Evidence for distinct L-methionine catabolic pathways in the yeast Geotrichum candidum and the bacterium Brevibacterium linens

Tracing experiments were carried out to identify volatile and nonvolatile L-methionine degradation intermediates and end products in the yeast Geotrichum candidum and in the bacterium Brevibacterium linens, both of which are present in the surface flora of certain soft cheeses and contribute to the ripening reactions. Since the acid-sensitive bacterium B. linens is known to produce larger amounts and a greater variety of volatile sulfur compounds (VSCs) than the yeast G. candidum produces, we examined whether the L-methionine degradation routes of these microorganisms differ. In both microorganisms, methanethiol and alpha-ketobutyrate are generated; the former compound is the precursor of other VSCs, and the latter is subsequently degraded to 2,3-pentanedione, which has not been described previously as an end product of L-methionine catabolism. However, the L-methionine degradation pathways differ in the first steps of L-methionine degradation. L-Methionine degradation is initiated by a one-step degradation process in the bacterium B. linens, whereas a two-step degradation pathway with 4-methylthio-2-oxobutyric acid (MOBA) and 4-methylthio-2-hydroxybutyric acid (MHBA) as intermediates is used in the yeast G. candidum. Since G. candidum develops earlier than B. linens during the ripening process, MOBA and MHBA generated by G.candidum could also be used as precursors for VSC production by B. linens.     

The nature of the reaction of an essential tyrosine residue of bovine heart mitochondrial ATPase with 4-chloro-7-nitrobenzofurazan and related compounds

The import and processing of cytochrome c1 and the iron sulfur protein of the cytochrome b-c1 complex were studied in Zajdela hepatoma ascites cells. Both peptides were synthesized as larger percursor molecules which were approximately 2-3 kDa and 5-6 kDa larger than the mature forms of apocytochrome c1 and apo-iron sulfur protein, respectively. Comparison of these precursors to those reported for functionally homologous peptides in yeast and Neurospora indicate significant size changes have occurred in mammals. Rhodamine 6G, a specific vital stain for mitochondria, is a potent inhibitor of precursor processing in isolated hepatoma cells. Both precursor to cytochrome c1 and precursor to FeS accumulate in the soluble and particulate fractions obtained by digitonin treatment of tumor cells treated with Rhodamine 6G. Appearance of the mature peptides was abolished. The precursors are unstable, however, and disappear from the cytosolic and membrane fractions during a 10 min chase. Comparison of the effects of Rhodamine 6G and carbonylcyanide m-chlorophenylhydrazone on precursor processing shows that: (a) Rhodamine 6G is a more effective inhibitor of processing, (b) it has less of an inhibitory effect on cellular protein synthesis, and (c) it inhibits processing under conditions in which it appears to have little influence on coupled respiration in whole cells. The data suggest that the most likely mode of action of Rhodamine 6G is on the matrix processing step.     

Cell-free synthesis of mitochondrial ATPase inhibitor precursor and its transport into yeast mitochondria

ATPase inhibitor protein, which blocks mitochondrial ATPase activity by forming an enzyme-inhibitor complex, was found to be synthesized as a larger precursor in a cell-free translation system directed by yeast mRNA. Other protein factors, which stabilize latent ATPase by binding to the enzyme-inhibitor complex, were also found to be formed as larger precursors. The precursor of ATPase inhibitor protein was transported into isolated yeast mitochondria and was cleaved to the mature peptide in the mitochondria. Impaired mitochondria lacking phosphorylation activity could not convert the precursor to the mature form. Neither antimycin A nor oligomycin alone exhibited a marked effect on the transport-processing of the precursor by intact mitochondria. However, when antimycin A was added with oligomycin, the transport-processing was markedly inhibited. The processing was also strongly inhibited by an uncoupler, carbonylcyanide p-trifluoro-methoxyphenyl hydrazone. The inhibition by the uncoupler was not relieved by ATP added externally. It is concluded that the transport-processing of precursor proteins requires intact mitochondria with a potential difference across the inner membrane.     

Precursor protein is readily degraded in mitochondrial matrix space if the leader is not processed by mitochondrial processing peptidase

It is not known why leader peptides are removed by the mitochondrial processing peptidase after import into the matrix space. The leaders of yeast aldehyde dehydrogenase (pALDH) and malate dehydrogenase were mutated so that they would not be processed after import. The recombinant nonprocessed precursor of yeast pALDH possessed a similar specific activity as the corresponding mature form but was much less stable. The nonprocessed pALDH was transformed into a yeast strain missing ALDHs. The transformed yeast grew slowly on ethanol as the sole carbon source showing that the nonprocessed precursor was functional in vivo. Western blot analysis showed that the amount of precursor was 15-20% of that found in cells transformed with the native enzyme. Pulse-chase experiments revealed that the turnover rate for the nonprocessed precursor was greater than that of the mature protein indicating that the nonprocessed precursor could have been degraded. By using carbonyl cyanide m-chlorophenylhydrazone, we showed that the nonprocessed precursor was degraded in the matrix space. The nonprocessed precursor forms of precursor yeast malate dehydrogenase and rat liver pALDH also were degraded in the matrix space of HeLa cell mitochondria faster than their corresponding mature forms. In the presence of o-phenanthroline, an inhibitor of mitochondrial processing peptidase, the wild type precursor was readily degraded in the matrix space. Collectively, this study showed that the precursor form is less stable in the matrix space than is the mature form and provides an explanation for why the leader peptide is removed from the precursors.     

Cytochrome b is necessary for the effective processing of core protein I and the iron-sulfur protein of complex III in the mitochondria

The effect of cytochrome b on the assembly of the subunits of complex III into the inner mitochondrial membrane has been studied in a mutant of yeast (W-267, Box 6-2) that lacks a spectrally detectable cytochrome b and synthesizes a shortened form of apocytochrome b. We recently reported that several cytochrome b-deficient mutants contained significantly diminished amounts of core proteins I and II as well as the iron-sulfur protein, but contained equal amounts of cytochrome c1 compared to the wild type (K. Sen and D. S. Beattie, Arch. Biochem. Biophys. 242, 393-401, 1985). In the present study, the time course of processing of precursors of both core protein I and the iron-sulfur protein which had accumulated in cells treated with the uncoupler carbonyl m-chlorophenyl hydrazone (CCCP) was noted to be significantly lower in the mutant compared to the wild type. The amounts of the mature forms of these proteins in mitochondria pulse labeled under different conditions was also considerably decreased at all times studied. The synthesis of both proteins appeared to be unaffected in the mutant, as the precursor forms of both proteins accumulated to the same extent when processing in vivo was blocked by CCCP. Furthermore, translation of RNA in a reticulocyte lysate in vitro indicated that the messenger RNAs for both proteins were present in the mutant and translated with equal efficiency. The import into isolated mitochondria of the precursor forms of the iron-sulfur protein synthesized in the cell-free system was also decreased in the mutant mitochondria. In addition, the precursor form was bound to the exterior of the mitochondrial membrane where it was sensitive to digestion with proteases. By contrast, the synthesis and processing of cytochrome c1 appeared to be unaffected in these mutants. These results suggest that cytochrome b is necessary for the proper processing and assembly of both core protein I and the iron-sulfur protein, but not for cytochrome c1, into complex III of the inner mitochondrial membrane.     

Evidence of a novel dipeptidyl aminopeptidase in mammalian GH(3) cells: new insights into the processing of peptide hormone precursors

We investigated whether yeast signals could regulate hormone processing in mammalian cells. Chmeric genes coding for the prepro region of yeast alpha-factor and the functional hormone region of anglerfish somatostatin was expressed in rat pituitary GH(3) cells. The nascent prepro-alpha-factor-somatostatin peptides disappeared from cells with a half-life of 30 min, and about 20% of unprocessed precursors remained intracellular after a 2 h chase period. Disappearance of propeptide was insensitive to lysosomotropic agents, but was inhibited at 15 degrees C or 20 degrees C, suggesting that the hybrid propeptides were not degraded in the secretory pathway to the trans Golgi network or in lysosomes. It appeared that while most unprocessed precursors were constitutively secreted into the medium, a small portion were processed at their paired dibasic sites by prohormone-processing enzymes located in trans Golgi network/secretory vesicles, resulting in the production of mature somatostatin peptides. To test this hypothesis, we investigated the processing pattern of two different hybrid precursors: the 52-1 hybrid precursor, which has a Glu-Ala spacer between the prepro region of alpha-factor and somatostatin, and the 58-1 hybrid precursor, which lacks the Glu-Ala spacer. Processing of metabolically labeled hybrid propeptides to smaller somatostatin peptides was assessed by HPLC. When pulse-labeled cells were chased for up to 2 h, 68% of the initially synthesized propeptides were secreted constitutively. About 22% of somatostatin-related products were proteolytically processed to mature somatostatin, of which 38.7% were detected intracellularly after 2 h. From N-terminal peptide sequence determination of somatostatin-related products in GH(3)-52 and GH(3)-58 cells, we found that both hybrid precursors were accurately cleaved at their dibasic amino acid sites. Notably, we also observed that the Glu-Ala spacer sequence was removed from 52-1 hybrid precursors. The latter result strongly suggests that a novel dipeptidyl aminopeptidase activity - a yeast STE13-like enzyme - is present in the post-trans Golgi network compartment of GH(3) cells. The data from these studies indicate that mechanisms which control protein secretion are conserved between yeast and mammalian cells.     

The MAS-encoded processing protease of yeast mitochondria. Interaction of the purified enzyme with signal peptides and a purified precursor protein

The matrix of yeast mitochondria contains a chelator-sensitive protease that removes matrix-targeting signals from most precursor proteins transported into this compartment. The enzyme consists of two nonidentical subunits that are encoded by the nuclear genes MAS1 and MAS2. With the aid of these cloned genes, we have now overexpressed the active holoenzyme in yeast, purified it in milligram amounts, and studied its biochemical and physical properties. Atomic absorption analysis shows that the purified enzyme lacks significant amounts of zinc, manganese, or cobalt; if none of these metal ions is added during the assay, the enzyme is catalytically inactive but can still cleave substoichiometric amounts of substrate. The amino-terminal sequences of the two mature subunits were determined; comparison with the deduced amino acid sequences of the corresponding precursors revealed that the MAS1 and MAS2 subunits are synthesized with prepeptides composed of 19 and 13 residues, respectively, which have similar sequences. The enzyme is inhibited competitively by chemically synthesized matrix-targeting peptides; the degree of inhibition correlates with the peptides' targeting efficacy. Matrix-targeting peptides containing the cleavage site of the corresponding authentic precursor protein are cleaved correctly by the purified enzyme. A purified artificial precursor protein bound to the holoenzyme can be photocross-linked to the MAS2 subunit.     

Transport of proteins across the mitochondrial outer membrane. A precursor form of the cytoplasmically made intermembrane enzyme cytochrome c peroxidase.

Cytochrome c peroxidase, a cytoplasmically made enzyme located between the inner and outer membrane of yeast mitochondria, is synthesized as larger precursor in a reticulocyte cell-free lysate as well as in pulsed yeast spheroplasts. When the pulsed spheroplasts are chased, the precursor is converted to the mature apoprotein. When the in vitro synthesized precursor is incubated with isolated yeast mitochondria in the absence of protein synthesis, it is cleaved to the mature form; the mature form co-sediments with the mitochondria and is resistant to externally added proteases. These results, in conjunction with those reported earlier (Maccecchini, M.-L., Rudin, Y., Blobel, G., and Schatz, G. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 343-347) suggest that the mechanism of protein transport into the mitochondrial intermembrane space is quite similar to that of protein transport into the matrix or the inner membrane.     

Transcription and processing of intervening sequences in yeast tRNA genes.

Genes for yeast tRNATyr and tRNAPhe have been sequenced (Goodman, Olson and Hall, 1977; Valenzuela et al., 1978) which contain additional nucleotides (intervening sequences) within the middle of the gene that are not present in the mature tRNA. We have isolated precursors to rRNATyr and tRNAPhe from a yeast temperature-sensitive mutant (at the rna1 locus) which accumulates only certain precursor tRNAs at the nonpermissive temperature. The tRNATyr and tRNAPhe precursors were analyzed by oligonucleotide mapping; they each contain the intervening sequence and fully matured 5' and 3' termini. Furthermore, these precursors were used as substrates to search for an enzymatic activity which can remove the intervening sequences and religate the ends. We have shown that wild-type yeast contains such an activity, and that this activity specifically removes the intervening sequences to produce mature-sized RNAs.     

RNA quality control: degradation of defective transfer RNA

The distinction between stable (tRNA and rRNA) and unstable (mRNA) RNA has been considered an important feature of bacterial RNA metabolism. One factor thought to contribute to the difference between these RNA populations is polyadenylation, which promotes degradation of unstable RNA. However, the recent discovery that polyadenylation also occurs on stable RNA led us to examine whether poly(A) might serve as a signal for eliminating defective stable RNAs, and thus play a role in RNA quality control. Here we show that a readily denaturable, mutant tRNA(Trp) does not accumulate to normal levels in Escherichia coli because its precursor is rapidly degraded. Degradation is largely dependent on polyadenylation of the precursor by poly(A) polymerase and on its removal by polynucleotide phosphorylase. Thus, in the absence of these two enzymes large amounts of tRNA(Trp) precursor accumulate. We propose that defective stable RNA precursors that are poorly converted to their mature forms may be polyadenylated and subsequently degraded. These data indicate that quality control of stable RNA metabolism in many ways resembles normal turnover of unstable RNA.     

Presence of the cytosolic factor stimulating the import of precursor of mitochondrial proteins in rabbit reticulocytes and rat liver cells

Previously we purified a cytosolic factor that stimulates the import of the extrapeptide (the synthetic peptide of the presequence of ornithine aminotransferase) into the mitochondrial matrix (Ono, H., and Tuboi, S., 1988, J. Biol. Chem. 263, 3188-3193). In this work this cytosolic factor was shown also to stimulate the import of the precursors of ornithine aminotransferase, a large subunit of succinate dehydrogenase, and sulfite oxidase. The amounts of these precursors bound to the outer mitochondrial membrane were increased by this cytosolic factor, suggesting that the cytosolic factor participates in the recognition step in the import process of the precursor protein. When the cytosolic factor was applied to an ATP-agarose column, the import-stimulating activity was recovered entirely in the unadsorbed fraction. Immunochemical studies showed that in these conditions the 70-kDa heat shock-related protein (Hsp 70) was present exclusively in the fraction adsorbed to the ATP-agarose column. The cytosolic factor is thus different from the 70-kDa heat shock-related protein, which was identified as a factor required for the import of mitochondrial proteins in yeast. The cytosolic factor was also detected in the cytosol of rat liver cells, and a considerable amount of this factor was recovered from rat liver mitochondria by washing them with high salt buffer, suggesting that the cytosolic factor has affinity to the outer mitochondrial membrane and binds to its receptor on the membrane. From these results, we conclude that the cytosolic factor forms a complex with the precursor of mitochondrial protein and then this complex binds to the outer mitochondrial membrane, probably via the receptor of the cytosolic factor.     

A yeast mutant temperature-sensitive for mitochondrial assembly is deficient in a mitochondrial protease activity that cleaves imported precursor polypeptides.

We have previously described two yeast mutants which, at elevated temperature, stop growing and accumulate precursors to several imported mitochondrial proteins. We now show that one of these mutants (mas 1) is deficient in a matrix-located protease activity which cleaves the pre-sequences from mitochondrial precursor proteins. Isolated mas 1 mitochondria catalyze oxidative phosphorylation, exhibit respiratory control and import mitochondrial precursor polypeptides, but are defective in removing transient pre-sequences from imported precursors. The phenotype of the mas 1 mutant suggests that the matrix-located processing protease is essential for growth and for mitochondrial assembly.     

A precursor to a minor species of yeast tRNASer contains an intervening sequence.

Certain tRNAs in S. cerevisiae (tRNATyr and tRNAPhe) arise via precursor molecules which are mature at the 5' and 3' termini but contain intervening sequences adjacent to the anticodon (Knapp et al., 1978; O'Farrell et al., 1978). In addition to these molecules, precursors to several other tRNAs accumulate in a temperature-sensitive mutant (ts136) at the nonpermissive temperature. We have analyzed one of these species and shown that it is a precursor to a minor species of tRNASer. This precursor is also mature at both termini and contains an intervening sequence of 19 nucleotides adjacent to the hypermodified A residue 3' to the anticodon. The sequence can be arranged in a secondary structure in which the anticodon stem is extended by additional base-pairing, and contains the sites of excision and ligation within two looped regions. Support for this structure was provided by analysis of the products of limited digestion with RNAase T1. recently Piper (1978) reported the isolation of a minor species of tRNASer which decodes UCG. He found this species to be structurally heterogeneous and determined that the less abundant form corresponds to the tRNA which is altered in the recessive lethal SUP-RL1 amber suppressor. Our data now suggest that the more abundant form may be restricted to reading UCA in vivo; thus mutation of the minor species would result in complete loss of UCG-decoding ability and explain the recessive lethality of SUP-RL1. We have shown that the precursor which accumulates in ts136 corresponds exclusively to this minor tRNASerUCG species. Our results suggest that this may be the only gene for tRNASer in yeast which contains an intervening sequence.     

Processing of a wheat light-harvesting chlorophyll a/b protein precursor by a soluble enzyme from higher plant chloroplasts

A processing activity has been identified in higher plant chloroplasts that cleaves the precursor of the light-harvesting chlorophyll a/b- binding protein (LHCP). A wheat LHCP gene previously characterized (Lamppa, G.K., G. Morelli, and N.-H. Chua, 1985. Mol. Cell Biol. 5:1370- 1378) was used to synthesize RNA and subsequently the labeled precursor polypeptide in vitro. Incubation of the LHCP precursors with a soluble extract from lysed chloroplasts, after removal of the thylakoids and membrane vesicles, resulted in the release of a single 25-kD peptide. In contrast, when the LHCP precursors were used in an import reaction with intact pea or wheat chloroplasts, two forms (25 and 26 kD) of mature LHCP were found. The peptide released by the processing activity in the organelle-free assay comigrated with the lower molecular mass form of mature LHCP produced during import. Properties of the processing activity suggest that it is an endopeptidase. Chloroplasts from both pea and wheat, two divergent higher plants, contain the processing enzyme, suggesting its physiological importance in LHCP assembly into the thylakoids. We discuss the implications of LHCP precursor processing by a soluble enzyme that may be in the stromal compartment.     

Import of cytochrome b2 to the mitochondrial intermembrane space: the tightly folded heme-binding domain makes import dependent upon matrix ATP.

Cytochrome b2 is synthesized as a precursor in the cytoplasm and imported to the intermembrane space of yeast mitochondria. We show here that the precursor contains a tightly folded heme-binding domain and that translocation of this domain across the outer membrane requires ATP. Surprisingly, it is ATP in the mitochondrial matrix rather than external ATP that drives import of the heme-binding domain. When the folded structure of the heme-binding domain is disrupted by mutation or by urea denaturation, import and correct processing take place in ATP-depleted mitochondria. These results indicate that (1) cytochrome b2 reaches the intermembrane space without completely crossing the inner membrane, and (2) some precursors fold outside the mitochondria but remain translocation-competent, and import of these precursors in vitro does not require ATP-dependent cytosolic chaperone proteins.     

Processing of precursor interleukin 1 beta and inflammatory disease

The processing of precursor interleukin 1 beta (IL1 beta) by elastase, cathepsin G, and collagenase, the major proteases released at sites of inflammation, was investigated using recombinant pro-IL1 beta. Each of these proteases cleaved the 31-kDa inactive precursor to a form similar in size and specific activity (greater than 10(8) units/mg) to the 17-kDa mature protein isolated from activated monocytes. Elastase, collagenase, and cathepsin G cleaved the IL1 beta precursor at distinct sites which are amino-terminal to the monocyte-processing site, Ala-117 (Cameron, P., Lumjuco, G., Rodkey, J., Bennett, C., and Schmidt, J. A. (1985) J. Exp. Med. 162, 790-801). Amino-terminal sequencing of the products of digestion by elastase and cathepsin G determined that resultant active IL1 beta proteins contained an additional 13 or 3 amino acids relative to mature IL1 beta. Synovial fluid collected from patients with inflammatory polyarthritis and bronchoalveolar lavage fluid from patients with sarcoidosis supplied similar processing activity(s). Control fluids from patients who had no symptoms of inflammatory disease did not exhibit processing activity. Lavage fluids that processed precursor IL1 beta were demonstrated to contain cathepsin G and/or elastase activity, whereas controls were negative. Because a significant fraction of IL1 beta may be secreted from monocytes as the inactive 31-kDa precursor (Hazuda, D. J., Lee, J. C., and Young, P. R. (1988) J. Biol. Chem. 263, 8473-8479, Bomford, R., Absull, E., Hughes-Jenkins, C., Simpkin, D., and Schmidt, J. (1987) Immunology 62, 543-549, and Mizel, S. B. (1988) in Cellular and Molecular Aspects of Inflammation Poste, G., and Crooke, S., eds) pp. 75-93, Plenum Publishing Corp., New York), these results suggest that in vivo the IL1 beta precursor can be processed after secretion by any of several proteases released at inflammatory sites.     

Secondary methylation of yeast ribosomal precursor RNA.

The timing of methylation of the ribosomal sequences of ribosomal precursor RNA (pre-rRNA) from the yeast Saccharomyces carlsbergensis was investigated by fingerprint analysis of the methylated oligonucleotides derived from the various precursors. From the total of 37 ribose and 6 base-methyl groups found in 26-S rRNA, the two copies of the base-methylated nucleoside m3U as well as the doubly methylated sequence Um-Gm psi are not yet present in 37-S RNA, the predominant common precursor of 26-S and 17-S rRNA. Introduction of these methyl groups into the ribosomal sequences appears to take place at the level of 29-S pre-rRNA, the immediate precursor to 26-S rRNA. From the total of 18 ribose-methylated and 6 base-methylated nucleosides found in 17-S rRNA, the latter group (one copy of m7G, the m62A-m62A- sequence and the hypermodified methylated nucleoside "mX") is completely missing in 37-S pre-rRNA. The methyl group of m7G is introduced into 18-S pre-rRNA, the direct precursor of 17-S rRNA, in the nucleus. The -m62A-m62A- sequence is methylated after transport of the 18-S pre-rRNA to the cytoplasm prior to the final maturation into 17-S rRNA.     

Isolation and characterization of the Schizosaccharomyces pombe cDNA encoding the mitochondrial endonuclease(1)

We isolated the cDNA of the fission yeast mitochondrial endonuclease SpNUC1, which consists of 322 amino acids and has a significant homology with the budding yeast NUC1 and mammalian endonuclease G. Comparison of the cDNA sequence with the genomic sequence showed that the gene consists of three exons and two introns and spans 1.31 kb. The enzyme localization in mitochondria was demonstrated by expressing the SpNUC1-green fluorescent protein fusion in the yeast. The endonuclease was activated by truncation of the amino-terminal region of the protein, indicating that the enzyme is encoded as an inactive precursor. The active enzyme degraded single-stranded DNA and RNA, the activity being dependent on Mg(2+) (Mn(2+)).     

Chemical composition of the yeast ascospore wall. The second outer layer consists of chitosan

In a preceding paper (Briza, P., Winkler, G., Kalchhauser, H., and Breitenbach, M. (1986) J. Biol. Chem. 261, 4288-4294), we reported the presence of dityrosine in the outer layers of yeast ascospore walls. Both outer layers seen in electron micrographs of yeast ascospore walls are sporulation-specific. Here we show that the second of these two outer layers consists of chitosan. In intact spores, it is shielded from staining with primulin by the outermost layer. However, in purified spore walls, the second layer is brightly stained by primulin, and hydrolysates of such preparations contain about 10% glucosamine relative to spore wall dry weight. The spore wall material staining with primulin is resistant to chitinase, but readily degraded by treatment with HNO2. Acetylation prior to HNO2 treatment completely prevents its degradation. A partial acid hydrolysate of spore walls contains predominantly soluble poly-beta-(1,4)-glucosamine as determined by 13C NMR spectroscopy. By these criteria, the glucosamine polymer of yeast ascospore walls is chitosan. As spore walls treated with alkali lack the inner layers but contain chitosan and as chitosan is not exposed at the surface of the spore, we conclude that it is localized in the second outer layer of the spore wall.     

Transport of proteins to the mitochondrial intermembrane space: the ''sorting'' domain of the cytochrome c1 presequence is a stop-transfer sequence specific for the mitochondrial inner membrane.

The presequence of yeast cytochrome c1 (an inner membrane protein protruding into the intermembrane space) contains a matrix-targeting domain and an intramitochondrial sorting domain. This presequence transports attached subunit IV of cytochrome c oxidase into the intermembrane space (van Loon et al. (1987) EMBO J., 6, 2433-2439). In order to determine how this fusion protein reaches the intermembrane space, we studied the kinetics of its import into isolated mitochondria or mitoplasts and its accumulation in the various submitochondrial compartments. The imported, uncleaved fusion precursor and a cleavage intermediate were bound to the inner membrane and were always exposed to the intermembrane space; they were never found at the matrix side of the inner membrane. In contrast, analogous import experiments with the authentic subunit IV precursor, or the precursor of the iron-sulphur protein of the cytochrome bc1 complex also an inner membrane protein exposed to the intermembrane space), readily showed that these precursors were initially transported across both mitochondrial membranes. We conclude that the intramitochondrial sorting domain within the cytochrome c1 presequence prevents transport of attached proteins across the inner, but not the outer membrane: it is a stop-transfer sequence for the inner membrane. Since the presequence of the iron-sulphur protein lacks such 'stop-transfer' domain, it acts by a different mechanism.