Genetic Analysis of the Mode of Interplay between an ATPase Subunit and Membrane Subunits of the Lipoprotein-Releasing ATP-Binding Cassette
http://www.100md.com
《细菌学杂志》
Institute of Molecular and Cellular Biosciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-0032, Japan,Department of Life Science, Rikkyo University, 3-34-1 Nishi-ikebukuro, Toshima-ku, Tokyo 171-8501, Japan
ABSTRACT
The LolCDE complex, an ATP-binding cassette (ABC) transporter, releases lipoproteins from the inner membrane, thereby initiating lipoprotein sorting to the outer membrane of Escherichia coli. The LolCDE complex is composed of two copies of an ATPase subunit, LolD, and one copy each of integral membrane subunits LolC and LolE. LolD hydrolyzes ATP on the cytoplasmic side of the inner membrane, while LolC and/or LolE recognize and release lipoproteins anchored to the periplasmic leaflet of the inner membrane. Thus, functional interaction between LolD and LolC/E is critically important for coupling of ATP hydrolysis to the lipoprotein release reaction. LolD contains a characteristic sequence called the LolD motif, which is highly conserved among LolD homologs but not other ABC transporters of E. coli. The LolD motif is suggested to be a region in contact with LolC/E, judging from the crystal structures of other ABC transporters. To determine the functions of the LolD motif, we mutagenized each of the 32 residues of the LolD motif and isolated 26 dominant-negative mutants, whose overexpression arrested growth despite the chromosomal lolD+ background. We then selected suppressor mutations of the lolC and lolE genes that correct the growth defect caused by the LolD mutations. Mutations of the lolC suppressors were mainly located in the periplasmic loop, whereas ones of lolE suppressors were mainly located in the cytoplasmic loop, suggesting that the mode of interaction with LolD differs between LolC and LolE. Moreover, the LolD motif was found to be critical for functional interplay with LolC/E, since some LolD mutations lowered the ATPase activity of LolCDE without affecting that of LolD.
INTRODUCTION
Bacterial lipoproteins possess an N-terminal Cys, which is modified by fatty acids (17). In gram-negative bacteria, lipoproteins are anchored to either the inner or outer membrane via N-terminal fatty acids (16). Lipoproteins are synthesized as precursors in the cytoplasm and then translocated across the inner membrane by Sec machinery, followed by modification on the periplasmic side of the inner membrane into mature lipoproteins (16). Lipoproteins destined for the outer membrane are further transported across the periplasmic space by the Lol system (24), which is composed of five proteins (LolABCDE). The LolCDE complex in the inner membrane releases outer membrane-directed lipoproteins from the inner membrane in an ATP-dependent manner (15, 27, 28), leading to the formation of a water-soluble complex comprising one molecule each of lipoprotein and LolA in the periplasm (10, 21). The LolA-lipoprotein complex then interacts with outer membrane receptor LolB, which catalyzes the anchoring of lipoproteins to the outer membrane (11, 22). In contrast, inner membrane lipoproteins with Asp at position 2 avoid recognition by the LolCDE complex and therefore remain in the inner membrane (4, 9, 19, 23, 30).
ATP-binding cassette (ABC) transporters generally have at least 10 transmembrane stretches (6), whereas both LolC and LolE span the membrane only four times and have a large domain exposed to the periplasm between the first and second transmembrane segments (27). In addition, the sequences of LolC and LolE are similar to each other, with 26% identity. The LolCDE complex differs mechanistically from all other ABC transporters in that it does not catalyze the trans-bilayer movement of substrates but releases lipoproteins from one leaflet of a lipid bilayer (15, 27). This might be why the LolCDE complex has a total of only eight transmembrane segments. The LolCDE complex is considered to be an interesting ABC exporter variant (3).
LolD contains the consensus sequences of ABC proteins, i.e., Walker A, Walker B, and ABC signature motifs (27). The crystal structure of MJ0796, a methanococcal LolD homolog exhibiting 43.7% sequence identity, showed a very similar tertiary fold to those of the ATPase subunits of other ABC transporters (20, 32). Therefore, it is plausible that the ATP-hydrolyzing mechanism of LolD is essentially the same as those of other ABC proteins (3, 5). In addition to the motifs characteristic of ABC proteins, LolD contains a sequence called the LolD motif, which is not conserved among other ABC proteins of E. coli but is highly conserved among LolD homologs. The LolD motif is located between the Walker A and ABC signature (LSGGQ) motifs. Judging from the crystal structure of vitamin B12 transporter BtuCD (8), the LolD motif is speculated to be very close to the cytoplasmic loops of LolC and LolE. Moreover, genetic and biochemical evidence indicates that the region corresponding to the LolD motif is involved in the communication between the integral membrane subunits and the ATPase subunits of maltose transporter MalFGK2 (7, 13).
To determine the mechanism underlying the coupling of ATP hydrolysis and lipoprotein release by the LolCDE complex, we conducted site-specific mutagenesis of all 32 residues constituting the LolD motif and isolated dominant-negative mutants. We then selected suppressor mutants of LolC and LolE that correct the growth defects caused by the LolD mutations. The location of the LolC and LolE suppressor mutations strongly suggests that the two proteins play different roles in the lipoprotein release reaction.
MATERIALS AND METHODS
Bacterial strains, plasmids, and media. Escherichia coli K-12 strains MC4100 (1), JM109 (31), JM83 (31), and DLP79-36 (23) were used. The plasmids used were pKM202 carrying lolDH (encoding hexahistidine-tagged LolD) under Ptac, pKM402 carrying lolC and lolDH under PBAD, and pKM301 carrying lolE under Ptac (9). To construct pKM202, PCR was performed with pKM201 (9) as a template and a pair of oligonucleotides, 5'-ACGATGAGCTCGAAGGAGATATAAATATGAATAAGATCCTGTTGCAATGC-3' and 5'-CACTAAGCTTAATGATGATGATGATGATGTTCTAACTCCGCCCCCATCAGGCTCAG TTCCGCC-3'. The amplified DNA was digested with SacI and HindIII and then cloned into the same site of pTTQ18 (Amersham Biosciences). E. coli strains were grown at 30°C on LB medium (12). When required, ampicillin and chloramphenicol were added at concentrations of 50 and 25 μg/ml, respectively.
Random mutagenesis of the 32 residues in the LolD motif. All of the amino acid residues constituting the LolD motif were mutagenized by PCR using a QuikChange site-directed mutagenesis kit (Stratagene), with pKM202 as a template and a pair of primers (see Table S1 in the supplemental material). The plasmids were amplified in E. coli XL1-Blue (Stratagene) and then introduced into MC4100 cells. Transformants were replicated on LB agar supplemented with 1 mM isopropyl--D-thiogalactopyranoside (IPTG), and clones that exhibited a growth defect on IPTG-containing medium were selected after overnight incubation at 30°C. Introduction of the LolD motif mutations into pKM402 was also performed with a QuikChange site-directed mutagenesis kit and the primers listed in Table S2 in the supplemental material.
Isolation of suppressor mutants. To isolate suppressor mutants of LolE, pKM301 was randomly mutagenized in vivo using E. coli mutator strain XL-1 Red (Stratagene) according to the manufacturer's instructions. The mutagenized plasmids were introduced into JM109 carrying pKM402 derivatives. To isolate suppressor mutants of LolC, derivatives of pKM402 with a single-amino-acid substitution in the LolD motif were mutagenized in XL-1 Red and then introduced into JM109 carrying pKM301. After overnight incubation at 30°C, colonies that grew on LB agar supplemented with 0.2% arabinose and 40 μM IPTG were selected, and then the pKM402 derivatives carried by these colonies were isolated and then reintroduced into JM109 harboring pKM301 to confirm the suppression of the defective LolD mutations. Prior to the isolation of suppressor mutants, the concentrations of IPTG and arabinose in LB agar plates were carefully determined. It was found that the JM109 cells harboring wild-type pKM402 (PBAD-lolC-lolD) and wild-type pKM301 (Ptac-lolE) grew on LB agar containing 0.2% arabinose and 40 μM IPTG, whereas those harboring a pKM402 derivative carrying defective lolD and wild-type pKM301 did not.
Sequence analysis. Mutant plasmids were sequenced with a dye terminator cycle sequencing kit and a CEQ8000 multicapillary DNA sequencing system (Beckman Coulter).
ATPase activity of purified LolD and LolCDE. His-tagged LolD was overproduced in DLP79-36 cells harboring a pKM202 derivative. After disruption of the cells by French pressure cell, LolD was purified from cytoplasmic fractions on a metal affinity column packed with TALON (CLONTECH) resin equilibrated with 50 mM Tris-HCl (pH 7.5) containing 100 mM NaCl and 10% glycerol. After adsorption of LolD, the resin was washed with the same buffer supplemented with 10 mM imidazole. LolD was eluted with the buffer supplemented with 250 mM imidazole. LolD fractions were combined and dialyzed against 50 mM Tris-HCl (pH 7.5) containing 10% glycerol and then further purified by MonoQ column (Amershm Biosciences). LolD was recovered in a pass-through fraction.
The LolCDE complex containing His-tagged LolD was overproduced in JM83 cells harboring a pKM402 derivative and pKM301. Total membrane fractions were prepared after disruption of cells by French pressure cell and solubilized on ice for 30 min with 50 mM Tris-HCl (pH 7.5), 5 mM MgCl2, 1% n-dodecyl--D-maltopyranoside (DDM), and 10% glycerol containing or not containing 2 mM ATP, followed by centrifugation at 100,000 x g for 30 min. LolCDE in the supernatant was adsorbed to TALON resin equilibrated with the same buffer supplemented with 10 mM imidazole and then developed by a linear gradient of imidazole (10 to 250 mM). LolCDE was eluted at 60 mM imidazole and then dialyzed against 50 mM Tris-HCl (pH 7.5), 0.01% DDM, and 10% glycerol with or without 2 mM ATP.
To reconstitute LolCDE into proteoliposomes, E. coli phospholipids (0.8 mg) and LolCDE (8 μg) were incubated on ice for 10 min in 100 μl of 50 mM Tris-HCl (pH 7.5), containing 5 mM MgSO4, 100 mM NaCl, and 1.2% sucrose monocaprate. The mixture was diluted with 9 volumes of 50 mM Tris-HCl (pH 7.5) containing 5 mM MgSO4 and 100 mM NaCl and then dialyzed overnight against the same buffer. Reconstituted proteoliposomes were recovered by centrifugation at 100,000 x g for 2 h and then resuspended in 100 μl of the dialysis buffer as reported (9).
ATPase activities of LolD (5 μg) and LolCDE (90 μl proteoliposomes) were determined in 50 mM Tris-HCl (pH 7.5) containing 5 mM MgSO4, 100 mM NaCl, and 2 mM ATP according to the method described in reference 2. The reaction mixture containing LolD was supplemented with 10% glycerol.
SDS-PAGE and immunoblotting. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting were carried out as described previously (29). Antibodies specific for LolC, LolD, and LolE were raised in rabbits against the synthetic peptides as previously reported (14).
RESULTS
Random mutagenesis of the 32 residues in the LolD motif. The LolD motif located between the Walker A and ABC signature 1 motifs (Fig. 1A) is highly conserved among the homologs of gram-negative bacteria (Fig. 1B) but not among other E. coli ABC proteins. To determine the function of this region, each residue constituting this motif was subjected to random mutagenesis. Plasmids carrying the mutagenized lolD gene under Ptac were constructed by PCR and then introduced into E. coli MC4100. Transformants were replicated onto LB agar supplemented with or without 1 mM IPTG, and those unable to grow on IPTG-containing media were selected. For 12 of the 32 residues, a total of 26 mutants were obtained (Fig. 1A), which inhibited the growth of MC4100 cells upon induction with IPTG even though the cells carried chromosomal lolCDE genes. No such mutant was obtained for the other 20 residues, although their mutagenesis was successful since the plasmids isolated from five independent colonies obtained in each mutagenesis experiment carried different mutations.
Effects of LolD derivatives on the growth of E. coli on liquid media. MC4100 cells harboring a plasmid encoding one of the 26 LolD mutants were grown on LB liquid medium supplemented with or without 1 mM IPTG (Fig. 2A). The levels of the respective LolD mutants were not significantly different from the level of wild-type LolD expressed from the plasmid (Fig. 2B). When the plasmid encoded wild-type LolD, IPTG had no effect on the growth of cells. In contrast, induction of LolD mutants with IPTG inhibited growth to various degrees (closed symbols for R85 to P111 in Fig. 2A). Among them, D101N and D101R caused the strongest inhibition of growth. It has been revealed that mislocalization of the major outer membrane lipoprotein Lpp causes immediate growth arrest since a covalent linkage between Lpp in the inner membrane and peptidoglycan is lethal to cells (26). However, the two D101 mutants caused severe growth arrest even when the cells did not express Lpp (data not shown). It is not clear at present why the two D101 mutants strongly inhibited growth in an Lpp-independent manner. In contrast, the growth inhibition by the other mutants was dependent on Lpp (data not shown), suggesting that the expression of these LolD mutants causes mislocalization of Lpp.
LolE mutations that suppress defective LolD mutations. We next examined whether or not mutations of the lolE gene correct the growth defects caused by LolD mutants. In the case of more than one species of mutant being obtained for a certain target residue, we selected a mutant that was most effectively overproduced and purified as a soluble protein for future biochemical analyses. Two mutants were selected as H97 mutants because they were especially easy to purify. Thus, the 13 LolD mutants shown in Table 1 were selected. These lolD mutations were introduced into pKM402 carrying the lolC and lolD genes under PBAD. JM109 cells harboring pKM402 and pKM301 carrying the lolE gene under Ptac grew normally on LB agar supplemented with arabinose and IPTG. On the other hand, cells harboring a pKM402 derivative carrying any one of the 13 lolD mutations and pKM301 did not grow, indicating that the LolD mutants remain dominant negative even when LolC and LolE are overexpressed (see wild-type and NS sectors in Fig. 3A and 4A, respectively). Plasmid pKM301 was next randomly mutagenized in E. coli mutator strain XL1-Red and then transformed into JM109 harboring a pKM402 derivative. For each LolD mutant, 104 to 105 transformants were plated on LB agar containing 0.2% arabinose and 40 mM IPTG, followed by incubation overnight at 30°C. Depending on the species of the LolD mutation, from zero to some dozens of transformants appeared on the LB agar plates. Plasmids were prepared from these colonies and then retransformed into JM109 harboring a pKM402 derivative. These steps were repeated twice, with 27 transformants finally obtained. Nucleotide sequencing of the plasmids prepared from these 27 clones revealed that 2 had no mutation in the lolE gene. These clones were not investigated further. The other 25 plasmids carried single-amino-acid substitutions in lolE, but some carried the same mutation to the same lolD mutation. Thus, a total of 17 lolE mutations were obtained as suppressor mutations for the 13 LolD mutants (Table 1). Suppression of the respective LolD mutations by these 17 pKM301 derivatives is shown in Fig. 3A. Lysates were prepared from the cells shown in Fig. 3A and then subjected to immunoblotting with antibodies against LolC, LolD, and LolE (Fig. 3B). Some clones expressed LolE at a considerably reduced level, whereas the levels of LolC and LolD remained normal. Among the 17 suppressor mutations, some mutations were obtained as suppressors for different lolD mutations, thereby further reducing the number of different lolE suppressors to 13, most of which were located in the cytoplasmic loop and the second transmembrane segment (Table 1 and Fig. 5B). LolE suppressors 7 and 13 carried the same mutation (Table 1) and appeared to be expressed at reduced levels (Fig. 3B). It seems to be possible that the mutation reduced the reactivity to anti-LolE antibodies.
Allele specificity of suppression. Since different plasmids carried lolD and lolE genes, we examined whether or not a certain lolE suppressor mutant isolated for a given lolD mutation suppresses other lolD mutations. To examine this, 13 lolD mutations carried on pKM402 derivatives and 13 lolE suppressor mutations carried on pKM301 derivatives were variously combined and transformed into JM109 cells. The growth of transformants was then examined on LB agar containing 0.2% arabinose and 40 mM IPTG overnight at 30°C. All lolE mutants suppressed multiple lolD alleles (Table 2), although suppression by the lolE G281D mutant was limited to mutations in the C-terminal region of the LolD motif. Growth inhibition by the A104M mutation of lolD was severe (Fig. 2A), but most lolE mutations suppressed this lolD mutation (Table 2). Mutations in the N-terminal half of the LolD motif were suppressed by fewer species of lolE mutants than mutations in the C-terminal half. The H96P mutation of LolD was suppressed only by the originally isolated lolE (L305S) mutant. In contrast, this L305S lolE mutant suppressed a number of lolD mutations. Importantly, no suppressor was isolated for the Q94L mutation of lolD. This Q is a highly conserved residue in ABC transporters and has an important function (18). The region containing this Q is called the Q-loop (8). Substitution of this Q may be fatal for the LolD function, and therefore no suppressor was isolated.
Isolation of lolC mutations that suppress defective lolD mutations. We next attempted to isolate lolC suppressor mutants using pKM402 derivatives constructed for the isolation of lolE suppressor mutants. Derivatives of pKM402 carrying the wild-type lolC and the dominant-negative lolD alleles were subjected to random mutagenesis in E. coli XL1-Red and then introduced into JM109 harboring wild-type pKM301 (lolE).
For each LolD mutant, 104 to 105 transformants were plated on LB agar containing 0.2% arabinose and 40 μM IPTG and then incubated overnight at 30°C. As mentioned for the isolation of lolE suppressors, plasmids were prepared from colonies grown on LB agar plates and then retransformed into JM109 harboring pKM301. These steps were repeated twice, and 20 transformants were finally obtained as candidates carrying LolC suppressors. Immunoblotting of whole-cell extracts revealed that three transformants expressed neither LolC nor LolD. These three transformants were therefore excluded. Nucleotide sequencing of the plasmids prepared from the other 17 transformants revealed that 13 had a single mutation in lolC without an additional mutation in lolD (Table 1), whereas 4 were excluded because of no additional mutation. Suppression of the respective lolD mutations by the 13 lolC mutations is shown in Fig. 4A. Whole-cell extracts of the 13 transformants shown in Fig. 4A were subjected to immunoblotting with antibodies against LolC, LolD, and LolE (Fig. 4B). The three Lol proteins were expressed in all of the cells examined, although their levels slightly varied depending on the mutations.
Some lolC mutants were obtained as suppressors for different lolD mutations: i.e., the R182H mutant of lolC was isolated as a suppressor of the R85Y and H97N mutations of lolD and the L335P mutant of lolC was isolated as a suppressor of the H97N and A104M mutations of lolD. Thus, 11 different lolC mutants were isolated as suppressors for eight lolD mutations. Surprisingly, all lolC suppressor mutations were located in either the periplasmic loop or the periplasmic side of transmembrane segments (Fig. 5A and Table 1). These mutation sites of LolC were distantly located from the LolD motif. Allosteric effects of mutations most likely account for the lolC suppressors. In any event, the distinct distributions of mutation sites in lolC and lolE suppressors imply that the mode of interaction with LolD differs between LolC and LolE, despite their homologous primary and secondary structures. Since one plasmid carried the suppressor lolC gene and a mutant lolD gene, examination of the allele specificity of suppression was not immediately possible.
LolD motif mutations impair the ATPase activity of LolCDE. To clarify the function of the LolD motif, effects of the LolD mutation on the ATPase activity were examined. For this, six LolD mutants including two D101 mutants were overproduced from pKM202 derivatives in DLP79-36 cells, which do not express Lpp (23). His-tagged LolD mutants were purified from cytoplasmic fractions. LolCDE complexes containing the respective LolD derivatives were overproduced from pKM402 derivative in JM83, which also harbored pKM301. Mutant and wild-type LolCDE complexes were similarly overproduced in this strain and purified from membranes after solubilization with 1% DDM. The levels of LolC and LolE copurified with the respective His-tagged LolD mutant were not significantly different from those copurified with wild-type LolD when ATP was present during the purification procedure (Fig. 6A). However, the ATPase activities of LolD derivatives varied depending on mutation positions: i.e., three mutations in the N-terminal regions of the LolD motif had no effect while the C-terminal three mutations severely impaired the activity. In marked contrast, ATPase activities of mutant LolCDE complexes were significantly retarded irrespective of mutation positions (Fig. 6A).
It has been reported that ATP stabilizes LolCDE after its solubilization with a detergent such as sucrose monocaprate (27). When DDM was used to solubilize membranes, the wild-type LolCDE complex was stably isolated even in the absence of ATP (Fig. 6B). On the other hand, two LolD mutants, D101N and D101R, were obtained without LolC and LolE in the absence of ATP. These properties of the two D101 mutants may be related to the severe growth arrest even when cells do not express Lpp. Taken together, these results indicate that the LolD motif is indeed involved in the functional communication between LolD and membrane subunits. Mutations in this region therefore generally cause retardation of ATPase activities of LolCDE complex or reduction in stability.
DISCUSSION
The LolCDE complex releases lipoproteins anchored to the outer leaflet of the inner membrane (27). Therefore, recognition of lipoproteins by LolC and/or LolE occurs on the periplasmic side of the lipid bilayer. In contrast, the ATP hydrolysis required to drive the release reaction takes place on the cytoplasmic side of the membrane and is due to LolD. To understand the mechanism underlying the lipoprotein release reaction, the mode of communication between LolC/LolE and LolD needs to be clarified. Moreover, it remains to be clarified whether or not LolC and LolE play different roles in the lipoprotein release reaction. Since the LolD motif is specifically conserved in LolD homologs, its relationship to the specific function of LolCDE has been speculated (27). Recent X-ray crystallographic analysis of the ABC transporter BtuCD complex (8) revealed that the region corresponding to the LolD motif is located at the site of contact with the integral membrane subunit, suggesting that this region is the site for intersubunit communication.
Random mutagenesis at the respective residues of the LolD motif yielded 26 dominant-negative mutants. While the C-terminal half of the LolD motif was more tolerant of amino acid substitution, the N-terminal half yielded various dominant-negative mutants (Fig. 1). The dominant-negative property of the LolD motif mutants, except two D101 mutants, disappeared when the cells did not express Lpp. Lpp is the most abundant outer membrane lipoprotein in E. coli, and its accumulation in the inner membrane causes immediate growth arrest due to the formation of a covalent linkage between the C-terminal Lys of Lpp and peptidoglycan (26). These results therefore suggest that LolD mutants cause mislocalization of Lpp in the inner membrane. It is not clear at present why the two D101 mutants inhibited growth even in the absence of Lpp. The dominant-negative property of LolD mutants indicates that these mutants compete with wild-type LolD expressed from the chromosome for the formation of a complex with wild-type LolC and LolE. Thus, the functional LolCDE complex is absent under these conditions. Importantly, all LolD mutants retained the dominant-negative property when wild-type LolC and LolE were overproduced from plasmids (Fig. 3 and 4), indicating that the defective LolCDE complex present in a large amount inhibits the wild-type LolCDE complex present in a small amount. How do these defective LolCDE complexes exhibit the dominant-negative phenotype One possibility is that the defective complex occupies LolA, thereby inhibiting the transfer of lipoproteins from the wild-type LolCDE complex to LolA. Another possibility is that the defective LolCDE complex tightly binds lipoproteins and thus never releases them. In either case, Lpp remains in the inner membrane and causes growth arrest, as shown previously for LolD mutants having mutations in the Walker A, B, or ABC signature motif (9, 27).
We obtained a number of LolC and LolE mutants that suppress dominant-negative LolD mutations, indicating that the LolD motif region is involved in the intersubunit interaction and therefore affected by LolC/LolE mutations. On the other hand, an LolC or LolE mutant that suppresses the Q94L mutation of LolD was not isolated. This Gln residue is highly conserved in the ATPase domains of ABC transporters. Furthermore, the crystal structure of GlcV, an ABC transporter for glucose uptake by Sulfolobus solfataricus (25), revealed that this Gln residue interacts with Mg2+. Q94 of LolD may be essential for the activity because this residue is directly involved in ATP hydrolysis.
The amino acid sequences and membrane topologies of LolC and LolE are similar to each other, whereas their suppressor mutations were located in different domains. LolC suppressors had mutations in the region exposed to the periplasm or on the periplasmic side of transmembrane segments (Fig. 5A). On the other hand, most mutations of the LolE suppressors were located in the cytoplasmic loop or the middle of a transmembrane segment (Fig. 5B). These results strongly suggest that these two membrane subunits interact differently with LolD and presumably play distinct roles. It has been suggested that MalF and MalG, integral membrane subunits of maltose transporter MalFGK2, play different roles because identical amino acid substitutions at the corresponding positions resulted in different phenotypes (13). Suppression by most LolC mutants is probably indirect because their mutations are located on the periplasmic loop or the periplasmic surface. Allosteric effects are most likely responsible for the LolC suppressors. On the other hand, LolE suppressors may have mutations in the contact site for LolD. We assume that most LolC and LolE suppressors isolated here correct defective LolD mutations and thus allow the formation of the functional LolCDE complex, although some suppressors might merely neutralize toxic LolD mutations by allowing the formation of a nonfunctional complex.
Taking advantage of the fact that different plasmids encoded LolD and LolE, the specificity of suppression was examined with various combinations of LolD and LolE mutants (Table 1). None of the LolE mutants suppressed the Q94L mutation of LolD. Only one LolE mutant, L305S, was a suppressor of the H96P mutation of LolD. These results suggest that the center of the LolD motif is critically important for the function. It is likely that H96 of LolD closely interacts with the second cytoplasmic loop of LolE containing Leu at position 305. On the other hand, other LolD mutations including the very strong mutation D101R were suppressed by multiple LolE mutants. Furthermore, mutations near the center of the LolD motif, Y93H and H97N, were also suppressed by multiple LolE mutations, highlighting the very distinct role of Q94.
The ATPase activities of the LolD mutants and the LolCDE complexes containing the LolD mutant revealed that the LolD motif region is involved in the communication between LolD and LolC/E. The three mutations Y93H, H96P, and H97N had no effect on the ATPase activity of the LolD subunit, whereas these mutations significantly lowered the ATPase activity of the LolCDE complex (Fig. 6), suggesting that interaction with membrane subunits became abnormal due to the mutations and therefore inhibited ATP hydrolysis. The two mutants D101N and D101R were defective in the interaction with membrane subunits and required ATP for copurification of LolC/E.
Here we isolated a number of LolC and LolE mutants that suppress various mutants of LolD. These mutants will be very powerful tools for clarifying the mechanisms underlying the lipoprotein release reaction, which involves transmembrane transfer of energy obtained through ATP hydrolysis.
ACKNOWLEDGMENTS
We wish to thank Rika Ishihara for technical support.
Y.I. and H.M. contributed equally to this study.
REFERENCES
Casadaban, M. J. 1976. Transposition and fusion of the lac genes to selected promoters in Escherichia coli using bacteriophage lambda and Mu. J. Mol. Biol. 104:541-555.
Chifflet, S., A. Torriglia, R. Chiesa, and S. Tolosa. 1988. A method for the determination of inorganic phosphate in the presence of labile organic phosphate and high concentration of protein: application to lens ATPases. Anal. Biochem. 168:1-4.
Davidson, A. L., and J. Chen. 2004. ATP-binding cassette transporters in bacteria. Annu. Rev. Biochem. 73:241-268.
Hara, T., S. Matsuyama, and H. Tokuda. 2003. Mechanism underlying the inner membrane retention of E. coli lipoproteins caused by Lol avoidance signals. J. Biol. Chem. 278:40408-40414.
Higgins, C. F., and K. J. Linton. 2004. The ATP switch model for ABC transporters. Nat. Struct. Mol. Biol. 11:918-926.
Holland, I. B., and M. A. Blight. 1999. ABC-ATPases, adaptable energy generators fuelling transmembrane movement of a variety of molecules in organisms from bacteria to humans. J. Mol. Biol. 293:381-399.
Hunke, S., M. Mourez, M. Jehanno, E. Dassa, and E. Schneider. 2000. ATP modulates subunit-subunit interactions in an ATP-binding cassette transporter (MalFGK2) determined by site-directed chemical cross-linking. J. Biol. Chem. 275:15526-15534.
Locher, K. P., A. T. Lee, and D. C. Rees. 2002. The E. coli BtuCD structure: a framework for ABC transporter architecture and mechanism. Science 296:1091-1098.
Masuda, K., S. Matsuyama, and H. Tokuda. 2002. Elucidation of the function of lipoprotein-sorting signals that determine membrane localization. Proc. Natl. Acad. Sci. USA 99:7390-7395.
Matsuyama, S., T. Tajima, and H. Tokuda. 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.
Matsuyama, S., N. Yokota, and H. Tokuda. 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.
Miller, J. H. 1992. A short course in bacterial genetics: a laboratory manual and handbook for Escherichia coli and related bacteria. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Mourez, M., M. Hofnung, and E. Dassa. 1997. Subunit interactions in ABC transporters: a conserved sequence in hydrophobic membrane proteins of periplasmic permeases defines an important site of interaction with the ATPase subunits. EMBO J. 16:3066-3077.
Narita, S., K. Kanamaru, S. Matsuyama, and H. Tokuda. 2003. A mutation in the membrane subunit of an ABC transporter LolCDE complex causing outer membrane localization of lipoproteins against their inner membrane-specific signals. Mol. Microbiol. 49:167-177.
Narita, S., K. Tanaka, S. Matsuyama, and H. Tokuda. 2002. Disruption of lolCDE, encoding an ATP-binding cassette transporter, is lethal for Escherichia coli and prevents release of lipoproteins from the inner membrane. J. Bacteriol. 184:1417-1422.
Pugsley, A. P. 1993. The complete general secretory pathway in gram-negative bacteria. Microbiol. Rev. 57:50-108.
Sankaran, K., and H. C. Wu. 1994. Lipid modification of bacterial prolipoprotein. J. Biol. Chem. 269:9701-19706.
Schneider, E., and S. Hunke. 1998. ATP-binding-cassette (ABC) transport systems: functional and structural aspects of the ATP-hydrolyzing subunits/domains. FEMS Microbiol. Rev. 22:1-20.
Seydel, A., P. Gounon, and A. P. Pugsley. 1999. Testing the ‘+2 rule’ for lipoprotein sorting in the Escherichia coli cell envelope with a new genetic selection. Mol. Microbiol. 34:810-821.
Smith, P. C., N. Karpowich, L. Millen, J. E. Moody, J. Rosen, P. J. Thomas, and J. F. Hunt. 2002. ATP binding to the motor domain from an ABC transporter drives formation of a nucleotide sandwich dimer. Mol. Cell 10:139-149.
Tajima, T., N. Yokota, S. Matsuyama, and H. Tokuda. 1998. Genetic analyses of the in vivo function of LolA, a periplasmic chaperone involved in the outer membrane localization of Escherichia coli lipoproteins. FEBS Lett. 439:51-54.
Tanaka, K., S. Matsuyama, and H. Tokuda. 2001. Deletion of lolB, encoding an outer membrane lipoprotein, is lethal for Escherichia coli and causes accumulation of lipoprotein localization intermediates in the periplasm. J. Bacteriol. 183:6538-6542.
Terada, M., T. Kuroda, S. Matsuyama, and H. Tokuda. 2001. Lipoprotein-sorting signals evaluated as the LolA-dependent release of lipoproteins from the cytoplasmic membrane of Escherichia coli. J. Biol. Chem. 276:47690-47694.
Tokuda, H., and S. Matsuyama. 2004. Sorting of lipoproteins to the outer membrane in E. coli. Biochim. Biophys. Acta 1693:5-13.
Verdon, G., S. V. Albers, B. W. Dijkstra, A. J. M. Driessen, and A.-M. W. H. Thunnissen. 2003. Crystal structures of the ATPase subunit of the glucose ABC transporter from Sulfolobus solfataricus: nucleotide-free and nucleotide-bound conformations. J. Mol. Biol. 330:343-358.
Yakushi, T., T. Tajima, S. Matsuyama, and H. Tokuda. 1997. Lethality of the covalent linkage between mislocalized major outer membrane lipoprotein and the peptidoglycan of Escherichia coli. J. Bacteriol. 179:2857-2862.
Yakushi, T., K. Masuda, S. Narita, S. Matsuyama, and H. Tokuda. 2000. A new ABC transporter mediating the detachment of lipid-modified proteins from membranes. Nat. Cell Biol. 2:212-218.
Yakushi, T., N. Yokota, S. Matsuyama, and H. Tokuda. 1998. LolA-dependent release of a lipid-modified protein from the inner membrane of Escherichia coli requires nucleoside triphosphate. J. Biol. Chem. 273:32576-32581.
Yamada, H., S. Matsuyama, H. Tokuda, and S. Mizushima. 1989. A high concentration of SecA allows proton motive force-independent translocation of a model secretory protein into Escherichia coli membrane vesicles. J. Biol. Chem. 264:18577-18581.
Yamaguchi, K., F. Yu, and M. Inouye. 1988. A single amino acid determinant of the membrane localization of lipoproteins in E. coli. Cell 53:423-432.
Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103-119.
Yuan, Y. R., S. Blecker, O. Martsinkevich, L. Millen, P. J. Thomas, and J. F. Hunt. 2001. The crystal structure of the MJ0796 ATP-binding cassette. Implications for the structural consequences of ATP hydrolysis in the active site of an ABC transporter. J. Biol. Chem. 276:32313-32321.(Yasuko Ito, Hitomi Matsuz)
ABSTRACT
The LolCDE complex, an ATP-binding cassette (ABC) transporter, releases lipoproteins from the inner membrane, thereby initiating lipoprotein sorting to the outer membrane of Escherichia coli. The LolCDE complex is composed of two copies of an ATPase subunit, LolD, and one copy each of integral membrane subunits LolC and LolE. LolD hydrolyzes ATP on the cytoplasmic side of the inner membrane, while LolC and/or LolE recognize and release lipoproteins anchored to the periplasmic leaflet of the inner membrane. Thus, functional interaction between LolD and LolC/E is critically important for coupling of ATP hydrolysis to the lipoprotein release reaction. LolD contains a characteristic sequence called the LolD motif, which is highly conserved among LolD homologs but not other ABC transporters of E. coli. The LolD motif is suggested to be a region in contact with LolC/E, judging from the crystal structures of other ABC transporters. To determine the functions of the LolD motif, we mutagenized each of the 32 residues of the LolD motif and isolated 26 dominant-negative mutants, whose overexpression arrested growth despite the chromosomal lolD+ background. We then selected suppressor mutations of the lolC and lolE genes that correct the growth defect caused by the LolD mutations. Mutations of the lolC suppressors were mainly located in the periplasmic loop, whereas ones of lolE suppressors were mainly located in the cytoplasmic loop, suggesting that the mode of interaction with LolD differs between LolC and LolE. Moreover, the LolD motif was found to be critical for functional interplay with LolC/E, since some LolD mutations lowered the ATPase activity of LolCDE without affecting that of LolD.
INTRODUCTION
Bacterial lipoproteins possess an N-terminal Cys, which is modified by fatty acids (17). In gram-negative bacteria, lipoproteins are anchored to either the inner or outer membrane via N-terminal fatty acids (16). Lipoproteins are synthesized as precursors in the cytoplasm and then translocated across the inner membrane by Sec machinery, followed by modification on the periplasmic side of the inner membrane into mature lipoproteins (16). Lipoproteins destined for the outer membrane are further transported across the periplasmic space by the Lol system (24), which is composed of five proteins (LolABCDE). The LolCDE complex in the inner membrane releases outer membrane-directed lipoproteins from the inner membrane in an ATP-dependent manner (15, 27, 28), leading to the formation of a water-soluble complex comprising one molecule each of lipoprotein and LolA in the periplasm (10, 21). The LolA-lipoprotein complex then interacts with outer membrane receptor LolB, which catalyzes the anchoring of lipoproteins to the outer membrane (11, 22). In contrast, inner membrane lipoproteins with Asp at position 2 avoid recognition by the LolCDE complex and therefore remain in the inner membrane (4, 9, 19, 23, 30).
ATP-binding cassette (ABC) transporters generally have at least 10 transmembrane stretches (6), whereas both LolC and LolE span the membrane only four times and have a large domain exposed to the periplasm between the first and second transmembrane segments (27). In addition, the sequences of LolC and LolE are similar to each other, with 26% identity. The LolCDE complex differs mechanistically from all other ABC transporters in that it does not catalyze the trans-bilayer movement of substrates but releases lipoproteins from one leaflet of a lipid bilayer (15, 27). This might be why the LolCDE complex has a total of only eight transmembrane segments. The LolCDE complex is considered to be an interesting ABC exporter variant (3).
LolD contains the consensus sequences of ABC proteins, i.e., Walker A, Walker B, and ABC signature motifs (27). The crystal structure of MJ0796, a methanococcal LolD homolog exhibiting 43.7% sequence identity, showed a very similar tertiary fold to those of the ATPase subunits of other ABC transporters (20, 32). Therefore, it is plausible that the ATP-hydrolyzing mechanism of LolD is essentially the same as those of other ABC proteins (3, 5). In addition to the motifs characteristic of ABC proteins, LolD contains a sequence called the LolD motif, which is not conserved among other ABC proteins of E. coli but is highly conserved among LolD homologs. The LolD motif is located between the Walker A and ABC signature (LSGGQ) motifs. Judging from the crystal structure of vitamin B12 transporter BtuCD (8), the LolD motif is speculated to be very close to the cytoplasmic loops of LolC and LolE. Moreover, genetic and biochemical evidence indicates that the region corresponding to the LolD motif is involved in the communication between the integral membrane subunits and the ATPase subunits of maltose transporter MalFGK2 (7, 13).
To determine the mechanism underlying the coupling of ATP hydrolysis and lipoprotein release by the LolCDE complex, we conducted site-specific mutagenesis of all 32 residues constituting the LolD motif and isolated dominant-negative mutants. We then selected suppressor mutants of LolC and LolE that correct the growth defects caused by the LolD mutations. The location of the LolC and LolE suppressor mutations strongly suggests that the two proteins play different roles in the lipoprotein release reaction.
MATERIALS AND METHODS
Bacterial strains, plasmids, and media. Escherichia coli K-12 strains MC4100 (1), JM109 (31), JM83 (31), and DLP79-36 (23) were used. The plasmids used were pKM202 carrying lolDH (encoding hexahistidine-tagged LolD) under Ptac, pKM402 carrying lolC and lolDH under PBAD, and pKM301 carrying lolE under Ptac (9). To construct pKM202, PCR was performed with pKM201 (9) as a template and a pair of oligonucleotides, 5'-ACGATGAGCTCGAAGGAGATATAAATATGAATAAGATCCTGTTGCAATGC-3' and 5'-CACTAAGCTTAATGATGATGATGATGATGTTCTAACTCCGCCCCCATCAGGCTCAG TTCCGCC-3'. The amplified DNA was digested with SacI and HindIII and then cloned into the same site of pTTQ18 (Amersham Biosciences). E. coli strains were grown at 30°C on LB medium (12). When required, ampicillin and chloramphenicol were added at concentrations of 50 and 25 μg/ml, respectively.
Random mutagenesis of the 32 residues in the LolD motif. All of the amino acid residues constituting the LolD motif were mutagenized by PCR using a QuikChange site-directed mutagenesis kit (Stratagene), with pKM202 as a template and a pair of primers (see Table S1 in the supplemental material). The plasmids were amplified in E. coli XL1-Blue (Stratagene) and then introduced into MC4100 cells. Transformants were replicated on LB agar supplemented with 1 mM isopropyl--D-thiogalactopyranoside (IPTG), and clones that exhibited a growth defect on IPTG-containing medium were selected after overnight incubation at 30°C. Introduction of the LolD motif mutations into pKM402 was also performed with a QuikChange site-directed mutagenesis kit and the primers listed in Table S2 in the supplemental material.
Isolation of suppressor mutants. To isolate suppressor mutants of LolE, pKM301 was randomly mutagenized in vivo using E. coli mutator strain XL-1 Red (Stratagene) according to the manufacturer's instructions. The mutagenized plasmids were introduced into JM109 carrying pKM402 derivatives. To isolate suppressor mutants of LolC, derivatives of pKM402 with a single-amino-acid substitution in the LolD motif were mutagenized in XL-1 Red and then introduced into JM109 carrying pKM301. After overnight incubation at 30°C, colonies that grew on LB agar supplemented with 0.2% arabinose and 40 μM IPTG were selected, and then the pKM402 derivatives carried by these colonies were isolated and then reintroduced into JM109 harboring pKM301 to confirm the suppression of the defective LolD mutations. Prior to the isolation of suppressor mutants, the concentrations of IPTG and arabinose in LB agar plates were carefully determined. It was found that the JM109 cells harboring wild-type pKM402 (PBAD-lolC-lolD) and wild-type pKM301 (Ptac-lolE) grew on LB agar containing 0.2% arabinose and 40 μM IPTG, whereas those harboring a pKM402 derivative carrying defective lolD and wild-type pKM301 did not.
Sequence analysis. Mutant plasmids were sequenced with a dye terminator cycle sequencing kit and a CEQ8000 multicapillary DNA sequencing system (Beckman Coulter).
ATPase activity of purified LolD and LolCDE. His-tagged LolD was overproduced in DLP79-36 cells harboring a pKM202 derivative. After disruption of the cells by French pressure cell, LolD was purified from cytoplasmic fractions on a metal affinity column packed with TALON (CLONTECH) resin equilibrated with 50 mM Tris-HCl (pH 7.5) containing 100 mM NaCl and 10% glycerol. After adsorption of LolD, the resin was washed with the same buffer supplemented with 10 mM imidazole. LolD was eluted with the buffer supplemented with 250 mM imidazole. LolD fractions were combined and dialyzed against 50 mM Tris-HCl (pH 7.5) containing 10% glycerol and then further purified by MonoQ column (Amershm Biosciences). LolD was recovered in a pass-through fraction.
The LolCDE complex containing His-tagged LolD was overproduced in JM83 cells harboring a pKM402 derivative and pKM301. Total membrane fractions were prepared after disruption of cells by French pressure cell and solubilized on ice for 30 min with 50 mM Tris-HCl (pH 7.5), 5 mM MgCl2, 1% n-dodecyl--D-maltopyranoside (DDM), and 10% glycerol containing or not containing 2 mM ATP, followed by centrifugation at 100,000 x g for 30 min. LolCDE in the supernatant was adsorbed to TALON resin equilibrated with the same buffer supplemented with 10 mM imidazole and then developed by a linear gradient of imidazole (10 to 250 mM). LolCDE was eluted at 60 mM imidazole and then dialyzed against 50 mM Tris-HCl (pH 7.5), 0.01% DDM, and 10% glycerol with or without 2 mM ATP.
To reconstitute LolCDE into proteoliposomes, E. coli phospholipids (0.8 mg) and LolCDE (8 μg) were incubated on ice for 10 min in 100 μl of 50 mM Tris-HCl (pH 7.5), containing 5 mM MgSO4, 100 mM NaCl, and 1.2% sucrose monocaprate. The mixture was diluted with 9 volumes of 50 mM Tris-HCl (pH 7.5) containing 5 mM MgSO4 and 100 mM NaCl and then dialyzed overnight against the same buffer. Reconstituted proteoliposomes were recovered by centrifugation at 100,000 x g for 2 h and then resuspended in 100 μl of the dialysis buffer as reported (9).
ATPase activities of LolD (5 μg) and LolCDE (90 μl proteoliposomes) were determined in 50 mM Tris-HCl (pH 7.5) containing 5 mM MgSO4, 100 mM NaCl, and 2 mM ATP according to the method described in reference 2. The reaction mixture containing LolD was supplemented with 10% glycerol.
SDS-PAGE and immunoblotting. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting were carried out as described previously (29). Antibodies specific for LolC, LolD, and LolE were raised in rabbits against the synthetic peptides as previously reported (14).
RESULTS
Random mutagenesis of the 32 residues in the LolD motif. The LolD motif located between the Walker A and ABC signature 1 motifs (Fig. 1A) is highly conserved among the homologs of gram-negative bacteria (Fig. 1B) but not among other E. coli ABC proteins. To determine the function of this region, each residue constituting this motif was subjected to random mutagenesis. Plasmids carrying the mutagenized lolD gene under Ptac were constructed by PCR and then introduced into E. coli MC4100. Transformants were replicated onto LB agar supplemented with or without 1 mM IPTG, and those unable to grow on IPTG-containing media were selected. For 12 of the 32 residues, a total of 26 mutants were obtained (Fig. 1A), which inhibited the growth of MC4100 cells upon induction with IPTG even though the cells carried chromosomal lolCDE genes. No such mutant was obtained for the other 20 residues, although their mutagenesis was successful since the plasmids isolated from five independent colonies obtained in each mutagenesis experiment carried different mutations.
Effects of LolD derivatives on the growth of E. coli on liquid media. MC4100 cells harboring a plasmid encoding one of the 26 LolD mutants were grown on LB liquid medium supplemented with or without 1 mM IPTG (Fig. 2A). The levels of the respective LolD mutants were not significantly different from the level of wild-type LolD expressed from the plasmid (Fig. 2B). When the plasmid encoded wild-type LolD, IPTG had no effect on the growth of cells. In contrast, induction of LolD mutants with IPTG inhibited growth to various degrees (closed symbols for R85 to P111 in Fig. 2A). Among them, D101N and D101R caused the strongest inhibition of growth. It has been revealed that mislocalization of the major outer membrane lipoprotein Lpp causes immediate growth arrest since a covalent linkage between Lpp in the inner membrane and peptidoglycan is lethal to cells (26). However, the two D101 mutants caused severe growth arrest even when the cells did not express Lpp (data not shown). It is not clear at present why the two D101 mutants strongly inhibited growth in an Lpp-independent manner. In contrast, the growth inhibition by the other mutants was dependent on Lpp (data not shown), suggesting that the expression of these LolD mutants causes mislocalization of Lpp.
LolE mutations that suppress defective LolD mutations. We next examined whether or not mutations of the lolE gene correct the growth defects caused by LolD mutants. In the case of more than one species of mutant being obtained for a certain target residue, we selected a mutant that was most effectively overproduced and purified as a soluble protein for future biochemical analyses. Two mutants were selected as H97 mutants because they were especially easy to purify. Thus, the 13 LolD mutants shown in Table 1 were selected. These lolD mutations were introduced into pKM402 carrying the lolC and lolD genes under PBAD. JM109 cells harboring pKM402 and pKM301 carrying the lolE gene under Ptac grew normally on LB agar supplemented with arabinose and IPTG. On the other hand, cells harboring a pKM402 derivative carrying any one of the 13 lolD mutations and pKM301 did not grow, indicating that the LolD mutants remain dominant negative even when LolC and LolE are overexpressed (see wild-type and NS sectors in Fig. 3A and 4A, respectively). Plasmid pKM301 was next randomly mutagenized in E. coli mutator strain XL1-Red and then transformed into JM109 harboring a pKM402 derivative. For each LolD mutant, 104 to 105 transformants were plated on LB agar containing 0.2% arabinose and 40 mM IPTG, followed by incubation overnight at 30°C. Depending on the species of the LolD mutation, from zero to some dozens of transformants appeared on the LB agar plates. Plasmids were prepared from these colonies and then retransformed into JM109 harboring a pKM402 derivative. These steps were repeated twice, with 27 transformants finally obtained. Nucleotide sequencing of the plasmids prepared from these 27 clones revealed that 2 had no mutation in the lolE gene. These clones were not investigated further. The other 25 plasmids carried single-amino-acid substitutions in lolE, but some carried the same mutation to the same lolD mutation. Thus, a total of 17 lolE mutations were obtained as suppressor mutations for the 13 LolD mutants (Table 1). Suppression of the respective LolD mutations by these 17 pKM301 derivatives is shown in Fig. 3A. Lysates were prepared from the cells shown in Fig. 3A and then subjected to immunoblotting with antibodies against LolC, LolD, and LolE (Fig. 3B). Some clones expressed LolE at a considerably reduced level, whereas the levels of LolC and LolD remained normal. Among the 17 suppressor mutations, some mutations were obtained as suppressors for different lolD mutations, thereby further reducing the number of different lolE suppressors to 13, most of which were located in the cytoplasmic loop and the second transmembrane segment (Table 1 and Fig. 5B). LolE suppressors 7 and 13 carried the same mutation (Table 1) and appeared to be expressed at reduced levels (Fig. 3B). It seems to be possible that the mutation reduced the reactivity to anti-LolE antibodies.
Allele specificity of suppression. Since different plasmids carried lolD and lolE genes, we examined whether or not a certain lolE suppressor mutant isolated for a given lolD mutation suppresses other lolD mutations. To examine this, 13 lolD mutations carried on pKM402 derivatives and 13 lolE suppressor mutations carried on pKM301 derivatives were variously combined and transformed into JM109 cells. The growth of transformants was then examined on LB agar containing 0.2% arabinose and 40 mM IPTG overnight at 30°C. All lolE mutants suppressed multiple lolD alleles (Table 2), although suppression by the lolE G281D mutant was limited to mutations in the C-terminal region of the LolD motif. Growth inhibition by the A104M mutation of lolD was severe (Fig. 2A), but most lolE mutations suppressed this lolD mutation (Table 2). Mutations in the N-terminal half of the LolD motif were suppressed by fewer species of lolE mutants than mutations in the C-terminal half. The H96P mutation of LolD was suppressed only by the originally isolated lolE (L305S) mutant. In contrast, this L305S lolE mutant suppressed a number of lolD mutations. Importantly, no suppressor was isolated for the Q94L mutation of lolD. This Q is a highly conserved residue in ABC transporters and has an important function (18). The region containing this Q is called the Q-loop (8). Substitution of this Q may be fatal for the LolD function, and therefore no suppressor was isolated.
Isolation of lolC mutations that suppress defective lolD mutations. We next attempted to isolate lolC suppressor mutants using pKM402 derivatives constructed for the isolation of lolE suppressor mutants. Derivatives of pKM402 carrying the wild-type lolC and the dominant-negative lolD alleles were subjected to random mutagenesis in E. coli XL1-Red and then introduced into JM109 harboring wild-type pKM301 (lolE).
For each LolD mutant, 104 to 105 transformants were plated on LB agar containing 0.2% arabinose and 40 μM IPTG and then incubated overnight at 30°C. As mentioned for the isolation of lolE suppressors, plasmids were prepared from colonies grown on LB agar plates and then retransformed into JM109 harboring pKM301. These steps were repeated twice, and 20 transformants were finally obtained as candidates carrying LolC suppressors. Immunoblotting of whole-cell extracts revealed that three transformants expressed neither LolC nor LolD. These three transformants were therefore excluded. Nucleotide sequencing of the plasmids prepared from the other 17 transformants revealed that 13 had a single mutation in lolC without an additional mutation in lolD (Table 1), whereas 4 were excluded because of no additional mutation. Suppression of the respective lolD mutations by the 13 lolC mutations is shown in Fig. 4A. Whole-cell extracts of the 13 transformants shown in Fig. 4A were subjected to immunoblotting with antibodies against LolC, LolD, and LolE (Fig. 4B). The three Lol proteins were expressed in all of the cells examined, although their levels slightly varied depending on the mutations.
Some lolC mutants were obtained as suppressors for different lolD mutations: i.e., the R182H mutant of lolC was isolated as a suppressor of the R85Y and H97N mutations of lolD and the L335P mutant of lolC was isolated as a suppressor of the H97N and A104M mutations of lolD. Thus, 11 different lolC mutants were isolated as suppressors for eight lolD mutations. Surprisingly, all lolC suppressor mutations were located in either the periplasmic loop or the periplasmic side of transmembrane segments (Fig. 5A and Table 1). These mutation sites of LolC were distantly located from the LolD motif. Allosteric effects of mutations most likely account for the lolC suppressors. In any event, the distinct distributions of mutation sites in lolC and lolE suppressors imply that the mode of interaction with LolD differs between LolC and LolE, despite their homologous primary and secondary structures. Since one plasmid carried the suppressor lolC gene and a mutant lolD gene, examination of the allele specificity of suppression was not immediately possible.
LolD motif mutations impair the ATPase activity of LolCDE. To clarify the function of the LolD motif, effects of the LolD mutation on the ATPase activity were examined. For this, six LolD mutants including two D101 mutants were overproduced from pKM202 derivatives in DLP79-36 cells, which do not express Lpp (23). His-tagged LolD mutants were purified from cytoplasmic fractions. LolCDE complexes containing the respective LolD derivatives were overproduced from pKM402 derivative in JM83, which also harbored pKM301. Mutant and wild-type LolCDE complexes were similarly overproduced in this strain and purified from membranes after solubilization with 1% DDM. The levels of LolC and LolE copurified with the respective His-tagged LolD mutant were not significantly different from those copurified with wild-type LolD when ATP was present during the purification procedure (Fig. 6A). However, the ATPase activities of LolD derivatives varied depending on mutation positions: i.e., three mutations in the N-terminal regions of the LolD motif had no effect while the C-terminal three mutations severely impaired the activity. In marked contrast, ATPase activities of mutant LolCDE complexes were significantly retarded irrespective of mutation positions (Fig. 6A).
It has been reported that ATP stabilizes LolCDE after its solubilization with a detergent such as sucrose monocaprate (27). When DDM was used to solubilize membranes, the wild-type LolCDE complex was stably isolated even in the absence of ATP (Fig. 6B). On the other hand, two LolD mutants, D101N and D101R, were obtained without LolC and LolE in the absence of ATP. These properties of the two D101 mutants may be related to the severe growth arrest even when cells do not express Lpp. Taken together, these results indicate that the LolD motif is indeed involved in the functional communication between LolD and membrane subunits. Mutations in this region therefore generally cause retardation of ATPase activities of LolCDE complex or reduction in stability.
DISCUSSION
The LolCDE complex releases lipoproteins anchored to the outer leaflet of the inner membrane (27). Therefore, recognition of lipoproteins by LolC and/or LolE occurs on the periplasmic side of the lipid bilayer. In contrast, the ATP hydrolysis required to drive the release reaction takes place on the cytoplasmic side of the membrane and is due to LolD. To understand the mechanism underlying the lipoprotein release reaction, the mode of communication between LolC/LolE and LolD needs to be clarified. Moreover, it remains to be clarified whether or not LolC and LolE play different roles in the lipoprotein release reaction. Since the LolD motif is specifically conserved in LolD homologs, its relationship to the specific function of LolCDE has been speculated (27). Recent X-ray crystallographic analysis of the ABC transporter BtuCD complex (8) revealed that the region corresponding to the LolD motif is located at the site of contact with the integral membrane subunit, suggesting that this region is the site for intersubunit communication.
Random mutagenesis at the respective residues of the LolD motif yielded 26 dominant-negative mutants. While the C-terminal half of the LolD motif was more tolerant of amino acid substitution, the N-terminal half yielded various dominant-negative mutants (Fig. 1). The dominant-negative property of the LolD motif mutants, except two D101 mutants, disappeared when the cells did not express Lpp. Lpp is the most abundant outer membrane lipoprotein in E. coli, and its accumulation in the inner membrane causes immediate growth arrest due to the formation of a covalent linkage between the C-terminal Lys of Lpp and peptidoglycan (26). These results therefore suggest that LolD mutants cause mislocalization of Lpp in the inner membrane. It is not clear at present why the two D101 mutants inhibited growth even in the absence of Lpp. The dominant-negative property of LolD mutants indicates that these mutants compete with wild-type LolD expressed from the chromosome for the formation of a complex with wild-type LolC and LolE. Thus, the functional LolCDE complex is absent under these conditions. Importantly, all LolD mutants retained the dominant-negative property when wild-type LolC and LolE were overproduced from plasmids (Fig. 3 and 4), indicating that the defective LolCDE complex present in a large amount inhibits the wild-type LolCDE complex present in a small amount. How do these defective LolCDE complexes exhibit the dominant-negative phenotype One possibility is that the defective complex occupies LolA, thereby inhibiting the transfer of lipoproteins from the wild-type LolCDE complex to LolA. Another possibility is that the defective LolCDE complex tightly binds lipoproteins and thus never releases them. In either case, Lpp remains in the inner membrane and causes growth arrest, as shown previously for LolD mutants having mutations in the Walker A, B, or ABC signature motif (9, 27).
We obtained a number of LolC and LolE mutants that suppress dominant-negative LolD mutations, indicating that the LolD motif region is involved in the intersubunit interaction and therefore affected by LolC/LolE mutations. On the other hand, an LolC or LolE mutant that suppresses the Q94L mutation of LolD was not isolated. This Gln residue is highly conserved in the ATPase domains of ABC transporters. Furthermore, the crystal structure of GlcV, an ABC transporter for glucose uptake by Sulfolobus solfataricus (25), revealed that this Gln residue interacts with Mg2+. Q94 of LolD may be essential for the activity because this residue is directly involved in ATP hydrolysis.
The amino acid sequences and membrane topologies of LolC and LolE are similar to each other, whereas their suppressor mutations were located in different domains. LolC suppressors had mutations in the region exposed to the periplasm or on the periplasmic side of transmembrane segments (Fig. 5A). On the other hand, most mutations of the LolE suppressors were located in the cytoplasmic loop or the middle of a transmembrane segment (Fig. 5B). These results strongly suggest that these two membrane subunits interact differently with LolD and presumably play distinct roles. It has been suggested that MalF and MalG, integral membrane subunits of maltose transporter MalFGK2, play different roles because identical amino acid substitutions at the corresponding positions resulted in different phenotypes (13). Suppression by most LolC mutants is probably indirect because their mutations are located on the periplasmic loop or the periplasmic surface. Allosteric effects are most likely responsible for the LolC suppressors. On the other hand, LolE suppressors may have mutations in the contact site for LolD. We assume that most LolC and LolE suppressors isolated here correct defective LolD mutations and thus allow the formation of the functional LolCDE complex, although some suppressors might merely neutralize toxic LolD mutations by allowing the formation of a nonfunctional complex.
Taking advantage of the fact that different plasmids encoded LolD and LolE, the specificity of suppression was examined with various combinations of LolD and LolE mutants (Table 1). None of the LolE mutants suppressed the Q94L mutation of LolD. Only one LolE mutant, L305S, was a suppressor of the H96P mutation of LolD. These results suggest that the center of the LolD motif is critically important for the function. It is likely that H96 of LolD closely interacts with the second cytoplasmic loop of LolE containing Leu at position 305. On the other hand, other LolD mutations including the very strong mutation D101R were suppressed by multiple LolE mutants. Furthermore, mutations near the center of the LolD motif, Y93H and H97N, were also suppressed by multiple LolE mutations, highlighting the very distinct role of Q94.
The ATPase activities of the LolD mutants and the LolCDE complexes containing the LolD mutant revealed that the LolD motif region is involved in the communication between LolD and LolC/E. The three mutations Y93H, H96P, and H97N had no effect on the ATPase activity of the LolD subunit, whereas these mutations significantly lowered the ATPase activity of the LolCDE complex (Fig. 6), suggesting that interaction with membrane subunits became abnormal due to the mutations and therefore inhibited ATP hydrolysis. The two mutants D101N and D101R were defective in the interaction with membrane subunits and required ATP for copurification of LolC/E.
Here we isolated a number of LolC and LolE mutants that suppress various mutants of LolD. These mutants will be very powerful tools for clarifying the mechanisms underlying the lipoprotein release reaction, which involves transmembrane transfer of energy obtained through ATP hydrolysis.
ACKNOWLEDGMENTS
We wish to thank Rika Ishihara for technical support.
Y.I. and H.M. contributed equally to this study.
REFERENCES
Casadaban, M. J. 1976. Transposition and fusion of the lac genes to selected promoters in Escherichia coli using bacteriophage lambda and Mu. J. Mol. Biol. 104:541-555.
Chifflet, S., A. Torriglia, R. Chiesa, and S. Tolosa. 1988. A method for the determination of inorganic phosphate in the presence of labile organic phosphate and high concentration of protein: application to lens ATPases. Anal. Biochem. 168:1-4.
Davidson, A. L., and J. Chen. 2004. ATP-binding cassette transporters in bacteria. Annu. Rev. Biochem. 73:241-268.
Hara, T., S. Matsuyama, and H. Tokuda. 2003. Mechanism underlying the inner membrane retention of E. coli lipoproteins caused by Lol avoidance signals. J. Biol. Chem. 278:40408-40414.
Higgins, C. F., and K. J. Linton. 2004. The ATP switch model for ABC transporters. Nat. Struct. Mol. Biol. 11:918-926.
Holland, I. B., and M. A. Blight. 1999. ABC-ATPases, adaptable energy generators fuelling transmembrane movement of a variety of molecules in organisms from bacteria to humans. J. Mol. Biol. 293:381-399.
Hunke, S., M. Mourez, M. Jehanno, E. Dassa, and E. Schneider. 2000. ATP modulates subunit-subunit interactions in an ATP-binding cassette transporter (MalFGK2) determined by site-directed chemical cross-linking. J. Biol. Chem. 275:15526-15534.
Locher, K. P., A. T. Lee, and D. C. Rees. 2002. The E. coli BtuCD structure: a framework for ABC transporter architecture and mechanism. Science 296:1091-1098.
Masuda, K., S. Matsuyama, and H. Tokuda. 2002. Elucidation of the function of lipoprotein-sorting signals that determine membrane localization. Proc. Natl. Acad. Sci. USA 99:7390-7395.
Matsuyama, S., T. Tajima, and H. Tokuda. 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.
Matsuyama, S., N. Yokota, and H. Tokuda. 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.
Miller, J. H. 1992. A short course in bacterial genetics: a laboratory manual and handbook for Escherichia coli and related bacteria. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Mourez, M., M. Hofnung, and E. Dassa. 1997. Subunit interactions in ABC transporters: a conserved sequence in hydrophobic membrane proteins of periplasmic permeases defines an important site of interaction with the ATPase subunits. EMBO J. 16:3066-3077.
Narita, S., K. Kanamaru, S. Matsuyama, and H. Tokuda. 2003. A mutation in the membrane subunit of an ABC transporter LolCDE complex causing outer membrane localization of lipoproteins against their inner membrane-specific signals. Mol. Microbiol. 49:167-177.
Narita, S., K. Tanaka, S. Matsuyama, and H. Tokuda. 2002. Disruption of lolCDE, encoding an ATP-binding cassette transporter, is lethal for Escherichia coli and prevents release of lipoproteins from the inner membrane. J. Bacteriol. 184:1417-1422.
Pugsley, A. P. 1993. The complete general secretory pathway in gram-negative bacteria. Microbiol. Rev. 57:50-108.
Sankaran, K., and H. C. Wu. 1994. Lipid modification of bacterial prolipoprotein. J. Biol. Chem. 269:9701-19706.
Schneider, E., and S. Hunke. 1998. ATP-binding-cassette (ABC) transport systems: functional and structural aspects of the ATP-hydrolyzing subunits/domains. FEMS Microbiol. Rev. 22:1-20.
Seydel, A., P. Gounon, and A. P. Pugsley. 1999. Testing the ‘+2 rule’ for lipoprotein sorting in the Escherichia coli cell envelope with a new genetic selection. Mol. Microbiol. 34:810-821.
Smith, P. C., N. Karpowich, L. Millen, J. E. Moody, J. Rosen, P. J. Thomas, and J. F. Hunt. 2002. ATP binding to the motor domain from an ABC transporter drives formation of a nucleotide sandwich dimer. Mol. Cell 10:139-149.
Tajima, T., N. Yokota, S. Matsuyama, and H. Tokuda. 1998. Genetic analyses of the in vivo function of LolA, a periplasmic chaperone involved in the outer membrane localization of Escherichia coli lipoproteins. FEBS Lett. 439:51-54.
Tanaka, K., S. Matsuyama, and H. Tokuda. 2001. Deletion of lolB, encoding an outer membrane lipoprotein, is lethal for Escherichia coli and causes accumulation of lipoprotein localization intermediates in the periplasm. J. Bacteriol. 183:6538-6542.
Terada, M., T. Kuroda, S. Matsuyama, and H. Tokuda. 2001. Lipoprotein-sorting signals evaluated as the LolA-dependent release of lipoproteins from the cytoplasmic membrane of Escherichia coli. J. Biol. Chem. 276:47690-47694.
Tokuda, H., and S. Matsuyama. 2004. Sorting of lipoproteins to the outer membrane in E. coli. Biochim. Biophys. Acta 1693:5-13.
Verdon, G., S. V. Albers, B. W. Dijkstra, A. J. M. Driessen, and A.-M. W. H. Thunnissen. 2003. Crystal structures of the ATPase subunit of the glucose ABC transporter from Sulfolobus solfataricus: nucleotide-free and nucleotide-bound conformations. J. Mol. Biol. 330:343-358.
Yakushi, T., T. Tajima, S. Matsuyama, and H. Tokuda. 1997. Lethality of the covalent linkage between mislocalized major outer membrane lipoprotein and the peptidoglycan of Escherichia coli. J. Bacteriol. 179:2857-2862.
Yakushi, T., K. Masuda, S. Narita, S. Matsuyama, and H. Tokuda. 2000. A new ABC transporter mediating the detachment of lipid-modified proteins from membranes. Nat. Cell Biol. 2:212-218.
Yakushi, T., N. Yokota, S. Matsuyama, and H. Tokuda. 1998. LolA-dependent release of a lipid-modified protein from the inner membrane of Escherichia coli requires nucleoside triphosphate. J. Biol. Chem. 273:32576-32581.
Yamada, H., S. Matsuyama, H. Tokuda, and S. Mizushima. 1989. A high concentration of SecA allows proton motive force-independent translocation of a model secretory protein into Escherichia coli membrane vesicles. J. Biol. Chem. 264:18577-18581.
Yamaguchi, K., F. Yu, and M. Inouye. 1988. A single amino acid determinant of the membrane localization of lipoproteins in E. coli. Cell 53:423-432.
Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103-119.
Yuan, Y. R., S. Blecker, O. Martsinkevich, L. Millen, P. J. Thomas, and J. F. Hunt. 2001. The crystal structure of the MJ0796 ATP-binding cassette. Implications for the structural consequences of ATP hydrolysis in the active site of an ABC transporter. J. Biol. Chem. 276:32313-32321.(Yasuko Ito, Hitomi Matsuz)