Figure 2. Verification of unusual ancestral plastid-targeted proteins.

(Panel A) lists the ten proteins selected for experimental characterisation and their most probable previous localisation prior to their establishment in the ochrophyte plastid, based on the first 50 nr BLAST hits. Exemplar alignments and single-gene tree topologies for some of these proteins are shown in Figure 2—figure supplements 1–4. (Panel B) shows the localisation of GFP constructs for copies of two proteins with an unambiguous plastid localisation (a pyrophosphate-dependent PFK, which localises to the pyrenoid, and a novel plastid protein, with cosmopolitan distribution across the plastid) and one protein with a periplastid localisation (a predicted peroxisomal membrane protein) from the diatom Phaeodactylum tricornutum, the diatom endosymbiont of the dinoflagellate Glenodinium foliaceum and the eustigmatophyte Nannochloropsis gaditana, expressed in P. tricornutum. All scale bars = 10 μm. Expression constructs for seven additional P. tricornutum proteins and three additional N. gaditana proteins with multipartite plastid localisations are shown in Figure 2—figure supplements 5 and 6, and control images (wild-type cells, and cells expressing untargeted eGFP) are shown in Figure 2—figure supplement 7.

DOI: http://dx.doi.org/10.7554/eLife.23717.006

Figure 2—figure supplement 1. Exemplar ochrophyte plastid protein alignments.

This figure shows untrimmed GeneIOUS alignments for two ancestral HPPGs of unusual provenance. In each case the full length of the protein (labelled i) and N-terminal region only (ii) are shown, demonstrating the broad conservation of the N-terminus position. Sequences for which exemplar targeting constructs (Phaeodactylum tricornutum, Nannochloropsis gaditana, Glenodinium foliaceum) were generated are shown at the top of each alignment.

DOI: http://dx.doi.org/10.7554/eLife.23717.007

Figure 2—figure supplement 2. Tree of ochrophyte glycyl-tRNA synthetase sequences.

This tree shows the consensus unrooted Bayesian topology for a 95 taxa x 487 aa alignment of glycyl tRNA synthetase sequences. The font colour of each sequence corresponds to the taxonomic origin (see legend below for details) and are labelled with the taxonomic identifiers previously defined in Table S1. Sequences labelled with chl_ possess apparent plastid targeting sequences recognisable by CASH lineage plastids. The ancestral ochrophyte plastidic isoform, of apparent chlamydiobacterial origin, is labelled with a blue ellipse. Black circles at each node denote posterior probabilities of 1.0 in Bayesian inferences with three different substitution matrices (GTR, Jones, and WAG), and grey circles indicate posterior probabilities of 0.8 with at least two of these matrices. Support values for all remaining nodes, is provided using both Bayesian analysis (above line) and RAxML tree (below line), using three substitution matrices, as defined in the figure legend.

DOI: http://dx.doi.org/10.7554/eLife.23717.008

Figure 2—figure supplement 3. Tree of ochrophyte pyrophosphate dependent phosphofructo-1- kinase sequences.

This tree shows the consensus Bayesian topology inferred for a 94 taxa x 449 aa alignment of pyrophosphate-dependent PFK, with taxa and support values shown as per Figure 2—figure supplement 2. The ancestral ochrophyte plastid isoform, of probable aplastidic stramenopile origin, is labelled with a cyan ellipse.

DOI: http://dx.doi.org/10.7554/eLife.23717.009

Figure 2—figure supplement 4. Tree of a novel ochrophyte plastid-targeted protein.

This tree shows the consensus Bayesian topology inferred for a 16 taxa x 103 aa alignment of a plastid-targeted protein seemingly restricted to ochrophytes and one dinoflagellate lineage. Taxa are labelled and support values are shown as per Figure 2—figure supplement 2.

DOI: http://dx.doi.org/10.7554/eLife.23717.010

Figure 2—figure supplement 5. Multipartite Phaeodactylum plastid-targeted proteins.

This figure shows the localisation of GFP overexpression constructs for copies of seven proteins from the diatom Phaeodactylum tricornutum that are of non-plastid origin, but show multipartite localization to the plastid and one other organelle (the mitochondria, or in the case of the ‘ER heat shock protein’ to the endoplasmic reticulum).

DOI: http://dx.doi.org/10.7554/eLife.23717.011

Figure 2—figure supplement 6. Heterologous expression constructs of multipartite plastid-targeted proteins.

This figure shows the localisation of GFP overexpression constructs for copies of two proteins from the dinotom Glenodinium foliaceum (Panel A), and three proteins from the eustigmatophyte Nannochloropsis gaditana (Panel B) that are of non-plastid origin, but show multipartite localisation to the plastid and one other organelle, per Figure 2—figure supplement 5.

DOI: http://dx.doi.org/10.7554/eLife.23717.012

Figure 2—figure supplement 7. Exemplar control images for confocal microscopy.

This figure shows fluorescence patterns for wild-type Phaeodactylum tricornutum cells (i), and transformant Phaeodactylum cells expressing GFP that has not been fused to any N-terminal targeting sequence (ii), both visualised under the same conditions used for all other cultures.

DOI: http://dx.doi.org/10.7554/eLife.23717.013