Skip to main content
. 2017 May 12;6:e23717. doi: 10.7554/eLife.23717

Figure 7. Ancient and bidirectional connections between the ochrophyte plastid and mitochondria.

(Panel A) shows Mitotracker-Orange stained P. tricornutum lines expressing GFP fusion constructs for the N-terminal regions of histidyl- and prolyl-tRNA synthetase sequences from P. tricornutum and the eustigmatophyte Nannochloropsis gaditana. Targeting constructs for an additional four dual-targeted proteins in P. tricornutum and one dual-targeted protein in G. foliaceum, alongside Mitotracker-negative and wild type control images, are shown in Figure 7—figure supplement 1. (Panel B) profiles the predicted evolutionary origins of the 34 ancestral dual-targeted HPPGs, as inferred by BLAST top hit and single-gene tree analysis. Data supporting the thresholds used to identify probable dual-targeted HPPGs in silico are supplied in Figure 7—figure supplement 2. (Panel C) shows seven classes of tRNA synthetase for which only two copies were inferred in the genome of the last common ochrophyte ancestor. Evolutionary origins are inferred from combined BLAST top hit and single-gene tree analysis for dual-targeted proteins, and from BLAST top hit analysis alone for cytoplasmic proteins. In five cases the dual-targeted isoform is inferred to be of ultimate red algal origin, indicating that a protein derived from the endosymbiont has functionally replaced the endogenous host mitochondria-targeted copy.

DOI: http://dx.doi.org/10.7554/eLife.23717.035

Figure 7.

Figure 7—figure supplement 1. Experimental verification of additional ochrophyte dual-targeted proteins.

Figure 7—figure supplement 1.

(Panel A) shows Mitotracker-orange stained Phaeodactylum tricornutum lines expressing four additional dual-targeted proteins (glycyl-, leucyl-, and methionyl-tRNA synthetases, and a predicted mitochondrial GroES-type chaperone) from Phaeodactylum tricornutum, and a dual-targeted histidyl-tRNA synthetase from Glenodinium foliaceum. (Panel B) shows control images that confirm an absence of crosstalk between GFP and Mitotracker: wild-type Phaeodactylum cells stained with Mitotracker, and cells expressing the Glenodinium histidyl-tRNA synthetase–GFP fusion construct and visualised with the Mitotracker laser and channel in the absence of Mitotracker stain.
Figure 7—figure supplement 2. Comparison of different in silico targeting prediction programmes for the identification of dual-targeted ochrophyte proteins.

Figure 7—figure supplement 2.

(Panel A) shows Mitofates scores for ochrophyte proteins verified experimentally to be dual-targeted in this and a previous study (Gile et al., 2015). (Panel B) shows Mitofates scores for all ochrophyte proteins for which a subcellular localisation has been identified in previous studies. The red lines in each graph show the Mitofates default cutoff (0.385) and the green lines indicate our chosen cutoff (0.35). (Panel C) compares different in silico targeting prediction algorithms with respect to predicted mitochondrial localization by experimentally validated localization. Mitofates strikes the best balance between high true positives and low false positives.