Báo cáo khoa học: Protein transport in chloroplasts ) targeting to the intermembrane space pptx

Protein transport in chloroplasts
)
targeting to the
intermembrane space
Lea Vojta, Ju
¨
rgen Soll and Bettina Bo
¨
lter
University of Munich, Department of Botany, Munich, Germany
Protein import into organelles not only requires trans-
location into the organelle, but also sorting to the
various subcompartments. Chloroplasts are highly
structured organelles that contain three distinct mem-
brane systems (i.e. the outer and inner envelope mem-
brane and the photosynthetic thylakoid membranes) as
well as three soluble subcompartments, the thylakoid
lumen, the stroma and the intermembrane space. The
large majority of chloroplast proteins are encoded by
nuclear genes, synthesized in the cytosol and post-
translationally imported into the organelle [1,2]. In the
cytosol, preproteins associate with different molecular
chaperones (e.g. Hsp70 or Hsp90). This interaction
determines the primary receptor at the chloroplast
surface that is chosen by the preprotein chaperone
complex [3].
Toc64 recognizes Hsp90-associated preproteins,
which are released from Hsp90 and transferred to
Toc34 in an ATP dependent process [3]. Toc34 func-
tions as a primary receptor of Hsp70-associated as well
as monomeric precursor proteins. Toc34 receptor func-

tion is regulated by GTP binding and hydrolysis [4–7].
Toc34
GDP
interacts with the second G-protein in the
Toc complex, Toc159, and simultaneously transfers the
preprotein to Toc159. Toc159 action facilitates pre-
protein translocation through the Toc75 channel [8].
Translocation of stromal or thylakoid preproteins
occurs simultaneously through the Toc and Tic trans-
locon. Translocation across the inner envelope mem-
brane requires ATP, probably for the action of
stromal molecular chaperones [9,10].
Beside this standard import pathway, which is taken
by the majority of chloroplast preproteins, several spe-
cialized translocation pathways have been described.
These can be distinguished in general by the differ-
ences in ATP-, Toc-receptor or presequence require-
ment. For example, Tic32 and QORH, two proteins of
the chloroplasts inner envelope, do not contain an
N-terminal targeting sequence, but Tic32 and QORH
are targeted to chloroplasts by internal sequence infor-
mation present at the N- or C-terminus, respectively
[11,12]. In addition, Tic32 translocation requires
<20 lm ATP in contrast to stromal preproteins or
the precursor of Tic110, which require > 20 lm ATP.
This indicates that translocation of Tic32 does not
involve the action of stromal chaperones. Insertion of
Keywords
intermembrane space; MGD1; Tic; Tic22;
Toc

Correspondence
J. Soll, University of Munich, Botany,
Menzinger Strasse 67, 80638 Munich,
Germany
Fax: +49 89 17861185
Tel: +49 89 17861245
E-mail:
(Received 29 May 2007, revised 31 July
2007, accepted 1 August 2007)
doi:10.1111/j.1742-4658.2007.06023.x
The import of proteins destined for the intermembrane space of chloro-
plasts has not been investigated in detail up to now. By investigating
energy requirements and time courses, as well as performing competition
experiments, we show that the two intermembrane space components Tic22
and MGD1 (E.C. 2.4.1.46) both engage the Toc machinery for crossing the
outer envelope, whereas their pathways diverge thereafter. Although
MGD1 appears to at least partly cross the inner envelope, Tic22 very likely
reaches its mature form in the intermembrane space without involving
stromal components. Thus, different pathways for intermembrane space
targeting probably exist in chloroplasts.
Abbreviation
LSU, large subunit of RubisCO.
FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS 5043
preproteins into the outer envelope is less well charac-
terized and no components have yet been identified,
except for the import of the precursor of Toc75. Toc75
is made with a cleavable N-terminal presequence and
uses the Toc and Tic translocon in a specialized import
pathway [13]. The Toc75 preprotein contains a bipar-
tite targeting signal. The N-terminal domain is respon-

sible for chloroplast targeting and translocation
initiation. Translocation is halted when this domain is
translocated across the Tic complex only to become
processed by the stromal processing peptidase. The
protein subsequently retracts from the Tic translocon
and is redirected to the outer envelope [13]. Toc75
itself could play a role in the insertion of the outer
envelope protein OEP14 [14].
In the present study, we describe the import charac-
teristics of two proteins, namely Tic22 and the MGDG
synthase (MGD1, E.C. 2.4.1.46), localized in the inter-
membrane space between the outer and the inner enve-
lope. Tic22 is a subunit of a soluble intermembrane
space complex, which facilitates the transfer of prepro-
teins from the Toc to the Tic translocon [15]. The
import behaviour of pea Tic22 has previously been
described by Kuoranov et al. [16]; thus, we used this
protein as a reference for intermembrane space target-
ing. However, in the course of our studies, we obtained
contrasting results to those reported by Kouranov
et al. [16], which led us to a refined model for the
import of pTic22. MGDG synthases are proposed to
be associated with either the inner or the outer enve-
lope, dependent on the plant species studied [17–19].
Two different enzyme types are found in chloroplasts;
type B enzymes (MDG2 ⁄ 3) appear to be in the outer
envelope membrane [20], whereas further studies in
Arabidopsis strongly indicated that type A MGDG
synthase, represented by MGD1, is an intermembrane
space component and bound to the intermembrane

face of the inner envelope, although this has not been
demonstrated unequivocally [20]. Our results show that
both pTic22 and pMGD1 use the Toc translocon, but
they differ in their ATP-requirement for translocation,
indicating clear differences in the final translocation
steps. Furthermore, pMGD1 import is greatly stimu-
lated by the addition of potassium phosphate in the
import reaction. These data suggest that chloroplasts
have established a number of specialized translocation
pathways.
Results
As a starting point to investigate the import character-
istics of the intermembrane space proteins Tic22
and MGD1, in vitro import experiments into isolated
chloroplasts were performed using
35
S-labelled precur-
sor proteins. Both pTic22 and pMGD1 were imported
and processed to a smaller mature form in the presence
of ATP (Fig. 1). Upon protease treatment after com-
pletion of the import assay, the organellar surface
bound pMGD1 was completely removed as expected,
although a significant amount of pTic22 was protease
resistant (Fig. 1A). This phenomenon was consistently
observed, indicating that the rate of translocation
exceeds the rate of processing for pTic22. The pTic22
translation product is completely degraded by the pro-
tease thermolysin, indicating that the precursor is not
protease resistant per se (data not shown). Recognition
and translocation of pTic22 is completely dependent

on the N-terminal cleavable presequence. An N-termi-
nal truncation of 60 amino acids, most likely repre-
senting the entire targeting signal, was neither bound,
nor translocated into isolated chloroplasts (Fig. 1A,
lanes 4–6). The import of QORH into the inner enve-
lope of chloroplasts was shown to depend on targeting
information present in the C-terminus of the pre-
protein [11]. Therefore, we constructed a pTic22DC
mutant in which the carboxy-terminal amino acids
were deleted but still contained the N-terminal target-
ing signal. PTic22DC imported with an efficiency simi-
lar to the wild-type protein (Fig. 1A, lanes 7–9),
indicating that the N-terminal presequence is both nec-
essary and sufficient for recognition and translocation
[16].
The import yield for pMGD1 was consistently low
and the running behaviour of the processed mature
form was partially distorted by the large subunit of
RubisCO (LSU) at 54 kDa, resulting in a sharp band
in front of LSU, and a smear of radioactively-labelled
protein mixed with LSU (Fig. 1B, lanes 3–6, indicated
by an asterisk). We have demonstrated earlier that the
presence of KPi buffer could greatly stimulate the
import yield of the inner envelope protein IEP96 by an
unknown mechanism [21]. When we used 80 mm KPi
in the import reaction of pMGD1, both binding and
translocation were stimulated by several-fold (Fig. 1C,
lanes 3–5), whereas the import of pTic22 was not influ-
enced (data not shown). These initial data already indi-
cate subtle, but clearly distinguishable differences in

the import characteristics for these two intermembrane
space localized preproteins.
First, we wanted to verify the localization of MGD1
to the inter envelope space in chloroplasts. An immu-
noblot was carried out using purified inner and outer
envelope membranes from pea chloroplasts and anti-
sera against Tic110 and Toc75 as a marker for the
localization (Fig. 1D). MGD1 was found to be in
almost equal distribution in both membrane fractions,
Intermembrane space targeting in chloroplasts L. Vojta et al.
5044 FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS
whereas the marker proteins Toc75 and Tic110 were
largely confined to their respective localization, indicat-
ing that MGD1 spans the intermembrane space and is
in contact with both envelope membranes. Inner and
outer envelopes are separated by treatment of intact
chloroplasts with hypertonic buffer during the fraction-
ation procedure; this may result in the observed distri-
bution behaviour. However, the MGD1 population in
the inner envelope membrane behaves differently from
that in the outer envelope upon treatment with 6 m
urea or 0.1 m Na
2
CO
3
. Although MGD1 present in
the inner envelope is recovered almost exclusively in
the urea or Na
2
CO

3
insoluble membrane fraction,
MGD1 in the outer envelope is partly or largely recov-
ered in the Na
2
CO
3
or urea soluble fraction, respec-
tively. MGD1 contains no obvious hydrophobic
transmembrane a-helices. Therefore, we propose that
MGD1 binds to the outer leaflet of the inner envelope
by hydrophobic interactions, which is in accordance
with its behaviour with respect to urea and Na
2
CO
3
extraction. The protein might span the intermembrane
space and simultaneously interact with the inner leaflet
of the outer envelope (but less strongly than with the
inner membrane), which could explain the different
behaviour upon treatment with high salt concentra-
tions and basic pH.
To determine these differences more clearly, import
experiments were carried out into isolated chloroplasts
that contained two different radioactively-labelled
preproteins simultaneously, either pTic22 or pMGD1
together with pSSU, a stromal marker, respectively, or
a mixture of the two intermembrane space proteins.
This was performed for better comparison of the
import behaviour. Under these conditions, we can

exclude any differences in the treatment of the sam-
ples. No difference in the import efficiency was
observed with respect to the number of precursor pro-
teins present in one sample (data not shown). The
results from these experiments were quantified and a
representative example of each is shown as a gel image
A
B
D
C
Fig. 1. (A) AtTic22 is imported into pea chloroplasts. In vitro syn-
thesized [
35
S]pTic22 (lanes 1–3), [
35
S]mTic22 (lanes 4–6) and
[
35
S]Tic22DC (lanes 7–9) were incubated with isolated intact chlo-
roplasts at 25 °C for 20 min, in a standard import reaction contain-
ing 3 m
M ATP. After import, samples were reisolated on a Percoll
cushion and treated with thermolysin (Th) (lanes 3, 6 and 9). The
results were analyzed by SDS ⁄ PAGE. Lanes 1, 4 and 7 represent
10% of the translation product used for the import reactions. The
positions of pTic22, mTic22 and Tic22DC are indicated by arrows.
(B) Import of atMGD1 into pea chloroplasts. In vitro synthesized
[
35
S]pMGD1 was incubated with isolated intact pea chloroplasts at

25 °C for 20 min, in a standard import reaction. Lane 1 represents
10% of the translation product used for the import. In lane 2, trans-
lation product was treated with thermolysin. Import was performed
in the absence (lanes 3 and 4) or presence (lanes 5 and 6) of 3 m
M
ATP. After import, chloroplasts were reisolated on a Percoll cushion
and subjected to the treatment with 0.5 lg thermolysin (Th) per lg
chlorophyll (lanes 4 and 6). Untreated samples are shown in lanes
3 and 5. The results were analyzed by SDS ⁄ PAGE. (C) Import of
pMGD1 performed in the presence of 80 m
M KPi. Precursor protein
(pMGD1), mature form (mMGD1) and typical thermolysin degrada-
tion pattern (Th) are indicated. Lane 1 represents 10% of the trans-
lation product. Import was performed in the absence (lanes 2 and
3) or presence (lanes 4 and 5) of 3 m
M ATP. After import chloro-
plasts were reisolated on a Percoll cushion and subjected to the
treatment with 0.5 lg thermolysin (Th) per lg chlorophyll (lanes 3
and 5). Untreated samples are shown in lanes 2 and 4. The results
were analyzed by SDS ⁄ PAGE. The mature form of MGD1, which is
compressed by LSU without addition of KPi, is marked with an
asterisk. (D) Extraction of MGD1 from inner and outer envelope
vesicles from pea by 0.1
M Na
2
CO
3
,6M urea or 1 M NaCl. Chloro-
plast envelopes were pelleted at 256 000 g for 10 min at 4 °C using
a Himac CS150GX centrifuge and S150AT rotor (Hitachi, Tokyo,

Japan) and resuspended in either 0.1
M Na
2
CO
3
(pH 11.5) (lanes 3,
4, 9 and 10), 6
M urea (lanes 1, 2, 7 and 8) or 1 M NaCl (lanes 5, 6,
11 and 12) for 20 min at room temperature, followed by centrifuga-
tion at 256 000 g for 10 min at 4 °C. The pellet and the supernatant
were analyzed by SDS ⁄ PAGE and immunoblotting. a-MGD1,
a-Tic22, a-Tic110 and a-Toc75 were used for immunodecoration.
L. Vojta et al. Intermembrane space targeting in chloroplasts
FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS 5045
(Fig. 2). Initially, we tested ATP dependence and the
kinetics of the import reaction using radiolabelled pre-
cursor proteins depleted of ATP by gel-filtration as
well as chloroplasts stored in the dark to deplete intra-
organellar ATP (Fig. 2A). Subsequently, import reac-
tions were carried out in the dark under dim green
safelight, which does not support photosynthetic ATP
production. We consistently observed that pTic22 was
imported and processed even in the absence of exoge-
nous ATP. The yield of pTic22 import in the absence
of ATP varied between 20–50% of that in the presence
of ATP. The import of pMGD1 was efficient only at
C
AB
Fig. 2. Comparison of ATP- and time-
demands for import of Tic22, MGD1 and

SSU. Import into intact pea chloroplasts
was performed under standard conditions,
by incubating in vitro synthesized
[
35
S]pTic22 and [
35
S]pMGD1 with chloro-
plasts corresponding to 20 lg chlorophyll
at 25 °C. Parallel imports combining
[
35
S]pMGD1 and [
35
S]pSSU, [
35
S]pTic22 and
[
35
S]pSSU and [
35
S]pTic22 and [
35
S]pMGD1
in the same reaction were performed. After
import, chloroplasts were reisolated on a
Percoll cushion and all samples were trea-
ted with thermolysin. The results were
analyzed by SDS ⁄ PAGE. The respective
precursor and mature forms are indicated

by arrow heads. (A) ATP scale import into
intact pea chloroplasts was performed using
increasing concentrations of ATP from 0 to
3000 l
M for 15 min at 25 °C. The top three
panels represent gel images; the bottom
panel depicts the quantification graph. For
quantification, import at 3 m
M ATP was
taken as maximal import rate. (B) Time-scale
import into intact pea chloroplasts was per-
formed using increasing times as indicated
and 3 m
M ATP at 25 °C. ATP- and time-
dependent import reactions from five inde-
pendent experiments were quantified and
the results presented graphically. The top
three panels represent gel images; the bot-
tom panel depicts the quantification graph.
Import after 20 min was calculated as maxi-
mal import rate. (C) Stroma was isolated
from pea chloroplasts and incubated with
radioactively-labelled translation products
[
35
S]pTic22, [
35
S]pMGD1 and [
35
S]pSSU for

90 min at 26 °C. Reactions were stopped by
addition of Laemmli buffer and samples
were analyzed by SDS ⁄ PAGE. Lanes 1, 3
and 5 represent 2 lL of the corresponding
translation products, and lanes 2, 4 and 6
represent 2 lL of the translation product
which was incubated with isolated stromal
fraction, respectively. Precursor and mature
forms of MGD1 and SSU, appearing after
processing by a stromal processing assay,
are indicated by arrow heads.
Intermembrane space targeting in chloroplasts L. Vojta et al.
5046 FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS
ATP concentrations above 100 lm. The import of the
stromal marker pSSU, which imports constantly with
high yield, clearly requires also exogenous ATP. Prom-
inent amounts of mSSU accumulated at 25 lm ATP,
which we consistently observed for this highly import
competent preprotein. Import increased almost linearly
up to 0.5–1 mm ATP.
Likewise, the import kinetics clearly differ between
the three preproteins. Although pSSU imports extre-
mely rapidly, and mSSU is detectable already after a
few seconds of import (zero time point) and continues
linearly only for up to 5 min, the import of both
pTic22 and pMGD1 is much slower. Import becomes
detectable only after 1–2 min and then continues line-
arly for up to 20 min (Fig. 2B). Another indication for
the different import pathways of pTic22 and pMGD1,
respectively, is precursor cleavage by the stromal pro-

cessing peptidase. Whereas the control protein pSSU
as well as pMGD1 are processed by the SPP in a stro-
mal processing assay, pTic22 remains intact (Fig. 2C).
As already indicated by the results presented in
Fig. 1 and corroborated by those shown in Fig. 2, clear
differences in the import behaviour can be determined
not only between pTic22 and pMGD1, respectively,
but also between each of the two and pSSU. Another
requirement for the import competence of chloroplasts
for preproteins such as pSSU is the presence of
protease sensitive translocon components in the outer
chloroplasts envelope. In an attempt to determine the
involvement of such protease sensitive components in
the import pathway of pTic22 and pMGD1, we treated
isolated chloroplasts with the protease thermolysin,
which removes exposed parts of translocon components
such as Toc159, Toc64 and Toc34 (Fig. 3). The import
yield of pSSU into protease treated chloroplasts
dropped to approximately 20–30%, which corresponds
well to the results described earlier. The import effi-
ciency of pTic22 and pMGD1 were consistently less
susceptible to protease pretreatment of organelles, and
residual imports rates vary between 20–45% for pTic22
and 40–60% for MGD1, respectively (Fig. 3, gel
images are presented on the left side, quantification is
depicted on the right hand side). These data suggest
that import of all preproteins tested is mediated by
proteinaceous components of the outer membrane.
Although differences were observed in the import
behaviour between the intermembrane space proteins

pTic22 and pMGD1 in comparison with the stromal
precursor pSSU, the similarities that were observed
raised the possibility that targeting to the intermem-
brane space may involve subunits of the general
import pathway. In an initial attempt to test this
Fig. 3. Import of pTic22 and pMGD1 is
reduced by thermolysin pretreatment of
chloroplasts. Gel images are depicted on
the left side. Precursor and mature forms
are indicated by arrow heads. A graphical
presentation is shown of the influence of
thermolysin pretreatment of chloroplasts on
the import of pTic22, pMGD1 and pSSU in
the presence of ATP, derived by 2D densi-
tometry evaluation (AIDA image analyser) of
five independently performed experiments
for each protein. For import, intact chloro-
plasts were used that were either pretreat-
ed or not treated with 1 mg of thermolysin
per 1 mg of chlorophyll. Import of
[
35
S]pTic22, [
35
S]pMGD1 and [
35
S]pSSU into
intact pea chloroplasts corresponding to
15 lg of chlorophyll was performed for
15 min at 25 °C for pTic22 and pMGD1, and

5 min for pSSU. After import, chloroplasts
were either subjected to thermolysin post-
treatment (+Th) or not (–Th). Import without
pre- and post-treatment was considered to
be the maximal import rate.
L. Vojta et al. Intermembrane space targeting in chloroplasts
FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS 5047
possibility, we conducted import experiments in the
presence or absence of an excess of unlabelled hetero-
logously expressed soluble chloroplast precursor
protein, the 33 kDa subunit of the oxygen evolving
complex pOE33 (Fig. 4). Unlabelled pOE33, but not
its mature form mOE33, efficiently competed for the
import of
35
S-labelled pSSU; the maximum inhibition
of approximately 90% was reached at a competitor
concentration of 2 lm (Fig. 4, middle panel). The
import efficiency of pTic22 also clearly decreased in
the presence of pOE33. However, at 2 lm competitor,
we still observed a 50% import yield and, at the high-
est competitor concentration tested (10 lm), import
yield was still approximately 30% (Fig. 4, upper
panel). This result is clearly distinct from those previ-
ously obtained [16], which indicate that pTic22 import
is not competed for by standard chloroplasts prepro-
teins such as pSSU (see below). The import of
pMGD1 was reduced to only approximately 50% at
the highest competitor concentration tested and inhibi-
tion was barely detectable at 2 lm pOE33 (Fig. 4,

lower panel). However, in every case, the reduction of
import yield depended on the precursor from of OE33,
as can be deduced from the gel images and the quanti-
fication data. In no case did we observe any significant
effect of the mature form of OE33 on the import of
any of the three preproteins. A slight decrease of
import efficiency was observed at 10 lm mOE33, but
this effect was much weaker than at the same concen-
trations of the precursor form.
In an effort to address the involvement of known
Toc subunits more directly, we expressed the soluble
domain of one of the receptor proteins, Toc34, and
used this peptide as a competitor for import (Fig. 5).
Toc34 or the deletion Toc34DTM, which does not con-
tain the transmembrane anchor and can therefore serve
as a soluble receptor, interact with preproteins but not
their mature forms in solution [5]. To this end, purified
Toc34DTM was preloaded with 3 mm GTP in the
import mix for 10 min at 4 °C. Subsequently,
35
S-
labelled preproteins pTic22, pMGD1 and pSSU were
added to the mixture and incubation continued for
10 min. The import reaction was initiated by addition
of chloroplasts and carried out as described above.
Soluble Toc34DTM competed for the import of all
three preproteins tested (Fig. 5A). Import inhibition
by Toc34DTM was approxiately 60% for pMGD1
and approxiately 50% for both pTic22 and pSSU, as
Fig. 4. pTic22 and pMGD1 compete with

pOE33 for import into chloroplasts. Increas-
ing concentrations of overexpressed protein
pOE33 or its mature form mOE33, as indi-
cated, were added into the import mix prior
to import of [
35
S]pTic22, [
35
S]pMGD1 and
[
35
S]pSSU into intact pea chloroplasts corre-
sponding to 15 lg chlorophyll. Import reac-
tion was performed for 12 min at 25 °C for
pTic22 and pMGD1 and 5 min for the pSSU
control. After import, chloroplasts were sub-
jected to thermolysin post-treatment. Repre-
sentative gel images are depicted on the
left-hand side, and quantifications on the
basis of five independently performed com-
petition experiments are shown on the right.
Import without competitor was considered
to be the maximal import rate.
Intermembrane space targeting in chloroplasts L. Vojta et al.
5048 FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS
indicated in the autoradiographs as well as in the
quantification data (Fig. 5A). Furthermore, we could
demonstrate that Toc34DTM interacts directly with
each of the three preproteins (Fig. 5B). To do so, over-
expressed purified Toc34DTM was coupled to Ni-NTA

matrix (see Experimental procedures) and preloaded
with 1 mm GTP. Subsequently, radioactively-labelled
translation products of pTic22, pMGD1 and pSSU
were incubated with the matrix for 45 min. Unbound
preproteins were washed off and bound preproteins
eluted with 250 mm imidazole containing buffer. In
every case, precursor protein was detected in the imid-
azole eluate, indicating a direct interaction with
Toc34DTM. This interaction was not observed when
we used the mature form mTic22 or mSSU (data not
shown) or the empty Ni-NTA matrix incubated solely
with radioactively-labelled preproteins. Taken together,
these results indicate that the two intermembrane space
proteins, pTic22 and pMGD1, are deduced to share
common import components with pSSU.
In a further attempt to test this idea, we used a
chemical cross-link approach (Fig. 6). Chloroplasts
were incubated with radioactively-labelled precursor
proteins under conditions that allow binding and inser-
tion into the translocon but not complete translocation
(i.e. in the presence of 3 mm ATP at 4 °C). After
preincubation, preproteins were cross-linked using
0.5 mm dithiobis-succinimdyl
-
proprionate. Chloroplasts
were then solubilized with 1% SDS and coimmunopre-
cipitation was performed using antisera against the
translocon subunits Toc34, Toc75, Tic110 and the
outer envelope protein OEP16 as a control (Fig. 6A).
The intermembrane space preproteins pTic22 and

pMGD1 appears to interact strongly with Toc34 and
AB
Fig. 5. Import of pTic22 is inhibited by the soluble domain of the receptor protein Toc34. (A) Increasing concentrations (0–10 lM) of overex-
pressed soluble receptor Toc34DTM, 3 m
M ATP and 3 mM GTP, were added to the import mix prior to import of [
35
S]pTic22, [
35
S]pMGD1
and [
35
S]pSSU into intact pea chloroplasts corresponding to 15 lg of chlorophyll. The import reaction was performed for 12 min at 25 °C for
pTic22 and pMGD1 and 5 min for the pSSU control. After import, chloroplasts were subjected to thermolysin post-treatment (+Th) or not
()Th). After import of MGD1, all samples were post-treated with thermolysin. Representative gel images are shown on the upper left-hand
side. The results were quantified on the basis of five independently performed competition experiments for each preprotein. The quanifica-
tion graph is depicted at the bottom. Import without competitor was considered to be the maximal import rate. (B) pTic22 and pMGD1 inter-
act with the soluble domain of the receptor protein Toc34. For each separate experiment, 300 lg of overexpressed soluble receptor
Toc34DTM was coupled to 10 lL of Ni-NTA matrix and preloaded with 1 m
M GTP. Ni-NTA matrix without bound Toc34DTM was used as
negative control. [
35
S]pTic22, [
35
S]pMGD1 and [
35
S]pSSU were added to the column in binding buffer and incubated for 45 min. The flow
through after incubation (Ft), the third wash of the matrix (W) and the elution with 300 m
M imidazole (E) were analyzed by SDS ⁄ PAGE. Tp
represents 10% of the translation product used in each experiment.
L. Vojta et al. Intermembrane space targeting in chloroplasts

FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS 5049
Toc75 in the outer membrane but not with OEP16.
This result is identical to the data obtained for the
control precursor pSSU, which is clearly established as
a substrate for the general import pathway. The stro-
mal precursor pSSU also interacted strongly with the
inner envelope translocon subunit Tic110, whereas
pTic22 interaction was weak and that of pMGD1
barely detectable. The interaction of pTic22 and
Tic110 might be explained by Tic22 being a compo-
nent of the inner envelope translocon, although a
direct interaction with Tic110 has not yet been shown
[15]. Therefore, this weak interaction might not neces-
sarily indicate a role of Tic110 in the translocation of
pTic22.
Besides the translocation pore, the outer chloroplast
envelope contains Toc75-III, commonly called Toc75,
which constitutes the import channel for the general
import pathway, an evolutionary more ancient isoform
named Toc75-V [22]. The function of Toc75-V, which
constitutes approximately 10% of the total Toc75-like
proteins present in chloroplasts, is not yet clear. There-
fore, we considered whether pTic22 and pMGD1
might also use this channel protein. An identical cross-
link approach to that described above was used
(Fig. 6B). Although we could again detect a cross-link
product between Toc75-III and pTic22, pMGD1 and
pSSU, no interaction of Toc75-V could be found with
any of the three preproteins. Because the coimmuno-
precipitation using Toc75-III and Toc75-V antisera

were carried out from identical samples, we conclude
that Toc75-V plays no role in the translocation of
pTic22 and pMGD1.
The cross-linking of pTic22 and pMGD1 to Toc34
and Toc75 might be nonspecific because these two
translocon subunits are very abundant polypeptides in
the chloroplast outer envelope. Therefore, we repeated
the cross-linking experiments in the presence of an
excess of the soluble chloroplast preprotein pOE33 to
compete for specific binding, whereas the control
experiment contained an equal amount of mOE33
A
C
B
Fig. 6. Chemical cross-linking and immunoprecipitation of pTic22 and pMGD1 to the major components of the translocation machinery. (A)
[
35
S]pTic22, [
35
S]pMGD1 and [
35
S]pSSU were incubated with intact pea chloroplasts corresponding to 20 lg of chlorophyll for 8 min on ice.
After reisolation on a Percoll cushion and subsequent washing, chloroplasts with bound precursor proteins were subjected to cross-linking
using 0.5 m
M dithiobis-succinimdyl-proprionate. Immunoprecipitation was performed after lysis of chloroplasts, centrifugation and solubiliza-
tion of the membranes. Antibodies raised against Toc34, Toc75, Tic110, and OEP16 were used and incubated for 1 h at room temperature.
Antibodies were bound to protein A-sepharose. Ten percent of the flow through after incubation with protein A-sepharose (Ft), 10% of the
third wash (W), and the elution with Laemmli sample buffer (E) were analyzed by SDS ⁄ PAGE. TL indicates 10% of the translation product
used for each experiment. (B) Cross-linking and immunoprecipitation were performed under the same conditions, using antibodies against
two Arabidopsis Toc75-isoforms: atToc75(III) and atToc75(V). (C) Cross-linking and immunoprecipitation were performed in the presence of

10 l
M mOE33 or 10 lM pOE33 in the import mixture.
Intermembrane space targeting in chloroplasts L. Vojta et al.
5050 FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS
(Fig. 6C). The interaction of both pTic22 and pMGD1
was largely abolished in the presence of the competitor
pOE33 but not in the presence of mOE33. Taken
together, we conclude that the intermembrane space
constituents pTic22 and pMGD1 are bona fide sub-
strates of the Toc translocon in chloroplasts.
Discussion
The data presented in the present study suggest that
proteins located in the intermembrane space of chloro-
plasts use components of the general import pathway
in the outer envelope membrane, but nevertheless
clearly show distinctive import behaviour compared to
stromal precursors as well as to each other. We investi-
gated the import characteristics of two intermembrane
space localized proteins, Tic22 and MGDG synthase
(MGD1). The first noticable difference was the concen-
tration of exogenously added ATP that was required
for import. Whereas stromal proteins such as pSSU
need > 0.1 mm ATP for complete translocation, most
likely for the action of stromal chaperones, pTic22 was
already imported at a concentration of less than 20 lm
ATP, indicating that no stromal components are
involved in this process. By contrast, the import of
pMGD1 required more than 100 lm ATP to be effi-
cient, suggesting that this precursor might reach the
stroma before being released to its final destination in

the intermembrane space. This was corroborated by
stromal processing assays in which pMGD1 was pro-
cessed by the SPP, whereas pTic22 was not (Fig. 2C).
With respect to the Toc75 import and processing
features, it is feasible that MGD1 also is partly trans-
located to the stroma, where the transit peptide is
cleaved off by the stromal processing peptidase, and is
then released to its final localization in the intermem-
brane space.
Another difference between pTic22, pMGD1 and
pSSU is their import rate. The stromal precursor
reaches its destination within seconds, whereas the
intermembrane space proteins require 1–2 min. Again,
pTic22 and pMGD1 show distinctive features: the pro-
cessing of pTic22 appears to be very slow compared to
translocation (i.e. because the precursor becomes resis-
tant to externally added protease), whereas pMGD1
import and processing occur at similar rates. In addi-
tion to the differences in ATP requirements and the
stromal processing assay, the import kinetics definitely
indicate that the two preproteins not only use different
pathways, but also are processed by different prote-
ases. The protease responsible for maturation of
pTic22 appears to be located in the intermembrane
space but this requires further investigation.
Competition experiments using pOE33 and
Toc34DTM, as well as cross-linking and immunopre-
cipitation assays, clearly show that both pTic22 and
pMGD1 engage the Toc complex. We could demon-
strate interaction with the receptor Toc34 and the gen-

eral import pore Toc75. This contradicts previously
published studies [16] that showed no competition of a
stromal precursor (i.e. the authors used overexpressed
pSSU as competitor, which should not make a differ-
ence because both pSSU and pOE33 engage the Toc
translocon) with pTic22. In the previous study [16],
however, import rates of pTic22 were generally very
low (approximately 5%) so that the competition effect
clearly demonstrated in the present study might not
have been detectable. Import kinetics have to be estab-
lished for each precursor individually to determine
biochemically relevant data in further experiments.
Therefore, import experiments to elucidate the effect
of competitors were conducted only for 12 min, which
is within the linear time course of pTic22 and pMGD1
import (cf. Fig. 2), whereas Kouranov et al. [16] incu-
bated the import reaction for much longer. This might
be one reason to explain their negative competition
data. Even under our optimal conditions, the import
of pTic22 and pMGD1 is much slower than that of
pSSU, which makes the competition of pSSU import
more visible due to the greater difference of the import
rate with and without competitor, respectively. Fur-
thermore, it is possible that not all components
involved in pSSU translocation play a role in the
import of pTic22 and pMGD1, and therefore the com-
petition by pOE33 is not as pronounced as it is for
pSSU. Nevertheless, the effect of pOE33 on import of
pTic22 and pMGD1 is clearly apparent.
Taken together, our experiments indicate that

pTic22 and pMGD1 use the general import pathway
to traverse the outer envelope and diverge at the level
of the intermembrane space ⁄ inner envelope. The
results of the present study clearly indicate distinct
import pathways not only for proteins destined for
stromal or membrane compartments, but also for the
intermembrane space of chloroplasts.
Experimental procedures
In vitro transcription and translation
The coding region for Tic22 from Arabidopsis thaliana
(At4g33350) was cloned into the vector pSP65 (Promega,
Madison, WI, USA) under the control of the SP6 promoter.
The coding region for MGD1 from A. thaliana (At4g38170)
was cloned into the vector pET21d (Merck, Darmstadt,
Germany) under the control of the T7 promoter. Mature
L. Vojta et al. Intermembrane space targeting in chloroplasts
FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS 5051
forms of these proteins were produced by the removal of the
sequences encoding the transit peptides (177 bp for Tic22
and 321 bp for MGD1) in the same vectors, using standard
PCR protocols. Tic22DC was produced by removal of the
coding sequence corresponding to 225 C-terminal amini
acids of the preprotein. Transcription of linearized plasmids
was carried out in the presence of SP6 (for Tic22) or T7
polymerases (for MGD1) using chemicals obtained from
MBI Fermentas (St Leon-Rot, Germany). Translation was
carried out using the Flexi Rabbit Reticulocyte Lysate
System or the TNT Coupled Reticulocyte Lysate System
(Promega) in the presence of [
35

S]methionine ⁄ cysteine mix-
ture (MGD1) or [
35
S]cysteine (Tic22) for radioactive label-
ling. After translation, the reaction mixture was centrifuged
at 50 000 g for 20 min at 4 °C using a Himac CS150GX
centrifuge (Hitachi, Tokyo, Japan) and S150AT rotor
and the postribosomal supernatant was used for import
experiments.
Chloroplast isolation and protein import
Chloroplasts were isolated from leaves of 9–11-day-old pea
seedlings (Pisum sativum var. Arvica) as described previ-
ously [23]. Prior to import, ATP was depleted from chlo-
roplasts and the in vitro translation product. Intact
chloroplasts were left on ice in the dark for 30 min. Trans-
lation products were treated with 0.5 U apyrase per 10 lL
of translation product at 25 °C for 15 min, or passed
through Micro Bio-Spin Chromatography Columns (Bio-
Rad, Hercules, CA, USA). A standard import reaction con-
tained chloroplasts equivalent to 15–20 lg of chlorophyll in
100 lL of import buffer [330 mm sorbit, 50 mm Hepes ⁄
KOH (pH 7.6), 3 mm MgSO
4,
10 mm methionine, 10 mm
cysteine, 20 mm K-gluconate, 10 mm NaHCO
3
, 2% BSA
(w ⁄ v)], up to 3 mm ATP and
35
S-labelled translation prod-

ucts in the maximal amount of 10% (v ⁄ v). The import
reactions were initiated by the addition of translation prod-
uct and carried out for 20 min at 25 °C, unless indicated
otherwise. Reactions were terminated by separation of
chloroplasts from the reaction mixture by centrifugation
through 40% (v ⁄ v) Percoll cushion. Chloroplasts were
washed once, and import products were separated by
SDS ⁄ PAGE. Radiolabelled proteins were analyzed by a
phosphoimager or by exposure on X-ray films.
Chloroplasts were treated before or after import with the
protease thermolysin. For pretreatment, 1 mg thermolysin
per mg chlorophyll was applied for 30 min on ice. The
reaction was terminated by reisolation on a Percoll density
gradient in the presence of 5 mm EDTA [24]. For post-
treatment, 0.5 lg of thermolysin per lg chlorophyll was
applied for 20 min on ice. The reaction was stopped by the
addition of 5 mm EDTA, sedimenting the chloroplasts and
resuspending them in Laemmli buffer [50 mm Tris pH 6.8,
100 mm b-MeOH, 2% (w/v) SDS, 0.1% bromophenol blue
(w/v), 10% glycerol (v/v)].
Import competition experiments were performed by the
addition of up to 10 lm of purified competitor protein
pOE33, as well as its mature form mOE33, into the import
mixture prior to import. Fifteen micrograms of chlorophyll
per reaction were used and the import reaction lasted 5
(pSSU) to 12 min (Tic22, MGD1) at 25 °C. Competition
for import by the cytosolic domain of Toc34 receptor was
performed in the presence of 3 mm GTP and up to 10 lm
Toc34DTM in the standard import mixture. First,
Toc34DTM was preincubated with GTP in the import mix-

ture for 10 min on ice. Subsequently, radioactively-labelled
translation product was added and incubated for another
10 min to allow the interaction of preprotein with
Toc34DTM. Finally, chloroplasts equivalent to 15 lgof
chlorophyll were added and the reaction was incubated for
10–12 min at 25 °C for Tic22 and MGD1 and 5 min for
pSSU.
Overexpression and purification of pOE33-His6,
mOE33-His6 and Toc34DTM-His6 for competition
experiments
Transformed Escherichia coli BL21(DE3) cells were grown
in LB medium containing 100 lgÆmL
)1
of ampicilin (and
1mm MgSO
4
and 0.4% glucose for mOE33) to an D
600 nm
of 0.6. Expression was induced by 1 mm isopropyl thio-b-
d-galactoside and cells were grown for 3 h at 37 °C.
pOE33 and mOE33 were purified from inclusion bodies
under denaturing conditions via Ni-affinity chromatogra-
phy and eluted by decreasing the pH. Refolding of the pro-
teins was accomplished using stepwise dialysis against 6, 4,
2 and 0 m urea (in 50 mm Tris, pH 8.0), respectively.
Toc34DTM was expressed in a soluble form and purified
under native conditions and elution by 250 mm imidazole.
The protein was always used fresh and diluted so that the
final imidazole concentration in the import reaction did not
exceed 30 mm.

Binding of Toc34 DTM to precursor proteins
Three hundred micrograms of purified Toc34DTM were
coupled to 10 lL of Ni-NTA matrix in binding buffer
(50 mm NaCl, 50 mm Na
i
PO
4
, 0.5% BSA, pH 7.9) for
45 min, rotating at room temperature. The prepared matrix
was preincubated with 1 mm GTP, and subsequently
10–12 lL of a radioactively-labelled translation product
were applied in the reaction containing 1 mm GTP, 2 mm
MgCl
2
,20mm Tris ⁄ HCl (pH 7.6), 50 mm NaCl and 0.5%
BSA. The incubation lasted 45–50 min. The matrix was
subsequently washed three times with wash buffer (50 mm
NaCl, 50 mm NaPi, 30 mm imidazole, pH 7.9) and eluted
with 50 lL of elution buffer (50 mm NaCl, 50 mm NaPi,
300 mm imidazole, pH 7.9). The obtained fractions were
analysed by SDS ⁄ PAGE.
Intermembrane space targeting in chloroplasts L. Vojta et al.
5052 FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS
Chemical cross-linking and immunoprecipitation
After import, chloroplasts were reisolated on a Percoll
cushion, washed, and chemical cross-linking was performed
by incubation of chloroplasts with 0.5 mm dithiobis-
succinimdyl-proprionate (Pierce, Mu
¨
nchen, Germany) in

330 mm sorbit, 50 mm Hepes ⁄ KOH (pH 7.6) and 0.5 mm
CaCl
2
, for 15 min at 4 °C. The reaction was stopped by the
addition of 125 mm glycin and further incubation at 4 °C
for 15 min. Chloroplasts were washed twice in 330 mm sor-
bit, 50 mm Hepes ⁄ KOH (pH 7.6) and 0.5 mm CaCl
2
and
finally lysed in hypotonic buffer [20 mm Hepes ⁄ KOH
(pH 7.6), 5 mm EDTA] for 30 min on ice. A total mem-
brane fraction was recovered by centrifugation at 256 000 g
for 30 min using a Himac (Hitachi) CS150GX centrifuge
and rotor S150AT. Membranes were solublized in 1% SDS
(w ⁄ v), 25 mm Hepes ⁄ KOH (pH 7.6), 150 mm NaCl, diluted
ten-fold in the above buffer in the absence of SDS, centri-
fuged for 2 min at 20 000 g using an Eppendorf (Hamburg,
Germany) 5417R centrifuge and EL033 rotor and the
supernatant used for immunoprecipitation with antisera
against Toc75(III), Toc75(V), Toc34, Tic110 and OEP16.
All antibodies are directed against the corresponding pro-
teins from pea. The affinity matrix was washed with 30
bed-volumes of wash buffer before elution with Laemmli
sample buffer in the presence of b-mercaptoethanol to split
the cross-link products.
Acknowledgements
This work was supported by the Deutsche Forschungs-
gemeinschaft, the Fonds der Chemischen Industrie and
the excellence cluster CIPSM.
References

1 Gutensohn M, Fan E, Frielingsdorf S, Hanner P, Hou
B, Hust B, Klosgen RB (2006) Toc, Tic, Tat-et al.:
structure and function of protein transport machineries
in chloroplasts. J Plant Physiol 163, 333–347.
2 Soll J & Schleiff E (2004) Protein import into chloro-
plasts. Nat Rev Mol Cell Biol 5, 198–208.
3 Qbadou S, Becker T, Mirus O, Tews I, Soll J & Schleiff
E (2006) The molecular chaperone Hsp90 delivers pre-
cursor proteins to the chloroplast import receptor
Toc64. EMBO J 25, 1836–1847.
4 Becker T, Jelic M, Vojta A, Radunz A, Soll J & Schleiff
E (2004) Preprotein recognition by the Toc complex.
EMBO J 23, 520–530.
5 Schleiff E, Soll J, Sveshnikova N, Tien R, Wright S,
Dabney-Smith C, Subramanian C & Bruce BD (2002)
Structural and guanosine triphosphate ⁄ diphosphate
requirements for transit peptide recognition by the cyto-
solic domain of the chloroplast outer envelope receptor,
Toc34. Biochemistry 41, 1934–1946.
6 Schnell DJ & Hebert DN (2003) Protein translocons:
multifunctional mediators of protein translocation
across membranes. Cell 112, 491–505.
7 Sveshnikova N, Soll J & Schleiff E (2000) Toc34 is a
preprotein receptor regulated by GTP and phosphoryla-
tion. Proc Natl Acad Sci USA 97, 4973–4978.
8 Schleiff E, Jelic M & Soll J (2003) A GTP-driven motor
moves proteins across the outer envelope of chloro-
plasts. Proc Natl Acad Sci USA 100, 4604–4609.
9 Kessler F & Blobel G (1996) Interaction of the protein
import and folding machineries in the chloroplast. Proc

Natl Acad Sci USA 93, 7684–7689.
10 Nielsen E, Akita M, Davila-Aponte J & Keegstra K
(1997) Stable association of chloroplastic precursors
with protein translocation complexes that contain pro-
teins from both envelope membranes and a stromal
Hsp100 molecular chaperone. EMBO J 16, 935–946.
11 Miras S, Salvi D, Ferro M, Grunwald D, Garin J,
Joyard J & Rolland N (2002) Non-canonical transit
peptide for import into the chloroplast. J Biol Chem
277, 47770–47778.
12 Nada A & Soll J (2004) Inner envelope protein 32 is
imported into chloroplasts by a novel pathway. J Cell
Sci 117, 3975–3982.
13 Inoue K, Baldwin AJ, Shipman RL, Matsui K, Theg
SM & Ohme-Takagi M (2005) Complete maturation of
the plastid protein translocation channel requires a type
I signal peptidase. J Cell Biol 171, 425–430.
14 Tu SL, Chen LJ, Smith MD, Su YS, Schnell DJ & Li
HM (2004) Import pathways of chloroplast interior pro-
teins and the outer-membrane protein OEP14 converge
at Toc75. Plant Cell 16, 2078–2088.
15 Kouranov A, Chen X, Fuks B & Schnell DJ (1998)
Tic20 and Tic22 are new components of the protein
import apparatus at the chloroplast inner envelope
membrane. J Cell Biol 143, 991–1002.
16 Kouranov A, Wang H & Schnell DJ (1999) Tic22 is
targeted to the intermembrane space of chloroplasts
by a novel pathway. J Biol Chem 274, 25181–25186.
17 Awai K, Marechal E, Block MA, Brun D, Masuda T,
Shimada H, Takamiya K, Ohta H & Joyard J (2001)

Two types of MGDG synthase genes, found widely in
both 16:3 and 18:3 plants, differentially mediate galac-
tolipid syntheses in photosynthetic and nonphotosyn-
thetic tissues in Arabidopsis thaliana. Proc Natl Acad Sci
USA 97, 10960–10965.
18 Jarvis P, Dormann P, Peto CA, Lutes J, Benning C
& Chory J (2000) Galactolipid deficiency and abnor-
mal chloroplast development in the Arabidopsis MGD
synthase 1 mutant.
Proc Natl Acad Sci USA 97,
8175–8179.
19 Miege C, Marechal E, Shimojima M, Awai K, Block
MA, Ohta H, Takamiya K, Douce R & Joyard J
(1999) Biochemical and topological properties of type
A MGDG synthase, a spinach chloroplast envelope
L. Vojta et al. Intermembrane space targeting in chloroplasts
FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS 5053
enzyme catalyzing the synthesis of both prokaryotic
and eukaryotic MGDG. Eur J Biochem 265,
990–1001.
20 Benning C & Ohta H (2005) Three enzyme systems for
galactoglycerolipid biosynthesis are coordinately regu-
lated in plants. J Biol Chem 280, 2397–2400.
21 Hirsch S & Soll J (1995) Import of a new chloroplast
inner envelope protein is greatly stimulated by potas-
sium phosphate. Plant Mol Biol 27, 1173–1181.
22 Eckart K, Eichacker L, Sohrt K, Schleiff E, Heins L &
Soll J (2002) A Toc75-like protein import channel is
abundant in chloroplasts. EMBO Rep 3, 557–562.
23 Waegemann K & Soll J (1991) Characterization of the

protein import apparatus in isolated outer envelopes of
chloroplasts. Plant J 1, 149–158.
24 Waegemann K & Soll J (1995) Characterization and iso-
lation of the chloroplast protein import machinery.
Meth Cell Biol 50, 255–267.
Intermembrane space targeting in chloroplasts L. Vojta et al.
5054 FEBS Journal 274 (2007) 5043–5054 ª 2007 The Authors Journal compilation ª 2007 FEBS