The fungus Ustilago maydis and humans share disease-related proteins that are not found in Saccharomyces cerevisiae

Abstract

Background: The corn smut fungus Ustilago maydis is a well-established model system for molecular phytopathology. In addition, it recently became evident that U. maydis and humans share proteins and cellular processes that are not found in the standard fungal model Saccharomyces cerevisiae. This prompted us to do a comparative analysis of the predicted proteome of U. maydis, S. cerevisiae and humans.

Results: At a cut off at 20% identity over protein length, all three organisms share 1738 proteins, whereas both fungi share only 541 conserved proteins. Despite the evolutionary distance between U. maydis and humans, 777 proteins were shared. When applying a more stringent criterion (> or = 20% identity with a homologue in one organism over at least 50 amino acids and > or = 10% less in the other organism), we found 681 proteins for the comparison of U. maydis and humans, whereas the both fungi share only 622 fungal specific proteins. Finally, we found that S. cerevisiae and humans shared 312 proteins. In the U. maydis to H. sapiens homology set 454 proteins are functionally classified and 42 proteins are related to serious human diseases. However, a large portion of 222 proteins are of unknown function.

Conclusion: The fungus U. maydis has a long history of being a model system for understanding DNA recombination and repair, as well as molecular plant pathology. The identification of functionally un-characterized genes that are conserved in humans and U. maydis opens the door for experimental work, which promises new insight in the cell biology of the mammalian cell.

Figure 1

Global comparison of the predicted proteome sequence of the fungi U. maydis, S. cerevisiae and H. sapiens. Based on the Smith-Waterman comparison of all proteins from genome A against all proteins of genomes B the median e-value and the median %-identity value were calculated. Note that U. maydis is more similar to humans than to its fungal cousin, whereas the proteome of S. cerevisiae is more closely to that of the corn smut.

Figure 2

Comparison of the proteome of U. maydis, S. cerevisiae and H. sapiens using cut-off criteria of 20 % identity over the total protein length.

Figure 3

Genome-wide protein homology correlation of U. maydis against humans and S. cerevisiae using the % identity over entire protein length. Depicted are predicted U. maydis proteins that show > 20% identity to one partner and > 10% less identity to the other. Most proteins (2159) are conserved at similar levels in S. cerevisiae and H. sapiens (B). However, 587 proteins are more closely related to S. cerevisiae (A), whereas an even larger number of 651 are more similar to humans (C). Note that this analysis show "% identity over the total protein length" and thus does not include proteins that share sequence similarity only within a domain.

Figure 4

Total number of predicted proteins that show significant similarity (> 20% identity to one partner and > 10% less identity to the other) over their entire length (grey bar) or within domains of at least 50 amino acids (black bar).

Figure 5

Proteins of the U. maydis – H. sapiens and S. cerevisiae – H. sapiens homology sets, which human counterpart have a predicted role in diseases. The classification was done according to main disease class in the genetic association database [47]. Note that U. maydis contains numerous genes that are thought to be involved in psychological disorders, such as ERCC1 and ERCC4.