CompletereconstitutionofanATP-binding cassette
transporter LolCDEcomplexfromseparately isolated
subunits
Kyoko Kanamaru*
,
, Naohiro Taniguchi, Shigehiko Miyamoto, Shin-ichiro Narita
and Hajime Tokuda
Institute of Molecular and Cellular Biosciences, University of Tokyo, Japan
Escherichia coli has at least 90 species of lipoproteins
[1], which have the N-terminal Cys modified with thio-
ether-linked diacylglycerol and an amino-linked acyl
chain [2]. Most lipoproteins are present in the outer
membrane, but there are some in the inner membrane.
Sorting of lipoproteins depends on the species of
the residue at position 2 [3–5], and is catalyzed by the
Lol system, composed of five Lol proteins [6]. The
LolCDE complex in the inner membrane belongs to
the ATP-bindingcassette (ABC) transporter super-
family, and mediates detachment of lipoproteins from
the inner membrane [7]. This results in the formation
of a complex between lipoprotein and LolA [8], a peri-
plasmic molecular chaperone for lipoproteins. LolB in
the outer membrane then accepts lipoproteins from
LolA and incorporates them into the outer membrane
[9]. Inner membrane-specific lipoproteins, which have
Asp at position 2, avoid the action of LolCDE,
thereby remaining in the inner membrane [10]. Such
a LolCDE avoidance function of Asp depends on
Keywords
ABC transporter; lipoprotein; LolCDE;
reconstitution
Correspondence
H. Tokuda, Institute of Molecular and
Cellular Biosciences, University of Tokyo,
1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-0032,
Japan
Fax: +81 3 5841 8464
Tel: +81 3 5841 7830
E-mail: htokuda@iam.u-tokyo.ac.jp
*Present address
Department of Biological Mechanisms and
Functions, Graduate School of Bioagricultu-
ral Sciences, Nagoya University, Nagoya,
Japan
†
These authors contributed equally to this
work
(Received 6 November 2006, revised
11 April 2007, accepted 17 April 2007)
doi:10.1111/j.1742-4658.2007.05832.x
The LolCDEcomplexof Escherichia coli belongs to the ATP-binding cas-
sette transporter superfamily and mediates the detachment of lipoproteins
from the inner membrane, thereby initiating lipoprotein sorting to the
outer membrane. The complex is composed of one copy each of membrane
subunits LolC and LolE, and two copies of ATPase subunit LolD. To
establish the conditions for reconstituting the LolCDEcomplexfrom sepa-
rately isolated subunits, the ATPase activities of LolD and LolCDE were
examined under various conditions. We found that both LolD and
LolCDE were inactivated on incubation at 30 °C in a detergent solution.
ATP and phospholipids protected LolCDE, but not LolD. Furthermore,
phospholipids reactivated LolCDE even after its near complete inactiva-
tion. LolD was also protected from inactivation when membrane subunits
and phospholipids were present together, suggesting the phospholipid-
dependent reassembly ofLolCDE subunits. Indeed, the functional lipo-
protein-releasing machinery was reconstituted into proteoliposomes with
E. coli phospholipids and separately purified LolC, LolD and LolE. Prein-
cubation with phospholipids at 30 °C was essential for the reconstitution
of the functional machinery from subunits. Strikingly, the lipoprotein-
releasing activity was also reconstituted from LolE and LolD without
LolC, suggesting the intriguing possibility that the minimum lipoprotein-
releasing machinery can be formed from LolD and LolE. We report here
the completereconstitutionof a functional ATP-bindingcassette transpor-
ter fromseparately purified subunits.
Abbreviations
ABC, ATP-binding cassette; BN, blue native; DDM, n-dodecyl-b-
D-maltopyranoside; His-tag, hexahistidine tag.
3034 FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS
phosphatidylethanolamine in the inner membrane [11].
It has been proposed that a steric and electrostatic
interaction between Asp at position 2 and phosphatidyl-
ethanolamine is responsible for the LolCDE avoidance
mechanism [11].
ABC transporters have four domains, two mem-
brane domains and two nucleotide-binding domains.
These domains are frequently present in separate sub-
units in bacteria, whereas eukaryotic ABC transporters
generally have these domains in a single polypeptide
chain [12]. The LolCDEcomplexof E. coli is com-
posed of one copy each of membrane subunits LolC
and LolE, and two copies of ATPase subunit LolD [7].
Both LolC and LolE are assumed to span the mem-
brane four times and to have a periplasmic region
comprising 200 amino acids. The two proteins are
similar to each other, the sequence identity being 26%.
However, both LolC and LolE are required for the
growth of E. coli [13]. As lipoproteins are present on
the outer leaflet of the inner membrane, LolC and ⁄ or
LolE, but not LolD, are responsible for the recogni-
tion of lipoproteins. It is of great interest how the
membrane and ATP-bindingsubunits communicate
with each other, as this is essential for the transfer of
substrate-binding information from LolC ⁄ LolE to
LolD, and that of ATP energy from LolD to LolC ⁄
LolE.
We recently reported the isolation of several LolC
and LolE mutants that suppress dominant negative
mutants of LolD [14]. Interestingly, the suppressor
mutations of LolE were mostly located in the cytoplas-
mic and transmembrane regions, whereas those of
LolC were found in the periplasmic domain, suggesting
that LolC and LolE interact differently with LolD and
play different roles in the LolCDE complex. To under-
stand the mechanism of LolCDE, the mode of commu-
nication between the respective membrane subunits
and LolD needs to be clarified. It is therefore import-
ant to establish conditions for the complete reconstitu-
tion of the LolCDEcomplexfromseparately isolated
subunits. However, this has been reported only for
OpuA of Lactococcus lactis [15] and Bacillus subtilis
[16], although the functional reassembly ofan ABC
transporter from a membrane complex comprising two
subunits and an ATPase subunit has been reported
[17,18]. L. lactis OpuA is composed of two copies of a
translocator subunit with a substrate-binding domain
and two copies ofan ATPase subunit. The L. lactis
OpuA complex disassembles and reassembles upon a
decrease and increase, respectively, in the glycerol con-
centration of the buffer [15]. To investigate the role
of two substrate-binding domains, hetero-oligomeric
OpuA complexes were formed by decreasing and then
increasing the glycerol concentration of a solution con-
taining OpuA mixtures. The hetero-oligomeric OpuA
thus formed was then reconstituted into proteolipo-
somes [15]. As this method was not adaptable to
B. subtilis OpuA, all subunitsof the B. subtilis OpuA
were separatelyisolated and then successfully reassoci-
ated in detergent solution [16].
Here, we report that a functional LolCDE complex
could be reconstituted fromseparately purified LolC,
LolD and LolE. Moreover, we found that the lipo-
protein release activity could be reconstituted from
LolD and LolE without LolC.
Results
ATPase activities ofLolCDE and LolD
LolD possessing a hexahistidine tag (His-tag) at the C-
terminus and the LolCDEcomplex containing LolC
with a His-tag at the C-terminus were overproduced
and purified using a TALON metal affinity resin. LolD
was purified from the cytosol as a soluble protein,
and LolCDE was purified after solubilization of
membranes with 1% n-dodecyl-b-d-maltopyranoside
(DDM). The initial rates of ATP hydrolysis were then
determined in a DDM solution containing various
concentrations of ATP. The K
m
values thus determined
were 0.11 ± 0.02 mm (n ¼ 4) and 0.43 ± 0.02 mm
(n ¼ 3) for LolCDE and LolD, respectively, where n
represents the number of determinations. The V
max
values were 0.38 ± 0.03 (n ¼ 4) and 0.43 ± 0.05
(n ¼ 3) lmol ATP hydrolyzedÆmin
)1
Æmg
)1
LolCDE
and LolD, respectively. The reported ATPase activities
of ABC transporters vary significantly between 0.01
and 20 lmolÆmin
)1
Æmg
)1
protein [19]. Turnover num-
bers were 0.9 ± 0.08 and 0.19 ± 0.02 mol ATP
hydrolyzedÆs
)1
Æmol
)1
LolCDE and LolD, respectively.
The LolCDEcomplex contained two molecules of
LolD. However, the turnover numbers were still higher
with LolCDE than with LolD, even after correction
for LolD molecules. The ATPase activity of the
LolCDE complex was essentially the same whether or
not a His-tag was attached to LolD [20] or LolC.
LolD was monomeric (see below) and did not exhibit
cooperativity in the hydrolysis of ATP in a DDM
solution (data not shown).
Inactivation and reactivation of the ATPase
activity of LolCDE
It was previously found that the LolCDE complex
in n-octyl-b-d-glucopyranoside was quickly inacti-
vated even when ATP or phospholipid was added. We
K. Kanamaru et al. Reconstitutionof the LolCDEcomplexfrom subunits
FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS 3035
therefore used sucrose monocaprate for purification
and reconstitutionofLolCDE [7]. However, ATP was
required for the stabilization ofLolCDE in this deter-
gent. We then found that the LolCDEcomplex could
be stably purified with 1% DDM not only in the pres-
ence but also in the absence of ATP, leading to the
isolation of a unique liganded LolCDEcomplex [20].
Purified LolCDE could be stored frozen in 0.01%
DDM without generation of precipitates. LolCDE
was reconstituted by incubation with phospholipids
in a solution containing 1.2% sucrose monocaprate,
followed by dialysis and dilution [20]. Similarly, the
maltose transportercomplex MalFGK
2
was solubilized
with 1% DDM, purified in 0.01% DDM, and then
reconstituted into proteoliposomes by the octylgluco-
side dilution method [21].
To construct the completereconstitution system of
the LolCDEcomplexfromisolated subunits, it seemed
important to examine in detail the stability of LolD
and LolCDE in a DDM solution. The ATPase activity
of LolCDE in a DDM solution was stable on ice for
at least 2 h even in the absence of ATP. However,
incubation at 30 °C was found to cause a rapid
decrease in the ATPase activity ofLolCDE (Fig. 1A).
In contrast, no inactivation occurred when ATP or
E. coli phospholipids were present during incubation.
Blue native (BN)-PAGE revealed that LolCDE, which
has a molecular mass of 140 kDa, migrated to a
position corresponding to a molecular mass of
180 kDa (lane 1), whereas no material was detected
at this position when LolCDE was incubated at 30 °C
for 60 min (lane 2). It seems likely that the major frac-
tion ofLolCDE did not enter the gel because of disas-
sembly and ⁄ or denaturation induced by incubation
with detergent. On the other hand, when ATP was pre-
sent during incubation, LolCDE migrated to a position
corresponding to a slightly lower molecular mass
( 170 kDa) (lane 3) than in the case of the nonincu-
bated sample (lane 1). ATP binding to LolD seemed to
cause differences in the migration position of LolCDE.
When LolCDE was incubated in a DDM solution at
30 °C for 60 min, the rate of ATP hydrolysis decreased
to only about 15% of that determined before incuba-
tion (compare the open and closed circles in Fig. 1C).
This decreased ATPase activity may represent the acti-
vity of LolD alone because of the disassembly of
LolCDE. The inactivated LolCDE was then mixed with
E. coli phospholipids and further incubated for the
specified times. The incubation with phospholipids
caused recovery of the activity ofLolCDE to about 50%
and 80% of the original level after 10 min (squares)
and 120 min (closed triangles), respectively, suggesting
that disassembled LolCDE was reassembled.
A
C
B
Fig. 1. Inactivation and reactivation of LolCDE. The LolCDE com-
plex was overproduced from plasmids pNASCH and pKM501.
LolD was overproduced from pKM202. (A) LolCDE (3 lg) was
incubated at 30 °C for the specified times in 105 lLof50m
M
Tris ⁄ HCl (pH 7.5) containing 10% glycerol and 0.3% DDM. Where
specified, 8 mg mL
)1
E. coli phospholipids (PL) or 2 mM ATP were
also present during the incubation. ATP hydrolysis was examined
by the addition of 2 m
M ATP and 2 mM MgSO
4
, as described
under Experimental procedures. (B) LolCDE (3 lg) was analyzed
by BN-PAGE as described under Experimental procedures. Lane 1:
LolCDE before incubation. Lane 2: LolCDE after incubation with
no supplementation. Lane 3: LolCDE after incubation with 2 m
M
ATP. The migration positions of molecular mass markers (M) are
indicated in kDa. (C) ATPase activity was examined with LolCDE
incubated at 30 °C for 60 min as in (A) (closed circles) or not incu-
bated (open circles). After 60 min of incubation at 30 °C, LolCDE
was further incubated with E. coli phospholipids (8 mgÆmL
)1
) for
10 min (open squares), 20 min (open reverse triangles), 30 min
(closed reverse triangles), 40 min (open triangles), or 60 min
(closed triangles), and then subjected to ATPase assay at the indi-
cated times.
Reconstitution of the LolCDEcomplexfromsubunits K. Kanamaru et al.
3036 FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS
LolD purified as a soluble protein from the cytoplas-
mic fraction was also inactivated when incubated in
the DDM solution (compare the open and closed cir-
cles in Fig. 2A,C,E). Unlike in the case of LolCDE,
the presence of phospholipids alone did not protect
LolD (compare the open and closed circles in
Fig. 2B,D,F). Neither the addition of LolC or LolE,
nor the addition of both in the absence of phospho-
lipids, protected the ATPase activity of LolD (compare
the open and closed triangles in Fig. 2A.,C,E). On the
other hand, the addition of LolC (Fig. 2B) or LolC ⁄
LolE (Fig. 2F) in the presence of phospholipids pre-
vented inactivation of LolD to some extent (compare
the open and closed triangles).
Taken together, these results suggest that the mem-
brane subunits stabilize the ATPase subunit LolD in
the presence of phospholipids. It was also strongly sug-
gested that the membrane subunits interact with LolD
in the presence of phospholipids even when they are
added separately.
Reconstitution of the functional
lipoprotein-releasing machinery from subunits
The four domains of bacterial ABC transporters are
frequently located in different subunits. Complete
reconstitution of ABC transporters from separate sub-
units has been reported only for OpuA [15,16],
although reassembly ofan ATPase homodimer with a
heterodimer of the membrane subunit has been repor-
ted [17,18]. The results shown in Figs 1 and 2 sugges-
ted a functional interaction between LolC ⁄ LolE and
LolD. We therefore examined the reconstitution of
lipoprotein-releasing activity from the three subunits
(Fig. 3). The efficiency of lipoprotein release from pro-
teoliposomes is usually low even with the LolCDE
complex, presumably because the orientation of the
reconstituted proteins is random, thereby leaving
a major fraction of lipoproteins incompetent with
regard to release [7,10]. Nevertheless, reconstitution of
LolCDE revealed important aspects of the lipoprotein
release reaction [10,11,20]. When the LolCDE complex
was used, lipoprotein-releasing activity was reconstitu-
ted whether incubation with phospholipids was per-
formed on ice or at 30 °C (Fig. 3A). In marked
contrast, incubation at 30 °C was absolutely essential
for reconstituting the activity fromseparately purified
LolC, LolD and LolE. To our surprise, the lipo-
protein-releasing activity was also reconstituted from
LolD and LolE without LolC. The reconstituted lipo-
protein-releasing activity was dependent on LolA. On
the other hand, the activity was hardly reconstituted
from LolC and LolD.
The Ala fi Pro mutation at position 40 of LolC
causes the outer membrane localization of lipoproteins
possessing the inner membrane retention signal [22].
This may indicate the importance of LolC for lipopro-
tein sorting. The two Asp residues at positions 2 and 3
of lipoproteins function as typical inner membrane
retention signals, and are found in native inner mem-
brane lipoproteins [5]. We examined whether or not
the active machinery lacking LolC releases Pal with
Fig. 2. Protection of LolD by membrane subunits. His-tagged LolD,
LolC and LolE were overproduced from pKM202, pNASCH and
pNASEH, respectively. The ATPase activity of LolD (4.5 lg) before
incubation (open circles) or after incubation at 30 °C for 60 min
(closed circles) was determined in 50 m
M Tris ⁄ HCl (pH 7.5) contain-
ing 10% glycerol and 0.3% DDM as described under Experimental
procedures. Where specified (closed triangles), incubation was car-
ried out in the presence of LolC (A, B) or LolE (C, D), or both (E, F)
with (B, D, F) or without (A, C, E) 8 mgÆmL
)1
E. coli phospholipids
(PL). The open triangles in each panel represent the activity deter-
mined before incubation in the presence of the specified mem-
brane subunits with or without phospholipids.
K. Kanamaru et al. Reconstitutionof the LolCDEcomplexfrom subunits
FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS 3037
this signal (Fig. 3B). The signal remained inner mem-
brane-specific, and Pal(DD) was not released from any
of the three machineries, i.e. LolD ⁄ E, LolC ⁄ D ⁄ E and
LolCDE. The release of lipoproteins from these machi-
neries was sensitive to orthovanadate (Fig. 3C), which
is a specific inhibitor ofLolCDE [7]. Taken together,
these results indicate that the minimum lipoprotein-
releasing machinery can be reconstituted with LolD
and LolE without LolC.
Assembly of the LolCDEcomplexfrom separately
isolated subunits
To examine the formation of lipoprotein-releasing Lol
complexes, separatelyisolated Lol proteins were mixed
as indicated, incubated on ice or at 30 °C with or with-
out phospholipids, and then subjected to analysis by
gel filtration chromatography (Fig. 4). When LolC,
LolE and LolD were separately examined, they were
eluted at positions corresponding to respective mono-
mers even after incubation at 30 °C with phospholipids
(Fig. 4A,B,C). The LolCDEcomplex was eluted at a
position corresponding to 160 kDa (Fig. 4D). When
LolC, LolD and LolE were mixed and incubated at
30 °C in the absence of phospholipids, the three Lol
proteins remained at the respective monomer positions
(Fig. 4E). Incubation of LolC, LolD, LolE and phos-
pholipids together on ice caused the formation of a
small amount of the LolCDE complex, which was elu-
ted at a position corresponding to the intact LolCDE
complex (Fig. 4F). In contrast, incubation of these
three Lol proteins with phospholipids at 30 °C caused
the formation of substantial amounts of the LolCDE
complex (Fig. 4G).
Incubation of LolD with either LolC (Fig. 4I) or
LolE (Fig. 4J) at 30 °C in the presence of phospho-
lipids also caused elution of a small amount of Lol
proteins at fractions corresponding to 160 kDa,
indicating that LolCD and LolDE complexes are
formed. These results suggest that both LolC and
LolE can directly interact with LolD, although the
formation of LolCD (Fig. 4I) and LolDE (Fig. 4J)
complexes was significantly less efficient than that of
the LolCDEcomplex (Fig. 4G). The LolDE complex
exhibited a low Pal-releasing activity, whereas the
activity of the LolCD complex was not detected
(Fig. 3). These results suggest that the two membrane
subunits play different roles in the lipoprotein release
reaction.
To determine the subunit stoichiometry of com-
plexes formed in vitro, the amounts of Lol proteins
were quantitated and corrected with regard to the
respective molecular masses. The LolCDE complex
formed in vitro (Fig. 4F,G) had essentially the same
subunit stoichiometry as the intact LolCDE complex
(Fig. 4D). LolD contents in LolCD (Fig. 4I) and
LolDE (Fig. 4J) complexes were slightly higher than
expected. It is not clear at present whether these com-
plexes are composed of two copies of the membrane
subunit and three copies of LolD (molecular mass ¼
164–168 kDa) or two copies each of the membrane
subunit and LolD (molecular mass ¼ 138–142 kDa),
although an ABC transporter is generally composed of
two membrane domains and two nucleotide-binding
domains.
Discussion
Bacterial ABC transporters frequently have four
domains in separate subunits [12]. It was previously
suggested that LolC and LolE interact differently with
LolD and play different roles in the LolCDE complex
A
B
C
Fig. 3. Reconstitutionof the lipoprotein-releasing machinery from
isolated subunits. (A) LolD (177 pmol), LolC (88 pmol), and LolE
(88 pmol) were mixed in various combinations, and then incubated
with 2 lg of Pal and 0.8 mg of E. coli phospholipids for 60 min
either on ice or at 30 °C in 1.2% sucrose monocaprate solution. To
reconstitute proteoliposomes, the mixtures were then subjected to
dilution and dialysis as described under Experimental procedures.
As a control, the LolCDEcomplex was also reconstituted. Reconsti-
tuted proteoliposomes were collected and subjected to the release
reaction in the presence of LolA and ATP as described under
Experimental procedures. (B) Pal(DD) was also reconstituted as in
(A), and the ability of proteoliposomes to release Pal(DD) was
examined. (C) Proteoliposomes were reconstituted with the indica-
ted Lol proteins and Pal as in (A). The release of Pal was then
examined in the presence and absence of 1 m
M orthovanadate.
Reconstitution of the LolCDEcomplexfromsubunits K. Kanamaru et al.
3038 FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS
[14]. Moreover, several mutants have been isolated for
each subunit [7,10,14,20]. We therefore wanted to
establish the conditions for reconstituting the func-
tional complexfromseparatelyisolated subunits. How-
ever, so far, only OpuA has been reported to be
reconstituted from separate subunits. Our previous
attempt to reconstitute the LolCDEcomplexfrom sub-
units was also unsuccessful. Here, we found a rather
simple method; incubation ofsubunits at 30 °C in the
presence of phospholipids leads to the reconstitution
of the functional LolCDEcomplex (Figs 3 and 4).
On the other hand, various membrane apparatuses,
including ABC transporters such as maltose permease
[21,23], histidine permease [24] and the LolCDE
complex [7], and a Sec protein translocase [25,26],
have been reconstituted at low temperature. This tem-
perature-dependent reconstitution is caused by tem-
perature-dependent assembly of Lol subunits in the
presence of phospholipids (Fig. 4). It has been repor-
ted that the components of maltose permease aggre-
gate upon separate overproduction [27], whereas
the three subunitsof the LolCDEcomplex could be
A
B
C
D
E
F
G
H
I
J
Fig. 4. In vitro assembly of Lol subunits.
LolC (88 pmol), LolD (176 pmol) and LolE
(88 pmol) were incubated on ice or at 30 °C
for 60 min in 100 lLof20m
M Tris ⁄ HCl
(pH 7.5) containing 10% glycerol, 5 m
M
MgCl
2
,2mM ATP, 0.8 mg E. coli phospholi-
pids (PL) and 0.01% DDM as described
under Experimental procedures. Where spe-
cified, phospholipids (D, E), ATP (H), LolC (J)
or LolE (I) were omitted. The reaction mix-
ture was then subjected to gel filtration
chromatography (Superose 6, 10 ⁄ 300 GL),
on a column that had been equilibrated with
20 m
M Tris ⁄ HCl (pH 7.5) containing 10%
glycerol and 0.01% DDM. The column was
developed with the same buffer at a rate of
0.5 mLÆmin
)1
. Aliquots of fractions (0.5 mL)
were analyzed by SDS ⁄ PAGE and CBB
staining after precipitation with trichloro-
acetic acid. The amounts of the respective
Lol proteins were densitometrically deter-
mined in the specified fractions and correc-
ted with regard to the respective molecular
masses. The molecular amounts of LolD
and LolE are indicated, taking the amount of
LolC as 1. The elution positions of molecular
mass markers are indicated above the gel.
As controls, isolated LolC, LolE, LolD and
LolCDE were also analyzed (A–D), and their
elution positions are indicated by open
arrowheads.
K. Kanamaru et al. Reconstitutionof the LolCDEcomplexfrom subunits
FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS 3039
separately overproduced and reassembled to form the
functional complex. The temperature-dependent assem-
bly ofsubunits may be characteristic of the LolCDE
complex.
The integrity of the LolCDEcomplex at 30 °C was
found to be strictly dependent on phospholipids. The
complex rapidly lost its activity when incubated at
30 °C in a DDM solution (Fig. 1). This inactivation
was completely prevented by the addition of phospho-
lipids or ATP. BN-PAGE suggested that the LolCDE
complex disassembles and denatures in the absence of
protective agents upon incubation. Phospholipids reac-
tivated LolCDE, presumably by mediating the reas-
sembly of the three subunits. ATP did not reactivate
LolCDE, suggesting that phospholipids and ATP
protect LolCDE through different mechanisms. ATP
binding to the nucleotide-binding domains of ABC
transporters has been proposed to yield the closed
dimer [28,29], which is likely to be more resistant to
inactivation. Inactivation of LolD on incubation in the
DDM solution was also prevented when both phos-
pholipids and membrane subunits were present (Fig. 2).
Overproduced LolD was isolatedfrom the cytosol as a
soluble protein, and remained active unless it was incu-
bated in the DDM solution. It seems possible that
DDM at 30 °C has a weak denaturing effect, which is
prevented by the phospholipid-dependent interaction
with membrane subunits.
It has been proposed that LolCDE recognizes the
N-terminal Cys of lipoproteins together with attached
diacylglycerol and an N-linked acyl chain [11]. There-
fore, the structure recognized by LolCDE resembles
that of phospholipid. This may be related to the
strong phospholipid dependence of LolCDE, although
LolCDE does not export phospholipids.
LolC was found to be dispensable for the reconsti-
tution of the minimum lipoprotein-releasing machin-
ery (Fig. 3). This was unexpected, because both LolC
and LolE are required for the growth of E. coli [13].
The isolation of defective mutants of various Lol pro-
teins revealed that efficient lipoprotein sorting to the
outer membrane is essential for the growth of E. coli
[20,30,31], which possesses more than 80 species of
outer membrane-specific lipoproteins [1]. On the other
hand, only Pal was reconstituted into proteolipo-
somes. This may be the reason why the lack of LolC
caused a marginal defect in the release activity of
proteoliposomes. It is likely that both LolC and LolE
are essential in vivo, because a large amount of lipo-
proteins should be rapidly sorted to the outer mem-
brane. Our data suggest that the lack of LolC
decreases the affinity for lipoproteins (unpublished
results). These seem to be unfavorable for the efficient
outer membrane sorting of lipoproteins in vivo,
whereas the defect was only marginal in the reconsti-
tuted proteoliposomes.
Both the membrane topology and amino acid
sequence (26% identity) are similar between LolC and
LolE, whereas the two proteins seem to play different
roles [14]. The results shown here suggest that the lipo-
protein-binding site is present in LolE, which is cur-
rently under investigation. Lol proteins are highly
conserved in various Gram-negative bacteria. How-
ever, some bacteria, such as Bordetella pertussis and
Neisseria meningitidis, possess a single species of mem-
brane subunit [32], suggesting that the lipoprotein-
releasing apparatus is composed of a homodimer of
the membrane subunit and a homodimer of LolD in
these bacteria.
Experimental procedures
Materials
Escherichia coli phospholipids were obtained from Avanti
Polar Lipids (Alabaster, AL) and washed with acetone as
previously reported [33]. b-d-Fructopyranosyl-a-d-glucopyr-
anoside monodecanoate (sucrose monocaprate) and DDM
were purchased from Dojindo Laboratories (Kumamoto,
Japan).
Overproduction of Lol proteins
Lol proteins were overproduced in E. coli JC7752
(supE hsdS met gal lacY fhua DtolB-pal) [34] harboring the
specified plasmids listed in Table 1. When the culture
absorbance at 660 nm reached 0.5, the expression of Lol
proteins from Ptac and the araBAD operon promoter
(P
BAD
) was induced at 30 °C for 2 h by the addition of
1mm isopropyl-b-d-thiogalactopyranoside and 0.2% arabi-
nose, respectively. Unless otherwise specified, the LolCDE
complex was purified from cells harboring pNASCH and
pKM501.
Table 1. Plasmids used in this study. ‘-his’ represents a hexahisti-
dine tag attached to the C-terminus of the respective Lol protein.
Plasmid Protein Promoter Reference
pKM202 LolD-his Ptac [14]
pKM301 LolE Ptac [10]
pKM402 LolC, LolD-his P
BAD
[10]
pKM501 LolD, LolE Ptac This study
pNASC LolC P
BAD
This study
pNASCH LolC-his P
BAD
This study
pNASE LolE P
BAD
This study
pNASEH LolE-his P
BAD
This study
Reconstitution of the LolCDEcomplexfromsubunits K. Kanamaru et al.
3040 FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS
Construction of plasmids
To construct pKM501 carrying lolD and lolE under the
control of tacPO and lacIq, the corresponding region of
pJY310 [7] was amplified by PCR using a pair of oligonu-
cleotides, 5¢-GAGCTCGAAGGAGATATAAATATGAAT
AAGATCCTGTTGCAATGC-3¢ and 5¢-AAGCCTGCAG
TTTTTGTTCCACCAATATCAAACCC-3¢. The amplified
DNA was digested with SacI and PstI, and then inserted
into the same restriction site of pTTQ18 [35].
To construct pNASC carrying lolC under P
BAD
,a
1.2 kbp EcoRI–PstI fragment of pKM101 [10] was cloned
into the same site of pMAN885EH [36].
To construct pNASCH carrying a gene that encodes
LolC with a His-tag at its C-terminus, PCR was performed
with pJY310 as a template and a pair of oligonucleotides,
5¢-GATGAATTCGGAGGTTTAAATTTATGTACCAAC
CTGTCGCTCTATTTA-3¢ and 5¢-CAATTCAAGCTTAA
TGATGATGATGATGATGCTCCAGTTCATAACGTAA
AGCCTCAGCGG-3¢. The amplified DNA was digested
with Eco RI and HindIII, and then cloned into the same site
of pMAN885EH.
To construct pNASE carrying lolE under P
BAD
,a
1.3 kbp EcoRI–PstI fragment of pKM301 [10] was cloned
into the same site of pMAN885EH.
To construct pNASEH carrying a gene that encodes
LolE with a His-tag at its C-terminus, PCR was performed
with pJY310 as a template and a pair of oligonucleotides,
5¢-GATGAATTCGGAGGTTTAAATTTATGGCGATGC
CTTTATCGTTATTAA-3¢ and 5¢-CAATTCAAGCTTAA
TGATGATGATGATGATGCTCCAGCTGGCCGCTAAG
GACTCGCGCAG-3¢. The amplified DNA was digested
with Eco RI and HindIII, and then cloned into the same site
of pMAN885EH.
Isolation of Lol proteins
JC7752 cells overproducing Lol proteins were converted
into spheroplasts, and then disrupted by passage through
a French pressure cell (10 000 lbÆin
)1
). Lysates were fract-
ionated into total membrane fractions and supernatants
by centrifugation at 100 000 g for 60 min using a rotor
type 50.2 Ti in Optima L-60 ultracentrifuge (Beckman
Coulter, Fulleston, CA). To purify Lol proteins and Lol
protein complexes, total membranes at 5 mgÆmL
)1
were
solubilized on ice for 30 min with 50 mm Tris ⁄ HCl
(pH 7.5) containing 10% glycerol, 5 mm MgCl
2
and 1%
DDM. A solubilized supernatant was obtained by centrif-
ugation at 100 000 g for 30 min using rotor type 50.2 Ti
in Optima L-60, and then applied to a 1 mL TALON col-
umn (Clontech Laboratories, Mountain View, CA) that
had been equilibrated with 50 mm Tris ⁄ HCl (pH 7.5) con-
taining 10% glycerol, 100 mm NaCl, and 0.01% DDM.
Lol proteins and their complexes were eluted with a linear
gradient of imidazole (0–250 mm). His-tagged LolD was
purified from supernatants of cell lysates and then purified
on a TALON column as described above, except for the
absence of DDM.
ATPase activity
ATP hydrolysis by LolCDE (3 lg) or LolD (4.5 lg) was
determined in 105 lLof50mm Tris ⁄ HCl (pH 7 .5) con-
taining 10% glycerol and 0.3% DDM. The assay was
started at 30 °C by the addition of 2 mm ATP and 2 mm
MgCl
2
. Aliquots (15 lL) of the reaction mixture were
withdrawn at the indicated time points, and then mixed
with the same volume of 12% SDS to stop the hydrolysis.
The amounts of inorganic phosphate were determined
according to a previously reported method [37]. In some
experiments, ATP hydrolysis by LolCDE and LolDE was
examined after their reconstitution into proteoliposomes
with or without Pal.
Page
SDS ⁄ PAGE was performed according to Laemmli [38].
Immunoblotting [39] was performed as described. BN-PAGE
was carried out according to a previously reported method
[40]. The cathode buffer contained 0.002% Coomassie Bril-
liant Blue G-250 and 0.01% DDM was included in the
sample buffer.
Reconstitution of the LolCDEcomplexfrom its
subunits
Reconstitution of the LolCDEcomplex into proteolipo-
somes was performed as described previously [20]. To
form the complexfromisolated subunits, specified Lol
proteins were incubated for 1 h on ice or at 30 °C with
0.8 mg of E. coli phospholipids and 2 lg of Pal in 100 lL
of 50 mm Tris ⁄ HCl (pH 7.5) containing 2 mm MgSO
4
,
100 mm NaCl and 1.2% sucrose monocaprate. The mix-
ture was diluted with 900 lLof50mm Tris ⁄ HCl (pH 7.5)
containing 2 mm MgSO
4
and 100 mm NaCl, and then dia-
lyzed against 1000 mL of the same buffer at 4 °C over-
night. Reconstituted proteoliposomes were collected by
centrifugation at 100 000 g for 2 h using a TLA55 rotor
in a Beckman ultracentrifuge Optima MAX, and then sub-
jected to the Pal release assay at 30 °C for 15 min in the
presence of 4 lg of LolA and 2 mm ATP as previously
reported [7]. The reaction mixtures were fractionated into
proteoliposomes and supernatants by centrifugation at
100 000 g for 2 h using a TLA55 rotor in a Beckman
ultracentrifuge Optima MAX. Pal in the pellets and sup-
ernatants was analyzed by SDS ⁄ PAGE and immunoblot-
ting with antibodies to Pal. Unless otherwise specified,
1 ⁄ 50 of the pellet material and 1 ⁄ 3 of the supernatant
material were applied to the gel.
K. Kanamaru et al. Reconstitutionof the LolCDEcomplexfrom subunits
FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS 3041
In vitro assembly of Lol subunits
The complete reaction mixture contained 88 pmol of LolC,
176 pmol of LolD and 88 pmol of LolE in 100 lLof
20 mm Tris ⁄ HCl (pH 7.5) containing 10% glycerol, 5 mm
MgCl
2
,2mm ATP, 0.8 mg of E. coli phospholipids,
and 0.01% DDM. Where specified, LolC or LolE was
omitted or incubation was carried out in the absence of
phospholipids or ATP. The reaction mixture was incubated
on ice or at 30 °C for 60 min, and then subjected to gel fil-
tration chromatography (Superose 6, 10 ⁄ 300GL, GE
Healthcare, Chalfont St Giles, UK) on a column that had
been equilibrated with 20 mm Tris ⁄ HCl (pH 7.5) containing
10% glycerol and 0.01% DDM. The column was developed
with the same buffer at a rate of 0.5 mLÆmin
)1
. Fractions
of 0.5 mL were collected, and aliquots [1 ⁄ 3] was analyzed
by SDS ⁄ PAGE and Coomassie Brilliant Blue staining after
precipitation with trichloroacetic acid. The amounts of Lol
proteins were densitometrically determined in specified gel
filtration chromatography fractions and corrected with
regard to the respective molecular masses.
Acknowledgements
We wish to thank Rika Ishihara for technical support.
This work was supported by grants to H. Tokuda
from the Ministry of Education, Science, Sports and
Culture of Japan.
References
1 Miyadai H, Tanaka-Masuda K, Matsuyama S &
Tokuda H (2004) Effects of lipoprotein overproduction
on the induction of DegP (HtrA) involved in quality
control in the Escherichia coli periplasm. J Biol Chem
279, 39807–39813.
2 Sankaran K & Wu HC (1994) Lipid modification of
bacterial prolipoprotein. Transfer of diacylglyceryl
moiety from phosphatidylglycerol. J Biol Chem 269,
19701–19706.
3 Yamaguchi K, Yu F & Inouye M (1988) A single amino
acid determinant of the membrane localization of lipo-
proteins in E. coli. Cell 53, 423–432.
4 Seydel A, Gounon P & Pugsley AP (1999) Testing the
‘2 rule’ for lipoprotein sorting in the Escherichia coli cell
envelope with a new genetic selection. Mol Microbiol
34, 810–821.
5 Terada M, Kuroda T, Matsuyama S & Tokuda H
(2001) Lipoprotein sorting signals evaluated as the
LolA-dependent release of lipoproteins from the inner
membrane of Escherichia coli. J Biol Chem 276,
47690–47694.
6 Tokuda H & Matsuyama S (2004) Sorting of lipo-
proteins to the outer membrane in E. coli. Biochim
Biophys Acta 1693, 5–13.
7 Yakushi T, Masuda K, Narita S, Matsuyama S &
Tokuda H (2000) A new ABC transporter mediating the
detachment of lipid-modified proteins from membranes.
Nat Cell Biol 2, 212–218.
8 Matsuyama S, Tajima T & Tokuda H (1995) A novel
periplasmic carrier protein involved in the sorting and
transport of Escherichia coli lipoproteins destined for
the outer membrane. EMBO J 14, 3365–3372.
9 Matsuyama S, Yokota N & Tokuda H (1997) A novel
outer membrane lipoprotein, LolB (HemM), involved in
the LolA (p20)-dependent localization of lipoproteins to
the outer membrane of Escherichia coli. EMBO J 16,
6947–6955.
10 Masuda K, Matsuyama S & Tokuda H (2002) Elucida-
tion of the function of lipoprotein-sorting signals that
determine membrane localization. Proc Natl Acad Sci
USA 99, 7390–7395.
11 Hara T, Matsuyama S & Tokuda H (2003) Mechanism
underlying the inner membrane retention of E. coli lipo-
proteins caused by Lol avoidance signals. J Biol Chem
278, 40408–40414.
12 Holland IB & Blight MA (1999) ABC-ATPases, adapta-
ble energy generators fuelling transmembrane movement
of a variety of molecules in organisms from bacteria to
humans. J Mol Biol 293, 381–399.
13 Narita S, Tanaka K, Matsuyama S & Tokuda H (2002)
Disruption of lolCDE, encoding an ATP-binding-
cassette transporter, is lethal for Escherichia coli and
prevents the release of lipoproteins from the inner mem-
brane. J Bacteriol 184, 1417–1422.
14 Ito Y, Matsuzawa H, Matsuyama S, Narita S &
Tokuda H (2006) Genetic analysis of the mode of inter-
play between an ATPase subunit and membrane sub-
units of the lipoprotein-releasing ATP-binding cassette
transporter LolCDE. J Bacteriol 188, 2856–2864.
15 Biemans-Oldehinkel E & Poolman B (2003) On the role
of the two extracytoplasmic substrate-binding domains
in the ABC transporter OpuA. EMBO J 22, 5983–5993.
16 Horn C, Bremer E & Schmitt L (2005) Functional over-
expression and in vitro reassociation of OpuA, as osmo-
tically regulated ABC-transport complexfrom Bacillus
subtilis. FEBS Lett 579, 5765–5768.
17 Liu PQ & Ames GF (1998) In vitro disassembly and
reassembly ofan ABC transporter, the histidine per-
mease. Proc Natl Acad Sci USA 95, 34595–33500.
18 Sharma S, Davis JA, Ayvaz T, Traxler B & Davidson
A (2005) Functional reassembly of the Escherichia coli
maltose transporter following purification of a MalF–
MalG subassembly. J Bacteriol 187, 2908–2911.
19 Schneider E & Hunke S (1998) ATP-binding-cassette
(ABC) transporter systems: functional and structural
aspects of the ATP-hydrolyzing subunits ⁄ domains.
FEMS Microbiol Rev 22, 1–20.
20 Ito Y, Kanamaru K, Taniguchi N, Miyamoto S &
Tokuda H (2006) A novel ligand bound ABC
Reconstitution of the LolCDEcomplexfromsubunits K. Kanamaru et al.
3042 FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS
transporter, LolCDE, provides insights into the molecu-
lar mechanisms underlying membrane detachment of
bacterial lipoproteins. Mol Microbiol 62, 1064–1075.
21 Davidson AL & Nikaido H (1991) Purification and
characterization of the membrane-associated compo-
nents of the maltose transport system from Escherichia
coli. J Biol Chem 266, 8946–8951.
22 Narita S, Kanamaru K, Matsuyama S & Tokuda H
(2003) A mutation in the membrane subunit ofan ABC
transporter LolCDEcomplex causing outer membrane
localization of lipoproteins against their inner mem-
brane-specific signals. Mol Microbiol 49, 167–177.
23 Davidson AL & Nikaido H (1990) Overproduction,
solubilization, and reconstitutionof the maltose trans-
port system from Escherichia coli. J Biol Chem 265,
4254–4260.
24 Panagiotidis CH, Reyes M, Sievertsen A, Boos W &
Shuman HA (1993) Characterization of the structural
requirements for assembly and nucleotide binding of
an ATP-bindingcassette transporter. The maltose trans-
port system of Escherichia coli. J Biol Chem 268,
23685–23696.
25 Akimaru J, Matsuyama S, Tokuda H & Mizushima S
(1991) Reconstitutionof a protein translocation system
containing purified SecY, SecE, and SecA from Escheri-
chia coli. Proc Natl Acad Sci USA 88, 6545–6549.
26 Brundage L, Hendrick J, Schiebel E, Driessen AJM &
Wickner W (1990) The purified E. coli integral mem-
brane protein SecY ⁄ E is sufficient for reconstitution of
SecA-dependent precursor protein translocation. Cell
62, 649–657.
27 Liu CE & Ames GF (1997) Characterization of trans-
port through the periplasmic histidine permease using
proteoliposomes reconstituted by dialysis. J Biol Chem
272, 859–866.
28 Davidson AL & Chen J (2004) ATP-binding cassette
transporters in bacteria. Annu Rev Biochem 73, 241–268.
29 Higgins CF & Linton KJ (2004) The ATP switch
model for ABC transporters. Nat Struct Mol Biol 11,
918–926.
30 Miyamoto A, Matsuyama S & Tokuda H (2001)
Mutant of LolA, a lipoprotein-specific molecular cha-
perone of Escherichia coli, defective in the transfer of
lipoproteins to LolB. Biochem Biophys Res Commun
287, 1125–1128.
31 Wada R, Matsuyama S & Tokuda H (2004) Targeted
mutagenesis of five conserved tryptophan residues of
LolB involved in membrane localization of Escherichia
coli lipoproteins. Biochim Biophys Res Commun 323,
1069–1074.
32 Narita S & Tokuda H (2006) An ABC transporter
mediating the membrane detachment of bacterial lipo-
proteins depending on their sorting signals. FEBS Lett
580, 1164–1170.
33 Tokuda H, Shiozuka K & Mizushima S (1990) Recon-
stitution of translocation activity for secretory proteins
from solubilized components of Escherichia coli. Eur
J Biochem 192, 583–589.
34 Bouveret E, Derouiche R, Rigal A, Lloubes R,
Lazdunski C & Benedetti H (1995) Peptidoglycan-
associated lipoprotein–TolB interaction. A possible key
to explaining the formation of contact sites between the
inner and outer membranes of Escherichia coli. J Biol
Chem 270, 11071–11077.
35 Stark MJ (1987) Multicopy expression vectors carrying
the lac repressor gene for regulated high-level expression
of genes in Escherichia coli. Gene 51, 255–267.
36 Yakushi T, Tajima T, Matsuyama S & Tokuda H
(1997) Lethality of the covalent linkage between mislo-
calized major outer membrane lipoprotein and the
peptidoglycan of Escherichia coli. J Bacteriol 179,
2857–2862.
37 Chifflet S, Torriglia A, Chiesa R & Tolosa S (1988) A
method for the determination of inorganic phosphate in
the presence of labile organic phosphate and high con-
centrations of protein: application to lens ATPases.
Anal Biochem 168, 1–4.
38 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
39 Matsuyama S, Fujita Y & Mizushima S (1993) SecD is
involved in the release of translocated secretory proteins
from the cytoplasmic membrane of Escherichia coli.
EMBO J 12, 265–270.
40 Scha
¨
gger H, Cramer WA & von Jagow G (1994)
Analysis of molecular masses and oligomeric states of
protein complexes by blue native electrophoresis and
isolation of membrane protein complexes by two-
dimensional native electrophoresis. Anal Biochem 217,
220–230.
K. Kanamaru et al. Reconstitutionof the LolCDEcomplexfrom subunits
FEBS Journal 274 (2007) 3034–3043 ª 2007 The Authors Journal compilation ª 2007 FEBS 3043
. Complete reconstitution of an ATP-binding cassette
transporter LolCDE complex from separately isolated
subunits
Kyoko Kanamaru*
,
, Naohiro Taniguchi,. the complete reconstitution system of
the LolCDE complex from isolated subunits, it seemed
important to examine in detail the stability of LolD
and LolCDE