J. Phycol. 53, 790–803 (2017) © 2017 Phycological Society of America DOI: 10.1111/jpy.12540 PHYLOGENETIC POSITION OF THE CORAL SYMBIONT OSTREOBIUM (ULVOPHYCEAE) INFERRED FROM CHLOROPLAST GENOME DATA1 Heroen Verbruggen, Vanessa R. Marcelino School of BioSciences, University of Melbourne, Melbourne, Victoria 3010, Australia Michael D. Guiry AlgaeBase, Ryan Institute, National University of Ireland, Galway H91 TK33, Ireland Ma. Chiela M. Cremen, and Christopher J. Jackson 2 School of BioSciences, University of Melbourne, Melbourne, Victoria 3010, Australia The green algal genus Ostreobium is an important Abbreviations: bp, base pairs; kbp, kilo base pairs; symbiont of corals, playing roles in reef LBA, long branch attraction; Mb, mega base pairs; decalcification and providing photosynthates to the ML, maximum likelihood; ORF, open reading coral during bleaching events. A chloroplast genome frame of a cultured strain of Ostreobium was available, but low taxon sampling and Ostreobium’s early-branching nature left doubt about its phylogenetic position. Ostreobium is an early-branching genus in the Here, we generate and describe chloroplast genomes siphonous green algae (order Bryopsidales), a group from four Ostreobium strains as well as Avrainvillea of seaweeds composed of a single giant multinucle- mazei and Neomeris sp., strategically sampled early- ate cell (Vroom and Smith 2003). The genus is unu- branching lineages in the Bryopsidales and sual among the Bryopsidales in being microscopic Dasycladales respectively. At 80,584 bp, the chloroplast and endolithic, usually burrowing into marine lime- genome of Ostreobium sp. HV05042 is the most stone substrates. It is best known living in association compact yet found in the Ulvophyceae. The with corals, forming a green layer in the skeleton Avrainvillea chloroplast genome is ~94 kbp and underneath the living coral tissue (Verbruggen and contains introns in infA and cysT that have nearly Tribollet 2011). Ostreobium plays a major role in car- complete sequence identity except for an open bonate dissolution of live and dead corals, eroding reading frame (ORF) in infA that is not present in as much as a kilogram of reef carbonate per square cysT. In line with other bryopsidalean species, it also meter per year (Tribollet 2008, Grange et al. 2015). contains regions with possibly bacteria-derived ORFs. It can also confer benefits to the coral hosts by pro- The Neomeris data did not assemble into a canonical viding them with photosynthates and protecting circular chloroplast genome but a large number of them from high-light stress (Schlichter et al. 1995, contigs containing fragments of chloroplast genes Inoue et al. 1997, Fine and Loya 2002). and showing evidence of long introns and intergenic Due to the importance of their ecological roles in regions, and the Neomeris chloroplast genome size reef ecosystems, endolithic algae have been studied was estimated to exceed 1.87 Mb. Chloroplast quite intensively in recent years. One of the key phylogenomics and 18S nrDNA data showed strong findings of recent work is that the Ostreobium lin- support for the Ostreobium lineage being sister to the eage, previously thought to be composed of only a remaining Bryopsidales. There were differences in handful of species, is in fact a complex of multiple branch support when outgroups were varied, but the undescribed lineages (Gutner-Hoch and Fine 2011, overall support for the placement of Ostreobium was del Campo et al. 2016, Marcelino and Verbruggen strong. These results permitted us to validate two 2016, Sauvage et al. 2016). suborders and introduce a third, the Ostreobineae. An earlier phylogenetic study of the siphonous Key index words: green algae suggested that Ostreobium was sister to Bryopsidales; chloroplast genome; the two main suborders of the Bryopsidales: Bryo- coral symbiosis; Dasycladales; molecular phylogenet- Ostreobium psidineae and Halimedineae (Verbruggen et al. ics; ; phylogenomics 2009). This study, however, was based on only five genes, of which only three were sampled for Ostreo- bium, and maximum likelihood bootstrap support 1Received 1 November 2016. Accepted 13 December 2016. First for its position was low (66% for the Bryopsidineae Published Online 10 April 2017. Published Online 12 May 2017, + Halimedineae grouping). Adding more data can Wiley Online Library (wileyonlinelibrary.com). 2Author for correspondence: e-mail [email protected]. help to resolve such phylogenetic uncertainties, with Editorial Responsibility: L. Graham (Associate Editor) both gene and taxon sampling playing important 790 OSTREOBIUM CHLOROPLAST PHYLOGENOMICS 791 roles. Furthermore, it has been demonstrated that sequences for Bryopsidales that were included in the datasets phylogenomic analysis with only a few taxa can yield are listed in Table 1. high support values for the wrong tree topology, Choosing an appropriate outgroup was difficult because especially in cases where systematic bias is an issue depending on taxon and gene sampling, different groups have been inferred to be the sister group of the Bryopsidales, (Hedtke et al. 2006). With the knowledge that usually with low support. To add to the problem, the class Ostreobium is a complex of multiple lineages, we can Ulvophyceae (to which the Bryopsidales belong) was non- now take advantage of that diversity by sequencing monophyletic in chloroplast genome phylogenies (Fucıkova chloroplast genomes from multiple Ostreobium spe- et al. 2014, Leliaert and Lopez-Bautista 2015, Sun et al. 2016, cies to reduce phylogenetic artifacts and more accu- Turmel et al. 2016) while nuclear data suggested monophyly rately infer the position of the genus among the (Cocquyt et al. 2010). We have chosen to use four outgroup lineages (Table 1) and have run analyses on various combina- siphonous green algae. tions of these (specified below). High-throughput sequencing methods permit Samples and cultures. Avrainvillea mazei HV02664 and Neo- addressing problematic phylogenetic relationships meris sp. HV02668 were collected from Tavernier (Florida, with increased sampling both of taxa and molecular USA) and Islamorada (Florida, USA), respectively, and pre- markers. Chloroplast genome sequencing and phy- served in silica-gel. To obtain Ostreobium tissue, two fragments logenomic analysis have been used to resolve, for of coral skeleton were collected. Both were taken from coral example, difficult relationships among red algae skeletons (Diploastrea heliopora) collected in the Kavieng area of Papua New Guinea. Samples HV05007 and HV05042 were (Costa et al. 2016), core Chlorophyta (Fucıkova collected at 4 and 13 m depth, respectively, and both showed et al. 2014, Sun et al. 2016, Turmel et al. 2016), a pronounced green band in the skeleton. The fragments Trebouxiophyceae (Lemieux et al. 2014a) and were kept in unenriched sterilized seawater at 24°C and Chlorophyceae (Turmel et al. 2008). The first com- under dim light for several months until Ostreobium filaments plete chloroplast genomes of siphonous green started growing out of the limestone. When the filaments ~ algae, including Ostreobium quekettii, have been reached 1 cm, they were harvested for DNA extraction. We recently sequenced (Leliaert and Lopez-Bautista did not attempt to establish pure cultures. Molecular work and DNA sequencing. Total genomic DNA 2015, del Campo et al. 2016, Marcelino et al. 2016). was extracted using a modified CTAB protocol (Cremen et al. While these provide a good starting dataset to 2016). Two types of library preparation and sequencing were resolve the phylogenetic placement of the Ostreobium used. For A. mazei HV02664, a library was prepared from lineage, many early-branching lineages have not yet ~350 bp fragments using a TruSeq Nano LT kit and sequenced been sampled. This results in long branches which on the Illumina HiSeq 2000 platform (paired end, 100 bp). For the remaining strains, libraries were prepared from a would otherwise be broken up by additional taxon ~ sampling, leading to issues such as long branch 500 bp size fraction with a Kapa Biosystems kit and sequenced on the Illumina NextSeq platform (paired end, 150 bp). attraction (LBA), a phylogenetic artifact in which Assembly and annotation. Sequence reads from both Ostreo- long branches cluster together even if they are not bium strains were assembled using IDBA-UD 1.1.1 (Peng et al. closely related (Bergsten 2005). 2012), SPAdes 3.8.1 (Bankevich et al. 2012), CLC Genomics Here, we report additional chloroplast genomes Workbench 7.5.1 (http://www.clcbio.com) and MEGAHIT strategically obtained to break up the long branches 1.0.6 (Li et al. 2016). Contigs matching to Ulvophycean in siphonous green algae in order to infer the phy- chloroplast genomes reference sequences were imported into logenetic position of the Ostreobium lineage and the Geneious 8.0.5 (http://www.geneious.com), where complete- ness and circularity were evaluated by comparing the contigs early diversification of Bryopsidales. The goals of from different assemblers manually. Assembly graphs were this study are: (i) to obtain draft chloroplast gen- visually assessed in Bandage 0.8.0 (Wick et al. 2015) and read omes of multiple Ostreobium strains and from two mapping was carried out in Geneious 8.0.5. other early-branching families in the Bryopsidales Because one of the limestone samples turned out to con- (Dichotomosiphonaceae) and Dasycladales (Dasy- tain multiple Ostreobium species, metagenome binning was cladaceae), (ii) infer the phylogenetic position