Influence of the mature portion of a precursor protein on the mitochondrial signal sequence.

Most mitochondrial proteins are synthesized with an N-terminal signal sequence that targets these proteins to various compartments within the mitochondria. Signal sequences have been shown to be functional by fusing them to a nonmitochondrial passenger protein and observing import. In many cases, a signal sequence has been fused to passenger proteins, such as dihydrofolate reductase, and import occurred. There are, though, several unexplained instances in which a signal sequence was attached to a passenger protein and import was not observed. In this study, the N-terminal 23 residues of the matrix enzyme rhodanese could import several passenger proteins but were unable to import the mature form of mitochondrial aldehyde dehydrogenase (mALDH). However, if these same 23 residues were fused to the middle portion of mALDH, import was recovered, suggesting that the rhodanese signal sequence and N terminus of mALDH were incompatible for import. Circular dichroism data indicated that a peptide corresponding to the region of fusion between rhodanese and mALDH had less structure than corresponding peptides from imported fusion proteins, suggesting that mALDH may alter the helix in the rhodanese signal sequence, thus preventing import.

Most mitochondrial proteins are synthesized with an N-terminal signal sequence that targets these proteins to various compartments within the mitochondria. Signal sequences have been shown to be functional by fusing them to a nonmitochondrial passenger protein and observing import. In many cases, a signal sequence has been fused to passenger proteins, such as dihydrofolate reductase, and import occurred. There are, though, several unexplained instances in which a signal sequence was attached to a passenger protein and import was not observed. In this study, the N-terminal 23 residues of the matrix enzyme rhodanese could import several passenger proteins but were unable to import the mature form of mitochondrial aldehyde dehydrogenase (mALDH). However, if these same 23 residues were fused to the middle portion of mALDH, import was recovered, suggesting that the rhodanese signal sequence and N terminus of mALDH were incompatible for import. Circular dichroism data indicated that a peptide corresponding to the region of fusion between rhodanese and mALDH had less structure than corresponding peptides from imported fusion proteins, suggesting that mALDH may alter the helix in the rhodanese signal sequence, thus preventing import.
The majority of mitochondrial proteins are encoded in the nucleus and translated in the cytoplasm with an N-terminal signal sequence necessary for mitochondrial import (1,2). It is believed that the signal sequence binds to various cytosolic factors that maintain the precursors in a loosely folded conformation necessary for translocation through the mitochondrial outer and inner membranes (3)(4)(5). Following import into the matrix space, the signal sequence is removed by the mitochondrial processing peptidase (6,7).
Signal sequences are rich in positively charged residues and have the theoretical ability to form an amphiphilic ␣-helix (8,9). A more detailed structural evaluation of four cleavable and five noncleavable signal sequences by two-dimensional NMR has revealed that these sequences can form a stable N-terminal helix in a micellar environment (10 -16). Additionally, removal of this helix from several precursors abolished import, demonstrating the importance of this helix (17,18).
The identification and role of the N-terminal signal sequence for mitochondrial import have been based mainly on a number of experiments with fusion proteins consisting of a signal sequence joined to a nonmitochondrial passenger protein. The first set of such experiments involved the fusion of the signal sequence of the cytochrome c oxidase subunit IV, an inner membrane protein, to dihydrofolate reductase (DHFR). 1 The resulting fusion protein was imported into the matrix and processed by the mitochondrial processing peptidase (18). Additionally, the 32-amino acid signal sequence from the matrix enzyme ornithine transcarbamoylase was also fused to DHFR, which resulted in import (19). However, there are several unexplained examples in which fusion of a signal sequence to a passenger protein did not result in import. For instance, a fusion protein consisting of the signal sequence from the matrix enzyme manganese superoxide dismutase attached to DHFR was imported, whereas fusion of the same signal sequence to the cytosolic protein invertase did not result in import (20). Also, fusion of the signal sequence of the ␦-subunit of mitochondrial F 1 -ATPase to bacterial ␤-glucuronidase did not allow for import unless 27 residues of the mature F 1 -ATPase were included in the fusion (21). A similar situation occurred when the 20-residue F 1 -ATPase ␤-subunit signal sequence was fused to ␤-galactosidase. Import did not occur unless an additional 147 residues from the mature form of the F 1 -ATPase ␤-subunit were fused to this protein. However, this 20-residue signal sequence could direct invertase to the mitochondria (22).
We constructed several fusion proteins consisting of the signal sequence of rhodanese, a nonprocessed matrix enzyme, attached to various passenger proteins. While the N-terminal 23 residues of rhodanese could direct several different passengers to the mitochondria, these residues did not allow for import of the mature form of the mitochondrial rat liver aldehyde dehydrogenase (ALDH). Although several possible explanations exist for the lack of import, we propose that a likely explanation was that the passenger protein altered the secondary structure of the signal sequence, thereby abolishing import.

EXPERIMENTAL PROCEDURES
Construction of Chimeric Proteins-The various constructs used in these studies are illustrated in Fig. 1. cDNA encoding rhodanese had been cloned into the plasmid pT7-7 as described previously (23). ⌬51-Rhod and ⌬27-Rhod were constructed by amplifying residues 52 to the stop codon of rhodanese or residues 28 to the stop codon of rhodanese with flanking 5Ј-NdeI and 3Ј-BamHI sites using the polymerase chain reaction (PCR) and rhodanese as a template. The oligonucleotides used for ⌬51-Rhod were 5Ј-TTTCATATGCACGTGCCTGGCGCGTCCTTC-3Ј and 5Ј-TTTGGATCCTCAGGCCTTCCCACTCTTCCC-3Ј. The oligonucleotides used for ⌬27-Rhod were 5Ј-TTTCATATGAGCCTTCGGGT-GCTGGACGC-3Ј and 5Ј-TTTGGATCCTCAGGCCTTCCCACTCT-TCCC-3Ј. The resulting PCR products were digested with NdeI and BamHI and cloned into pT7-7. 55 Rhod/mALDH was constructed by amplifying the entire pT7-7 vector and cDNA encoding the N-terminal * This work was supported in part by Grant AA05812 from the National Institute on Alcohol Abuse and Alcoholism. This is Journal Paper 15100 from the Purdue University Agricultural Experiment Station. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Transcription and Translation-In vitro transcription and translation were performed using the TNT T7 Coupled Reticulocyte Lysate System according to manufacturer's instructions (Promega) with [ 35 S]methionine (Amersham Life Science, Inc.) as the labeled amino acid.
In Vitro Import of Precursor Proteins into Isolated Mitochondria-Rat liver mitochondrial isolation and in vitro import were performed as described previously (23)(24)(25). Briefly, 2-4 l of translated protein in reticulocyte lysate was incubated with 25 l of isolated mitochondria (7 mg of protein/ml) in import buffer (24) with a final volume of 100 l at 30°C for 30 min. Subsequently, half of the reaction mixture was separated into a mitochondrial pellet and supernatant by centrifugation at 10,000 ϫ g for 5 min. The mitochondrial pellet was resuspended in import buffer and analyzed by SDS-polyacrylamide gel electrophoresis. This represented protein associated with the mitochondria (ϪPK lanes in Fig. 2). The other half of the import reaction was treated with 4 l of proteinase K (2 mg/ml) (Sigma) at 0°C for 15 min to digest protein that was not imported. Mitochondria were reisolated using centrifugation at 10,000 ϫ g and treated with phenylmethylsulfonyl fluoride (2 l of a 200 mM solution) to stop the proteinase K reaction. Subsequently, samples were subjected to SDS-polyacrylamide gel electrophoresis. Import was quantitated using a PhosphorImager (Bio-Rad).
Urea Denaturation of Proteins Translated in Rabbit Reticulocyte Lysate-Protein translated in rabbit reticulocyte lysate was denatured in urea as described previously (26). Protein was translated in rabbit reticulocyte lysate for 1 h. Subsequently, 1 volume of saturated (NH 4 ) 2 SO 4 was added to the lysate. After a 30-min incubation on ice, precipitated protein was collected by centrifugation at 15,000 ϫ g for 10 min. Following centrifugation, the pellet was resuspended in a ureacontaining buffer (8 M urea, 20 mM Tris-HCl, pH 7.6, 20 mM dithiothreitol). Ten l of the urea-denatured protein was subsequently used in import reactions as described above.
Circular Dichroism-Circular dichroism spectra were obtained on a Jasco J-600 spectropolarimeter (28). The samples were typically scanned from 350 to 290 nm at 25°C with a path length of 0.1 cm. The buffer was 50 mM phosphate, pH 5.2. For spectra that included TFE, the buffer was mixed with TFE in a 4:1 proportion (v/v) to give a 20% TFE solution. Base-line spectra for each solvent were obtained prior to those for the peptide spectra. Peptide concentrations were determined from the UV absorbance of the tryptophan residue and for CD spectra, they were in a range of 20 -25 M. Helical content was estimated using a nonlinear least squares fit of the experimental spectra to standard values derived from a data base of protein CD spectra (29).
Miscellaneous-Isoelectric points were calculated using Gene Runner 3.04. The PCR reagents were purchased from Perkin-Elmer. Restriction enzymes and T4 DNA ligase were obtained from New England Biolabs Inc. DNA sequencing was performed using Sequenase 2.0 from U. S. Biochemical Corp. (30). The plasmid pGEM-7zf was purchased from Promega. SDS-polyacrylamide gel electrophoresis was run according to Laemmli (31).

RESULTS
The N Terminus of Rhodanese Contains Information Necessary for Import-Rhodanese is a mitochondrial matrix enzyme that possesses a noncleaved signal sequence (23,32). The crystal structure of rhodanese revealed that residues 11-22 form an ␣-helix (33), and two-dimensional NMR showed that residues 4 -21 in a peptide corresponding to the N-terminal 23 residues of rhodanese can form an ␣-helix in a micellar environment (28). To determine if these N-terminal residues were sufficient for import and that the targeting information did not lie in other regions of the protein, the N-terminal 51 and 27 residues of rhodanese were removed (referred to as ⌬51-Rhod and ⌬27-Rhod, respectively), and import of these truncated proteins was examined. As predicted, deletion of this N-terminal region from rhodanese abolished import in both cases (data not shown).
To determine if the N-terminal 23 residues of rhodanese would allow import to occur, residues 24 -56 were deleted within rhodanese, referred to as ⌬24 -56-Rhod in Fig. 1. As was expected, ⌬24 -56-Rhod was imported as we showed previously (23), demonstrating that the N terminus of rhodanese allowed for import to occur (Fig. 2).
To further demonstrate that the N-terminal residues of rhodanese were essential for import, several fusion proteins were constructed to determine if the N terminus of rhodanese could direct a passenger protein to the mitochondria. The N-terminal 55, 26, or 23 residues of rhodanese were fused to the mature form of the mitochondrial rat liver ALDH (referred to as 55 Rhod/mALDH, 26 Rhod/mALDH, and 23 Rhod/mALDH in Fig. 1, respectively), or the 23 N-terminal residues of rhodanese were fused to DHFR (referred to as 23 Rhod/DHFR in Fig. 1). After incubation with isolated rat liver mitochondria, 55 Rhod/ mALDH and 23 Rhod/DHFR were imported as shown in Fig. 2.
Surprisingly, 23 Rhod/mALDH and 26 Rhod/mALDH were poorly imported. Full-length rat liver rhodanese has 297 amino acids, whereas ALDH has 500 amino acids. The possibility existed that the initial 23 or 26 residues of rhodanese were incapable of importing a protein the size of ALDH. Therefore, approximately 150 C-terminal residues were removed to truncate 26 Rhod/mALDH (referred to as 26 Rhod/350 ALDH in Fig. 1). The remaining 350 residues of mALDH approximate the size of rhodanese. However, import of this truncated ALDH fusion protein still did not occur.
Since structural data from crystallography and two-dimensional NMR indicated an ␣-helix existed through the initial 22 residues of rhodanese and since ⌬24 -56-Rhod and 23 Rhod/ DHFR were imported, it was expected that these residues should be sufficient to import mALDH. In fact, it was possible to fuse as few as 17 N-terminal residues of rhodanese to DHFR and observe import (Fig. 2). Thus, it appeared that fusion of rhodanese to mALDH made the signal sequence incompatible with import function.
To test whether the N terminus of mALDH adversely affected the rhodanese signal sequence in 23 Rhod/mALDH, this signal sequence was fused instead, for the convenience of creating a new restriction site, to residue 117 of mALDH (referred to as 23 Rhod/117 ALDH). Import was restored, showing that the 23 N-terminal residues of rhodanese could import the Nterminal truncated ALDH. To further show that these residues allowed import of ALDH, 9 C-terminal residues from the AS were included in the 23 Rhod/mALDH fusion to separate the rhodanese signal sequence from the N terminus of mALDH (referred to as 23 Rhod/9 AS/mALDH). As shown with 23 Rhod/ 117 ALDH, import was again recovered (Fig. 2). The amount of protein imported for each of the constructs listed above was calculated and is listed in Table I.
Misfolding May Not Be a Cause for Lack of Import-Since it appeared that some structural element existed at the area of fusion between the signal sequence of rhodanese and N-terminal mALDH that interfered with import, we examined the amino acid sequence in this region. A glycine residue was found to exist at the area of fusion between rhodanese and mALDH in both 23 Rhod/mALDH and 26 Rhod/mALDH. It was possible that these fusion proteins misfolded or that the presence of glycine may have allowed the signal sequence to be flexible so that it could interact with the passenger protein. To examine this, 8 M urea was added to 23 Rhod/mALDH and precursor ALDH after translation in rabbit reticulocyte lysate in an attempt to unfold these proteins. Subsequently, the urea-denatured protein was diluted into mitochondrial import reactions. Urea-denatured native pALDH was imported, while 23 Rhod/ mALDH was still not imported (data not shown). Therefore, misfolding did not appear to be the cause for the lack of import of 23 Rhod/mALDH.

The Region of Fusion between Rhodanese and mALDH Possesses Less Structure than Other Regions of Fusion-Since 23
Rhod/DHFR, 23 Rhod/117 ALDH and ⌬24 -56-Rhod were imported, whereas 23 Rhod/mALDH was not imported, it seemed possible that mALDH affected the structure of the rhodanese signal sequence. To test for the influence of the first few residues of mALDH on the secondary structure of the signal sequence, CD was performed on three different 22-residue peptides. The peptides represented the following: Rhod-mALDH, containing the region spanning residues 12-33 from the fusion of rhodanese with mALDH; Rhod-DHFR, containing the region spanning residues 12-33 from the fusion of rhodanese with  2. The import of the fusion proteins into rat liver mitochondria. Each of the fusion proteins was translated in rabbit reticulocyte lysate (Tx) and was incubated with isolated rat liver mitochondria for 30 min at 30°C. Half of the reaction was treated with proteinase K (ϩPK) and half of the reaction was left untreated (ϪPK). Imported protein is represented by the protein remaining after the mitochondria is treated with proteinase K, which destroys translated protein outside the mitochondria. Proteins were analyzed using SDSpolyacrylamide gel electrophoresis and phosphoimaging.

TABLE I Import of fusion proteins into mitochondria
Proteins translated in rabbit reticulocyte lysate were incubated with isolated rat liver mitochondria for 30 min, as shown in Fig. 2. The amount of protein imported into mitochondria was calculated by dividing the amount of protein remaining after proteinase K digestion by the total amount of protein added to the reaction. The amount of rhodanese imported was arbitrarily set at 100%, and the amount of import of the various constructs was compared with rhodanese. Data are presented as the average of four different experiments Ϯ 1 S.D. from the mean value. DHFR; and Rhod-Rhod, containing the region spanning residues 12-33 of rhodanese (Fig. 3).
As typically seen, the CD spectra showed little evidence of secondary structure when obtained in aqueous buffer alone (14). When 20% (v/v) TFE was used to induce the formation of secondary structure in the three peptides, differences were observed in the CD spectra (Fig. 4). In 20% TFE, the Rhod-mALDH peptide was estimated to be only 5% helical while the Rhod-Rhod and Rhod-DHFR peptides were estimated to be about 20% helical. These results indicate that the normal tendency of a portion of rhodanese to form a helical structure was disrupted by the residues of mALDH.

DISCUSSION
It is well established that N-terminal regions of precursor proteins carry information essential for mitochondrial targeting since hybrid proteins, which each consist of a fusion of a signal sequence to a nonmitochondrial passenger protein, are imported (18,19). However, there is increasing evidence that regions of the mature or passenger protein are also involved in targeting and import.
The mature portion of a precursor protein may be involved in many of the different interactions a precursor protein goes through on its route to its final destination (1,2). For instance, it is important that a protein remain unfolded for import to occur. If the mature protein folds too rapidly, import may be hampered. This appears to be a likely possibility since it was found that point mutations, which partially unfolded the passenger of a fusion protein consisting of the first 16 residues of yeast cytochrome oxidase subunit IV attached to DHFR, increased import efficiency. It was shown that an increased import rate correlated with destabilization of DHFR (34).
Additionally, the mature region of a precursor may play a role in binding to heat shock proteins that maintain the protein in a translocation-competent state. For example, both cytosolic and mitochondrial forms of aspartate aminotransferase exist. Since only the mitochondrial form interacted with hsp 70, which maintains proteins in their unfolded state, the cytosolic form folded much more rapidly than did the mitochondrial form in rabbit reticulocyte lysate. Deletion of residues from the N terminus of the mature form of mitochondrial aspartate aminotransferase eliminated the interaction of this protein with hsp 70, demonstrating the importance of the mature protein for interactions with anti-folding proteins (35).
The characteristics of amino acids within the mature region of a protein may affect the ability of a precursor to interact with import receptors and the membrane. Although positive charges are important in the signal sequence, positive charges in the mature region of mitochondrial proteins may also be important for import. Such charged residues may potentially interact with import receptors or negatively charged regions of the mitochondrial membranes. When a pairwise comparison of homologous cytosolic and mitochondrial proteins was performed, it was found that in 75% of the cases the pI of the mature form of the mitochondrial protein was higher than that of the corresponding cytosolic protein (36). It was argued that a more acidic mature protein required a more basic signal sequence (36).
The calculated pI, which approximates the value expected for the unfolded state of a protein, was determined for several proteins used in this study. It is interesting to note that the calculated pI of mALDH is 5.77, whereas the calculated pI of rhodanese is 7.82. The signal sequence of precursor pALDH has five positive charges, but the initial 23 residues of rhodanese only have three positive charges. The mature form of ALDH may require a more basic signal sequence for import to occur. In support of this idea, mutation of two arginine residues to glutamine in the ALDH signal sequence, which left only three remaining positive charges, drastically reduced import (37). Perhaps the less highly charged rhodanese signal sequence is better suited to import a more basic protein, and it is possible that the rhodanese signal sequence did not have enough of a positive charge to allow for the import of ALDH. Additionally, the calculated pI of DHFR, which has been found to serve as a good passenger protein in many studies, is 8.55. Perhaps DHFR is an ideal passenger because it is a basic protein.
The hydrophobicity of a passenger protein may also influence import. For instance, if the signal sequence of the ATPase subunit 9 was fused to bI4 RNA maturase, import occurred. FIG. 3. Peptides corresponding to the region of fusion of the signal sequence and passenger proteins of 23 Rhod/mALDH and 23 Rhod/DHFR. The amino acid sequences of peptides corresponding to the region of fusion between rhodanese and mALDH or between rhodanese and DHFR were used for CD and are shown. Additionally, as a control, a peptide corresponding to residues 12-33 of rhodanese was also used for CD. Even though rhodanese is not processed, the region of the protein following the 23 amino acid N-terminal signal sequence is designated as the "mature" protein.
FIG. 4. Circular dichroism spectra of peptides. CD spectra with mean residue ellipticity (MRE) of peptides corresponding to the natural rhodanese sequence (Rhod-Rhod) and the rhodanese-ALDH (Rhod-mALDH) and rhodanese-DHFR (Rhod-DHFR) fusion sites are shown. These spectra show that, under identical conditions, the Rhod-ALDH peptide displayed a significantly lower tendency than did the other two peptides to form secondary structure. Spectra were obtained in solutions that contained 20% (v/v) TFE and 50 mM phosphate buffer, pH 5.2, at 25°C. However, if hydrophobic stretches of amino acids derived from apocytochrome c were included between the signal sequence and maturase, import was drastically reduced. This reduction in import correlated with the number of hydrophobic stretches included in the fusions (38).
We and others have demonstrated that the mature protein and signal peptide must be compatible for efficient import to occur. Fusion of the 23 and 26 N-terminal residues of rhodanese did not allow for import of mALDH. However, the same 23 residues could direct a variety of other passenger proteins to the mitochondria, including mALDH if the N terminus of mALDH was deleted or if residues of the ALDH signal sequence were included in the fusion. It appeared that the N terminus of mALDH and the rhodanese signal sequence were incompatible for import.
It is possible that if 23 Rhod/mALDH misfolded, the signal sequence would not be available for mitochondrial targeting. Alternatively, the mature protein may have altered the structure of the signal sequence. The latter situation seems likely in our case since CD data indicated that the area of fusion between the rhodanese signal sequence and mALDH had much less structure than the area of fusion between rhodanese and DHFR or the comparable region in native rhodanese. A stable N-terminal helix is essential for import (17,39), and the loss of a stable helix in the nonimported fusion proteins may have affected their import competency.
Several unexplained instances exist in which fusion of a signal sequence to a passenger protein did not result in mitochondrial import. It is possible that the structure of the signal sequence was adversely affected by the mature protein in these cases. It may prove interesting to determine the structure of the region of fusion between the signal sequence and passenger protein of these nonimported proteins to determine if the signal sequence is altered by the passenger protein.