Báo cáo khoa học: Conservative sorting in a primitive plastid The cyanelle of Cyanophora paradoxa potx

Conservative sorting in a primitive plastid
The cyanelle of Cyanophora paradoxa
Juergen M. Steiner
1
, Juergen Bergho
¨
fer
2
, Fumie Yusa
1
, Johannes A. Pompe
1
, Ralf B. Klo
¨
sgen
2
and Wolfgang Lo
¨
ffelhardt
1
1 Max F. Perutz Laboratories, University Departments at the Vienna Biocenter, Department of Biochemistry and Molecular Cell Biology and
Ludwig Boltzmann Research Unit for Biochemistry
2 Martin-Luther-Universitaet Halle-Wittenberg, Institute for Plant Physiology, Vienna, Austria
The photosynthetic apparatus of cyanobacteria and
chloroplasts are organized in an analogous way, with
some differences in detail [1], and comprise stromal
proteins that temporarily attach to the thylakoid
surface, a number of integral thylakoid membrane
proteins, and some soluble or loosely membrane-asso-
ciated proteins of the thylakoid lumen. In addition,
there are certain proteins that possess a thylakoid

membrane anchor but expose the bulk of the poly-
peptide chain and the prosthetic group involved in
electron transfer to the lumen. In higher plants, the
chloroplast and nuclear genomes both contribute to
the complement of thylakoid proteins [2]. Nucleus-
encoded precursors to integral proteins may contain a
transit sequence only so that information for thylakoid
integration is contained in the mature protein: the cab
protein family is thought to use an exclusive mechan-
ism, the post-translational signal recognition particle
(SRP) pathway [3]. Others possess a bipartite pre-
sequence consisting of a stroma-targeting peptide fol-
lowed by a signal peptide-like hydrophobic domain.
Such proteins as CF
0
-II integrate via the spontaneous
(unassisted) pathway without the need for stroma fac-
tors receptors, ATP, and DpH [4]. Members of the
group of (predominately) lumenal proteins are passen-
gers of the Sec and DpH-dependent (or Tat) pathways,
corresponding to their relatively unfolded or folded
state during translocation, respectively. Identified
components of the chloroplast Sec machinery are
SecY ⁄ SecE (forming the translocation pore) and the
Keywords
Cyanophora paradoxa; cyanelles;
conservative sorting; Sec translocase; Tat
translocase
Correspondence
W. Lo

¨
ffelhardt, Max F. Perutz Laboratories,
University Departments at the Vienna
Biocenter, Department of Biochemistry and
Molecular Cell Biology and Ludwig
Boltzmann Research Unit for Biochemistry,
Dr Bohrgasse 9, 1030 Vienna, Austria
Fax: +43 14277 9528
Tel: +43 14277 52811
E-mail:
(Received 6 September 2004, revised 11
November 2004, accepted 17 December
2004)
doi:10.1111/j.1742-4658.2004.04533.x
Higher plant chloroplasts possess at least four different pathways for pro-
tein translocation across and protein integration into the thylakoid mem-
branes. It is of interest with respect to plastid evolution, which pathways
have been retained as a relic from the cyanobacterial ancestor (‘conserva-
tive sorting’), which ones have been kept but modified, and which ones
were developed at the organelle stage, i.e. are eukaryotic achievements as
(largely) the Toc and Tic translocons for envelope import of cytosolic pre-
cursor proteins. In the absence of data on cyanobacterial protein transloca-
tion, the cyanelles of the glaucocystophyte alga Cyanophora paradoxa for
which in vitro systems for protein import and intraorganellar sorting were
elaborated can serve as a model: the cyanelles are surrounded by a pepti-
doglycan wall, their thylakoids are covered with phycobilisomes and the
composition of their oxygen-evolving complex is another feature shared
with cyanobacteria. We demonstrate the operation of the Sec and Tat
pathways in cyanelles and show for the first time in vitro protein import
across cyanobacteria-like thylakoid membranes and protease protection of

the mature protein.
Abbreviations
GFP, green fluorescent protein; SRP, signal recognition particle.
FEBS Journal 272 (2005) 987–998 ª 2005 FEBS 987
translocation ATPase SecA which is a stromal protein
that binds to the pore during action. Energy source:
ATP. Inhibitor: sodium azide, acting on SecA. Proteins
that fold rather quickly or must acquire their pros-
thetic groups in the stroma prior to translocation enter
the Tat pathway, in most cases indicated by a ‘twin-
arginine’ motif preceding the hydrophobic core of their
signal sequences. The known components Tat A
(Hcf106), TatB (Tha4) and TatC are all membrane-
bound. Energy source is DpH at the thylakoid mem-
brane. Inhibitor is nigericin, dissipating DpH [5]. All
these data were collected using higher plant chloro-
plasts, with the exception of fucoxanthin ⁄ chloro-
phyll c-binding protein and the secondary plastids
(derived from endosymbiotic red algae) of the diatom
Odontella sinensis where indications for a SRP path-
way were obtained [6].
Cyanelles are the peptidoglycan-surrounded plastids
of glaucocystophyte algae [7], assumed to be the first
phototrophic eukaryotes [8]. The protein import appar-
atus of the cyanelle envelope seems to function in an
analogous way as for rhodoplasts and chloroplasts [9].
Are all four pathways for further protein routing
inside the chloroplast also operative in a primitive
plastid? The meaning of ‘conservative sorting’ [10], i.e.
the retainment of prokaryotic translocons in organelles

of endosymbiotic origin as first exemplified by the Sec
pathway, has changed during the past decade: the Tat
pathway was considered as an achievement of higher
plants until its occurrence in (cyano)bacteria was dem-
onstrated [5]. On the other hand, the spontaneous
pathway now seems to be restricted to chloroplasts
since related insertion processes in bacteria were found
to depend on the novel translocon component YidC
[11,12].
In cyanobacteria, there are dual Sec translocons, in
the thylakoid membrane as well as in the inner envel-
ope membrane [13]. Tat translocase is assumed to be
located in the inner envelope membrane of cyanobac-
teria: numerous periplasmic proteins were found where
the corresponding genes contained signal sequences
with the twin-arginine motif [14] and Tat signal pep-
tides directed green fluorescent protein (GFP) to the
periplasmic space of transgenic cyanobacteria [15]. On
the other hand, the Tat passengers known from chlo-
roplasts are largely absent from several completely
sequenced cyanobacterial genomes: the two extrinsic
proteins from the oxygen evolving complex, PsbP and
PsbQ, are replaced in cyanobacteria by the unrelated
proteins PsbV and PsbU [16], respectively, that both
lack the twin arginine motif in the precursors. Even
when a conserved lumenal protein like Hcf136 was
considered, the sorting signal appeared to have
changed after gene transfer to the nuclear genome:
the precursor ⁄ intermediate was shown to use the Sec
pathway in cyanobacteria and the Tat pathway in

higher plants [17].
One reason to investigate thylakoid transloca-
tion ⁄ integration in cyanelles is the bridge position of
these organelles between chloroplasts and free-living
cyanobacteria. Cyanelle thylakoids resemble the cyano-
bacterial ones in the composition of the OEC, the
presence of phycobilisomes [18] as antenna system and
the possibility of connections to the inner envelope
membrane [19]. In freeze-fracture experiments, cyanelle
thylakoids also behaved cyanobacteria-like and upon
isolation did not readily form closed vesicles as chloro-
plast thylakoids do [20]. In vitro experiments with
cyanobacterial thylakoids are hampered through this
problem: successful translocation is evidenced through
protease protection of the mature (processed) protein
which is not feasible when no tight vesicles can be pre-
pared (C. Robinson, personal communication). Assays
that cannot be done ‘in thylakoido’ with cyanobacteria
can be performed ‘in organello’ using intact, isolated
cyanelles. So phycobilisome assembly could be monit-
ored via integration of imported, labeled core linker
protein [18]. The Sec pathway in cyanelles, which was
made likely by the first demonstration of a functional
organellar-encoded secY gene [21], was corroborated
by determining the energetic requirements for cyto-
chrome c
6
, cytochrome c
550
and PsbO import. In this

paper, we demonstrate Rieske Fe ⁄ S protein as a Tat
passenger, and, for the first time with phycobilisome-
bearing thylakoids, we show protease protection of
translocated, processed thylakoid lumenal proteins.
Another advantage of the cyanelle system is that pas-
senger proteins can be studied that are different from
the established chloroplast import systems (i.e. that are
not imported in vivo into chloroplasts) as AtpI.
Results
Cytochrome c
6
uses the Sec pathway
Using a modified cyanelle isolation procedure, the effi-
ciency of homologous (envelope) import could be
greatly increased [9,19]. Import-competent cyanelles
from Cyanophora paradoxa efficiently took up the
15 kDa pre-apocytochrome c
6
and converted it into
the protease-protected mature form of 9.2 kDa, comi-
grating with the holoprotein (Fig. 1, left panel). The
time-course showed that the import was largely com-
pleted after 3 min. Prolonged incubation resulted in
eventual degradation of the mature protein, presuma-
bly through lumenal proteases [22]. No intermediate
Conservative sorting in cyanelles J. M. Steiner et al.
988 FEBS Journal 272 (2005) 987–998 ª 2005 FEBS
was observed under these assay conditions, indicating
that envelope import is the rate-determining step.
However, two-step processing, as reported for the

homologue from Chlamydomonas reinhardtii [23], could
be demonstrated when intermediate accumulated due
to the addition of the SecA inhibitor, sodium azide
(Fig. 1, right panel), whereas nigericin, the inhibitor of
the DpH-dependent Tat pathway had no effect (data
not shown). The effect of azide was most pronounced
after 3 min compared with the control assay. After
that time the intermediate is slowly imported and pro-
cessed, maybe because of intracyanellar ATP genera-
tion, as the azide-inhibition of SecA can be reverted by
a higher amount of ATP competing for the same bind-
ing sites [24]. As sodium azide might inhibit chloro-
plast and cyanelle proteases (our own experiments and
K. Cline, personal communication), it is somewhat dif-
ficult to compare the posterior time points. Sodium
azide also appeared to impede the overall import
process of pre-apocytochrome c
6
to some extent, since
a small amount of precursor protein was still bound to
the envelope (Fig. 1, right panel). For a small protein-
like cytochrome c
6
complete protease protection could
not be achieved [19]. We therefore extended our
experiments to a heterologous system where tight
thylakoid vesicles can be obtained. In vitro assays with
isolated spinach chloroplasts were performed including
competition experiments. Saturating amounts of the
OEC23 (PsbP) precursor from spinach were used to

block the DpH-dependent Tat pathway [5] and – in
parallel – saturating amounts of the spinach OEC33
(PsbO) precursor, a well known Sec-passenger [24] to
inhibit the Sec-pathway (Fig. 2). It could be clearly
shown that pre-apocytochrome c
6
from C. paradoxa
was readily imported into chloroplasts, processed to an
intermediate form in the stroma and then translocated
into the thylakoid lumen, where it was processed again
to its protease-protected mature form (Fig. 2). In the
heterologous system, the stroma-processing protease
cleavage site obviously was not properly recognized
leading to an intermediate of higher MW than that
observed in the cyanelle system (Fig. 1, right panel).
When the OEC23 precursor was used as a competitor,
the amount of mature protein was almost unchanged
and its location in the thylakoid lumen could be pro-
ven by protease protection (Fig. 2). In competition
experiments with the OEC33 precursor, the intermedi-
ate accumulated in the stroma fraction by a factor of
two (ImageQuant) and the amount of mature protein
was reduced by a factor of two compared to the con-
trol assay (Fig. 2). The results of homologous and
heterologous import experiments indicate that apocyto-
chrome c
6
is a Sec passenger in cyanelles (as its func-
tional homologue plastocyanin in higher plants) in
spite of the lack of protease protection due to system-

inherent problems.
Protease protection of lumenal proteins in
cyanobacterial-type thylakoid vesicles
In order to identify additional Sec passengers and,
eventually, to prove protease protection of a larger
lumenal protein in cyanelle thylakoids, we cloned the
psbO gene (GenBank accession number AJ784854) via
a PCR approach based on N-terminal sequence infor-
mation [25] and used the labeled precursor to study
the function of the cyanelle thylakoid translocons.
Figure 3 shows a typical import experiment in a time
Fig. 1. Time-course of import of
35
S-labeled pre-apocytochrome c
6
into isolated cyanelles. T, translation mix; – ⁄ +, without ⁄ with addition of
thermolysin; p, precursor; i, intermediate; m, mature protein. Left panel: control; right panel: plus 10 m
M sodium azide.
J. M. Steiner et al. Conservative sorting in cyanelles
FEBS Journal 272 (2005) 987–998 ª 2005 FEBS 989
course from 3 to 15 min. Compared to pre-apocyto-
chrome c
6
, thylakoid import was retarded in the
stroma, whereas envelope translocation was almost
completed after 3 min. The resulting intermediate form
of the OEC33 protein (iOEC33) appeared to be trans-
ported into the thylakoid lumen in a time-dependent
manner (Fig. 3). When sodium azide was added, mat-
uration of iOEC33 was completely stopped (Fig. 3).

These data do not unequivocally show if the mature
protein is internalized into the thylakoid lumen and
thus we wanted to demonstrate protease protection, at
least for larger proteins as OEC33. We established a
method to isolate tight thylakoid vesicles, which was
hitherto not achieved for phycobilisome-bearing,
cyanobacterial-type membranes. When the novel
cyanelle fractionation procedure was applied after an
import experiment, the mature OEC33 localized to the
thylakoid fraction (Fig. 4). Thermolysin treatment
digested the residual amount of precursor bound to
the peptidoglycan-containing envelope membranes,
which co-sedimented with the thylakoids and therefore
served as internal controls for protease activity,
whereas the internalized mature protein was protected
in the thylakoid lumen. Sodium azide decreased the
amount of precursor imported into the cyanelles
(Fig. 4), possibly due to inhibition of other ATP-
dependent processes [26]. iOEC33 accumulated in the
stroma and even a small proportion of mature PsbO
fractionated with the thylakoid membranes, possibly
due to the longer incubation time (25min.). When this
membrane pellet (containing also envelope membranes)
was treated with thermolysin, the bound precursor as
Fig. 2. Import of
35
S-labeled pre-apocytochrome c
6
of C. paradoxa into isolated spinach chloroplasts. Tr, translation mix; S, stroma; T–, thyl-
akoid membranes; T+, thylakoid membranes treated with thermolysin; p, precursor; i, intermediate; m, mature protein. Left panel, control;

middle panel, saturation of the Tat-pathway by the 23-kDa protein; right panel, saturation of the Sec-pathway by the 33-kDa protein.
Fig. 3. Time-course of import of
35
S-labeled
pre-OEC33 into isolated cyanelles. ivT,
translation mix; p, precursor; i, intermediate;
m, mature protein. (A) control; (B) plus
10 m
M sodium azide; control + TL, plus
thermolysin.
Conservative sorting in cyanelles J. M. Steiner et al.
990 FEBS Journal 272 (2005) 987–998 ª 2005 FEBS
well as the putative mature protein band disappeared,
indicating that PsbO protein generated in the presence
of azide is only bound to the thylakoid surface (or
rather arrested in a preliminary insertion stage that
allowed processing) but not internalized and thus not
protease-protected. Nigericin also caused a drop in over-
all import efficiency, but did not lead to accumulation
of the intermediate in the stroma. Furthermore, in con-
trast to the azide experiment, mature protein fractionat-
ing with the thylakoids was protease-protected (Fig. 4).
All these data point towards a functional Sec translocase
with an azide-sensitive SecA homologue in the cyanelle
thylakoid membrane and that lumenal cytochrome c
6
as
well as OEC33 protein had used this pathway.
The Rieske Fe/S protein is inserted into cyanelle
thylakoids via the DpH-dependent Tat pathway

In chloroplasts, Rieske Fe ⁄ S protein, a nuclear-enco-
ded subunit of the cytochrome b
6
⁄ f complex, appeared
in the stroma after in vitro import and only slowly
translocated further into the thylakoid membrane sys-
tem. It could also be shown, via competition experi-
ments and the sensitivity to nigericin, that thylakoid
translocation ⁄ integration of this protein in chloroplasts
takes place through the DpH-dependent Tat pathway
[27]. As the Rieske protein lacks a cleavable signal
peptide, its transport is mediated by the N-terminal
membrane anchor which does not contain the twin-
arginine motif typical of Tat transport signals. Instead,
higher plant as well as cyanelle Rieske proteins contain
a lysin-arginine sequence at this position, whereas
cyanobacterial Rieske proteins do possess the twin-
arginine motif. This renders the Rieske protein an
unusual Tat substrate and, concerning the evolutionary
position of the cyanelles, it was interesting which
pathway the authentic Rieske protein might use in the
cyanelle system. For that reason we cloned two petC
genes from C. paradoxa via a PCR approach (petC1,
GenBank accession number AJ784852 and petC2,
GenBank accession number AJ784853). We performed
in organello experiments with isolated intact cyanelles
and the precursor corresponding to petC1 in a time-
course manner in the presence⁄ absence of specific
translocase inhibitors (Fig. 5). A striking feature in the
targeting process of Rieske protein is the remarkably

slow sorting of the protein within the cyanelles to
its final destination, the thylakoid membrane system.
While the import of the Rieske precursor into the
stroma proceeded within 5 min (Fig. 5), only a minor
fraction (20%) reached the thylakoids and was cor-
rectly integrated during a total incubation time of
25 min (Fig. 6). The majority of the processed mature
protein of approximately 20 kDa accumulated in the
Fig. 4. Fractionation of isolated cyanelles after import of
35
S-labeled pre-OEC33. ivT, translation mix; C, intact cyanelles; S, stroma; T–, thyla-
koid membranes; T+, thylakoid membranes treated with thermolysin; p, precursor; i, intermediate; m, mature protein; azide, plus 10 m
M
sodium azide; nigericin, plus 2 lM nigericin.
Fig. 5. Import of
35
S-labeled pre-Rieske-protein into isolated cya-
nelles. ivT, translation mix; p, precursor; m, mature protein; azide,
plus 10 m
M sodium azide; nig, plus 2 lM nigericin.
J. M. Steiner et al. Conservative sorting in cyanelles
FEBS Journal 272 (2005) 987–998 ª 2005 FEBS 991
cyanelle stroma. Once the Rieske protein had
reached the thylakoids, it was to a large extent protected
against the activity of proteases that were added exter-
nally to the thylakoid vesicles, indicating that the
C-terminal hydrophilic domain had been completely
translocated into the lumenal space (Fig. 6). Only the
utmost N-terminal residues preceding the membrane
anchor domain remained accessible to thermolysin

resulting in a decrease in apparent molecular mass of
approximately 0.5 kDa. In the presence of nigericin,
thylakoid translocation of the Rieske protein was com-
pletely abolished, and the bulk of intracyanellar mature
protein localized with the stroma, while the thylakoid
membrane fraction contained only a minor amount of
loosely bound Rieske protein which disappeared after
thermolysin treatment. Sodium azide reduced the
amount of membrane-integrated Rieske protein to
about 30% of the control assay. This fraction of
mature protein was protease-protected and fully integ-
rated into the thylakoid membrane system through its
single membrane span. A slight downward shift in
polypeptide mobility was best noticeable here due to
the removal of about four amino acids from the
stroma-exposed N terminus by thermolysin treatment,
indicating correct integration into the cytochrome b
6
f
complex. In evaluating the azide effect on thylakoid
translocation of Rieske protein one should consider its
reported inhibitory action on numerous nucleotide-
binding proteins [28]: the azide-sensitive steps occur
very likely after the import of the apoprotein and prior
to membrane integration of the holoprotein [27]. Thus
our interpretation of the experiments shown in Fig. 6
is to name Rieske a bona fide Tat passenger: a similar
conclusion was made in the chloroplast system [27].
The first precursor to a cyanelle lumenal protein
(PsbU) containing the twin-arginine motif

PrePsbU from the red alga Cyanidium caldarium was
found to contain the twin-arginine motif in the thyla-
koid transfer domain of the presequence [29]: this was
the first incidence in algae containing ‘primitive’ plast-
ids and prompted us to clone the counterpart from
C. paradoxa based on sequence information from an
EST collection (S. Burey and H. Bohnert, unpublished
data). This completed the collection of cyanelle
OEC component genes (GenBank accession number
AJ784849) and presented another Tat passenger candi-
date (Fig. 7). Cyanelle import of prePsbU occurred as
readily as that of the other small cyanelle protein, pre-
cytochrome c
6
, with almost no envelope-bound precur-
sor or intermediate (Fig. 8). However, azide addition
did not result in the accumulation of intermediate,
excluding the Sec pathway. Nigericin did not produce
a clear-cut effect either (except also increasing the
amount of envelope-bound precursor). Small proteins
obviously can escape from cyanelle thylakoid vesicles,
thus mature PsbU localized to the stroma fraction in
comparable amounts irrespective of any inhibitor used
(Fig. 8). Considering these experimental difficulties
with regard to protease protection and the reported
DpH-independence of Tat translocation in C. reinhardtii
chloroplasts [30] we propose that PsbU, i.e. one of the
three cyanelle OEC proteins, is a Tat passenger.
Fig. 6. Fractionation of isolated cyanelles after import of the
35

S-labeled pre-Rieske-protein. ivT, translation mix; C, intact cyanelles; S,
stroma; T–, thylakoid membranes; T+, thylakoid membranes treated with thermolysin; p, precursor; m, mature protein; azide, plus 10 m
M
sodium azide; nigericin, plus 2 lm nigericin.
Conservative sorting in cyanelles J. M. Steiner et al.
992 FEBS Journal 272 (2005) 987–998 ª 2005 FEBS
AtpI: SRP-dependent or spontaneous thylakoid
integration?
As other primitive plastids, cyanelles encode a higher
number of ATP synthase subunits than higher plant
chloroplasts, e.g. atpG and atpD [32]. However, in con-
trast to all sequenced plastid genomes, atpI is a nuclear
gene in C. paradoxa [7]. This means that a precursor
exceeding Cab protein with respect to hydrophobicity
has to be imported into cyanelles, must cross the
stroma and insert into the thylakoid membranes. In the
case of Cab, the solution is the post-translational SRP
pathway involving a transit complex consisting of Cab,
SRP54 and SRP43 [3]. We cloned two atpI genes via a
PCR approach based on highly conserved domains of
the protein. They are listed in under the GenBank
accession numbers AJ784850 and AJ784851, respect-
ively. The closely related sequences comprise four to
five putative transmembrane regions as candidates for
binding to SRP54 [33] and a hydrophilic loop with
some resemblance to the ‘L18’ domain of Cab protein
(Fig. 9) which was shown to interact with SRP 43 [3].
Highly hydrophobic precursor proteins pose problems
upon in vitro import into isolated chloroplasts [34].
This also applies for cyanelle envelope translocation:

only a small fraction of AtpI is processed and internal-
ized (Fig. 10A) though a time course is noticeable. Cya-
nelle fractionation after incubation resulted in recovery
of substantial amounts of preAtpI in the thylakoid
(and envelope) fraction. Low amounts of mature pro-
tein were detected in the thylakoid fraction: here no
influence of added azide or nigericin became apparent.
Due to cleavage sites in the stromal loops thylakoid-
inserted AtpI was degraded upon thermolysin treat-
ment (Fig. 10B). With regard to the insertion pathway,
Sec and Tat do not seem to be involved. In order to
Domains: N H C
Nucleus-encoded:
pre-cytochrome c
6
SKKATFTAAATAAALLAASPVFA
pre-PsbO TTFGR
ALAAFVAAAGISFAGVSQANA
pre-PsbU KKGRR
EFVAAAGALFAAFAASPAAFA
Plastid-encoded:
pre-cytochrome c
550
MFNKNFWTSIIIGCLFCTITYSGVNA
p
re-
p
saF MRKLFLLMFCLSGLILTTDIRPVRA
Fig. 7. Thylakoidal signal peptides of intermediates (precursors) to
cyanelle proteins that are imported into or synthesized within the

organelle, respectively. Charged residues are underlined. C, C-ter-
minal domain; H, hydrophobic core (bold); N, N-terminal domain.
The signal sequences of nucleus-encoded precursors start with the
first amino acid after the putative SPP cleavage sites taken from
predictions of the
CHLOROP program [31].
Fig. 8. Fractionation of isolated cyanelles after import of the
35
S-labeled prepsbU-protein. ivT, translation mix; S, stroma; T–,
thylakoid membranes; T+, thylakoid membranes treated with
thermolysin; p, precursor; m, mature protein; azide, plus 10 m
M
sodium azide; nigericin, plus 2 lm nigericin.
Fig. 9. Sequence comparison of a hydrophilic loop between the
putative TM helices 3 and 4 of AtpI from C. paradoxa to the ‘L18’
domain of pea LHCP. Similarities are indicated by bold letters.
A
B
Fig. 10. (A) Time-course of import of the
35
S-labeled preatpI-protein
into isolated cyanelles. T, translation mix; p, precursor; m, mature
protein; 0°, incubation for 20 min on ice; –, control; +, plus thermo-
lysin. (B) Fractionation of isolated cyanelles after import of the
35
S-labeled preatpI-protein. ivT, translation mix; S, stroma; T–,
thylakoid membranes; T+, thylakoid membranes treated with
thermolysin; p, precursor; m, mature protein; azide, plus 10 m
M
sodium azide; nigericin, plus 2 lM nigericin.

J. M. Steiner et al. Conservative sorting in cyanelles
FEBS Journal 272 (2005) 987–998 ª 2005 FEBS 993
assess an SRP-based mechanism, the import efficiency
will have to be increased first. Spontaneous insertion is
less likely for a polytopic membrane protein as AtpI
but cannot be excluded at present.
Discussion
Two of the four pathways operating in thylakoid
translocation ⁄ integration in higher plant chloroplasts,
i.e. the Sec and the Tat translocase could be demon-
strated to function in the cyanelles of C. paradoxa.
With respect to the hydrophobicity of the core
domains signal sequences from cyanelle genes surpass
those from nuclear genes (Fig. 7). This has also been
observed with higher plants: the replacement (after
gene transfer) of leucine and phenylalanine by alanine
was interpreted as a measure to avoid interactions with
cytosolic SRP [35].
Interestingly, there seem to be relatively more
known Sec passengers in these primitive plastids than
Tat passengers, whereas the opposite was found for
chloroplasts [5]. There are several reasons for the
observed prevalence of the Sec pathway in cyanelles:
one of them is the replacement of evolutionary ancient,
i.e. cyanobacterial proteins by unrelated counterparts
in higher plants. The cyanelle-encoded cytochrome c
550
fulfills the function of the Tat passenger PsbP in the
OEC. By analogy to the other c-type cytochromes it
should use the Sec pathway. This was proven by

homologous and heterologous import experiments of a
construct containing the FNR transit sequence from
C. paradoxa at the N terminus (T. Ko
¨
cher and
J. Steiner, unpublished data). Second, Tat passengers
like polyphenol oxidase, PsbT, PsaN and others (with-
out a cyanobacterial homologue) might be absent from
cyanelles. Third, C. paradoxa with its peculiar gene
distribution between the nuclear and cyanelle genomes
allows to test the hypothesis that rapidly folding
(small) polypeptides without prosthetic groups can be
Sec passengers in cyanobacteria and in primitive plast-
ids (when they are plastom-encoded) but should be
Tat passengers as nuclear gene products in higher
plant chloroplasts. In the latter case, protein targeting
to the thylakoid lumen is much more time-consuming
and the intermediate should be rather tightly folded
upon arrival at the membrane [17]. The cyanelle-enco-
ded Hcf136 homolog resembles its cyanobacterial
counterpart in the absence of the twin arginine motive,
i.e. is a putative Sec passenger. PsbU, on the other
hand, an OEC component of cyanobacteria and chlo-
rophyll b-less algae, is nucleus-encoded in the latter.
Consequently, PsbU contains a twin-arginine motif in
its bipartite presequence and its translocation therefore
most likely is Tat-dependent. In this case, evolutionary
replacement through PsbQ in higher plants is not
accompanied by a change in the translocase used.
Isolated cyanobacterial thylakoid vesicles do not

allow in thylakoido import experiments or protease
protection of luminal proteins, in contrast to spinach
thylakoids [5,34]. Therefore it was a considerable pro-
gress to show protease protection at least for the
33 kDa PsbO protein after internalization into cyanelle
thylakoids. It is unknown, why these membranes are
still leaky for smaller proteins as cytochrome c
6
and
PsbU. Sodium azide appeared to be the diagnostic
inhibitor of choice for in cyanello experiments. Thyla-
koid import was retarded for small Sec passenger
proteins and completely blocked for larger ones,
respectively. This indicates high azide-sensitivity of
cyanelle SecA. Thus the Sec pathway can be excluded
when no significant azide effect is observed, e.g. in the
case of PsbU.
Nigericin completely abolished protease protection
of membrane-inserted Rieske protein. However, azide
addition resulted in a reduction of the amount of
correctly assembled cyanelle Rieske protein as was
observed with the chloroplast in organello system [27].
There it was shown that although the Rieske protein is
targeted exclusively by the DpH ⁄ Tat pathway, some
azide-sensitive stromal factors, such as the Cpn60
chaperonin [36] might play a role in correct folding
and ⁄ or attachment of the Fe ⁄ S cluster to the Rieske
mature (apo)protein. Recently, an interaction and ⁄ or
regulation partner for the Rieske protein has been
identified in Arabidopsis thaliana [37]. A thylakoid lum-

enal FKBP (immuno-suppressant FK506 binding pro-
tein) was isolated, whereof only the precursor, but not
the mature form, interacted with Rieske protein.
AtFKBP13 might serve as an ‘anchor’ chaperone that
holds the Rieske protein in the cytoplasm or in the
stroma so that excessive Rieske protein is not targeted
to the thylakoid, since its integration into the cyto-
chrome b6f complex underlies complex regulation and
coordination events. It might well be that sodium azide
blocks either one of those chaperones and ⁄ or other
factors necessary for the build-up of a functional cyto-
chrome b
6
f complex, and the resulting malfunctioning
or only partially moulded unit fails to be translocated
via the Tat-pathway. The effects of nigericin on thyla-
koid translocation of the other candidate Tat passen-
ger, PsbU, were not clear-cut. In this context it
should be noted that the Tat translocase in the alga
C. reinhardtii appeared to operate in the absence of a
DpH [30], whereas proton efflux at the expense of the
pH gradient was a prerequisite for Tat-dependent
translocation in higher plant chloroplasts [38].
Conservative sorting in cyanelles J. M. Steiner et al.
994 FEBS Journal 272 (2005) 987–998 ª 2005 FEBS
AtpI would be a candidate for the post-translational
SRP pathway in cyanelles. In Escherichia coli, its
homologue, subunit a, needs the assistance of YidC for
integration into the inner membrane [39]. Mitochondrial
Atp6 is also partially dependent upon Oxa1 in that

respect [40]. In chloroplasts, by analogy, the cotransla-
tional SRP pathway and the Albino3 translocon
might be involved. However, in the ac29 mutant of
C. reinhardtii, where one out of the two albino3 genes is
inactivated, no effect on ATP synthase (at least on the a
subunit) could be observed [41]. Thus far, there is no
evidence for ALB3 in the plastids of chlorophyll b-less
algae. In C. paradoxa, SecY is a component of two
different high molecular mass thylakoid-bound protein
complexes (F. Yusa and W. Lo
¨
ffelhardt, unpublished
data), a parallel to A. thaliana where ALB3 is also
contained [42]. Another crucial point will be to identify
SRP43 in nonchlorophytes, e.g. diatoms that integrate
fucoxanthin chlorophyll a ⁄ c-binding protein into their
thylakoid membranes [6]. On the other hand, reports on
vesicle transport of cab protein from the inner envelope
membrane to the thylakoid membrane in C. reinhardtii
[43] render the function of the transit complex and of
ALB3 questionable in this alga. However this vesicle
transport must be different from that described for
higher plant chloroplasts that can be visualized upon
lowering the temperature and is sensitive to microcystin
[44].
The fourth pathway might also exist in cyanelles:
SecE from C. paradoxa is not yet identified; in chloro-
plasts it was shown to use the spontaneous pathway
[11,12]. Chloroplast and cyanelle thylakoids are both
rich in galactolipids which obviously support unas-

sisted integration of proteins in contrast to the phos-
pholipids of the E. coli cytoplasmic membrane [12].
Experimental procedures
Materials
Cyanophora paradoxa LB555UTEX was grown as previ-
ously described [45]. In general, cells were harvested in the
exponential growth phase. Nucleic acids were isolated
according to published methods [45]. Spinach (Spinacia
oleracea) was purchased from the local market and kept
overnight at 4 °C before isolating chloroplasts. Pea seed-
lings (Pisum sativum) were grown for 8–10 days under a
16 h photoperiod.
Protein import into isolated chloroplasts
Precursor proteins were synthesized by in vitro transcription
of the corresponding cDNA clones and subsequent in vitro
translation in cell-free wheat germ lysates in the presence of
[
35
S]methionine. Intact chloroplasts were isolated from pea
or spinach leaves by Percoll gradient centrifugation and
were used in protein import experiments essentially as des-
cribed [46]. Competition experiments were performed with
precursor proteins that were obtained by overexpression in
Escherichia coli [47] and recovered from inclusion bodies by
solubilization in a buffer containing 7 m urea, 30 mm Hepes,
pH 8.0 and 2 mm EDTA. The solubilized proteins were
included in the import assays at concentrations up to 4 lm,
taking care that the concentration of urea in the assays
never exceeded 300 mm. Control assays contained the same
amount of buffer lacking any such solubilized protein.

Protein import into isolated cyanelles
Import-competent cyanelles were isolated as described in
[19]. They were used in protein transport experiments as
described in [18].
Isolation of cyanelle thylakoid membranes
The import reaction was stopped by the addition of 1 mL
ice-cold sorbitol resuspension medium (SRM) buffer
(50 mm Hepes, 0.33 m sorbitol, pH 8.0) followed by centri-
fugation at 800 g and 4 °C for 2 min. The cyanelle pellet
was washed in SRM and resuspended in 500 l L2· SRM
and incubated for 25 min at room temperature with 15 lL
of a 10 mgÆmL
)1
lysozyme stock solution in the presence of
protease inhibitors (Complete, Roche), which led to diges-
tion of the peptidoglycan wall, cyanelle lysis and release of
the phycobilisomes from the thylakoid membrane. After
centrifugation for 5 min at 9300 g in an Eppendorf centri-
fuge at 4 °C an aliquot of the deep-blue stromal superna-
tant was precipitated with 100% (v ⁄ v) acetone, the pellet
containing the thylakoids and the peptidoglycan-linked
outer envelope membrane was washed in 2· SRM and
finally resuspended in 500 lL2· SRM. An aliquot
(250 lL) was treated with thermolysin plus 10 mm CaCl
2
for 30 min on ice. After stopping the reaction with EDTA
all aliquots were pelleted by centrifugation.
Miscellaneous
Gel electrophoresis of proteins under denaturing conditions
was carried out according to [48]. Import data were ana-

lyzed using a PhosphoImager and the molecular dynam-
ics imagequant program (version 3.3), such that all the
signals remained in the linear detection range.
PCR, gene isolation
The nucleotide sequences determined via reverse translation
of highly conserved regions of the ATP synthase subunit
J. M. Steiner et al. Conservative sorting in cyanelles
FEBS Journal 272 (2005) 987–998 ª 2005 FEBS 995
CF0-IV (AtpI) and of the Rieske iron–sulfur protein (PetC)
were used to design degenerate primers.
AtpI: forward primer 5¢-GCNTAYTTYTAYGCNGG-3¢,
reverse primer 5¢-GGYTTNGTRAARTCYTC-3¢ (product
size: 111 bp); PetC: forward primer 5¢-CARGGNYTNAA
RGGNGAYCCNACNTA-3¢, reverse primer 5¢-TAYTGN
WSNCCRTGRCANGGRCA-3¢ (product size: 156 bp).
The PCR reaction mixture (50 lL) included 100 ng of
DNA from C. paradoxa, 0.1 lm concentration of each
primer species, 10 mm Tris ⁄ HCl (pH 8.3), 50 mm KCl,
1mm MgCl
2
, 0.2 mm dNTPs, and 1 U of Taq DNA
polymerase (Dynazyme, Finnzymes Oy, Espoo, Finland).
The following thermal cycle was used: Step 1, 96 °C for
5 min; Step 2, 94 ° C for 1 min; Step 3, 50 °C for 2 min;
Step 4, 72 °C for 3 min; Step 5, repeat steps 2–4 35 times;
Step 6, 72 °C for 7 min. The predominant PCR product
was cloned into pGEM-T and sequenced. After identifi-
cation of the correct products the fragments were labe-
led with the Digoxigenin Labeling ⁄ Detection System
(Boehringer Mannheim, Mannheim, Germany) and used

for screening of a C. paradoxa cDNA library in the vec-
tor k-ZAP II (Stratagene, La Jolla, CA, USA). Plaque
hybridization was performed under high stringency
conditions [49].
Full-length cDNAs were cloned into the vector pBAT
[50] to allow sufficient translation efficiency. psbO: forward
primer: 5¢-GARGGNYTNACNTAYGAYCA-3¢, reverse
primer: 5¢-AANGGNACNCKYTCNCCNCC-3¢. The for-
ward primer was designed using a peptide sequence
(EGLTYDQ) obtained via Edman-sequencing [25]. psbU:
forward primer: 5¢-AAGAATTCACGAGGCAGAAATG
GCGTTC-3¢, reverse primer: 5¢-AAGGATCCTGGGGAC
AGCAGAAACTTGG-3¢.
The psbU gene was isolated by PCR with a proof-reading
polymerase (Pfu) using data from a C. paradoxa EST lib-
rary (S. Burey and H. J. Bohnert, unpublished data) and
directly cloned into the pBAT-vector via its EcoRI and
BamHI sites.
Acknowledgements
We appreciate support from the Austrian Research
Fund (P15438-MOB, to W.L.). We thank Hans
Bohnert (Urbana, IL, USA) and Suzanne Burey for
providing EST data from Cyanophora paradoxa.
References
1 Hankamer B, Morris E, Nield J, Carne A & Barber J
(2001) Subunit positioning and transmembrane helix
organisation in the core dimer of photosystem II. FEBS
Lett 504, 142–151.
2 Herrmann RG (1998) Thylakoid membranes. A para-
digm for biogenetic and phylogenetic compexity. In

Photosynthesis, Mechanisms and Effects (Garab G, ed),
Vol. IV, pp. 3049–3056.
3 Eichhacker LA & Henry R (2001) Function of a chloro-
plast SRP in thylakoid protein export. Biochim Biophys
Acta 1541, 120–134.
4 Schleiff E & Klo
¨
sgen RB (2001) Without a little help
from ‘my’ friends: direct insertion of proteins into
chloroplast membranes? Biochim Biophys Acta 1541,
22–33.
5 Mori H & Cline K (2001) Post-translational protein
translocation into thylakoids by the Sec and DeltapH-
dependent pathways. Biochim Biophys Acta 1541, 80–90.
6 Lang M & Kroth P (2001) Diatom fucoxanthin chloro-
phyll a ⁄ c-binding protein (FCP) and land plant light-
harvesting proteins use a similar pathway for thylakoid
membrane insertion. J Biol Chem 276, 7985–7991.
7Lo
¨
ffelhardt W & Bohnert HJ (2001) The cyanelle (muro-
plast) of Cyanophora paradoxa: a paradigm for endosym-
biotic organelle evolution. In Symbiosis (Seckbach, J,
ed.), pp. 111–130. Kluwer Academic Publishers,
Dordrecht.
8 Martin WF, Stoebe B, Goremykin V, Hansmann S,
Hasegawa M & Kowallik KV (1998) Gene transfer to
the nucleus and the evolution of chloroplasts. Nature
393, 162–165.
9 Steiner JM & Lo

¨
ffelhardt W (2002) Protein import into
cyanelles. Trends Plant Sci 6, 72–77.
10 Smeekens S, Weisbeek P & Robinson C (1990) Protein
transport into and within chloroplasts. Trends Biochem
Sci 15, 73–76.
11 Steiner JM, Ko
¨
cher T, Nagy C & Lo
¨
ffelhardt W (2002)
Chloroplast SecE: evidence for spontaneous insertion
into the thylakoid membrane. Biochem Biophys Res
Commun 293, 747–752.
12 Jiang F, Yi L, Moore M, Chen M, Rohl T, van Wijk K-J,
de Gier J-WL, Henry R & Dalbey R (2002) Chloroplast
YidC homolog Albino3 can functionally complement the
bacterial YidC depletion strain and promote membrane
insertion of both bacterial and chloroplast thylakoid
proteins. J Biol Chem 277, 19281–19288.
13 Nakai M, Sugita D, Omata T & Endo T (1993) SecY
protein is localized in both the cytoplasmic and thyla-
koid membranes in the cyanobacterium Synechococcus
PCC 7942. Biochem Biophys Res Commun 193, 228–234.
14 Fulda S, Huang F, Nilsson F, Hagemann M & Norling B
(2000) Proteomics of Synechocystis sp. strain PCC 6803 –
identification of periplasmic proteins in cells grown at
low and high salt concentrations. Eur J Biochem 277,
5900–5907.
15 Spence E, Sarcina M, Ray N, Moller SG, Mullineaux CW

& Robinson C (2003) Membrane-specific targeting of
green fluorescent protein by the Tat pathway in the
cyanobacterium Synechocystis PCC6803. Mol Microbiol
48, 1481–1489.
Conservative sorting in cyanelles J. M. Steiner et al.
996 FEBS Journal 272 (2005) 987–998 ª 2005 FEBS
16 Nield J, Kruse O, Ruprecht J, da Fonseca P, Buchel C
& Barber J (2000) Three-dimensional structure of
Chlamydomonas reinhardtii and Synechococcus elongatus
photosystem II complexes allows for comparison of
their oxygen-evolving complex organization. J Biol
Chem 275, 27940–27946.
17 Hynds PJ, Plu
¨
cken H, Westhoff P & Robinson C (2000)
Different lumen-targeting pathways for nuclear-encoded
versus cyanobacterial ⁄ plastid-encoded Hcf136 proteins.
FEBS Lett 467, 97–100.
18 Steiner JM, Pompe JA & Lo
¨
ffelhardt W (2003) Charac-
terization of apcC, the nuclear gene for the phycobili-
some core linker polypeptide L
7:8
c
from the
glaucocystophyte alga Cyanophora paradoxa: import of
the precursor into cyanelles and integration of the
mature protein into intact phycobilisomes. Curr Genet
44, 132–137.

19 Steiner JM, Serrano A, Allmaier G, Jakowitsch J &
Lo
¨
ffelhardt W (2000) Cytochrome c
6
from Cyanophora
paradoxa: characterization of the protein and the cDNA
of the precursor and import into isolated cyanelles. Eur
J Biochem 267, 4232–4241.
20 Giddings TH, Wasman C & Staehelin LA (1983) Struc-
ture of the thylakoids and envelope membranes of the
cyanelles of Cyanophora paradoxa. Plant Physiol 71,
409–419.
21 Flachmann R, Michalowski C, Lo
¨
ffelhardt W &
Bohnert HJ (1993) SecY, an integral subunit of the
bacterial preprotein translocase is encoded by a plastid
genome. J Biol Chem 268, 7514–7519.
22 Adamska I (1998) Proteolytic enzymes in the chloro-
plast related to light stress conditions. In Photosynthesis:
Mechanisms and Effects (Garab, G, ed.), pp. 2035–2038.
Kluwer Academic Publishers, Dordrecht.
23 Merchant S & Dreyfuss BW (1998) Posttranslational
assembly of photosynthetic metalloproteins. Annu Rev
Plant Physiol Plant Mol Biol 49, 25–51.
24 Knott TG & Robinson C (1994) The SecA inhibitor,
azide, reversibly blocks the translocation of a subset of
lumenal proteins across the thylakoid membrane. J Biol
Chem 269, 7843–7846.

25 Shibata M, Kashino Y, Satoh K & Koike H (2001) Iso-
lation and characterization of oxygen-evolving thylakoid
membranes and photosystem II particles from a glauco-
cystophyte, Cyanophora paradoxa. Plant Cell Physiol 42,
733–741.
26 Mould RM, Knight JS, Bogsch E & Gray JC (1997)
Import and processing of the precursor of the Rieske
FeS protein of tobacco chloroplasts. Plant J 11, 1051–
1058.
27 Molik S, Karnauchov I, Weidlich C, Herrmann RG &
Klo
¨
sgen RB (2001) The Rieske Fe ⁄ S protein of the
cytochrome b
6
⁄ f complex in chloroplasts: missing link in
the evolution of protein transport pathways in chloro-
plasts? J Biol Chem 276, 42761–42766.
28 Muneyuki E, Makino M, Kamata H, Kagawa Y,
Yoshida M & Hirata H (1993) Inhibitory effect of
NaN
3
on the F
0
F
1
ATPase of submitochondrial parti-
cles as related to nucleotide binding. Biochim Biophys
Acta 1144, 62–68.
29 Ohta H, Okumura A, Okuyama S, Akiyama A, Iwai

M, Yoshihara S, Shen J-R, Kamo M & Enami I
(1999) Cloning and expression of the psbU gene and
functional studies of the recombinant 12 kDa protein
of photosystem II from the red alga Cyanidium
caldarium. Biochem Biophys Res Commun 260, 245–
250.
30 Finazzi G, Chasen C, Wollman F-A & de Vitry C
(2003) Thylakoid targeting of Tat passenger proteins
shows no DpH dependence in vivo. EMBO J 22 , 807–
815.
31 Emanuelsson O, Nielsen H & von Heijne G (1999)
ChloroP, a neural network-based method for predicting
chloroplast transit peptides and their cleavage sites. Pro-
tein Sci 8, 978–984.
32 Lo
¨
ffelhardt W, Bohnert HJ & Bryant DA (1997) The
complete sequence of the Cyanophora paradoxa cyanelle
genome. In Origins of Algae and Their Plastids
(Bhattacharya D, ed), pp. 149–162. Springer-Verlag,
Wien.
33 High S, Henry R, Mould RM, Valent Q, Meacock S,
Cline K, Gray JC & Luirink J (1997) Chloroplast
SRP54 interacts with a specific subset of thylakoid pre-
cursor proteins. J Biol Chem 272, 11622–11628.
34 Laidler V, Chaddock AM, Knott TG, Walker D &
Robinson C (1995) A SecY homolog in Arabidopsis
thaliana: sequence of a full length cDNA clone and
import of the precursor protein into chloroplasts. J Biol
Chem 268, 7514–7519.

35 Brink S, Bogsch E, Mant A & Robinson C (1997) Unu-
sual characteristics of amino-terminal and hydrophobic
domains in nuclear-encoded thylakoid signal peptides.
Eur J Biochem 245, 340–348.
36 Madueno F, Napier JA, Cejudo FJ & Gray JC (1992)
Import and processing of the precursor of the Rieske
FeS protein of tobacco chloroplasts. Plant Mol Biol 20,
289–299.
37 Gupta R, Mould RM, He Z & Luan S (2002) A chloro-
plast FKBP interacts with and affects the accumulation
of Rieske subunit of cytochrome bf complex. Proc Natl
Acad Sci USA 99, 15806–15811.
38 Alder NN & Theg SM (2003) Energetics of protein
transport across biological membranes: a study of the
thylakoid DpH-dependent cpTat pathway. Cell 112,
231–242.
39 Yi L, Jiang FL, Chen MY, Cain B , Bolhuis A & Dalbey RE
(2003) YidC is strictly required for membrane insertion of
subunits a and c of the F(1)F(0)ATP synthase and SecE
of the SecYEG translocase. Biochemistry 42, 10537–
10544.
J. M. Steiner et al. Conservative sorting in cyanelles
FEBS Journal 272 (2005) 987–998 ª 2005 FEBS 997
40 Hell K, Neupert W & Stuart R (2001) Oxa1p acts as a
general membrane insertion machinery for proteins
encoded by mitochondrial DNA. EMBO J 20, 1281–
1288.
41 Bellafiore S, Ferris P, Naver H, Go
¨
hre V & Rochaix J-D

(2002) Loss of Albino3 leads to the specific depletion of
the light-harvesting system. Plant Cell 14, 2303–2014.
42 Klostermann E, Droste General Helling I, Carde JP &
Schu
¨
nemann D (2002) The thylakoid membrane protein
ALB3 associates with the cpSec-translocase in Arabidop-
sis thaliana. Biochem J 368, 777–781.
43 Hoober JK & Eggink LL (2001) A potential role of
chlorophylls b and c in assembly of light-harvesting
complexes. FEBS Lett 489, 1–3.
44 Westphal S, Soll J & Vothknecht U (2001) A vesicle
transport system inside chloroplasts. FEBS Lett 506,
257–2611.
45 Jakowitsch J, Bayer MG, Maier TL, Lu
¨
ttke. A,
Gebhart UB, Brandtner M, Hamilton B, Neumann-
Spallart C, Michalowski CB, Bohnert HJ, Schenk HEA
&Lo
¨
ffelhardt W (1993) Sequence analysis of pre-ferre-
doxin-NADP
+
-reductase cDNA from Cyanophora
paradoxa specifying a precursor for a nucleus-encoded
cyanelle polypeptide. Plant Mol Biol 21, 1023–1033.
46 Clausmeyer S, Klo
¨
sgen RB & Herrmann RG (1993)

Protein import into chloroplasts: the hydrophilic lume-
nal proteins exhibit unexpected import and sorting
specificities in spite of structurally conserved transit pep-
tides. J Biol Chem 268, 13869–13876.
47 Michl D, Robinson C, Shackleton JB, Herrmann RG &
Klo
¨
sgen RB (1994) Targeting of proteins to the thyla-
koids by bipartite presequences: CF
0
II is imported by a
novel, third pathway. EMBO J 13, 1310–1317.
48 Laemmli UK (1970) Cleavage of structural proteins dur-
ing the assembly of the head of bacteriophage T4. Nat-
ure 227, 680–685.
49 Sambrook J, Fritsch EF & Maniatis T (1989) Molecular
Cloning: a Laboratory Manual, 2nd edn. Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, NY.
50 Anweiler A, Hipskind RA & Wirth T (1991) A strat-
egy for efficient in vitro translation of cDNAs using
the rabbit-globin leader sequence. Nucleic Acids Res
19, 3750.
Conservative sorting in cyanelles J. M. Steiner et al.
998 FEBS Journal 272 (2005) 987–998 ª 2005 FEBS