Tài liệu Báo cáo khoa học: Switching of the homooligomeric ATP-binding cassette transport complex MDL1 from post-translational mitochondrial import to endoplasmic reticulum insertion pptx

Switching of the homooligomeric ATP-binding cassette
transport complex MDL1 from post-translational
mitochondrial import to endoplasmic reticulum insertion
Simone Gompf
1
, Ariane Zutz
1
, Matthias Hofacker
1
, Winfried Haase
2
, Chris van der Does
1
and Robert Tampe
´
1
1 Institute of Biochemistry, Biocenter, Johann Wolfgang Goethe-University, Frankfurt am Main, Germany
2 Max-Planck Institute of Biophysics, Structural Biology, Frankfurt am Main, Germany
ATP-binding cassette (ABC) transporters belong to a
large family of membrane proteins found in all three
kingdoms of life. The chemical energy of ATP is used
to drive uphill transport of a broad range of solutes
across membranes [1–3]. ABC transporters have a
conserved domain organization consisting of two trans-
membrane domains (TMDs) and two nucleotide-bind-
ing domains (NBDs). The TMDs form a translocation
pore, whereas the NBDs catalyze ATP hydrolysis.
The ABC half-transporter multidrug resistance like
protein 1 (MDL1), composed of a TMD followed by a
NBD, is located in the inner mitochondrial membrane
(IMM) of Saccharomyces cerevisiae. It has been sug-

gested to be involved in the export of 6-mer to 20-mer
peptides, derived from proteolysis of nonassembled
inner membrane proteins by the m-AAA (i.e. matrix-
oriented ATPase associated with a variety of cellular
activities) protease [4]. It has been further reported
that MDL1 mediates resistance against oxidative stress
and can partially complement the function of ABC
transporter of mitochondria (ATM) 1 [5]. Deletion of
ATM1 in S. cerevisiae results in a severe growth defect
because ATM1 is essential for the biogenesis of cyto-
solic iron-sulfur (Fe-S) proteins [6].
Keywords
ABC transporter; ER targeting; membrane
protein trafficking transport ATPase;
mitochondrial import; mitochondrial
targeting sequence
Correspondence
R. Tampe
´
, Institute of Biochemistry,
Biocenter, Johann Wolfgang Goethe-
University, Max-von-Laue-Strasse 9,
D-60348 Frankfurt am Main, Germany
Fax: +49 (0) 69 798 29495
Tel: +49 (0) 69 798 29475
E-mail:
Website: chem.
uni-frankfurt.de
(Received 25 May 2007, revised 5 July
2007, accepted 20 August 2007)

doi:10.1111/j.1742-4658.2007.06052.x
The ATP-binding cassette transporter MDL1 of Saccharomyces cerevisiae
has been implicated in mitochondrial quality control, exporting degradation
products of misassembled respiratory chain complexes. In the present study,
we identified an unusually long leader sequence of 59 amino acids, which
targets MDL1 to the inner mitochondrial membrane with its nucleotide-
binding domain oriented to the matrix. By contrast, MDL1 lacking this lea-
der sequence is directed into the endoplasmic reticulum membrane with the
nucleotide-binding domain facing the cytosol. Remarkably, in both target-
ing routes, the ATP-binding cassette transporter maintains its intrinsic
properties of membrane insertion and assembly, leading to homooligomeric
complexes with similar activities in ATP hydrolysis. The physiological con-
sequences of both targeting routes were elucidated in cells lacking the mito-
chondrial ATP-binding cassette transporter ATM1, which is essential for
biogenesis of cytosolic iron-sulfur proteins. The mitochondrial MDL1 com-
plex can complement ATM1 function, whereas the endoplasmic reticulum-
targeted version, as well as MDL1 mutants deficient in ATP binding and
hydrolysis, cannot overcome the Datm1 growth phenotype.
Abbreviations
ABC, ATP-binding cassette; ATM, ABC transporter of mitochondria; ER, endoplasmic reticulum; 5-FOA, 5-fluoroorotic acid; IMM, inner
mitochondrial membrane; MDL1, multidrug resistance like protein 1; MTS, mitochondrial targeting signal; NBD, nucleotide-binding domain;
SC, synthetic complete; TIM, translocase of the inner mitochondrial membrane; TOM, translocase of the outer mitochondrial membrane;
TMD, transmembrane domain.
5298 FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS
Mitochondria contain approximately 800–1500 dif-
ferent proteins [7,8]. Although they include mtDNA
and a transcription ⁄ translation machinery, the vast
majority of mitochondrial proteins are encoded by
nuclear genes and synthesized as precursor proteins on
cytosolic ribosomes [9–13]. Several pathways of mito-

chondrial protein import have been characterized:
(a) the presequence pathway for matrix proteins;
(b) sorting and assembly of anchored mitochondrial
outer membrane proteins by transmembrane b-strands,
and (c) the carrier pathway for hydrophobic inner
membrane proteins [14].
The mitochondrial targeting signal (MTS) of pro-
teins directed to the IMM is recognized by receptors
of the translocase of the outer mitochondrial mem-
brane (TOM complex). The classical targeting signal is
represented by an N-terminal leader sequence of 20–
35 amino acids [15], enriched in basic, hydrophobic
and hydroxylated residues [16]. It has been suggested
that the leader peptide folds into a defined secondary
structure, which is essential for protein import, due to
the distribution of charged and apolar residues. The
N-terminal part of the MTS forms a positively charged
amphiphilic a-helix or b-sheet, whereas the C-terminal
region probably serves as a recognition site for matrix
proteases [15,17]. Positional amino acid preferences
have been found in the region immediately upstream
from the mature amino terminus [18]. In particular,
arginine can be enriched in position -2, -3, -10, and -11
relative to the cleavage site. The leader sequence inter-
acts with the TOM receptor, responsible for the trans-
location of the preproteins to the translocase of the
inner mitochondrial membrane (TIM)23 complex
located in the IMM. The presequence translocase-asso-
ciated motor is directly associated with TIM23 and
completes the translocation of the preprotein into the

matrix. There the presequence is removed by the mito-
chondrial processing peptidase. Subsequently, IMM
proteins are guided by a hydrophobic sorting sequence
that typically follows the positively charged pre-
sequence [19,20].
In the present study, we addressed the functional
role and physiological consequences of the unusual
long N-terminal leader sequence of MDL1. Full-
length MDL1 is targeted to the IMM, whereas the
leaderless ABC transporter is exclusively inserted into
endoplasmic reticulum (ER) membrane. Despite these
presequence-dependent trafficking routes, the mem-
brane insertion, the complex assembly, and the
ATPase function of MDL1 are preserved. The
physiological consequence of these two targeting
routes is addressed by in vivo complementation in
cells lacking the mitochondrial ABC transporter
ATM1, which is essential for the assembly of cyto-
solic Fe-S proteins.
Results
Targeting of MDL1 to the IMM
It has been postulated that MDL1 is involved in the
export of peptides generated (e.g. from misassembled
mitochondrially encoded respiratory chain subunits)
[4]. Unfortunately, the mechanism and transported
substrate remain largely elusive due to the intrinsic dif-
ficulties in studying mitochondrial export processes.
This is due to the fact that substrates are limited in the
matrix and their concentrations are very difficult to
control experimentally. In addition, substrates are

highly diluted after translocation into the external
medium. By contrast, many intracellular transport sys-
tems have been characterized in detail by means of
uptake assays; for example, the transporter associated
with antigen processing (TAP) [21,22] and TAP-like
(ABCB9) [23]. We therefore set out to target MDL1
from mitochondrial import to insertion into the ER
membrane in order to perform similar analyses.
An introduced ClaI restriction site and the endo-
genous BamHI site divided the MDL1 gene into three
cassettes, facilitating the exchange of segments between
different constructs. By means of the inducible GAL1-
promoter, the protein can be over-produced to a level
of approximately 1% of the total mitochondrial pro-
tein. This correlates with an over-expression compared
to native MDL1 of up to 100-fold. To determine the
localization of MDL1 in S. cerevisiae, mitochondria of
Dmdl1 ⁄ MDL1 cells were prepared by subcellular frac-
tionation. As shown in Fig. 1A, MDL1 is found in the
mitochondrial fraction even after over-expression.
Immunoblotting of marker proteins (TIM23 and
SEC61) confirms that the mitochondrial fraction
contains only traces of ER membranes. In addition,
we analyzed the subcellular localization of MDL1 by
immunogold labeling (Fig. 1B). As expected, MDL1
was detected exclusively in cristae membranes, demon-
strating that the nuclear encoded protein is post-transla-
tionally targeted to mitochondria. Identical results
were obtained using a C-terminally His-tagged version
of MDL1 (data not shown).

Post-translational maturation of MDL1
Mitochondrial ABC transporters do not exhibit signifi-
cant sequence similarities in their leader sequences. In
the case of MDL1, several algorithms for the predic-
tion of mitochondrial targeting sequences gave rather
S. Gompf et al. Membrane targeting on demand
FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS 5299
conflicting results. We therefore set out to examine
the post-translational modification experimentally.
After purification via a C-terminal His-tag from iso-
lated mitochondria (Fig. 2A), N-terminal sequence of
MDL1 was determined by Edman degradation. An
unusually long presequence of 59 amino acids was
identified. The N-terminus (position 2–6) of the iso-
lated, mature ABC transporter (ESDIAQ) matches
perfectly with residue S61 to Q65 (Swiss-Prot: P33310).
Surprisingly, we found that the glutamine expected at
position 60 (the newly generated N-terminus) had been
modified to a glutamate. Sequencing of the expression
construct and comparison with the protein data bank
confirmed the glutamine at position 60. We can further
exclude modifications during purification because
MDL1 was prepared from isolated mitochondria.
Taken together, we identified two post-translational
modifications of MDL1: first, cleavage after residue 59
in the mitochondrial matrix releasing a long prese-
quence and, second, an enzymatic deamidation of the
newly generated N-terminal glutamine to glutamate.
Such modification has been reported for cytosolic
proteins (N-end rule pathway) [24] and for at least two

mitochondrial proteins, TIM44 and COX4 [25,26]. To
date, it is not clear whether this modification is an
artifact of Edman degradation or whether this deami-
dation is catalyzed by a N-terminal amidase during
mitochondrial translocation.
We next examined the membrane targeting of
MDL1 lacking the mitochondrial leader sequence iden-
tified in the present study. Thus, MDL1(60-695) was
generated and expressed in S. cerevisiae. By contrast to
the full-length protein, we found leaderless MDL1
cofractionated with the ER marker SEC61 (Fig. 1A).
As a control, the mitochondrial marker TIM23 is found
only in the mitochondrial fraction, whereas SEC61 is
enriched in the ER fraction, but can also be found in
the mitochondrial fraction. In parallel, the subcellular
localization of leaderless MDL1 was confirmed by
immunogold labeling (Fig. 1C). MDL1 lacking the
MTS was detected in tubulo-vesicular membranes
resembling the yeast ER membrane. It is worth men-
tioning that the N-terminally tagged MDL1(60-695)
was also targeted to the ER, as demonstrated by sub-
cellular fractionation and immunogold labeling (data
not shown). This suggests that mistargeting is due to
of a lack of the leader sequence rather than the new
N-terminus. Collectively, these data demonstrate that
leaderless MDL1 is targeted to and inserted into the
ER membrane by a cryptic default pathway.
Directionality of membrane insertion
The orientation of the full-length and leaderless ABC
transporter in mitochondrial and ER membranes,

respectively, was examined by protease protection
A
BC
Fig. 1. Localization of full-length and leader-
less MDL1 in S. cerevisiae. ER and mito-
chondrial membranes were prepared from
cells over-expressing wild-type MDL1 and
leaderless MDL1(60-695) (A) and analyzed
by SDS ⁄ PAGE (10%) and immunoblotting
using antibodies specific for MDL1, the
mitochondrial maker TIM23 and the ER mar-
ker SEC61. Immunogold labeling of sections
through cells over-expressing wild-type
MDL1 (B) and leaderless MDL1(60-695) (C).
Full-length MDL1 is localized in the mito-
chondrial cristae membranes, whereas lead-
erless MDL1 is detected in tubulo-vesicular
membranes belonging to or deriving from
the endoplasmic reticulum. M, mitochon-
dria; N, nucleus; V, vacuole.
Membrane targeting on demand S. Gompf et al.
5300 FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS
assays. As expected, MDL1 targeted to the IMM was
resistant to trypsin digestion because it was shielded
by the outer mitochondrial membrane (Fig. 3A). To
determine the orientation of MDL1 in the IMM,
mitoplasts and inverted IMMs were prepared and
assayed for protease cleavage. A factor Xa cleavage
site was engineered at the C-terminus of MDL1 before
the His8-tag. Thus, if the C-terminus is accessible to

the protease, the His-tag will be cleaved off as detected
by immunoblotting with His-tag specific antibodies.
This way, MDL1 in mitoplasts was shown to be
protected against factor Xa cleavage, whereas the
C-terminus of MDL1 was accessible in the inverted
IMMs, demonstrating that the ABC transporter was
inserted with the NBDs oriented to the matrix
(Fig. 3B). By contrast to full-length MDL1 expressed
in the IMM, trypsin treatment of ER membranes con-
taining leaderless MDL1 resulted in a limited digestion
of MDL1 (Fig. 3C). Trypsin treatment digestion of
ABC half-transporters resulted in the cleavage of the
linker region between TMD and NBD [27]. Even at
5 lgÆmL
)1
of trypsin, the NBDs of MDL1 were specif-
ically cleaved, indicating that they were oriented in the
cytoplasm. In conclusion, the NBDs of mitochondrial
MDL1 are located in the matrix, whereas the NBDs of
the leaderless MDL1 targeted to the ER membrane
face the cytosol.
MDL1 forms homooligomeric complexes with
similar activities independent of the targeting
route
ABC half-transporters must assemble at least into
dimeric complexes to gain function. To analyze
whether both targeting routes are comparable in com-
plex assembly and ATPase activity, the full-length and
leaderless MDL1 were purified to homogeneity via
metal affinity chromatography, yielding approximately

20 lgÆg
)1
wet weight of yeast in both cases (Fig. 2).
After isolation from different cellular compartments,
we investigated complex formation of MDL1 by gel fil-
tration. Each fraction was subsequently analyzed by
SDS ⁄ PAGE and immunoblotting (Fig. 4A). The mito-
chondrial as well as ER-resident MDL1 forms homo-
oligomeric complexes of similar size. The broad
distribution is rather typical for digitonin solubilized
ABC transport complexes. Notably, no protein aggre-
gates were detected at the exclusion volume. Other
detergents resulted in MDL1 complexes, which rapidly
lost their ATPase activity [28]. To demonstrate that
the broad distribution is not due to misfolding, we per-
formed an alternative approach, where we investigated
the oligomeric state of MDL1 by Blue-Native electro-
phoresis (Fig. 4B). Full-length and leaderless MDL1
solubilized from yeast membranes migrate as defined
bands at approximately 250 kDa, which corresponds
to a homodimeric complex, as resolved by single parti-
cle electron microscopy analysis [28]. In summary,
MDL1 forms a homodimeric complex independent of
its subcellular targeting.
The ATPase activity of ABC half-transporters is
critically dependent on the complex formation. We
therefore compared the ATPase activity of MDL1 tar-
geted to different cellular compartments (Fig. 5A,B).
Mitochondrial MDL1 isolated from total membranes
was active in ATP hydrolysis with a K

m ATP
of
A
B
Fig. 2. Purification of MDL1. For purification of MDL1 (A) and
MDL1(60-695) (B) with C- or N-terminal His-tags, respectively, total
membranes were prepared from S. cerevisiae over-expressing the
protein. Membranes (10 mgÆmL
)1
) were solubilized in the presence
of 1% (w ⁄ v) digitonin. The protein was purified to homogeneity by
metal affinity chromatography. Pellet (P) and supernatant after solu-
bilization (S), flow-through (FT), and fractions with increasing
concentrations of imidazole were analyzed by SDS ⁄ PAGE (10%,
Coomassie Blue stained, upper panel) and immunoblotting with
anti-MDL1 serum (lower panel).
S. Gompf et al. Membrane targeting on demand
FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS 5301
120 ± 6 lm and a turnover rate k
cat
of 74 ± 1 ATPÆ
min
)1
(per monomer). In comparison, leaderless
MDL1 purified from microsomes showed a K
m ATP
of
200 ± 1 lm and a k
cat
of 77 ± 1 ATPÆmin

)1
(per
monomer). To exclude the possibility that the activity
is caused by contaminating ATPases, we expressed and
purified two MDL1 variants (E599Q and H631A),
each of which has a disrupted catalytic dyad.
MDL1(E599Q) and MDL1(H631A) show no ATPase
activity above background but are active in ATP bind-
ing [28]. We further examined whether MDL1 show
similar sensitivity towards vanadate inhibition in both
targeting routes. As shown in Fig. 5C,D, the ATPase
activity of MDL1 purified from mitochondria or ER
membranes was inhibited in a dose-dependent manner
by ortho-vanadate. Comparable to other ABC trans-
porters [29–31], the IC
50
values of 0.86 mm and
1.1 mm were determined for the mitochondrial and
ER-resident MDL1, respectively. Taken together, full-
length and leaderless MDL1 are comparable in respect
to assembly of homooligomeric complexes, ATPase
activity, and vanadate inhibition.
A
C
B
Fig. 3. Membrane orientation of mitochon-
drial and ER-resident MDL1. Mitochondrial
(A) and ER fractions (C) (30 lg each) con-
taining wild-type MDL1 and leaderless
MDL1(60-695), respectively, were incubated

with increasing concentrations of trypsin
(0–0.1 mgÆmL
)1
) and analyzed by
SDS ⁄ PAGE (10%) followed by immunoblot-
ting. (B) Mitoplasts and inverted IMMs
(IMV) (30 lg each) containing MDL1 were
incubated with factor Xa (0.5 lg) and
analyzed by SDS ⁄ PAGE (10%) and immuno-
blotting. In this case, MDL1 contains a
C-terminal His-tag separated by a factor Xa
cleavage site.
A
B
Fig. 4. Formation of homooligomeric com-
plexes of mitochondrial and ER-resident
MDL1. Purified MDL1 (upper panel) and
leaderless MDL1(60-695) (lower panel) were
analyzed by gel filtration on a Superdexä 200
PC 3.2 in the presence of 0.1% (w ⁄ v)
digitonin. Every second fraction (30 lL) was
analyzed by immunoblotting using an
MDL1-specific antibody (A). Total
membranes (10 mgÆmL
)1
) of cells expressing
MDL1 or MDL1(60-695) were solubilized
in presence of 1% (w ⁄ v) digitonin.
(B) Digitonin-solubilized proteins were
analyzed by Blue-Native electrophoresis and

immunoblotting using anti-MDL1 serum.
Apoferritin (443 kDa), b-amylase (200 kDa),
alcohol dehydrogenase (150 kDa), and
albumin (66 kDa) were used as markers.
Membrane targeting on demand S. Gompf et al.
5302 FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS
Uptake assays with isolated microsomes
containing MDL1
Leaderless MDL1 is targeted to ER membranes, where
the NBDs of the homooligomeric complex are oriented
to the cytosol. In this orientation, the ATP level and
substrates can effectively be controlled. To identify the
MDL1 substrate, we screened combinatorial peptide
libraries of different length X
n
(n ¼ 5–8, 11, 17, and
23, where X represents an equimolar distribution of all
19 amino acids except cysteine). These libraries have
been instrumental in deciphering the substrate specific-
ity of several eukaryotic and prokaryotic ABC trans-
porters [21,23,32,33]. In addition, we analyzed a set of
defined peptides expected to be a putative substrate of
MDL1. These include, for example, N-formylated pep-
tides or fragments of mitochondrially encoded gene
products, which have been identified as minor antigens
[34]. Systematic uptake assays with these peptidic sub-
strates, however, showed no MDL1-specific transport
activity, suggesting that MDL1 may be not a general
peptide transporter such as TAP or TAP-like, but
most likely transports a very specific or even modified

peptide.
Physiological function of MDL1 targeted to
different membranes
As shown in Fig. 6, MDL1 complements the severe
growth defect of Datm1 cells, indicating that the ABC
transporter can at least partially restore the assembly
of essential cytosolic Fe-S proteins. We next generated
a set of mutants defective in ATP binding and hydro-
lysis. Mutation of the conserved lysine in the Walk-
er A motif (K473A) is known to inhibit ATP binding,
whereas substitutions in the catalytic dyad (E599Q or
H631A) inhibit ATP hydrolysis [35–37]. Importantly,
these three mutants did not complement the Datm1
phenotype. Together with in vitro experiments these
data demonstrate for the first time that ATP binding
and hydrolysis are required for MDL1 function.
It has very recently been shown that the ATPase
activity of ATM1 is stimulated by cysteine-containing
peptides [38]. We therefore generated a cysteine-less
MDL1 and examined its function by in vivo comple-
mentation. Datm1 ⁄ MDL1(Cys-less) cells found to be
viable, demonstrating that cysteine residues are not
essential for MDL1 function. By contrast to wild-type
MDL1, the leaderless protein did not restore ATM1
function. Taken together, the function of MDL1 in
rescuing the cytosolic Fe-S cluster assembly machinery
requires ATP binding and hydrolysis and is strictly
coupled to its post-translational targeting to the mito-
chondrial membrane.
Discussion

Most mitochondrial proteins are synthesized by free
ribosomes in the cytosol. Once released into the cyto-
plasm with an N-terminal MTS, these preproteins are
imported into the mitochondria post-translationally
[39]. MTS usually consists of 20–35 residues and is
highly degenerated in primary sequence, but is rich in
basic, hydrophobic and hydroxylated residues and
Fig. 5. ATPase activity and vanadate inhibi-
tion of purified MDL1. ATPase activities
were measured as a function of ATP
concentration for 10 min at 30 °C with
0.5 l
M of purified protein. MDL1 (A) and
MDL1(60-695) (B) showed Michaelis–
Menten kinetics with a K
m ATP
of
120 ± 6 l
M and 200 ± 1 lM and a k
cat
of
74 ± 1 ATPÆmin
)1
and 77 ± 1 ATPÆmin
)1
(per MDL1 monomer), respectively.
Inhibition of ATPase activity of MDL1 (C)
and MDL1(60-695) (D) by different
concentrations of ortho-vanadate (given in
l

M). Based on the curve fit half-maximal
inhibitory concentrations (IC
50
) of 860 lM
and 1.1 mM were determined. All values are
derived from triplicate measurements.
S. Gompf et al. Membrane targeting on demand
FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS 5303
generally lacks acidic amino acids [16]. For post-trans-
lational targeting of MDL1 to the IMM, a 59 amino
acid long mitochondrial leader sequence was identified,
which is cleaved in the matrix. Subsequently, the
resulting N-terminal glutamine is converted to a gluta-
mate. Limited protease protection assays confirmed
that MDL1, even after over-expression, is efficiently
imported into mitochondria and properly inserted into
as well as assembled in the IMM with the NBDs
facing the mitochondrial matrix. Based on sequence
comparison, MDL1 should function as an exporter of
solutes to the intermembrane space. It is worth noting
that murine ABCB10, the closest homolog of MDL1,
also possesses an exceptionally long presequence of
105 amino acids [40]. Membrane topology algorithms
predict either five or six transmembrane helices for
MDL1. Protease accessibility assays and post-transla-
tional modifications revealed that the NBD and the
highly positively charged N-terminus of mature MDL1
are located in the mitochondrial matrix. Based on
these data, we propose that MDL1 comprises six
transmembrane helices.

In the present study, we addressed the functional
role of the unusually long leader sequence of MDL1 in
its subcellular targeting and physiological conse-
quences. By contrast to the full-length protein, which
is efficiently imported into mitochondria, leaderless
MDL1 is exclusively targeted to ER membranes.
Fig. 6. Physiological function of MDL1 vari-
ants analyzed by in vivo complementation.
Datm1 ⁄ ATM1 + MDL1 cells were plated on
SCD without uracil and tryptophan and used
for replica plating. Selection plates contain-
ing 5-FOA were incubated at 30 °C for
7 days. MDL1 can complement the severe
growth defect of Datm1 cells, whereas
mutants K473A, E599Q, H631A, inactive in
ATP binding or hydrolysis, as well as
MDL1(60-695) do not show complementa-
tion of ATM1. MDL1(Cys-less) is able to
take over the function of ATM1 and do not
affect growth.
Membrane targeting on demand S. Gompf et al.
5304 FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS
Protease accessibility assays demonstrated that the
NBDs of the ER-resident transporter are oriented to
the cytosol. The localization is not influenced by addi-
tional C- or N-terminal His-tags either for full-length
or leaderless MDL1.
To exclude that full-length and leaderless MDL1
have different activities, the proteins were purified to
homogeneity (Fig. 2). Remarkably, full-length and

leaderless MDL1 form homooligomeric complexes of
the same size and similar ATPase activities, K
m ATP
values of 120 lm and 200 lm and k
cat
values of
74 ATPÆmin
)1
and 77 ATPÆmin
)1
(per MDL1 subunit),
respectively. The ATPase activity of the transport
complex is in very good agreement with data of the
mitochondrial ABC transporter ATM1 [38] and the
purified NBD of MDL1 [36]. Both full-length and
leaderless MDL1 show sensitivity to ortho-vanadate,
similar to other ABC transporters [29–31].
MDL1 over-expressed in microsomes provides an
optimal setting to study the substrate specificity and
function of this sparingly characterized ABC trans-
porter. Based on a rather indirect assay, it has previ-
ously been concluded that MDL1 exports peptides of
6–20 amino acids in length [4]. To our surprise, no
transport activity was observed for microsomal MDL1
by screening combinatorial peptide libraries of differ-
ent lengths (X
n
, n ¼ 5–23 amino acids). Notably, this
approach has been crucial in the identification of the
substrate specificity of other peptide transporters

[21,23,32,33,41]. In addition, defined peptides favored
by the homologous TAP complex, such as the peptide
RRYQKSTEL, are not transported by MDL1.
Recently, a peptidic fragment, named COXI, of a
mitochondrially encoded subunit of the cytochrome
oxidase was identified to be presented on MHC class I
molecules of murine cells [42]. It was suggested that
COXI is transported from the matrix to the cytosol,
where the peptide is funneled into the pathway of
MHC class I antigen processing [34,43]. Thus,
N-terminal 7-, 9- and 12-mer fragments of COXI were
analyzed for an MDL1-dependent transport activity.
However, no uptake was detected. Taken together,
these findings suggest that MDL1, if indeed a peptide
transporter, is highly specific for a small set of peptides
or even modified peptides largely under-represented in
the peptide libraries. These systematic studies point to
an intriguing possibility that MDL1 may require addi-
tional factors for substrate transfer. Such factors may
be absent in uptake studies or in the libraries used.
Similar ATPase activities prove that the NBDs of both
MDL1 variants are correctly folded, although it can-
not be excluded that their TMDs are influenced by the
lipid compositions of the corresponding membranes.
Based on the important role of ATM1 in the biogen-
esis of cytosolic Fe-S proteins, Datm1 cells show a
severe growth phenotype. When Datm1 ⁄ ATM1 cells
are forced to loose the plasmid-encoded ATM1 (URA3
marker) by growth on 5-fluoroorotic acid (5-FOA),
Datm1 cells are almost nonviable. Multicopy expres-

sion of MDL1 (Datm1 ⁄ MDL1) can rescue this pheno-
type and cells are viable on fermentable carbon
sources [5]. This implies that ATM1 and MDL1 have
an overlapping function by which the growth pheno-
type of Datm1 cells is abrogated. However, by contrast
to ATM1 [38], no stimulation of the ATPase activity
of MDL1 was observed with thiol-containing peptides
of 10–15 residues in length (data not shown). A recent
report suggested that thiol-containing molecules are
first translocated by ATM1 and afterwards oxidized by
ERV1. These events are necessary for the maturation
of cytosolic and nuclear Fe-S proteins [38]. However, a
functional overlap between ATM1 and MDL1 with
regard to the translocation of thiol-containing peptides
appears to be very unlikely.
By analyzing several mutants, we demonstrated for
the first time that ATP binding and ATP hydrolysis
are required for the export function of MDL1. These
results are supported by data obtained in vitro showing
that the mutants K473A, E599Q and H631A are inac-
tive in ATP hydrolysis [28]. Notably, cysteine-less
MDL1 rescues ATM1 function, demonstrating that
cysteines are not essential for substrate translocation
across the IMM by MDL1. The conclusion that cyste-
ines are not involved in substrate translocation is also
in line with the observation that the ATPase activity
of MDL1 is not stimulated by cysteine-containing pep-
tides (see above). Leaderless MDL1, although correctly
assembled and fully active in ATP hydrolysis, does not
complement the growth phenotype of Datm1 cells. This

finding attests that the physiological function of the
ABC transporter MDL1 is intimately linked to its
correct targeting to the IMM.
Experimental procedures
Materials
A rabbit polyclonal antibody was generated against the
C-terminal 15 amino acids (KGGVIDLDNSVAREV) of
MDL1 from S. cerevisiae.
Cloning and expression of MDL1
The MDL1 gene from S. cerevisiae was divided into three
cassettes, separated by a newly generated silent ClaI restric-
tion site at S221 and the endogenous BamHI site at K422
S. Gompf et al. Membrane targeting on demand
FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS 5305
[28]. Cassette I includes the N-terminal part of the TMD
(M1 to A220), cassette II the C-terminal part of the TMD
(S221 to K422), and cassette III the NBD of MDL1 (D423
to V695). Furthermore, leaderless MDL1 (cassette IB, Q60
to A220) was generated. The different cassettes of MDL1
were amplified from genomic DNA (for sequences of the
primers, see Table 1). The corresponding PCR fragments
were cloned downstream of the GAL1-promoter in the
pYES2.1 ⁄ V5-His-TOPOÒ expression vector (Invitrogen,
Carlsbad, CA, USA) resulting in plasmids pMDL1 and
pMDL1(60-695). Using primers p1C(f) and p3(r), a similar
approach was applied to insert an N-terminal His10-tag fol-
lowed by leaderless MDL1 (60-695) resulting in pMDL1(60-
695,His). pMDL1(His), comprising four glycines, a
factor Xa cleavage site, and a His8-tag downstream of
MDL1, was generated with primers p1(f) and p3B(r).

Plasmids were transformed into Dmdl1 strain Y24137
(BY4743; Mat a ⁄ a; his3D1 ⁄ his3D1; leu2D0 ⁄ leu2D0; lys2D0 ⁄
LYS2; MET15 ⁄ met15D0; ura3D0 ⁄ ura3D0; YLR188w::
kanMX4 ⁄ YLR188w) [44]. Transformed cells were cultured
at 30 °C in synthetic complete (SC) medium in the presence
of 2% (w ⁄ v) glucose without uracil [45]. Cultures were
diluted to an A
600 nm
of 0.4 in SC medium containing 2%
(w ⁄ v) galactose and growth was continued for 12 h. Cells
were harvested by centrifugation and immediately used for
the isolation of total membranes [46] or mitochondria
[28,47]. Microsomes were separated from mitochondria by
centrifugation of the resulting supernatant at 100 000 g for
45 min at 4 °C (Ti45, Beckman Coulter, Fullerton, CA,
USA). Mitoplasts and inverted inner mitochondrial vesicles
are prepared as described [48,49]. The proteins were analyzed
by SDS ⁄ PAGE and immunodetection using the MDL1-
specific antibody. Protein concentrations were determined
using the Bradford assay (Pierce, Rockford, IL, USA).
Immunogold labeling
S. cerevisiae expressing full-length MDL1 or leaderless
MDL1(60-695), with and without the corresponding His-
tags, were fixed with 4% paraformaldehyde in 0.1 m
sodium cacodylate buffer (pH 7.2) supplemented with 0.8 m
sorbitol, 1 mm MgCl
2
and 1 mm CaCl
2
with or without 1%

glutardialdehyde. After 2 h, the fixative was exchanged for
cacodylate buffer containing decreasing concentrations of
sorbitol (0.5, 0.25, 0 m; three times 10-min incubation for
each concentration). Cells were treated with 1% sodium
meta-periodate, washed in water, and incubated in 0.05 m
NH
4
Cl. After 12 h, cells were washed again and enclosed in
agar-agar, which then was cut into small slices and passed
through increasing concentrations of ethanol for dehydra-
tion. Samples were stepwise infiltrated with LR White resin
Table 1. Primers used for generating MDL1 constructs. f, forward primer; r, reverse primer; mut, mutagenesis primer (exchanged bases
underlined).
Primer Sequence Site
a
p1(f) GGTACCACTAGTGCCGCCACCATGGTTGTAAGAAT KpnI, SpeI, NcoI
GATACGTCTTTGTAAAGG
p1B(f) GCCGCCACCATGCAATCAGACATTGCGCAAGGAAA
GAAGTCC
p1C(f) GCCGCCACCATGCACCATCACCATCACCATCACCAT
CACCATCAATCAGACATTGCGCAAGGAAAGAAGTCC
p1(r) GGCCACTATCGATGCATCAGATG ClaI
p2(f) CATCTGATGCATCGATAGTGGCC ClaI
p2(r) GTGTTTGGGCCGAGTGGGAT BamHI
b
p3(f) GCCATTGATTCGTCCGACTA BamHI
b
p3(r) TCTAGAAAGCTTTTATACTTCCCGGGCAACACTATT XbaI, HindIII
GTCC
p3B(r) TCTAGAAAGCTTTTAGTGATGGTGATGGTGATGGTG XbaI, HindIII

ATGCCGCCCTTCGATGCCGCCGCCGCCTACTTCCCGG
GCAACACTATTGTCC
p4(r) GCCACTAGTTGGAGCCCTTCC SpeI
p5(f) GGCACTAGTATGCAATCAGACATTGCG SpeI
pK473A(mut) CCATCAGGAAGCGGC
GCATCAACAATTG CGTCTTTG
pE599Q(mut) CTTATTTTAGAT
CAAGCAACCAGTGCC
pH 631 A(mut) CTATATCAATTGCA
GCGAGGCTTTCGACG
pC257S(mut) AAATTGACT
TCCGTAATGATG
pC464S(mut) GGTGAACACGTT
TCCGCTGTCGGTCCATCAGG
pC531S(mut) GGACAATATCCTCTAC
TCCATTCCGCCTGAAATTGC
pC552S(mut) CGTGCTATTGGAAAAGCTAAT
TCCACAAAATTTTTGGCC
a
Restriction endonuclease site introduced by primer.
b
The BamHI site is found up- or downstream of the primer.
Membrane targeting on demand S. Gompf et al.
5306 FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS
(London Resin Company Ltd, Reading, UK) and polymer-
ized for 30 h at 55 ° C. Thin sections were cut from the
resin bloc and transferred onto formvar-coated nickel grids.
For immunogold labeling, grids were placed on drops of
the respective solutions in the following order: saturated
sodium meta-periodate; water; NaCl ⁄ Pi containing 2%

glycine; NaCl ⁄ Pi; NaCl ⁄ Pi containing 1% BSA, 0.1%
Tween 20, NaCl ⁄ Pi, 0.1% BSA, 0.05% Tween 20. Sections
were incubated with the anti-MDL1 serum. After removal
of unbound antibodies, sections were incubated with sec-
ondary goat anti-rabbit serum coupled to gold particles
(diameter of 10 nm). Carefully washed slices were briefly
treated with 1% glutardialdehyde in NaCl ⁄ Pi and, after
contrasting with uranyl acetate and lead citrate, prepara-
tions were analyzed by electron microscopy (EM 208S, FEI
Company, Eindhoven, the Netherlands).
Blue-Native PAGE
Total membranes (10 mgÆmL
)1
) were solubilized in digito-
nin buffer [20 mm Tris ⁄ HCl pH 7.4, 50 mm NaCl, 10%
(v ⁄ v) glycerol, 1 mm EDTA, 1 mm phenylmethanesulfonyl
fluoride, 1% (w ⁄ v) digitonin (Calbiochem, Darmstadt, Ger-
many)] for 1 h at 4 °C under gentle rotation. Loading dye
(10 mm Bis-Tris pH 7, 50 mm e-amino-n-caproic acid, 5%
(w ⁄ v) Coomassie Blue (G) was added to solubilized mate-
rial after ultracentrifugation (100 000 g, 30 min, 4 °C) [28].
Blue-Native electrophoresis (gradient 6.0–16.5%) was per-
formed as previously described [50]. Apoferritin (443 kDa),
b-amylase (200 kDa), alcohol dehydrogenase (150 kDa),
and albumin (66 kDa) were used as markers.
Limited trypsin digestion and factor Xa cleavage
To determine the membrane orientation of MDL1 in iso-
lated organelles, 15 lg of organelles were incubated for
15 min on ice with increasing concentrations of trypsin (up
to 0.1 mgÆmL

)1
). Proteolysis was stopped by addition of tri-
chloroacetic acid to a final concentration of 7.5% (v ⁄ v).
After subsequent centrifugation, pellets were washed with
ice-cold acetone and resuspended in sample buffer [51]. For
factor Xa cleavage, mitoplasts and inverted IMMs (30 lg)
were incubated with 0.01 mgÆmL
)1
factor Xa in 20 mm
Tris ⁄ HCl pH 8.0, 100 mm NaCl, and 2 mm CaCl
2
for
30 min at 25 °C. The reaction was stopped by addition of
sample buffer and incubation for 10 min at 65 °C. The
accessibility to trypsin and factor Xa was determined by
immunodetection with MDL1-specific and anti-His-tag
(Novagen, San Diego, CA, USA) sera.
Purification of MDL1
Total membranes (10 mgÆ mL
)1
) were solubilized in buf-
fer A (20 mm Tris ⁄ HCl pH 8.0, 150 mm NaCl, 15% (v ⁄ v)
glycerol, EDTA-free complete protease inhibitor cocktail
(final concentration according to manufacturer, Roche,
Mannheim, Germany), 1% (w ⁄ v) digitonin (Calbiochem))
for 1 h at 4 °C under gentle rotation. Nonsolubilized mate-
rial was removed by ultracentrifugation (100 000 g, 30 min,
4 °C; Ti80, Beckman Coulter) and the soluble fraction was
loaded onto a 1 mL Ni
2+

-High-Trap Chelating column
(GE Healthcare, Piscataway, NJ, USA) equilibrated with
buffer B (20 mm Tris ⁄ HCl pH 8.0, 150 mm NaCl, 15%
(v ⁄ v) glycerol, 2 m m imidazole, 0.1% (w ⁄ v) digitonin).
After washing with buffer B containing 80 and 160 mm
imidazole, the protein was eluted in buffer B containing
400 mm imidazole.
Gel filtration
Full-length and leaderless MDL1 were analyzed by gel fil-
tration on a Superdexä 200 PC 3.2 (GE Healthcare) equili-
brated with SEC buffer (20 mm Tris ⁄ HCl pH 8.0, 150 mm
NaCl and 0.1% (w ⁄ v) digitonin); 60 lg of protein was
loaded at a flow rate of 50 lLÆmin
)1
.30lL fractions were
collected and analyzed by SDS ⁄ PAGE and immunoblotting
using anti-MDL1 serum. Ferritin (443 kDa), b-amylase
(200 kDa), and BSA (70 kDa) in SEC buffer without deter-
gent were used for calibration.
ATPase assays
The ATPase activity was essentially determined as described
[31]. 20 mm dithiothreitol was added to 1 lm purified
MDL1. The reaction was started by addition of ATP con-
taining buffer (20 mm Tris ⁄ HCl pH 8.0, 150 mm NaCl,
20 mm MgCl
2
, 0.1% (w ⁄ v) digitonin, 10 mm ATP traced
(370 000 : 1) with [c-
32
P]ATP (specific activity 110 TBqÆ

mmol
)1
; Hartmann Analytic, Braunschweig, Germany) in a
1 : 1 ratio at 30 °C. The reaction was stopped after 10 min
by adding 1 mL of 10 mm ammonium molybdate in 1 m
HCl. Subsequently, 15 lLof20mm H
3
PO
4
and 2 mL of a
butanol ⁄ cyclohexane ⁄ acetone (5 : 5 : 1) mixture were added.
After rigorous vortexing, the organic phase was extracted
and the radioactivity was quantified by liquid scintillation
b-counting (Beckman LS6500 Liquid Scintillation Counter;
Beckman Coulter Inc., Fullerton, CA, USA). K
m ATP
values
were derived by fitting the data to the Michaelis–Menten
equation. Specific inhibition of the ATPase activity was ana-
lyzed at various concentrations of vanadate, using the char-
coal adsorption method in combination with [c-
32
P]ATP
[52]. 0.5 lm of purified MDL1 was incubated with increasing
concentrations of ortho-vanadate. By addition of buffer
supplemented with ATP, the reaction was initiated and incu-
bated for 15 min at 30 °C. 750 lL of ice-cold 10% charcoal
in 10 mm EDTA were added to terminate the reaction. After
rigorous agitation, reactions were incubated for 3 h on ice to
allow maximal binding of free ATP to the charcoal. After

S. Gompf et al. Membrane targeting on demand
FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS 5307
centrifugation at 20 000 g for 15 min, the radioactivity of the
supernatant was measured by b -counting in the presence of
2 mL scintillation fluid. Data were fitted to a dose–response
equation and the half-maximal inhibitory concentration
(IC
50
) was calculated (Eqn 1).
activity ½%¼
100%
1 þ
orthoÀvanadate ½lM
IC
50
½lM

ð1Þ
Peptide transport
Combinatorial peptide libraries and defined peptides were
generated on a robot system by solid phase chemistry using
Fmoc [N-(9-fluorenyl) methoxycarbonyl] amino acids, as
previously described [41]. Peptides and peptide libraries were
radiolabeled and peptide transport was analyzed as described
[53] with the following modifications: microsomes (75 lgof
total protein) of the yeast strain Y06425 (BY4741; Mat a;
his3D1; leu2D0; met15D0; ura3D0; YLL048c::kanMX4),
expressing MDL1(60-695), were incubated with 1 lm
125
I-labeled peptides, 3 mm ATP or 5 U apyrase in 50 lLof

AP buffer (NaCl ⁄ Pi, 0.1 mm DTT, 5 mm MgCl
2
, pH 7.0) for
6 min at 30 °C. The reaction was stopped by addition of ice-
cold AP buffer (500 lL), supplemented with 10 mm EDTA.
After centrifugation and two washing steps, the radioactivity
in microsomes was measured by c-counting (Cobra II;
Packard Instrument Company, Meriden, CT, USA).
Complementation assay
DKY230 cells [5] were cultured on SCD plates without ura-
cil and tryptophan and used for replica plating. Selection
plates containing SCD without tryptophan were supple-
mented with 1 gÆL
)1
of 5-FOA. Resistant colonies appeared
within 5–7 days at 30 °C [54]. For 5-FOA in vivo screens,
several MDL1 mutants were generated by using MDL1
plasmid [5] as template. The mutants K473A, E599Q, and
H631A were generated using the mutagenesis primers
pK473A(mut), pE599Q(mut), and pH 631 A(mut), respec-
tively. To generate leaderless MDL1, primers p5(f) and
p4(r) were used. A Cys-less MDL1 construct was generated
with the mutagenesis primers pC257S(mut), pC464S(mut),
pC531S(mut) and pC552S(mut).
Acknowledgements
We are grateful to Drs Maja Chloupkova, David
M. Koeller (Oregon Health and Science University,
Portland, OR, USA) and Peter Ko
¨
tter (Goethe-Univer-

sity, Institute of Microbiology, Frankfurt, Germany) for
kindly providing yeast strains. We thank Dr Peter Ko
¨
t-
ter for helpful discussions regarding yeast genetics. This
work was supported by the Deutsche Forschungsgeme-
inschaft (DFG) ) SFB472 Molecular Bioenergetics.
References
1 Holland BI, Cole S, Kuchler K & Higgens C (2002)
ABC Proteins ) From Bacteria to Man. Academic Press,
London.
2 Davidson AL & Chen J (2004) ATP-binding cassette
transporters in bacteria. Annu Rev Biochem 73, 241–268.
3 Schmitt L & Tampe
´
R (2002) Structure and mechanism
of ABC transporters. Curr Opin Struct Biol 12, 754–760.
4 Young L, Leonhard K, Tatsuta T, Trowsdale J & Lan-
ger T (2001) Role of the ABC transporter Mdl1 in pep-
tide export from mitochondria. Science 291, 2135–2138.
5 Chloupkova M, LeBard LS & Koeller DM (2003)
MDL1 is a high copy suppressor of ATM1: evidence
for a role in resistance to oxidative stress. J Mol Biol
331, 155–165.
6 Lill R & Kispal G (2000) Maturation of cellular Fe-S
proteins: an essential function of mitochondria. Trends
Biochem Sci 25, 352–356.
7 Sickmann A, Reinders J, Wagner Y, Joppich C, Zahedi
R, Meyer HE, Schonfisch B, Perschil I, Chacinska A,
Guiard B et al. (2003) The proteome of Saccharomyces

cerevisiae mitochondria. Proc Natl Acad Sci USA 100,
13207–13212.
8 Taylor SW, Fahy E, Zhang B, Glenn GM, Warnock DE,
Wiley S, Murphy AN, Gaucher SP, Capaldi RA, Gibson
BW et al. (2003) Characterization of the human heart
mitochondrial proteome. Nat Biotechnol 21, 281–286.
9 Herrmann JM & Neupert W (2000) Protein transport
into mitochondria. Curr Opin Microbiol 3, 210–214.
10 Jensen RE & Dunn CD (2002) Protein import into and
across the mitochondrial inner membrane: role of the
TIM23 and TIM22 translocons. Biochim Biophys Acta
1592, 25–34.
11 Endo T, Yamamoto H & Esaki M (2003) Functional
cooperation and separation of translocators in protein
import into mitochondria, the double-membrane
bounded organelles. J Cell Sci 116, 3259–3267.
12 Koehler CM (2004) New developments in mitochondrial
assembly. Annu Rev Cell Dev Biol 20, 309–335.
13 Wiedemann N, Frazier AE & Pfanner N (2004) The
protein import machinery of mitochondria. J Biol Chem
279, 14473–14476.
14 Pfanner N, Wiedemann N, Meisinger C & Lithgow T
(2004) Assembling the mitochondrial outer membrane.
Nat Struct Mol Biol 11, 1044–1048.
15 Neupert W (1997) Protein import into mitochondria.
Annu Rev Biochem 66, 863–917.
16 Schatz G & Dobberstein B (1996) Common principles
of protein translocation across membranes. Science 271,
1519–1526.
17 Gavel Y & von Heijne G (1990) Cleavage-site motifs in

mitochondrial targeting peptides. Protein Eng 4, 33–37.
18 Hendrick JP, Hodges PE & Rosenberg LE (1989)
Survey of amino-terminal proteolytic cleavage sites in
Membrane targeting on demand S. Gompf et al.
5308 FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS
mitochondrial precursor proteins: leader peptides
cleaved by two matrix proteases share a three-amino
acid motif. Proc Natl Acad Sci USA 86, 4056–4060.
19 Glick BS, Brandt A, Cunningham K, Muller S, Hall-
berg RL & Schatz G (1992) Cytochromes c1 and b2 are
sorted to the intermembrane space of yeast mitochon-
dria by a stop-transfer mechanism. Cell 69, 809–822.
20 Stuart R (2002) Insertion of proteins into the inner
membrane of mitochondria: the role of the Oxa1 com-
plex. Biochim Biophys Acta 1592, 79–87.
21 Uebel S, Kraas W, Kienle S, Wiesmu
¨
ller KH, Jung G &
Tampe
´
R (1997) Recognition principle of the TAP
transporter disclosed by combinatorial peptide libraries.
Proc Natl Acad Sci USA 94, 8976–8981.
22 Urlinger S, Kuchler K, Meyer TH, Uebel S & Tampe
´
R
(1997) Intracellular location, complex formation, and
function of the transporter associated with antigen pro-
cessing in yeast. Eur J Biochem 245, 266–272.
23 Wolters JC, Abele R & Tampe

´
R (2005) Selective and
ATP-dependent translocation of peptides by the homo-
dimeric ATP binding cassette transporter TAP-like
(ABCB9). J Biol Chem 280, 23631–23636.
24 Baker RT & Varshavsky A (1995) Yeast N-terminal
amidase. A new enzyme and component of the N-end
rule pathway. J Biol Chem 270, 12065–12074.
25 Maarse AC, Van Loon AP, Riezman H, Gregor I,
Schatz G & Grivell LA (1984) Subunit IV of yeast cyto-
chrome c oxidase: cloning and nucleotide sequencing of
the gene and partial amino acid sequencing of the
mature protein. EMBO J 3, 2831–2837.
26 Weiss C, Oppliger W, Vergeres G, Demel R, Jeno P,
Horst M, de Kruijff B, Schatz G & Azem A (1999)
Domain structure and lipid interaction of recombinant
yeast Tim44. Proc Natl Acad Sci USA 96, 8890–8894.
27 Kashiwayama Y, Morita M, Kamijo K & Imanaka T
(2002) Nucleotide-induced conformational changes of
PMP70, an ATP binding cassette transporter on rat
liver peroxisomal membranes. Biochem Biophys Res
Commun 291, 1245–1251.
28 Hofacker M, Gompf S, Zutz A, Presenti C, Haase W,
van der Does C, Model K & Tampe
´
R (2007) Structural
and functional fingerprint of the mitochondrial ATP-
binding cassette transporter Mdl1 from Saccharomyces
cerevisiae. J Biol Chem 282, 3951–3961.
29 Urbatsch IL, Sankaran B, Bhagat S & Senior AE (1995)

Both P-glycoprotein nucleotide-binding sites are catalyt-
ically active. J Biol Chem 270, 26956–26961.
30 Urbatsch IL, Sankaran B, Weber J & Senior AE (1995)
P-glycoprotein is stably inhibited by vanadate-induced
trapping of nucleotide at a single catalytic site. J Biol
Chem 270, 19383–19390.
31 Gorbulev S, Abele R & Tampe
´
R (2001) Allosteric cross-
talk between peptide-binding, transport, and ATP
hydrolysis of the ABC transporter TAP. Proc Natl Acad
Sci USA 98, 3732–3737.
32 Detmers FJ, Lanfermeijer FC, Abele R, Jack RW,
Tampe
´
R, Konings WN & Poolman B (2000) Combina-
torial peptide libraries reveal the ligand-binding mecha-
nism of the oligopeptide receptor OppA of Lactococcus
lactis. Proc Natl Acad Sci USA 97, 12487–12492.
33 Doeven MK, Abele R, Tampe
´
R & Poolman B (2004)
The binding specificity of OppA determines the selectiv-
ity of the oligopeptide ATP-binding cassette transporter.
J Biol Chem 279 , 32301–32307.
34 Fischer Lindahl K, Hermel E, Loveland BE & Wang
CR (1991) Maternally transmitted antigen of mice: a
model transplantation antigen. Annu Rev Immunol 9,
351–372.
35 Smith PC, Karpowich N, Millen L, Moody JE, Rosen J,

Thomas PJ & Hunt JF (2002) ATP binding to the motor
domain from an ABC transporter drives formation of a
nucleotide sandwich dimer. Mol Cell 10, 139–149.
36 Janas E, Hofacker M, Chen M, Gompf S, van der Does
C & Tampe
´
R (2003) The ATP hydrolysis cycle of the
nucleotide-binding domain of the mitochondrial ATP-
binding cassette transporter Mdl1p. J Biol Chem 278,
26862–26869.
37 Zaitseva J, Jenewein S, Jumpertz T, Holland IB & Sch-
mitt L (2005) H662 is the linchpin of ATP hydrolysis in
the nucleotide-binding domain of the ABC transporter
HlyB. EMBO J 24 , 1901–1910.
38 Kuhnke G, Neumann K, Mu
¨
hlenhoff U & Lill R (2006)
Stimulation of the ATPase activity of the yeast mito-
chondrial ABC transporter Atm1p by thiol compounds.
Mol Membr Biol 23, 173–184.
39 Miyazaki E, Kida Y, Mihara K & Sakaguchi M (2005)
Switching the sorting mode of membrane proteins from
cotranslational endoplasmic reticulum targeting to post-
translational mitochondrial import. Mol Biol Cell 16,
1788–1799.
40 Graf SA, Haigh SE, Corson ED & Shirihai OS (2004)
Targeting, import, and dimerization of a mammalian
mitochondrial ATP binding cassette (ABC) transporter,
ABCB10 (ABC-me). J Biol Chem 279, 42954–42963.
41 Uebel S, Meyer TH, Kraas W, Kienle S, Jung G,

Wiesmu
¨
ller KH & Tampe
´
R (1995) Requirements for
peptide binding to the human transporter associated
with antigen processing revealed by peptide scans and
complex peptide libraries. J Biol Chem 270, 18512–
18516.
42 Morse MC, Bleau G, Dabhi VM, Hetu F, Drobetsky
EA, Lindahl KF & Perreault C (1996) The COI mito-
chondrial gene encodes a minor histocompatibility anti-
gen presented by H2-M3. J Immunol 156, 3301–3307.
43 Herget M & Tampe
´
R (2007) Intracellular peptide
transporters in human ) compartmentalization of the
‘peptidome’. Pflugers Arch 453, 591–600.
44 Brachmann CB, Davies A, Cost GJ, Caputo E, Li J,
Hieter P & Boeke JD (1998) Designer deletion strains
derived from Saccharomyces cerevisiae S288C: a useful
S. Gompf et al. Membrane targeting on demand
FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS 5309
set of strains and plasmids for PCR-mediated gene dis-
ruption and other applications. Yeast 14, 115–132.
45 Sherman F (1991) Getting started with yeast. Methods
Enzymol 194, 3–21.
46 Palsdottir H & Hunte C (2003) Purification of cyto-
chrome bc1 complex from yeast. In Membrane Protein
Purification and Crystallization: A Practical Guide

(Hunte C, von Jagow G & Scha
¨
gger H, eds), pp. 191–
203. Academic Press, San Diego, CA.
47 Meisinger C, Sommer T & Pfanner N (2000) Purifica-
tion of Saccharomcyes cerevisiae mitochondria devoid of
microsomal and cytosolic contaminations. Anal Biochem
287, 339–342.
48 Fritz S, Rapaport D, Klanner E, Neupert W & Wester-
mann B (2001) Connection of the mitochondrial outer
and inner membranes by Fzo1 is critical for organellar
fusion. J Cell Biol 152, 683–692.
49 Jascur T (1991) Import of precursor proteins into yeast
submitochondrial particles. Methods Cell Biol 34, 359–
368.
50 Scha
¨
gger H (2001) Blue-native gels to isolate protein
complexes from mitochondria. Methods Cell Biol 65,
231–244.
51 van der Does C, den Blaauwen T, de Wit JG, Manting
EH, Groot NA, Fekkes P & Driessen AJ (1996) SecA is
an intrinsic subunit of the Escherichia coli preprotein
translocase and exposes its carboxyl terminus to the
periplasm. Mol Microbiol 22, 619–629.
52 Tombline G & Fishel R (2002) Biochemical character-
ization of the human RAD51 protein. I. ATP hydroly-
sis. J Biol Chem 277, 14417–14425.
53 Chen M, Abele R & Tampe
´

R (2003) Peptides induce
ATP hydrolysis at both subunits of the transporter asso-
ciated with antigen processing. J Biol Chem 278, 29686–
29692.
54 Boeke JD, Trueheart J, Natsoulis G & Fink GR (1987)
5-Fluoroorotic acid as a selective agent in yeast mole-
cular genetics. Methods Enzymol 154 , 164–175.
Membrane targeting on demand S. Gompf et al.
5310 FEBS Journal 274 (2007) 5298–5310 ª 2007 The Authors Journal compilation ª 2007 FEBS