The HSP70 multigene family: Cloning, expression, and evolution

HSP70 genes are found in all organisms. In eukaryotes, these genes comprise a multigene family. I have demonstrated, by the isolation of several related genes, that the mouse genome contains such a family. Characterization of the isolated clones reveals that they consist of two inducible genes, one tissue-specific gene expressed only in testes, and four processed pseudogenes. Expression studies suggest that the family also contains two constitutively expressed genes (hsc70A and bip) and a second testis-specific gene.;Sequencing of these clones illuminated problems inherent in sequence analysis. Such problems include cloning artifacts and sequencing errors due to regions of high GC content. These were further demonstrated by the compilation of all known HSP70 genes where errors become obvious due to the extreme similarity between the genes.;The amino acid and corresponding nucleotide sequence compilations were used in the construction of dendrograms. These trees suggest that all prokaryotic and organelle HSP70 genes cluster, while all remaining eukaryotic HSP70 genes form a separate group. The clustering of archaebacteria with prokaryotes is inconsistent with the currently accepted three domain classification scheme of Woese which suggests that archaebacteria are more closely related to eukaryotes.;Consistent with the established endosymbiotic origin of organelles, chloroplast homologs are found most closely related to cyanobacteria HSP70 in trees constructed from both amino acid and nucleotide sequences. Similarly, in amino acid-based trees, mitochondrial homologs cluster with the purple bacteria. However, in nucleotide sequence-based trees, these sequences separate, with mitochondrial genes having high GC content clustering with other high GC content genes. The tendency of genes or genomes towards high or low GC (AT pressure) is well established. In the highly conserved HSP70 genes, this pressure is exerted predominantly at the third codon position where most changes are neutral. Deletion of the codon third base thus eliminates most of the effects of AT pressure. In trees constructed from these adjusted sequences, the mitochondrial homologs now cluster together. These results suggest that AT pressure significantly effects the outcome of trees constructed from nucleotide sequences, and that amino acid sequences are more reliable predictors of gene relationships and phylogeny.