Multiple plastid losses within photosynthetic stramenopiles revealed by comprehensive phylogenomics

Ochrophyta is a vast and morphologically diverse group of algae with complex plastids, including familiar taxa with fundamental ecological importance (diatoms or kelp), and a wealth of lesser-known and obscure organisms. The sheer diversity of ochrophytes poses a challenge for reconstructing their phylogeny, with major gaps in sampling and an unsettled placement of particular taxa yet to be tackled. We sequenced transcriptomes from 25 strategically selected representatives and used these data to build the most taxonomically comprehensive ochrophyte-centered phylogenomic supermatrix to date. We employed a combination of approaches to reconstruct and critically evaluate the relationships among ochrophytes. While generally congruent with previous analyses, the updated ochrophyte phylogenomic tree resolved the position of several taxa with previously uncertain placement, and supported a redefinition of the classes Picophagea and Synchromophyceae. Our results indicated that the heterotrophic plastid-lacking heliozoan Actinophrys sol is not a sister lineage of ochrophytes, as proposed recently, but rather phylogenetically nested among them, implying that it lacks a plastid due to loss. In addition, we found the heterotrophic ochrophyte Picophagus flagellatus to lack all hallmark plastid genes, yet to exhibit mitochondrial proteins that seem to be genetic footprints of a lost plastid organelle. We thus document, for the first time, plastid loss in two separate ochrophyte lineages. Furthermore, by exploring eDNA data we enrich the ochrophyte phylogenetic tree by identifying five novel uncultured class-level lineages. Altogether, our study provides a new framework for reconstructing trait evolution in ochrophytes and demonstrates that plastid loss is more common than previously thought.

Methods
For a majority of strains, cells grown in culture flasks were dislodged by scraping and collected by filtration over a 25mm 12 micron polycarbonate filter and flash frozen in liquid nitrogen. RNA was extracted using Machery Nagel plant RNA kit following manufacturers protocol except for cell lysis. This was carried out by bead beating for 5 minutes using a mixture of 0.1 and 0.5mm zirconia/silica beads (BioSpec) in a BioSpec Mini-Beadbeater. Sequencing libraries were constructed using the New England BioLabs Next Ultra RNA prep kit and sequenced on the Illumina HiSeq (4000) in paired-end mode (2×150bp) at either University of Maryland sequencing center or Genewiz (New Jersey, USA). Adapters and low-quality regions were removed using Trimmomatic. Trimmed reads were assembled using rnaSPAdes v3.13. In the case of O. luteus K-0444 and P. flagellatus RCC22, total RNA was extracted with TRI Reagent® (TR 118) (Molecular Research Center, Inc., Cincinnati, USA), following standard procedures. Sequencing libraries were prepared by Macrogen Inc. (Seoul, South Korea) using TruSeq Stranded mRNA LT Sample Prep Kit and transcriptome sequencing was performed with the Illumina NovaSeq 6000 platform in pair-end mode (2×151bp and 2×101bp for O. luteus and P. flagellatus, respectively). De novo transcriptome assemblies for O. luteus and P. flagellatus were obtained using Trinity v2.1.1. RNA of C. australica EC13 was extracted with a Purelink Plant RNA kit (ThermoFisher). Sequencing libraries were prepared and sequenced on Illumina HiSeq 2500 (Novogene, Hong Kong) and assembled with rnaSPAdes v3.13 Trinity 2.4.0. Transcripts were translated to protein sequences using TransDecoder (https://github.com/TransDecoder/). WinstonCleaner (https://github.com/kolecko007/WinstonCleaner) was used to identify and remove lowly expressed transcripts and cross-contamination from the assembled transcriptomes. Completeness of the genome or all transcriptomes generated as part of this study were assessed using BUSCO v. 5.5.0 with the Stramenopile gene set (Table S1).