Song and Ward, 2007). Additional sequences from Chaetoceros tenuissimus, Pseudo-nitzschia multistriata, and Skeletonema costatum were obtained by a blastp search of Phatr2_26029 against diatoms (taxid:2836) in the NCBI nr database. Outgroup sequences were identified as top non-diatom hits with a blastp search of Phatr2_26029 against the NCBI nr database. Amino acid transporter (AMT) sequences were obtained from Phycocosm (Bacillariophyta) using the TCDB accessions TC 1.A.11.2 (AMT1/2) and TC 1.A.11.4 (AMT -Rhesus type). Additionally, putative AMT sequences were identified from the non-assembled scaffolds of P. tricornutum (Phatr2_bd, https://mycocosm.jgi.doe.gov/Phatr2_bd/Phatr2_bd.home.html) using tblastn with other P. tricornutum AMTs as the query sequences. Pn. multistriata AMT sequences were retrieved with blastp search of Pn. multiseries sequences against NCBI nr against diatoms (taxid:2836). For construction of the phylogenetic tree, phosphate transporter sequences were identified in Phycocosm using TC 2.A.58 (Phosphate:Na+ Symporter or SLC34), TC 2.A.20 (Inorganic phosphate transporter, PiT or SLC20), and TC 2.A.1.9 (Phosphate: H+ symporter family) annotations and using accessions identified in (Cruz de Carvalho et al., 2016). Amino acid sequences and accessions can be downloaded from figshare (Smith, 2021).
Alignments and Phylogenetic TreeCLC Genomics Workbench (20.0.2) was used for all alignments and trees. Sequences were aligned with Gap cost settings as follows: Gap open cost = 10.0, Gap extension cost = 1.0, with the End gap cost = free. Alignment was run as Very accurate (slow). Short sequences, sequences with large gaps, and sequences with low coverage over consensus regions were removed from the alignments and long sequences were trimmed. Alignments used in construction of the NRT2 tree versions (see below) and the AMT tree can be accessed from figshare (Smith, 2021). The neighbor joining method was used for a start tree, the protein substitution model chosen was WAG, and Estimation was set to estimate topology. Bootstrap analysis was performed with 100 replicates, and only nodes with bootstrap support >50 are indicated.