The precursor of mitochondrial aspartate aminotransferase is imported into mitochondria faster than the homologous cytosolic isoenzyme with the same presequence attached

Evidence of evolutionary constraints that influences the sequence composition and diversity of mitochondrial matrix targeting signals

Mitochondrial targeting signals (MTSs) are responsible for trafficking nuclear encoded proteins to their final destination within mitochondria. These sequences are diverse, sharing little amino acid homology and vary significantly in length, and although the formation of a positively-charged amphiphilic alpha helix within the MTS is considered to be necessary and sufficient to mediate import, such a feature does not explain their diversity, nor how such diversity influences target sequence function, nor how such dissimilar signals interact with a single, evolutionarily conserved import mechanism. An in silico analysis of 296 N-terminal, matrix destined MTSs from Homo sapiens, Mus musculus, Saccharomyces cerevisiae, Arabidopsis thaliana, and Oryza sativa was undertaken to investigate relationships between MTSs, and/or, relationships between an individual targeting signal sequence and the protein that it imports. We present evidence that suggests MTS diversity is influenced in part by physiochemical and N-terminal characteristics of their mature sequences, and that some of these correlated characteristics are evolutionarily maintained across a number of taxa. Importantly, some of these associations begin to explain the variation in MTS length and composition.

The molecular evolution of pyridoxal-5'-phosphate-dependent enzymes

The pyridoxal-5-phosphate-dependent enzymes (B6 enzymes) that act on amino acid substrates are of multiple evolutionary origin. The numerous common mechanistic features of B6 enzymes thus are not historical traits passed on from a common ancestor enzyme but rather reflect evolutionary or chemical necessities. Family profile analysis of amino acid sequences supported by comparison of the available three-dimensional (3-D) crystal structures indicates that the B6 enzymes known to date belong to four independent evolutionary lineages of homologous (or more precisely paralogous) proteins, of which the alpha family is by far the largest. The alpha family (with aspartate aminotransferase as the prototype enzyme) includes enzymes that catalyze, with several exceptions, transformations of amino acids in which the covalency changes are limited to the same carbon atom that carries the amino group forming the imine linkage with the coenzyme (i.e., Calpha in most cases). Enzymes of the beta family (tryptophan synthase beta as the prototype enzyme) mainly catalyze replacement and elimination reactions at Cbeta. The D-alanine aminotransferase family and the alanine racemase family are the two other independent lineages, both with relatively few member enzymes. The primordial pyridoxal-5-phosphate-dependent enzymes apparently were regio-specific catalysts that first diverged into reaction-specific enzymes and then specialized for substrate specificity. Aminotransferases as well as amino acid decarboxylases are found in two different evolutionary lineages. Comparison of sequences from eukaryotic, archebacterial, and eubacterial species indicates that the functional specialization of most B6 enzymes has occurred already in the universal ancestor cell. The cofactor pyridoxal-5-phosphate must have emerged very early in biological evolution; conceivably, organic cofactors and metal ions were the first biological catalysts. In attempts to stimulate particular steps of molecular evolution, oligonucleotide-directed mutagenesis of active-site residues and directed molecular evolution have been applied to change both the substrate and reaction specificity of existent B6 enzymes. Pyridoxal-5-phosphate-dependent catalytic antibodies were elicited with a screening protocol that applied functional selection criteria as they might have been operative in the evolution of protein-assisted pyridoxal catalysis.

Protein import into subsarcolemmal and intermyofibrillar skeletal muscle mitochondria. Differential import regulation in distinct subcellular regions

To date, no studies have described the import of proteins in mitochondria obtained from skeletal muscle. In this tissue, mitochondria consist of the functionally and biochemically distinct intermyofibrillar (IMF) and subsarcolemmal (SS) subfractions, which are localized in specialized cellular compartments. This mitochondrial heterogeneity in muscle could be due, in part, to differential rates of protein import. To evaluate this possibility, the import of precursor malate dehydrogenase and ornithine carbamyltransferase proteins was investigated in isolated IMF and SS mitochondria in vitro. Import of these was 3-4-fold greater in IMF compared with SS mitochondria as a function of time. This could account for the higher malate dehydrogenase enzyme activity in IMF mitochondria. Divergent import rates in IMF and SS mitochondria likely result from a differential reliance on various components of the import pathway. SS mitochondria possess a greater content of the molecular chaperones hsp60 and Grp75, yet import is lower than in IMF mitochondria. On the other hand, adriamycin inhibition studies illustrated a greater reliance on acidic phospholipids (i.e. cardiolipin) for the import process in SS mitochondria. Matrix ATP levels were 3-fold higher in IMF mitochondria, but experiments in which ATP depletion was performed with atractyloside and oligomycin illustrated a dissociation between import rates and levels of ATP. In contrast, a close relationship was found between the rate of ATP production (i.e. mitochondrial respiration) and protein import. When respiratory rates in IMF and SS mitochondria were equalized, import rates in both subfractions were similar. These data indicate that 1) import rates are more closely related to the rate of ATP production than the steady state ATP level, 2) import into IMF and SS mitochondrial subfractions is regulated differently, and 3) mitochondrial heterogeneity within a cell type can be due to differences in the rates of protein import, suggesting that this step is a potentially regulatable event in determining the final mitochondrial phenotype.

Differential effects of molecular chaperones on refolding of homologous proteins

Three homologous aspartate aminotransferases with virtually identical spatial structures and pairwise amino acid sequence identities of > 40% differ markedly with respect to the yield of renaturation upon dilution from 6 M guanidine hydrochloride (mitochondrial << cytosolic < Escherichia coli). The enzymes also respond differently to molecular chaperones. GroEL/GroES, the Hsp60 homolog of E. coli, increased considerably the yield of renaturation of mitochondrial aspartate aminotransferase and to a lesser extent that of its cytosolic counterpart, but not that of the E. coli enzyme. DnaK/DnaJ/GrpE, the Hsp70 system of E. coli, also increased the yield of renaturation of mitochondrial aspartate aminotransferase. Apparently, specific features in the amino acid sequence or the folding pathway which are independent of the final secondary and tertiary structure determine the interactions of the folding proteins with the chaperone systems.

Homologous nuclear-encoded mitochondrial and cytosolic isoproteins. A review of structure, biosynthesis and genes

Mitochondrial and cytosolic proteins may be expected to differ in specific traits due to their different intracellular location. However, the identification of these differences between mitochondrial and cytosolic proteins is complicated by the heterogeneity of the two protein groups. These difficulties have been overcome by comparing traits of homologous genes, which are derived from a common ancestor gene, and their gene products. An earlier report [Hartmann, C., Christen, P. & Jaussi, R. (1991) Nature 352, 762-763] describing a positive net charge difference between the mature parts of nuclear-encoded mitochondrial proteins and their homologous cytosolic isoproteins, could be corroborated by extending the data collection. New data were gathered from computer databases and published studies. The average isoelectric points of the mitochondrial and cytosolic isoproteins are 7.5 and 6.5, respectively. Depending on the type of protein, the observed difference results from differences in the number of basic and/or acidic amino acid residues in the isoproteins. Probably both the conditions required for mitochondrial protein import and the local conditions within the organelle furthered the evolution of basic protein structures. The contribution of the mitochondrial targeting peptide to the positive charge of precursors of nuclear-encoded mitochondrial proteins is largest when the value of the isoelectric point of the mature protein is small. This mutual dependence of the charge of the targeting peptide and the mature protein part supports the notion that positive charge is essential for mitochondrial protein import. Several traits other than electric charge, i.e. codon usage, chromosome location, structural organization or regulation of the genes, do not show specific differences between the sets of the heterotopic isoproteins. There is no preference of gene location for any of the gene sets; only rarely are the genes for a mitochondrial and a cytosolic isoprotein located on the same chromosome. A variant of the 3' splice-site consensus exists in genes of nuclear-encoded mitochondrial proteins. This is most likely a consequence of the evolution of the genes in separate lineages before endosymbiosis led to the formation of mitochondria. Some of the original mRNA group II intron self-splicing functions of the endosymbiont seem to persist in part of the cytosolic splicing machinery and apparently require a specific consensus sequence [Juretic, N., Jaussi, R., Mattes, U. & Christen, P. (1987) Nucleic Acids Res.15, 10083-10086].

The mature form of imported mitochondrial proteins undergoes conformational changes upon binding to isolated mitochondria

Mature mitochondrial proteins (aspartate aminotransferase, malate dehydrogenase, hydroxyacyl coenzyme A dehydrogenase, creatine kinase) and cytosolic proteins (aldolase, glyceraldehyde-3-phosphate dehydrogenase) with a basic pI were found to bind to isolated mitochondria, electrostatic interactions being mainly responsible for their binding. Mitochondrial aspartate aminotransferase bound with a Kd' of 30 nM in 0.6 M sorbitol, 20 mM Hepes/KOH, pH 7.4, at 25 degrees C. Cytosolic aspartate aminotransferase and glutamate dehydrogenase (a protein located in the mitochondrial matrix) both with an acidic pI, did not bind to mitochondria. Treatment of mitochondria with proteinases did not affect the subsequent binding of imported mitochondrial proteins. Their association with both intact and proteinase-treated mitochondria resulted in a marked increase in their susceptibility toward proteinase K. In contrast, the basic cytosolic proteins tested bound only to intact mitochondria and thereby did not become more susceptible toward proteolytic attack. Treatment of mitochondria with adriamycin, a drug binding to acidic phospholipids, prevented the subsequent association of mitochondrial aspartate aminotransferase with mitochondria and the ensuing conformational labilization. Apparently, the mature moiety of imported mitochondrial proteins is partially unfolded upon interaction with the lipid component of the mitochondrial envelope. Both the binding of the mitochondrial proteins and their conformational labilization is independent of ATP and the electrochemical potential across the inner membrane.

Mitochondrial transport of mitoribosomal proteins, YmL8 and YmL20, in Saccharomyces cerevisiae

Two mitochondrial ribosomal (mitoribosomal) proteins, YmL8 and YmL20, of the yeast Saccharomyces cerevisiae and their derivatives were synthesized in vitro and their transport into isolated yeast mitochondria was examined. Of the two proteins, YmL20 possesses an N-terminal presequence of 18 amino acid residues, while YmL8 has no such presequence. Both proteins were found to be transported into isolated mitochondria in an energy-dependent manner. Furthermore, YmL20 protein without its N-terminal presequence was also transported, despite the fact that the presequence alone was capable of transporting a fused passenger protein, Chinese hamster dihydrofolate reductase (DHFR). Therefore, YmL20 protein appears to possess redundant transport signals in its structure. Similarly, YmL8 derivatives lacking either 40 or 86 amino acid residues from the N-terminus and/or 52 amino acid residues from the C-terminus were transported. In addition, the N-terminal segment of this protein was capable of transporting Chinese hamster DHFR into mitochondria, while its C-terminal segment was not. Thus, YmL8 protein also appears to possess two or more transport signals in its structure. Perhaps the presence of many basic amino acid residues in these proteins might, at least partly, contribute to their mitochondrial transport.

The in vitro-synthesized precursor and mature mitochondrial aspartate aminotransferase share the same import pathway in isolated mitochondria

Both the precursor and the mature form of mitochondrial aspartate aminotransferase were synthesized in a cell-free coupled transcription/translation system directed by the recombinant expression plasmid pOTS-pmAspAT and pOTS-mAspAT, respectively. Both newly synthesized forms of the protein were imported into isolated mitochondria, with the precursor correctly processed to the mature form. In both cases the import process showed resistance to externally added pronase and was abolished in mitochondria treated with the uncoupler carbonyl cyanide m-chlorophenylhydrazone. Moreover the imported products showed the same intramitochondrial localization as judged by a subfractionation procedure. In both cases import was time dependent and was completed in about 15 min. Finally a competitive inhibition of the import of the precursor of aspartate aminotransferase was found due to externally added purified aspartate aminotransferase.