Fructose-1,6-bisphosphate aldolase

Trypanosoma and Leishmania aldolases (Marchand et al., 1988; de Walque et al., 1999) constitute, in neighbor-joining analysis, a monophyletic group with the chloroplast aldolases from both algae and plants. The trypanosomatid enzymes form a sister goup to the enzymes of the rhodophyte Cyanidium caldarum and of the chlorophytes Dunaliella salina and Chlamydomonas reinhardtii. However this clustering is not robust (52% bootstrap support). The chloproplast aldolase from Euglena gracilis and the A and B isoenzymes of the malaria parasite Plasmodium falciparum are paraphyletic and group together at the base of the cytosolic eukaryotic aldolases, but this clustering was not supported by bootstrap analysis. Maximum likelihood analysis (Treepuzzle using the Blosum62 matrix as model) placed the two trypanosomatid sequences, together with the chloroplast sequences, the two aldolases from Cyanidium, and those from Plasmodium and Euglena, all within a single star-like subgroup. However, when gamma correction for unequal rates of evolution of sites was implemented, all protist sequences and all plant and algal aldolases (both cytosolic and chloroplast) collapsed into a single star-like subgroup which got a quartet puzzling support of 99%. When the Euglena sequence was removed from the dataset the Plasmodium and Trypanosomatid sequences clustered together and formed a single clade with the chloroplast sequences with high bootstrap / puzzle frequency support.

Click here for the combined neigbor-joining - maximum likelihood tree (complete dataset).

Click here for the neigbor-joining - maximum likelihood tree (dataset minus Euglena).

Details of the analysis

The T. brucei aldolase was compared with the SwissProt release 39 / TrEMBL release 17 database indexed at European Bioinformatics Institute (EBI, Hinxton UK), containing 671,000 protein sequences, using the NCBI BLASTP program and the BLOSUM 62 matrix (http://www2.ebi.ac.uk/blastall/) . Click here to inspect the BLASTP output file.

The best E values (0 and e-173) were obtained with itself (ALF_TRYBB) and with the sequence of the trypanosomatid Leishmania mexicana (Q9U5N6), followed by the aldolases from the plant Arabidopsis (Q9ZU52), the rhodophyte Cyanidium caldarium (Q9M633) and Nicotiana paniculata (Q9SXX5). Interestingly the BLASTP output only listed eukaryotic sequences. Therefore a second BLASTP search was carried out but this time against the SwissProt database. Again all sequences with a significant score were eukaryotic sequences and shared some 50% identity with the Trypanosoma sequence. The first prokaryotic sequences, those of Staphylococcus and the cyanobacterium Synechocystis gave a borderline score and the percentages of identity dropped to 31 and 32%, respectively. The first 35 sequences from the BlastP output were aligned using the "RunDBClustalW" option in the BLASTP output. The ClustalW alignment is availabe here for inspection. To this alignment were now added the Euglena chloroplast , the Leishmania aldolase , two Cyanidium caldarum and a Dunaliella salinum chloroplast sequences. From this alignment positions with gaps were removed and the alignment in Phylip format containing 40 taxa and 270 positions was saved to disk for further analysis.

Maximum likelihood mapping as implemented in PUZZLE version 4.0.1. indicated the presence in the dataset of a strong phylogenetic signal (only 1.7% star-like quartets while 95.7 % of the quartet trees were well-resolved).

Phylogenetic trees were created using neighbor-joining and maximum likelihood methods, with and without gamma correction for unequal rates of evolution of sites. Neigbor-joining analysis was able to divide the sequences into three major clades: chloroplast sequences, plant cytosolic sequences and aldolases of the remainder of the eukaryotes. The trypanosomaitds clustered witht the chloroplasts sequences, while the chloroplast sequence from Euglena and those of Plasmodium grouped with the other non-plant aldolases. However, there was no bootstrap support for any the deep-branching groups. Maximum likelihood analyses grouped all plant sequences (chloroplast and cytosol) as well as the sequences from Euglena, Plasmodium and the Trypanosomatidae together in a single star-like subtree.

Four-cluster likelihood mapping was used to determine the affinities of the two trypanosomatid sequences with the chloroplast clade. Ninety one percent of the quatret trees supported such clustering. Only 0.1 % suggested an affinity with the cytosolic plant aldolases, while 2.1 % showed an affinity with the remainder of the eukaryotes. A similar analysis carried out for Euglena chloroplast aldolase gave 41% percent of the quartet trees that grouped with the eukaryotes other than plants, 38% with the chloroplasts, and only 3% with the plant cytosolic sequences. In the third mapping experiment we evaluated the relatedness of the Euglena chloroplast aldolase with the trypanosomatid aldolases. In only 2% of the quartets Euglena and the trypanosomes clustered together, while 77% of the quartets favoured a clustering of Euglena aldolase with the enzymes from the plant cytosol.

Conclusion:

The Trypanosoma brucei and Leishmania mexicana aldolases are clearly more related to their homologues found in the chloroplasts of plants and algae than to any other eukaryotic or prokaryotic aldolase. This conclusion is supported not only by BLASTP analysis, or the pairwise distances between the trypanosomid sequences and their homologues from other groups, but also by four-cluster likelihood mapping en by phylogenetic analyses using neighbor-joining distance matrix analysis. Maximum likelihood robustly grouped the trypanosomatid sequences with those of plants, but was unable to decide whether they were more related to the chloroplast than to the plant cytosolic sequences. The Euglena chloroplast aldolase seems neither to be related to the trypanosomatid aldolases, nor to the aldolases from chloroplasts.