Let’s take a look at Tom20, the protein that acts as a receptor to snag the mitochondrial targeting sequence (MTS) and shuttle it to the membrane channel.
Tom20 has been studied in fungal and mammalian systems and shown to have the a domain structure as shown in Figure 1.
Figure 1. Domain organization of Tom20 in animals and fungi (modified from Fig 5 of reference listed below). N = amino terminus (the front-end of the protein) and C = carboxyl terminus (the back end of the protein).
The gray boxes represent regions that are disordered, the blue box represents a TPR domain that binds the MTS, and the black box represents the region that spans the outer membrane of the mitochondria. The line above represents the regions exposed to the cytoplasm.
So the black box region sits in the membrane and the blue box, with TPR domains, snags the MTS.
When scientists used the sequence of this protein to query databases for homologs in plants and protozoa, nothing was found. Could it be that plants and fungi don’t have Tom20?
Biochemical techniques were used is isolate a protein from mitochndria that was the same size as Tom20 and interacted with the membrane pore. Antibodies raised against this protein worked to block protein transport into the mitochondria. The gene for this protein was isolated and the protein’s structure was recently determined . The results are shown in Figure 2.
Figure 2. Domain organization of Tom20 in plants and algae (modified from Fig 5 of reference listed below). N = amino terminus (the front-end of the protein) and C = carboxyl terminus (the back end of the protein).
If you compare this plant Tom20 to animal/fungi Tom20, you’ll notice the same pattern of domains in reverse. This becomes even more clear if we try to visualize these proteins in the membrane (Figure 3).
Figure 3. Tom20 from animals/fungi (A.) compared to Tom20 from plants (B.) Yellow box represents membrane, with cytoplasm on top.
Notice that the pattern is the same, except that for animals, it is the back end of the protein that is anchored in the membrane while for plants and algae, it is the front-end of the protein that is anchored.
This reverse similarity is not restricted to domain organization, but also extends into the sequence of the two proteins:
Comparison of the plant and animal Tom20 transmembrane domains and the proximal cytosolic regions suggests striking structural similarities, but features within the sequences occur in reverse order. First, there are conserved glycine and aromatic residues within the region predicted to be in the bilayer of the outer mitochondrial membrane. Secondly, an aspartate residue is found at the cytosolic membrane interface. Thirdly, this aspartate is followed by a region of 13 residues in which charged residues (five basic, one acidic) predominate. 
The researchers summarize the significance of all this:
The transmembrane domain and its flanking regions of the plant and animal Tom20s only show clear sequence and structural similarities when viewed in reverse, and genetic mechanisms underlying protein evolution, such as duplication, cyclic permutations, and limited insertions and deletions, could not easily result in the sequence reversal that we observe in these two proteins. We therefore hypothesize that two distinct TPR protein ancestors existed prior to the split of the lineage giving rise to plants and protozoans from that giving rise to animals and fungi. These distinct ancestral TPR proteins independently gave rise to the Tom20 in plants and the Tom20 in animals and fungi. 
In other words, this is an example of convergent evolution at the molecular level.
1. , , , , Convergent evolution of receptors for protein import into mitochondria (2006) Current Biology, 16 (3), pp. 221-229.