MiGenes Mitopostulate

There are no absolute rules to distinguish which proteins function in mitochondria. Two examples illustrate the uncertainty in content of the mitochondrial proteome. In some cases it is clear that mitochondria contain a given enzyme activity, but it may be uncertain whether that activity is encoded by a particular protein sequence. In addition, the simple identification of a protein in a mitochondrial preparation by sequencing is not sufficient to confirm that this protein is an “authentic” mitochondrial protein. Some proteins are reproducibly found as contaminants in mitochondrial preparations. Thus, we have adopted the following “MitoPostulates” to classify proteins as mitochondrial:

 1. A protein is reported as mitochondrial if there is an assertion of a mitochondrial localization or sublocalization ( OM, IMS, IM, cristae, matrix) or a commonly accepted mitochondrial enzyme activity supported by a “strong” evidence code. Generally there should be a specific reference (PMID) supporting this assertion.

Note on evidence codes: codes in RED are considered “strong” evidence in favor of mitochondrial localization; other codes are considered “soft” or presumptive evidence that can be used to categorize a protein as mitochondria-related, but not truly mitochondrial. Manual curation is used to exclude some proteins from the list of mitochondrial proteins when the protein fails to satisfy MitoPostulate #3, below.

2. If a protein is reported as mitochondrial by MitoPostulate #1 in one organism, it may be classified as mitochondrial in a second organism if the following two conditions are met:

a. if blast searches show the two proteins are the best reciprocal homologs in the two organisms

b. if the homology is as good as 95% of the known homologous matches between the two organisms (This represents a stringency test for sequence similarity (evidence code ISS), which is otherwise considered a weak evidence code for classification purposes; we estimate that application of this rule should result in at most 5% false positives).

 3. If a protein is reported by a protein sequencing study as mitochondrial and if this protein has another well-defined location in the cell supported by GO terms and strong evidence codes, it is considered as mitochondrial related, not as mitochondrial until strong evidence of mitochondrial localization is obtained. This evidence should typically explain how the protein is targeted to mitochondria (e.g., a differential splice or differential translation start site).