Mutations in the passenger polypeptide can affect its partitioning between mitochondria and cytoplasm

Abstract

In this study, we report that the partitioning between mitochondria and cytoplasm of two variants, mCherry and DsRed Express (DRE), of the red fluorescent protein, DsRed, fused to one of the six matrix targeting sequences (MTSs) can be affected by both MTS and amino acid substitutions in DsRed. Of the six MTSs tested, MTSs from superoxide dismutase and DNA polymerase gamma failed to direct mCherry, but not DRE to mitochondria. By evaluating a series of chimeras between mCherry and DRE fused to the MTS of superoxide dismutase, we attribute the differences in the mitochondrial partitioning to differences in the primary amino acid sequence of the passenger polypeptide. The impairment of mitochondrial partitioning closely parallels the number of mCherry-specific mutations, and is not specific to mutations located in any particular region of the polypeptide. These observations suggest that both MTS and the passenger polypeptide affect the efficiency of mitochondrial import and provide a rationale for the observed diversity in the primary amino acid sequences of natural MTSs.

Introduction

The mitochondrial proteome has been estimated to consist of approximately 1,500 polypeptides, of which only 13 are encoded by mitochondrial DNA [1]. Therefore, almost all (>99%) mitochondrial proteins are synthesized on cytoplasmic ribosomes and subsequently imported into one of the four mitochondrial compartments (the outer membrane, the intermembrane space, the inner membrane, and the matrix). Many proteins destined for mitochondrial matrix possess a cleavable N-terminal matrix targeting sequence (MTS).

As far as mitochondrial protein import is concerned, it is widely believed that the MTS provides sufficient information to direct a protein across the outer and inner mitochondrial membranes. This notion provides little recognition for the role of the passenger polypeptide [2, 3]. Recent studies in yeast provide a more balanced view by recognizing the contribution of amino acid (aa) sequences of the passenger polypeptide that are immediately adjacent to MTS [4] or located at the very C-terminus of the artificial construct [5]. Nevertheless, it is currently unclear whether, and to what extent internal portions of the passenger polypeptide contribute to the efficiency of mitochondrial import. Also unclear is whether the mechanisms described for yeast are conserved in higher eukaryotes in general and in mammals in particular.

The long-term studies of mitochondria in live cells have been greatly facilitated by tagging these organelles with FPs, such as green fluorescent protein (GFP). The discovery of DsRed was met with considerable enthusiasm as this protein presented the potential for multicolor imaging in combination with GFP [6]. However, the wild type DsRed possesses a number of characteristics that makes it suboptimal as a fusion tag (e.g., oligomerization, low fluorescence in vivo, and slow maturation, which proceeds through a green fluorescent intermediate [7–9]). Therefore, much effort was invested in the generation of DsRed mutants with improved properties [9–11]. The latest addition to the family of red FPs resulting from these efforts, mCherry, is a monomeric protein endowed with red shifted fluorescence excitation and emission maxima, high brightness, photostability, and dramatically shortened maturation (t_0.5 at 37°C = 15 min vs. ≈ 10 h for DsRed [10, 12]). It differs from another DsRed mutant, DsRed Express (DRE) at 25 internal positions (88% identity). In addition, to improve the performance of the fusion proteins, seven N-terminal aa of DsRed were replaced with seven N-terminal aa of the enhanced green fluorescent protein (EGFP) followed by four aa spacer. Also, toward this goal three C-terminal aa of DsRed were replaced in mCherry with 10 C-terminal aa of EGFP (see [10] and Fig. 1). The performance of mCherry as a fusion tag with α-tubulin was found to be superior to that of mRFP1 [10]. However, to our knowledge, the utility of mCherry for mitochondrial labeling has not been reported. Therefore, the initial goal of this project was to evaluate the utility of mCherry as a mitochondrial tag.

Fig. 1.

Design and structure of proteins used in this study. (A) Amino acid alignments of DRE, mCherry, and mCherry short. The identical aa are indicated by asterisks, non-matching aa are indicated by periods. The position of M182 in DRE is indicated by the vertical arrow, and stretches of identical amino acids that represent overlaps in chimeras are indicated by horizontal lines above the sequence. The precise aa composition of chimeras can be derived by switching between aa sequences of two parental proteins at the region of overlap. E.g., the aa sequence for MD1 can be derived by switching from aa sequence of mCherry to that of DRE at MD1/DM1 region of overlap. Similarly, MD2 aa sequence can derived by switching between these two sequences at the MD2 region. Please note that core aa sequences of mCherry and DRE are different at only 25 positions. (B) The strategy for the construction of MD chimeras. Primers 1 and 4 are mCherryPCRf and DsExprRxba, respectively (Table 1). Primer 3, depending on the chimera number, was either MD1/DM1F, MD2F, MD3F, or MD4F. Similarly, depending on the chimera number, primer 4 was either MD1/DM1R, MD2R, MD3R, or MD4R. The DM1 chimera was constructed similarly using primers DsExprFnoATGbam, MD1/DM1R, MD1/DM1F, and mCherry PCRr as primers 1, 2, 3, and 4, respectively (Table 1 )

Materials and methods

Gene synthesis

The version of mCherry, that is codon optimized for expression in human cells was generated by PCR-assembly of synthetic overlapping 60-mer unphosphorylated oligonucleotides, which were synthesized by Integrated DNA Technologies (Coralville, IA) and used without further purification. The complete list of primers used in this study is found in the Table 1. The nucleotide sequence was generated by back translation of the published aa sequence [10] using Vector NTI software package (Invitrogen, Carlsbad, CA). The assembly was performed using recursive PCR essentially as described earlier [13] and its fidelity was verified by double-stranded DNA sequencing.

Table 1.

Primers used in this study

Name	Sequence*
mCherryPCRf	GCGGGATCCGTGAGCAAGGG
mCherry PCRr	CGCCTCGAGTCACTTGTACAGC
DsExprFnoATGbam	GCGGGATCCGCCTCCTCCGAGGACGTCAT
DsExprRxba	CGCTCTAGATTACAGGAACAGGTGGTGGCGGC
mCherrySHORTf	GCGGGATCCGCCTCCTCCGAGGACGTCATCAAGGAGTTCATGCGGTTC
mCherrySHORTr	CGCCTCGAGTTACAGGAACAGGTGGCTGTGCCGGCCCTCGGCCCG
MD1/DM1F	GAAGCACCCCGCCGACATCCCC
MD1/DM1R	GGGGATGTCGGCGGGGTGCTTC
MD2F	CGACGGCCCCGTAATGCAGAAGAAGAC
MD2R	GTCTTCTTCTGCATTACGGGGCCGTCG
MD3F	CTGAAGCTGAAGGACGGCGGCCACTAC
MD3R	GTAGTGGCCGCCGTCCTTCAGCTTCAG
MD4F	GCCAAGAAGCCCGTGCAGCT
MD4R	AGCTGCACGGGCTTCTTGGC
DsRes M182Kf	CAAGTCCATCTACAAGGCCAAGAAGCCC
DsRes M182Kr	GGGCTTCTTGGCCTTGTAGATGGACTTG
mCherry1	GCGGGATCCGTGAGCAAGGGCGAGGAGGACAACATGGCCATCATCAAGGAGTTCATG
mCherry2	CGTGGCCGTTCACGCTGCCCTCCATGTGCACCTTGAACCGCATGAACTCCTTGATGATGG
mCherry3	GGGCAGCGTGAACGGCCACGAGTTCGAGATCGAGGGCGAGGGCGAGGGCCGGCCCTACGA
mCherry4	GCCGCCCTTGGTCACCTTCAGCTTGGCGGTCTGGGTGCCCTCGTAGGGCCGGCCCTCGCC
mCherry5	TGAAGGTGACCAAGGGCGGCCCCCTGCCCTTCGCCTGGGACATCCTGAGCCCCCAGTTCA
mCherry6	ATGTCGGCGGGGTGCTTCACGTAGGCCTTGCTGCCGTACATGAACTGGGGGCTCAGGATG
mCherry7	GTGAAGCACCCCGCCGACATCCCCGACTACCTGAAGCTGAGCTTCCCCGAGGGCTTCAAG
mCherry8	TCACCACGCCGCCGTCCTCGAAGTTCATCACCCGCTCCCACTTGAAGCCCTCGGGGAAGC
mCherry9	CGAGGACGGCGGCGTGGTGACCGTGACCCAGGACAGCAGCCTCCAGGACGGCGAGTTCAT
mCherry10	GTCGCTGGGGAAGTTGGTGCCCCGCAGCTTCACCTTGTAGATGAACTCGCCGTCCTGGAG
mCherry11	GCACCAACTTCCCCAGCGACGGCCCCGTGATGCAGAAGAAGACAATGGGCTGGGAGGCCA
mCherry12	CCCTTCAGGGCGCCGTCCTCGGGGTACATCCGCTCGCTGCTGGCCTCCCAGCCCATTGTC
mCherry13	GAGGACGGCGCCCTGAAGGGCGAGATCAAGCAGCGGCTGAAGCTGAAGGACGGCGGCCAC
mCherry14	GCTTCTTGGCCTTGTAGGTGGTCTTCACCTCGGCGTCGTAGTGGCCGCCGTCCTTCAGCT
mCherry15	CACCTACAAGGCCAAGAAGCCCGTGCAGCTCCCCGGCGCCTACAACGTGAACATCAAGCT
mCherry16	CTCCACGATGGTGTAGTCCTCGTTGTGGCTGGTGATGTCCAGCTTGATGTTCACGTTGTA
mCherry17	AGGACTACACCATCGTGGAGCAGTACGAGCGGGCCGAGGGCCGGCACAGCACCGGCGGCA
mCherry18	CGCCTCGAGTCACTTGTACAGCTCGTCCATGCCGCCGGTGCTGTGCCGG

^*Restriction sites are in boldface, italicized, and underlined

Matrix targeting sequences

The MTSs for the mitochondrial isoform of human malate dehydrogenase (MDH2, aa 1–28), human cytochrome c oxidase subunit VIII (COX, aa 1–29), human superoxide dismutase (hSOD, aa 1–24), human ornithine transcarbamoylase (OTC, aa 1–32), human mitochondrial transcription factor A (TFAM, aa 1–42), and human DNA polymerase γ (PolG, aa 1–33) were generated by RT-PCR of total mRNA isolated from HEK293 cells using appropriate primers and verified by sequencing. The optimized translation initiation (Kozak) sequence GCCACC was introduced in front of each MTS to facilitate expression of the fusion proteins [14].

DNA constructs

Plasmids for the expression of fluorescent proteins (FPs) in mammalian cells were constructed using established methods [15]. All fusion constructs were assembled in the pMA898 (an adenovirus shuttle plasmid derived from pDC512, Microbix Biosystems, Toronto, Ontario, CA. Our unpublished results) under the control of the CMV promoter. The chimeras were constructed by overlap extension PCR [16] using primers listed in the Table 1. The primers were designed to the regions of aa identity between mCherry and DRE (Fig. 1) and are named accordingly (Table 1). The composition of the chimeras is listed in the Table 3. Each chimera name consists of two letters and a sequential number. The first and second letters indicate the source of N- and C-terminal fragments, respectively (e.g., MD1 indicates that this chimera is composed of an N-terminal portion that comes from mCherry and a C-terminal portion that comes from DsRed. Therefore, M and D in the chimera name stand for mCherry and DRE, respectively). The precise aa composition of the chimeras can be derived from Fig. 1 by switching between aa sequences of two proteins at the region of overlap. Since the junctions between mCherry and DRE in chimeras correspond to regions of aa identity (see Fig. 1), the two proteins overlap at the junction (see Table 3).

Table 3.

Composition of chimeras

Chimera	mCherry	DsRed express
DM1	77–236	2–82
MD1	2–87	72–225
MD2	2–151	128–225
MD3	2–178	169–225
MD4	2–196	187–225

For each chimera, the source protein and aa coordinates of the source fragment are indicated

Cell culture, transfections and confocal microscopy

Plasmid DNA was introduced into HeLa cells grown in DMEM medium by transient transfection using Polyfect^® (Qiagen, Valencia, CA) according to the manufacturer’s recommendations. Forty hours after transfection, mitochondria were stained with 400 μM MitoTracker Green FM^® (MT, Invitrogen, Carlsbad, CA) in DMEM for 15 min at 37°C in an atmosphere of 5% CO₂. Confocal microscopy was performed on live cells using a Leica DM RXE microscope and TCS SP2 confocal system in combination with a 63 × water immersion objective.

Data reduction

For each construct, three or more transfections were performed and subcellular localization of expressed FPs was scored in at least 50 randomly chosen cells in at least three different fields per transfection. The cells were scored into three bins: (i) cells with exclusive mitochondrial localization of FP, (ii) cells in which both mitochondrial and cytoplasmic localization of FP is observed, and (iii) cells with exclusively cytoplasmic localization of FP. To facilitate handling the large amount of data generated we used the mitochondrial propensity index (MPI). This index is calculated according to formula MPI = lg[(i + ii)/(ii/iii)], where lg is a logarithm with the base 10 and i, ii and iii are the numbers of cells in the corresponding bins. Importantly, while the base of logarithm was chosen to be 10 in this particular study, it can be varied to regulate the MPI’s sensitivity (reducing the base value would increase sensitivity). An MPI = 0 indicates equal propensity of the FP to localize to mitochondria and cytoplasm, an MPI > 0 indicates a higher propensity to localize to mitochondria, and an MPI < 0 indicates higher propensity to localize to cytoplasm.

Results and discussion

Gene synthesis and experimental setup

To optimize expression in HeLa cells, the nucleotide sequence of mCherry was designed to contain only codons that are most frequently used by human cells to encode polypeptides. However, several exceptions were made to allow for the removal of common restriction endonuclease sites from the coding sequence in order to facilitate the subsequent insertion of the synthetic gene into various cloning vectors. In this study, the subcellular partitioning of proteins was studied using confocal microscopy rather than conventional biochemical fractionation techniques. This choice was dictated by two factors. First, the biochemical fractionation results in “enriched” rather than “pure” organellar fractions. Second, fractionation requires large quantities of homogenous starting material, which is difficult to obtain using transient transfection for the delivery of recombinant DNA constructs. One potential problem which may complicate the interpretation of any experiment which involves fusion between a short MTS and an FP is leaky ribosomal scanning (LRS). LRS is a widespread phenomenon which results in translation initiation on both the first and downstream AUG codon(s) of mRNA. High abundance mRNAs such as those whose expression is driven by strong viral promoters, are particularly susceptible to LRS [17]. Therefore, to minimize the complications associated with expression of unfused FPs (generated by LRS), the first ATG codons were removed from both mCherry and DsRed with the aid of PCR.

Effects of MTS and passenger polypeptide on mitochondrial partitioning

In the initial experiment, the MTS of malate dehydrogenase, (MDH2, Fig. 2) was fused to mCherry and the subcellular distribution of this fusion construct was studied in HeLa cells. To our surprise, we observed a uniform cytoplasmic, instead of the anticipated mitochondrial, fluorescence in the transfected cells. The effect was recapitulated in 143B human osteosarcoma cells, 3T3 mouse fibroblasts, and A549 human non-small cell lung carcinoma cells (data not shown). To confirm this observation, we generated 12 fusion constructs by combining six different human MTSs with either mCherry or DRE. The scoring of the results of transfection of these constructs into HeLa cells resulted in the generation of the large amount of data, which were difficult to analyze. Therefore, the MPI (see Materials and Methods) was introduced, validated (see Fig. 3), and utilized. The results from 12 fusions can be summarized as follows: each of the six tested MTSs efficiently targeted DRE to mitochondria in HeLa cells. In contrast, mitochondrial partitioning of mCherry was impaired in a targeting-sequence-dependent manner (Fig. 3). The differences in the efficiency of mitochondrial partitioning of DRE versus mCherry by a given MTS can only be attributed to the differences in the primary aa sequences of these polypeptides. Three possibilities were considered:

Fig. 2.

The structure of targeting constructs (A) and the efficiency of their mitochondrial targeting (B). (A) CMV, the human cytomegalovirus promoter; MTS, one of the six mitochondrial targeting sequences used; FP, either DRE or mCherry; pA, polyadenylation site. (B) The red fluorescence (panels labeled FP) indicates the distribution of DRE or mCherry, the green fluorescence (panels labeled MT) shows mitochondrial distribution as revealed by MitoTracker Green Staining, and the panels labeled OL represent an overlay of two images. The areas of yellow color indicate mitochondrial localization of FP

Fig. 3.

Validation of MPI. Subcellular distribution (A) and MPIs (B) of a series of hSOD-FP fusions. Mito, mix, and cyto denote fractions of cells with purely mitochondrial, mixed mitochondrial and cytoplasmic, and purely cytoplasmic localization of FP, respectively. Note that MPI faithfully recapitulates the trend in subcellular distribution of FP

1. The N-terminal modification of mCherry resulted in the formation of fortuitous labile sites at the junction between the MTS and FP. The processing of these labile sites resulted in cleavage of the MTS and cytoplasmic localization of mCherry. This notion is inconsistent with our observation of efficient mitochondrial partitioning of EGFP, which has an MTS-FP junctions identical to those of mCherry, by all MTSs used in this study (results not shown);

2. The N-terminal modification of mCherry (“GFP-ization”, see [10]) resulted in the steric hindrance of mitochondrial presentation of MTS in some fusions.

3. AA differences in core domains of mCherry and DRE are primarily responsible for the observed differences in the efficiency of mitochondrial partitioning of these two proteins.

The contribution of the region proximal to the MTS

To evaluate the contribution of modified C-and N-termini in mCherry, this protein was refitted with DRE termini by means of PCR with primers mCherrySHORTf and mCherrySHORTr (Fig. 1 and Table 1), thus, generating mCherry Short. With two out of six MTSs tested, mCherry Short has a substantially higher MPI than mCherry. (Fig. 4). Importantly, even these improved MPIs remained well below those for DRE. Also, MPIs for fusions consisting of PolG or hSOD MTSs and mCherry Short (−1.6 and −1.3, respectively) failed to improve substantially over those with mCherry (−1.9 and −1.1, respectively. see Fig. 4 and primary data in Table 2). This shows that cytoplasmic partitioning of these constructs is not dependent on the very proximal and the very distal regions, which were changed in mCherry and restored in mCherry Short, but rather is dictated by internal mutations. Overall, these results suggest that the role of the very proximal [4] and the very distal [5] portions of the passenger polypeptide may be, at least in some cases, MTS-dependent (since N- and C-terminal “refitting” of mCherry increased MPIs with some, but not all, MTSs). Also, they suggest that the contribution of these regions in our experiments may be relatively minor (since no contribution was observed with PolG and hSOD MTSs). Therefore, it is unlikely that “GFP-ization”-induced steric hindrance plays a major role in the impairment of mitochondrial partitioning of at least for the fusions with PolG and hSOD MTSs (see point 2 above).

Fig. 4.

The MPIs of 18 fusions between DRE, mCherry, and mCherry Short and 6 MTSs. mCh, mCherry; DRE, DsRed Express; Short, mCherry Short. * indicates that calculated MPI of TFAM-DRE fusion is positive infinity. See Table 2 for primary data

Table 2.

Subcellular partitioning of fusion constructs

MTS	mCherry			mCherry short			Dsred Express
	Mito	Mix	Cyto	Mito	Mix	Cyto	Mito	Mix	Cyto
Cox	0	98.7 ± 1.3	1.3 ± 1.3	12.7 ± 4.3	87.3 ± 4.3	0	92 ± 1	8 ±1	0
OTC	48.3 ± 3.3	44.3 ± 2.6	7.3 ± 2.9	83.7 ± 7.8	16.3 ± 8	0	99 ± 1	1 ± 1	0
TFAM	7.3 ± 1.9	92.7 ± 1.9	0	4 ± 2	96 ± 2	0	100	0	0
MDH2	0	34.3 ± 11.9	65.7 ± 11.9	0.7 ± 0.7	98.3 ± 0.9	0	75 ± 4.7	24.5 ± 4.7	0.25 ± 0.25
hPolG	0	1.3 ± 1.3	98.7 ± 1.3	0	2.3 ± 1.2	97.7 ± 1.2	68.3 ± 5.4	31.7 ± 5.4	0
hSOD	0	7.7 ± 5	92.3 ± 5	0	5 ± 2.5	95 ± 2.5	67.7 ± 2.3	32.3 ± 2.3	0

The HeLa cells were transfected and the percentage of cells showing mitochondrial, cytoplasmic, or mixed localization of expressed FP was scored as described in Materials and Methods. The data are mean ± SEM of at least three independent transfections

Effect of mutations in passenger polypeptide on mitochondrial partitioning

The effect of mutations in the passenger polypeptide was studied using a chimera approach. First, two reciprocal chimeras between mCherry and DRE, DM1 and MD1, were constructed. Each chimera roughly consisted of the N-terminal 1/3 of one polypeptide and the C-terminal 2/3 of another (Table 3). The N-terminal portions of the chimeras differ at three aa positions in addition to differences introduced by N-terminal modification in mCherry (see Materials and Methods for details on chimeras). The MPIs of the fusions of these chimeras with three different MTSs indicate that the C-terminal portion of chimeras is an important determinant of mitochondrial partitioning. Indeed, the comparison of the MPIs of the fusions of mCherry, DRE, and MD1 to MTSs of COX, MDH2 and hSOD (Figs. 4 and 5A) reveals that MPIs of mCherry fusions are neutral, negative, and negative, respectively, while MPIs of both MD1 and DRE are positive. Therefore, in this experiment, MPIs of the whole MD1 chimera are more similar to those of the source of the C-terminal portion (i.e., DRE) than to those of the source of the N-terminal portion (i.e., mCherry), which underscores the important contribution of the C-terminal portion. Similarly, MPI of hSOD-DM1 fusion is negative (Fig. 5A) resembling that of hSOD-mCherry chimera (C-terminal portion) and dissimilar to that of hSOD-DRE fusion (N-terminal portion), which is positive (Fig. 4).

Fig. 5.

The MPIs of chimeric proteins. (A) reciprocal DM1 and MD1 chimeras were fused to three different MTSs. (B) MPIs of a series of chimeras with progressively increasing mCherry portion. Note that MPIs descrease as fraction of mCherry increases; (C) the effect of M182K mutation on mitochondrial localization of DRE and DM1 chimera

Furthermore, the study of a series of fusions between hSOD MTS and MD chimeras, in which the portion of mCherry (the C-terminal portion of the fusion) progressively increases, revealed that the MPI closely reflects and is inversely proportional to the fraction of mCherry in these fusions (Fig. 5B and Table 3). This observation strongly argues against the first two possibilities (sect. Effects of MTS and passenger polypeptide on mitochondrial partitioning) in the favor of the notion that it is aa differences in the core domains of mCherry and DRE that are primarily responsible for the observed differences in the efficiency of mitochondrial partitioning of these two proteins. It has been reported previously that protein unfolding is required for mitochondrial import and that stabilization of the folded state of the passenger polypeptide impairs its mitochondrial import [18, 19]. In a folded state, DsRed forms a very stable β-barrel structure that resists unfolding [8]. Therefore, it is conceivable that differences in maturation rates of DRE and mCherry are responsible for the differences in the efficiency of mitochondrial partitioning. Indeed, the fact that the M182K mutation, which has been described as the one that facilitates mCherry folding [10] negatively affects MPIs of both DRE and MD1 chimera suggests that it is accelerated folding of mCherry [10] that may, at least in part, be responsible for the impairment of mitochondrial partitioning of this protein by the MTS of hSOD (Fig. 5C). On the other hand, our results demonstrate that mutations other than M182K contribute to the impairment of mitochondrial partitioning of mCherry (Figs. 3 and 5). Although, the possibility that all these mutations contribute favorably to mCherry folding cannot be excluded, it seems unlikely. Therefore, it is important to note here that mutations that stabilize the β-barrel structure of mCherry against unfolding may have similar detrimental effect on the mitochondrial import of mCherry.

Our results are consistent with the following model which describes the relative contributions of the MTS and the passenger polypeptide:

1. Different MTSs afford different cytoplasmic residence times (before presentation to mitochondrial import complexes) for their passenger polypeptides and/or modulate recruitment of cytoplasmic chaperones. The increased cytoplasmic residence time may translate into more complete protein folding at the time of polypeptide presentation to the mitochondrial import machinery, which consequently may lead to impaired mitochondrial partitioning. Similarly, variations in chaperone recruitment may translate into differences in protein folding kinetics and/or mitochondrial presentation kinetics resulting in differences in the efficiency of mitochondrial partitioning [20].

2. The contribution of the passenger polypeptide to the efficiency of mitochondrial partitioning is determined by the rate of its folding [21, 22], and by the energy that is required for its unfolding during mitochondrial import [4, 5].

This model is very similar to the one recently proposed to describe the mitochondrial import of yeast fumarase [21]. This similarity suggests a potential conservation of mitochondrial import mechanisms among eukaryotes. Moreover, the observation that mitochondrial partitioning of the passenger polypeptide is extremely sensitive to both the sequence of MTS and sequence of this polypeptide itself suggests the rationale for the observed diversity of MTSs in vivo.

Practical implications

Our observations have three practical consequences as well. They show that mCherry is not suitable for studies aimed at identification of MTSs (and, possibly, of mitochondrial proteins) by the gene fusion approach. Indeed, half of the well-characterized MTSs used in this study failed to efficiently partitioning mCherry to mitochondria. Currently, there is a lack of an experimental framework to study a relationship between protein maturation and its mitochondrial targeting in vivo. This study suggests that the use of FPs with different maturation rates has a potential to fill in this void. It also suggests that for some applications, e.g., gene therapy of mitochondrial disorders, a combination of MTS and passenger polypeptide should be chosen carefully as it may affect the outcome of studies.

Abbreviations

aa

Amino acid(s)

COX

Cytochrome c oxidase

DRE

DsRed Express

FP

Fluorescent protein

GFP

Green fluorescent protein

hSOD

Human mitochondrial superoxide dismutase

LRS

Leaky ribosomal scanning

MDH2

Mitochondrial isoform of malate dehydrogenase

MPI

Mitochondrial propensity index

mRFP1

Monomeric red fluorescent protein

MTS

Matrix targeting sequence

OTC

Ornithine transcarbamylase

PolG

DNA polymerase γ

TFAM

Mitochondrial transcription factor A