen et al., 2022). The last-stage nauplius (LSN) is an unambiguously homologous and easily recognizable stage, of which there is never more than one for a given taxon (Olesen et al., 2022). The exuviae of the LSN (shed during metamorphosis to the y-cyprid), can serve as either a primary or complementary voucher specimen whether it is mounted on a glass slide or fixed and preserved in a liquid storage medium.
For example, the integrative description of Hansenocaris demodexwas based on a series of nearly identical specimens that were first reared and live-imaged and subsequently amplified and sequenced individually with a short 18S fragment (Olesen et al., 2022). To this was added crucial details of the y-cyprid as observed by both LM and SEM, the latter admittedly not relying on sequenced material, but on specimens confidently assigned to the species due to similarities of their preceding nauplii (in particular the LSN) to those of sequenced individuals. In future work, it is not unlikely that several data layers, including live images, LM images of voucher exuviae (of y-nauplii), SEM images (of y-cyprids and some LSN voucher exuviae, either naturally molted or remaining after DNA extraction), and DNA sequence data, will be harvested from a single specimen (Grygier et al., 2019). We recommend mounting naupliar exuviae in glycerin jelly as it solidifies at room temperature, thus easing maintenance in museum collections and tropical research facilities. Some LSN exuvial specimens may also be imaged by confocal laser-scanning microscopy (CLSM), although we have not attempted this yet for DNA-extracted specimens (for information on mounting exuviae for SEM, see Grygier et al., 2019; Olesen et al., 2022; Kolbasov et al., 2022).
Protocol 2 lumps together a series of culturing and sequencing strategies, all of which retain a less informative and less complete assortment of visual information than Protocol 1. It is typically applied to y-larvae with planktotrophic nauplii, such as y-naupliar Types A* and AE* and the planktotrophic y-nauplii from the Azores and the White Sea, as these cannot yet be reared in the lab. The morphological information is sufficient for screening morphotype affinities established by molecular data but, in our view, does not provide a comprehensive basis for species descriptions. Namely, different instars in the life cycle (e.g., y-nauplii and -cyprids) cannot easily be linked together, especially at localities Sesoko Island with many sympatric morphotypes. We also demonstrate that protocol 2 works on other Thecostraca larvae, e.g., an ascothoracid larva identifed as Baccalaureus sp. nests as a sister species to B. maldivensis (Fig. 1). Thus, we anticipate that procols 1 and 2 should work on other marine invertebrate larvae that molts and leaves exuviae.
Protocol 3, which involves the use of non-cultured y-larvae and has been used for all specimens with sequences currently deposited in GenBank, except for those of Olesen et al., (2022). It does not produce any images, and thus no vouchered data, and is therefore impractical for all levels of y-larva phylogenetics. At best, it retains anecdotal information about the specimens obtained prior to sequencing. Despite these limitations, the present work has allowed us to determining the identity of a least some of the y-larval material (“Facetotecta sp. 1-6”) used by Pérez-Losada et al., (2009), as some matches to the vouchered material sequenced herein from the same locality (Sesoko Island) are apparent. For example, their “Facetotecta spp. 1, 2, and 6” cluster within our Type A* and “Facetotecta sp. 4” is molecularly identical to Type D*. On the other hand, “Facetotecta sp. 5” and the Gallego et al’s (2012, 2015) two sequences from Antarctica are still ‘ghost’ sequences and remains unlinked to known, imaged Facetotecta. Although these specimens are unvouchered and lack any kind of published morphological information, they have played a critical role at higher taxonomic levels, helping to place Facetotecta systematically within Thecostraca and Pancrustacea (Pérez-Losada et al., 2002, 2009; Petrunina et al., 2014).