SEARCH

SEARCH BY CITATION

Keywords:

  • assembly;
  • chaperone;
  • lipopolysaccharide;
  • outer membrane;
  • phospholipid

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The Skp protein of Escherichia coli has been proposed to be a periplasmic molecular chaperone involved in the biogenesis of outer membrane proteins. In this study, evidence is obtained that Skp exists in two different states characterized by their different sensitivity to proteases. The conversion between these states can be modulated in vitro by phospholipids, lipopolysaccharides and bivalent cations. Skp is able to associate with and insert into phospholipid membranes in vitro, indicating that it may associate with phospholipids in the inner and/or outer membrane in vivo. In addition, it interacts specifically with outer membrane proteins that are in their non-native state. We propose that Skp is required in vivo for the efficient targeting of unfolded outer membrane proteins to the membrane.

Abbreviations
LPS

lipopolysaccharide

DOPG

1,2-dioleoyl-sn-glycero- 3-phosphoglycerol

DOPE

1,2-diol-eoyl-sn-glycero-3-phosphoethanolamine

INV

inside-out inner membrane vesicle.

The Skp protein (Skp stands for 17 kDa protein) of Eschericha coli has been localized in various studies in the cytoplasm [1–4], the outer membrane [5,6] and on the cell surface [7], and found to be associated with DNA, ribosomes and lipopolysaccharide (LPS). The cationic nature of Skp is probably responsible for the varying results in the subcellular localization experiments, as the protein can artificially associate with anionic compounds after disintegration of the cells. Skp was previously suggested to be localized in the periplasm, as the majority of the protein was released from spheroplasts, provided that bivalent cations were present during the fractionation procedure [8]. It is proposed that the cations prevent electrostatic interactions between Skp and the anionic structures in the cell envelope such as LPS. Consistent with an extracytoplasmic location of Skp, the protein is synthesized as a precursor containing an N-terminal leader peptide [8]. The function of Skp is not known at present. Homologs have been described in various gram-negative bacteria, even outside the Enterobacteriaceae [9], suggesting that it is an important protein. Amongst other possibilities, Skp could function in LPS biogenesis, as the skp gene maps at 4 min on the E. coli chromosome [4] in a cluster of genes involved in lipid A biosynthesis [10]. Interestingly, Skp has also been shown to compensate for SecA deficiency in the in vitro translocation of the precursor of outer membrane protein LamB into inverted inner membrane vesicles (INVs) [4]. Whereas the periplasmic location of Skp in vivo would be incompatible with a SecA-like function in the translocation of precursors across the cytoplasmic membrane, this observation could indicate a role at a later stage in the biogenesis of outer membrane proteins. Indeed, it was recently shown that Skp binds selectively to outer membrane proteins and that inactivation of the skp gene results in decreased amounts of these proteins [11]. In addition, a Tn10 insertion in the skp gene was shown to increase the activity of the σE regulon, which is indicative of an increased accumulation of misfolded outer membrane proteins in the periplasm [12]. These results suggest that Skp is a molecular chaperone, involved in protein folding in the periplasmic compartment of gram-negative bacteria and specifically required for the biogenesis of outer membrane proteins.

In the studies described in this paper, the association of Skp with outer membrane proteins and lipid compounds was investigated, and its role in the folding and assembly of outer membrane proteins was studied in an in vitro system.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Protease treatment of Skp

In vivo. An overnight culture of strain CE1224 [13] was diluted 1 : 20 into fresh LB medium and grown for 2 h at 37 °C under aeration. Cells from 1 mL culture samples were collected by 2 min centrifugation in an Eppendorf centrifuge, resuspended in 100 µL of 100 mm Tris/HCl, pH 8.0, and transferred to ice. Trypsin (5 µL of a solution of 2 mg·mL−1 trypsin in 100 mm Tris/HCl, pH 8.0, 50 mm MgCl2) and subsequently 5 µL of either 200 mm MgCl2 or 100 mm EDTA, pH 8.0, were added to the cells. After 30 min incubation at 37 °C, the protease was inhibited with phenylmethanesulfonyl fluoride (1 mm). Cells were collected by centrifugation, resuspended in 60 µL of sample buffer and incubated for 10 min at 100 °C to solubilize the proteins. Samples of volume 5 µL were used for SDS/PAGE [14] followed by Western immunoblotting with antiserum directed against Skp or β-lactamase. Proteins were electrotransferred to a nitrocellulose filter for 16 h at 30 V in transfer buffer [14.5 g Tris, 67.6 g glycine, 7.5 mL of 20% (w/v) SDS, 1200 mL of methanol in 6 L of water; the pH should be 8.3].

In vitro.A 1 µL sample of an S135 extract (in buffer B: 10 mm triethanolamine acetate, pH 7.5, 10 mm magnesium acetate, 22 mm ammonium acetate, 1 mm dithiothreitol) or of outer membrane or inner membrane preparation was added to 19 µL of buffer L (50 mm triethanolamine acetate, pH 7.5, 250 mm sucrose, 1 mm dithiothreitol) and incubated with proteinase K or trypsin (200 µg·mL−1) at 37 °C. After 30 min, the protease was inhibited with PMSF (1 mm) and protease-resistant proteins were solubilized in sample buffer and analyzed by SDS/PAGE and Western immunoblotting. Quantifications were performed with a laser densitometer (Molecular Dynamics).

Incubation of Skp with phospholipids, bivalent cations and LPS

Phospholipids [15] and LPS of chemotype S [16] were isolated from E. coli strains MC4100 [17] and U20 [18] respectively, as described. Phospholipid concentrations were determined by measuring the phosphate content [19]. The LPS content was determined by 3-deoxy-d-manno-octulosonic acid measurement [20]. Dried phospholipids (149 nmol) were resuspended in 15 µL of buffer L to prepare liposomes. Subsequently, 3 µL of either MgCl2 or EDTA (both 10 mm end concentration) was added, followed by 1 µL of LPS (1.9 nmol in water) or water. Finally, the total volume was brought to 20 µL with buffer L. The Skp protein, either in 10 µL of an S135 extract (in buffer B; ≈ 9 ng Skp) or as purified protein (30 ng in buffer B), was added and the total mixture was incubated for 10 min at 37 °C. A 20-µL aliquot of this mixture was added to 1 µL of proteinase K (50 µg·mL−1) and incubated for 15 min at 37 °C. The protease was inhibited with phenylmethanesulfonyl fluoride (1 mm), and proteins were solubilized in sample buffer and analyzed by SDS/PAGE and Western immunoblotting. A 10-µL aliquot of the mixture that was not treated with protease was used to demonstrate the total amount of Skp protein used.

In vitro translations

Isolation of S135 cell extracts and inner and outer membrane vesicles from strain MC4100 and in vitro transcription and translation reactions were performed as described [21,22]. Plasmids pJP29 [23] and pJP370 [24] were used to direct the synthesis of [35S]methionine-labeled precursor and quasi-mature (i.e. the complete mature protein extended with methionine and serine at the N terminus) forms of PhoE protein respectively. Plasmid pMalE A276G, a derivative of pBAR43 [25], was used to direct the synthesis of the precursor of β-lactamase and of a mutant form of the precursor of maltose-binding protein, containing a glycine instead of an alanine at position 276 of the mature region of MalE [26]. Routinely, puromycin (10 µm) was added 25 min after initiation of protein synthesis. After in vitro protein synthesis, Skp (12 µg·mL−1 in 50 mm triethanolamine acetate, pH 7.5, 50 mm potassium acetate, 5 mm magnesium acetate, 1 mm dithithreitol) was added to the translation mixture. After incubation for 2 min at 37 °C, co-immunoprecipitations were performed as described [27] using antiserum directed against Skp. To study the effect of Skp on the folding of PhoE protein, the in vitro synthesized quasi-mature PhoE protein was incubated with varying amounts of Skp. Subsequently, folding of PhoE into a trypsin-resistant conformation was studied as described [28]. Briefly, folding was initiated by the addition of an LPS/Triton X-100 mixture in buffer L [0.015% (w/v) Triton X-100, final concentration]. After 30 min incubation at 37 °C, trypsin was added to determine the total amount of PhoE protein folded into a trypsin-resistant folded monomer. To determine the half-life of the folding-competent state, PhoE protein was incubated with Skp at 37 °C, and, at various time points, aliquots of the mixture were incubated with LPS/Triton X-100 to induce folding as described above. The results were plotted as the log of trypsin-resistant PhoE against time.

Depletion of S135 extracts for Skp

Aliquots (30 µL) of antiserum directed against Skp protein were incubated for 20 min at room temperature with 10 mg of Protein A–Sepharose CL-4B (Pharmacia Biotech) in 400 µL Tris/HCl, pH 8.6, containing 0.15 m NaCl. The complexes were washed twice with buffer B and incubated for 1 h at 4 °C with 120 µL of S135 cell extract. Subsequently, the Protein A–Sepharose beads containing the bound antibodies and Skp were removed by centrifugation for 20 s in an Eppendorf centrifuge (14 000 r.p.m.), and 100 µL of the Skp-depleted S135 cell extract was recovered and used for in vitro protein sythesis. The efficiency of the depletion was determined by quantitive Western immunoblotting using a standard calibration curve of purified Skp protein.

Monolayer experiments

Monolayer surface pressure measurements were performed by the (platinum) Wilhelmy plate method [29] in a temperature-controlled box using a Cahn 2000 microbalance as described [30]. The subphase buffer used was 50 mm triethanolamineacetic acid, pH 7.5.

Pulse labeling of spheroplasts

An overnight culture of E. coli strain, MC4100 grown at 37 °C in minimal E-medium [31] containing 18 amino acids (without Met and Cys; 100 µm each) and 8 g·L−1 maltose to induce the synthesis of LamB, was diluted 1 : 50 in fresh medium and grown to an A578 of 0.8. Cells from 1 mL of culture were collected by centrifugation, resuspended in 250 µL of 100 mm Tris/HCl, pH 8.0, containing 18% (w/v) sucrose and carefully mixed with 250 µL of 0.2 mg·mL−1 lysozyme prepared in 8 mm EDTA, pH 8.0. Spheroplasts, which had been formed during 15 min incubation on ice, were collected by centrifugation after the addition of 15 µL of 0.5 m MgCl2 and resuspended in 0.2 mL of E-medium containing the 18 amino acids and maltose as described above, and 0.25 m sucrose. After 6 min preincubation at 30 °C, pulse-labeling was performed for 5 min with 25 µCi·mL−1[35S]methionine (> 1000 Ci·mmol−1· mL−1). MgCl2 was then added to a concentration of 15 mm, and soluble proteins were separated from the spheroplasts by centrifugation.

Immunoaffinity chromatography on anti-Skp columns

Immunoglobulins present in 75 µL each of a polyclonal rabbit Skp antiserum and a preimmune serum were coupled to 50 mg of Protein A–Sepharose CL-4B as described [32]. The resins were stored in buffer [25 mm Tris/HCl, pH 7.5, 2 mm magnesium acetate, 150 mm NaCl, 0.05% (v/v) Triton X-100] containing 0.02% (w/v) NaN3. Soluble proteins (0.4–0.8 mL) were applied to the immunoaffinity matrix equilibrated with E-medium containing 0.25 m sucrose. The gel suspension was mixed for 60 min at 4 °C by end-over-end rotation, and unbound material was removed by centrifugation. The matrix was then washed twice with 0.8 mL each of E-medium containing 0.25 m sucrose, and bound proteins were eluted with 0.7 mL of glycine/HCl buffer, pH 2.3, containing 500 mm NaCl and 0.3% Tween 20. The eluate was neutralized using 10–40 µL of 1.5 m Tris/HCl, pH 8.8, and either precipitated with 10% (w/v) trichloroacetic acid for subsequent SDS/PAGE or first denatured with SDS [1% (w/v) final concentration] and then indirectly immunoprecipitated using Protein A–Sepharose as described [33].

Purification of Skp

Strain HB101 [34] transformed with pGAH317 [3] was grown in 4 L batches of LB broth, supplemented with 200 mg·L−1 ampicillin until cells had reached an A578 of 1.4. They were washed once with 100 mm Tris/HCl, pH 6.8, resuspended in 30 mL of 100 mm Tris/HCl, pH 8.0, containing 20% (w/v) sucrose and incubated for 10 min on ice. The suspension was diluted with 2 vol. of distilled water and kept for another 10 min on ice. MgCl2 was added to a final concentration of 10 mm, and the fraction of periplasmic proteins was obtained in the supernatant after centrifugation at 2000 g for 25 min. Immediately before chromatography, which was performed at room temperature, Tris base, potassium acetate and magnesium acetate were added to the supernatant to yield buffer A (50 mm Tris, 50 mm potassium acetate, 15 mm magnesium acetate, pH 9.0, adjusted at room temperature with acetic acid). The solution was applied to 60 mL of CM Sepharose CL-6B, pre-equilibrated with buffer A. Elution was achieved with a linear 400 mL gradient prepared from buffer A and buffer A containing 300 mm potassium acetate. Fractions of 5 mL were collected into 1 mL of 1 m triethanolamine acetate, pH 7.5, each to quickly lower the pH. The purified Skp was more than 98% pure, as determined by SDS/PAGE.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Protease sensitivity of Skp

When cells are fractionated using standard biochemical techniques, i.e. in the absence of bivalent cations, a substantial amount of Skp co-fractionates with the membranes and a proportion is found in the soluble fraction [8]. The Skp present in membrane-free S135 extracts of E. coli cells was completely degraded when incubated with proteinase K or trypsin (results not shown). In contrast, Skp present in the membrane fraction was partially protected against these proteases (results not shown). This suggests that, as the result of association or integration into the lipid bilayer, Skp may exist in two different states, which differ in their sensitivity to proteases. Therefore we decided to investigate the protease sensitivity of the Skp protein in further detail. Interestingly, some of the cellular Skp was found to be trypsin-resistant (≈ 20%) when whole cells were incubated with the protease and with EDTA to allow access of the protease to the periplasm (Fig. 1A). No further degradation of Skp was observed when such incubations with trypsin were performed for 60 min instead of 30 min at 37 °C. Under these conditions, the C-terminal periplasmic domain of OmpA protein was completely digested (results not shown), confirming the entry of the protease into the periplasm. Skp therefore also appears to exist in vivo in two states with different protease sensitivities. Interestingly, hardly any Skp was released from the cells when they were incubated with EDTA in the absence of trypsin (Fig. 1A and B). In contrast, the periplasmic protein β-lactamase was almost completely released from the cells under these conditions (Fig. 1B). The vast majority of the cellular Skp apparently remains associated with the cells and only part of this population exists in a protease-resistant state.

image

Figure 1. Existence of Skp in vivo in two states which differ in protease sensitivity. (A) The protease sensitivity of Skp in vivo was determined by treating cells with trypsin (Tryp) in the presence of EDTA to permeabilize the outer membrane as described in Materials and Methods. Incubations of cells with Mg2+ or EDTA in the absence of trypsin served as controls. (B) Western immunoblot demonstrating the release of periplasmic β-lactamase (Bla), but not Skp, from permeabilized cells (EDTA).

Download figure to PowerPoint

The possibility that the membrane lipids modulate the protease sensitivity of Skp was subsequently investigated. Skp in the S135 extract could partially be converted into a protease-resistant state by adding it to liposomes, composed of total E. coli phospholipids, provided that Mg2+ was present (Fig. 2A; lanes 2 and 3). Approximately 24% of the total Skp protein could be converted into a protease-resistant conformation. Similar results were obtained with liposomes of the non-E. coli phospholipid 1,2-dioleoyl-sn-glycero-3-phosphocholine (data not shown). Inclusion of LPS in these experiments decreased the amount of Skp converted into a protease-resistant form (Fig. 2A; lane 4, ≈ 12% of the total Skp was still protease-resistant). Purified Skp could be converted into a protease-resistant form as well (Fig. 2B; lane 2, ≈ 12% of the total Skp was protease-resistant) and also in this case inclusion of LPS decreased the amount of Skp converted into the protease-resistant state (Fig. 2B; lane 4, 3%). Only very small amounts of Skp were converted into a proteinase K-resistant state when incubated with 10 mm Mg2+ but in the absence of phospholipids (Fig. 2C) or with phospholipid-free LPS (data not shown). Skp can apparently exist in a protease-sensitive and a protease-resistant state in vivo and in vitro and the conversion between these states is modulated in vitro by phospholipids, bivalent cations and LPS. Only very small amounts of the protease-resistant form of Skp could be reisolated together with the liposomes by ultracentrifugation (results not shown) probably because of the presence of Mg2+[8].

image

Figure 2. Modulation in vitro of the protease sensitivity of Skp by phospholipids (PL), LPS and bivalent cations. An Skp-containing S135 extract (A) or purified Skp (B) was incubated for 15 min at 37 °C with the indicated components and subsequently treated with proteinase K (PK). (C) Sensitivity of Skp (present in S135 extract) to proteinase PK in the presence of the indicated amount of Mg2+ but in the absence of phospholipids. Samples were analyzed by SDS/PAGE followed by Western immunoblotting with antiserum directed against Skp. In (A) and (B), half the amount of material was loaded on the gel in lane 1 as compared with that loaded in lanes 2–5.

Download figure to PowerPoint

Insertion of Skp into phospholipid monolayers

To investigate whether Skp interacts with phospholipids, monolayer studies were performed. Injection of purified Skp under a monolayer of the negatively charged phospholipid 1,2-dioleoyl-sn-glycero-3-phosphoglycerol (DOPG) gave a rapid increase in surface pressure (results not shown). However, this could be the result of non-specific ionic interactions, as similar results were obtained when another cationic protein, lysozyme, was injected under the monolayer, as reported previously [29]. However, in the case of lysozyme, the presence of bivalent cations (Mg2+) completely inhibited the surface pressure increase (Fig. 3A), but not that observed in the case of Skp (Fig. 3B). No increase in surface pressure was observed with the buffer in which Skp was solubilized (data not shown).

image

Figure 3. Insertion of Skp into phospholipid monolayers. Kinetics of the surface pressure increase induced by injection of lysozyme (A) or Skp (B) under DOPG monolayers. Before injection of the proteins (1.8 µg each; arrow), the monolayer was equilibrated with 10 mm magnesium acetate at 34 °C. (C) Insertion of Skp as a function of the initial surface pressure of DOPG with 10 mm magnesium acetate (•) or DOPE either in the absence (○) or presence (×) of magnesium acetate. Skp (3.6 µg) was injected under the monolayers, spread at the indicated inital surface pressure at 30 °C, and the surface pressure increase was plotted against the initial surface pressure.

Download figure to PowerPoint

Analysis of the insertion of Skp as a function of the initial surface pressure of the monolayers was carried out to detect any lipid specificity in the process (Fig. 3C). The limiting initial surface pressure was much higher for monolayers of the anionic phospholipid DOPG (≈ 41 mN·m−1) than for those of the zwitterionic phospholipid 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) (Fig. 3C). The biological membrane pressure (for erythrocyte membrane) has been determined to be between 31 and 34 mN·m−1[35], indicating that Skp has the ability to insert into a biological membrane. Insertion of Skp into monolayers of DOPG (data not shown) or DOPE was not affected by the absence or presence of Mg2+ (Fig. 3C). These results show that Skp is able to penetrate into a phospholipid monolayer, a process that is strongly stimulated by anionic phopholipids.

Modulation of complex-formation of Skp with outer membrane proteins by LPS

Recently, it has been shown that Skp interacts specifically with outer membrane proteins in vitro[11]. To investigate whether Skp also interacts with outer membrane proteins in a semi in vivo system, MC4100 cells were converted to spheroplasts and subsequently pulse-labeled with [35S]methionine. Periplasmic and outer membrane proteins synthesized during the pulse are secreted into the medium of the spheroplasts [36]. By immunoprecipitation with polyclonal antisera (Fig. 4A), Skp (lane 1), LamB (lane 2), OmpA (lane 3) and β-lactamase (Bla; lane 4) could be identified among the secreted proteins. When the spheroplast supernatant containing the secreted proteins was analyzed by immunoaffinity chromatography with Skp antibodies, a number of proteins were retained that could be eluted at pH 2.3 (lane 5). These proteins were not detected when preimmune serum instead of anti-Skp serum was used in the immunoaffinity chromatography (data not shown). Immunoprecipitation revealed the presence of the outer membrane proteins LamB and OmpA among those that bound firmly to the column (lanes 7 and 8). In contrast, the periplasmic enzyme β-lactamase was not retrieved in the eluate (lane 9), despite the fact that considerable amounts of it were present in the fraction loaded on to the column (lane 4). Hence we conclude that the outer membrane proteins LamB and OmpA were specifically complexed by Skp after their secretion by the spheroplasts and therefore retained by the immobilized Skp antibodies on the matrix.

image

Figure 4. Complex-formation between Skp and outer membrane proteins in vivo and in vitro. (A) A fraction of secreted proteins was prepared from pulse-labeled E. coli spheroplasts (medium) as described in Materials and Methods. Aliquots of 150 µL each were directly used for immunoprecipitation (Im. Prec.) with polyclonal antibodies directed against Skp (lane 1), LamB (lane 2), OmpA (lane 3) or β-lactamase (Bla; lane 4). Aliquots of 400 µL (lanes 5 and 6) and 800 µL (lanes 7–9) were used in immunoaffinity purifications using antibodies against Skp and subsequently eluted at pH 2.3 (eluate). Eluted proteins were either directly applied to SDS/PAGE (lane 5; t, total) or after immunoprecipitation as indicated. A fluorograph of the SDS/polyacrylamide gel is shown. (B) Translation mixtures containing in vitro synthesized precursor (pJP29) and mature (pJP370) PhoE proteins, chloramphenicol acetyltransferase (Cat) and the precursors of β-lactamase (Bla) and maltose-binding protein (MBP) were used in immunoprecipitations with antiserum directed against Skp (αSkp) as described in Materials and Methods. The lanes with total translation products (TL) contain only 5% of the total amount of translation mixtures used in the co-immunoprecipitations (Im. Prec). Immunoprecipitations using a preimmune serum (PI) were used to determine the amount of unspecifically precipitated proteins. (C) Co-immunoprecipitations of in vitro synthesized mature PhoE with Skp-specific antiserum after 5 min incubation at 37 °C of the Skp–PhoE complexes with LPS and/or Triton X-100 (TX-100) as indicated.

Download figure to PowerPoint

Co-immunoprecipitation experiments revealed that Skp can bind to in vitro synthesized outer membrane protein PhoE (Fig. 4B). The Skp concentration in the in vitro translation system had to be increased by adding purified protein in order to obtain complex-formation. The precursor (lane 2) as well as the mature form (lane 5) of PhoE protein could be co-precipitated efficiently (≈ 50% of the total amount of synthesized protein) with Skp when antiserum directed against Skp was used. Furthermore, no specific complex-formation was observed between Skp and the cytosolic protein chloramphenicol acetyltransferase (lane 2) or the precursors of the periplasmic proteins β-lactamase and maltose-binding protein [8] under these conditions (Fig. 4B). Similar amounts of these proteins were precipitated with the antiserum directed against Skp or with the preimmune serum. Rapid folding of the maltose-binding protein precursor into its native state can be excluded as an explanation for the absence of complex-formation, as the precursor used in these experiments was a mutant form containing a missense mutation (A276G), which drastically decreases the rate of folding [26]. Complex-formation between outer membrane proteins and Skp may occur directly between these proteins. However, one cannot exclude the possibility of interactions mediated via a third partner that might be present under these conditions.

The efficiency of the co-immunoprecipitation of PhoE protein with Skp decreased reproducibly by ≈ 50% when the Skp–PhoE complexes were preincubated for 5 min at 37 °C with LPS (30 nmol·mL−1 in either the absence or presence of Mg2+) and by 25% when incubated with 0.015% Triton X-100 (Fig. 4C). These results can be explained by complexation of Skp by LPS [7] and by the Triton X-100-induced folding of PhoE into an assembly-incompetent state [22] respectively. Interestingly, no PhoE protein at all was co-immunoprecipitated when similar preincubations were performed with 0.015% Triton X-100 and LPS simultaneously (Fig. 4C), conditions that have been shown to induce very efficient folding of PhoE protein into a folded monomer [28]. The ability to form a complex between Skp and outer membrane proteins is apparently affected by the conformational state of the outer membrane protein, and no complex is formed between folded monomers of PhoE and Skp. These results suggest that Skp plays a role at an early step in the biogenesis process, preceding the folding of outer membrane proteins.

Influence of Skp on PhoE folding

To investigate whether Skp has a role in the folding of PhoE protein, in vitro synthesized PhoE was incubated with various amounts of Skp and folding was initiated by the addition of purified LPS and 0.015% Triton X-100. Subsequently, trypsin was added to determine the amount of PhoE that was folded into a trypsin-resistant conformation (Table 1). A slight stimulation of the folding of PhoE was observed when the Skp concentration was raised to a certain optimum. However, progressive inhibition of folding was observed when the Skp concentration was increased further. Neither stimulation nor inhibition of PhoE folding was observed when Skp was replaced by the molecular chaperone SecB or by the cationic protein lysozyme at similar concentrations (results not shown). The inhibitory effect of Skp on folding is possibly due to complexation of the LPS, thereby decreasing the effective LPS concentration available for folding of PhoE. Indeed, efficient folding was restored when similar experiments were performed with increasing amounts of LPS (results not shown). The small stimulatory effect of Skp on PhoE folding suggests an involvement of Skp in the folding process of outer membrane proteins. However, immunodepletion of an S135 extract of Skp by more than 90% did not further decrease the folding efficiency (results not shown).

Table 1. Effect of Skp on folding of PhoE. The amount of Skp present during the preincubation with in vitro synthesized protein is shown in the first column. The amount of 9 pmol·mL−1 Skp is endogneously present in the S135 extract. The molar ratio of LPS over Skp during folding is shown. The amount of LPS during folding was 29.5 nmol·mL−1. Relative efficiency (Rel. eff.) of folding of PhoE is given in the final column. The highest folding efficiency obtained was defined as 1 and the other folding efficiencies were expressed as a fraction of this efficiency. Results are the average of two independent experiments.
Skp  (pmol·mL−1) LPS/Skp  ratio Rel. eff. of folding
956000.77
4610980.85
895711.00
1693010.87
3291550.78
809630.66
1609320.26

Skp had no effect on the half-life of the folding-competent state of in vitro synthesized PhoE protein, i.e. the capacity to fold into a folded monomer. The half-life remained between 70 and 80 min at either very low or very high concentrations of Skp (results not shown). Therefore we conclude that Skp does not stabilize a folding-competent state of PhoE protein.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The majority of Skp has previously been found to be localized in the periplasm of E. coli, as it was released from spheroplasts provided that bivalent cations were present [8]. However, we observed that Skp is able to interact with phospholipids and that it becomes protease-resistant in the presence of phospholipids and bivalent cations. This is probably due to a conformational change in Skp upon interaction with phospholipids and bivalent cations. The role of the bivalent cations remains obscure as a fraction of Skp was found to be protease-resistant in vivo after EDTA-induced permeabilization of the outer membrane. Hence, additional cellular components may be involved in the modulation of the protease sensitivity of Skp in vivo. An interaction of Skp with phospholipids was observed in monolayer studies, and the insertion of Skp into lipid monolayers was strongly stimulated by the presence of anionic phospholipids. A proportion of Skp may also be membrane-associated in vivo, which is in agreement with the observation that a small amount of Skp co-fractionates with the membranes in the presence of Mg2+[8]. However, the vast majority of Skp is not released from cells when these are permeabilized with EDTA, in contrast with a typical periplasmic protein. Possibly, the majority of Skp is associated with phospholipids under these conditions. Alternatively, the majority of Skp associates with other, as yet unknown, membrane components and part of the population is converted in a phospholipid-dependent manner to a protease-resistant form.

In spite of its periplasmic location in vivo, Skp has been shown to stimulate the translocation of the precursors of the outer membrane proteins LamB and OmpA [4], as well as of PhoE (M. Kleerebezem and J. Tommassen, unpublished observation) into INVs in vitro. Interestingly, this translocation-promoting activity of Skp was detected as a compensatory effect upon SecA depletion of the S135 extracts used for in vitro protein synthesis, but not upon SecB depletion. It seems unlikely that the export-promoting activity of this otherwise periplasmic protein is the result of Skp activities similar to those of SecA at the inner membrane-located translocase [37], as the sequence of Skp reveals no evidence of ATP-binding sites. However, like Skp, SecA has the ability to bind precursor proteins and to interact with and penetrate into phospholipid membranes [30]. We propose that these activities of Skp are responsible for the observed complementation of SecA function in vitro. Thus a precursor of an outer membrane protein in complex with Skp may be targeted to the INVs, because of the affinity of Skp for phospholipids. Subsequently, the precursor proteins, anchored via their leader peptide in the lipid phase, may reach the translocation sites by diffusion and interact with the translocase. In these experiments, sufficient SecA was present at these translocation sites to support further translocation, as the INVs were not depleted of SecA.

What therefore is the true physiological function of Skp in the periplasm? Recently, it has been proposed that it functions as a molecular chaperone involved in the folding of outer membrane proteins [11,12]. Cells containing an inactivated skp gene are viable but have reduced amounts of outer membrane proteins, and this phenotype was complemented by a plasmid-borne Skp protein [11]. However, the results described in this paper indicate that Skp is not an important factor that is essential for the folding or involved in the stabilization of a folding-competent state of outer membrane proteins. Skp is also not able to prevent aggregation of the outer membrane proteins PhoE or OmpA in vitro (H. de Cock, unpublished observations). We have previously shown that the folding of an in vitro synthesized PhoE protein into a folded monomer requires bivalent cations and lipopolysaccharides [28]. These folded monomers are assembly-competent intermediates, which can be targeted to outer membrane-assembly sites where the late steps in the biogenesis process, i.e. trimerization and insertion into the outer membrane, occur. We propose that Skp has a very early role in biogenesis and is required for the efficient targeting of unfolded outer membrane proteins directly after their translocation to sites at the periplasmic phase of the membrane. This process is reminiscent of the targeting of the outer membrane precursors to the INVs in the artificial translocation assays discussed above. At these membrane sites, the early steps in the folding and assembly of outer membrane proteins would occur. The outer membrane proteins may be targeted to sites at the periplasmic phase of the inner membrane. The observation that an assembly intermediate of outer membrane protein OmpA (immature-processed OmpA, imp-OmpA) was localized to the periplasmic phase of the inner membrane in vivo is in accordance with this supposition [38]. Furthermore, intermediates in the assembly of outer membrane protein LamB have been localized in the inner membrane in vivo[39,40]. Dissociation of outer membrane proteins from Skp may occur by an exchange reaction with de novo synthesized LPS and/or by initiation of the folding of the outer membrane protein. This is consistent with the observation that Skp–PhoE complexes dissociated in the presence of LPS or when folding was induced with Triton X-100 (Fig. 4C). The conversion between the protease-sensitive and protease-resistant states of Skp may be coupled to this association/dissociation cycle. Subsequently, the outer membrane proteins can fold into their native state and then be targeted to outer membrane assembly sites where the late steps of the biogenesis process can proceed, as was proposed previously [28]. In the absence of Skp, the efficiency of targeting to the inner membrane sites would be reduced, resulting in the accumulation of unfolded outer membrane proteins in the periplasm and activation of the σE regulon [12]. Thus channeling outer membrane proteins efficiently into the folding pathway prevents accumulation of unfolded outer membrane proteins in the periplasm. It was recently proposed that Skp may be required to remove or exchange the original LPS molecule associated with a folded intermediate of outer membrane proteins [12]. This exchange of LPS would trigger the insertion of assembly intermediates into the outer membrane and thus accelerate the transition to their ultimate folded conformation. In this model, Skp was proposed to act late in the biogenesis because of its capacity to bind LPS. In contrast, in our model, Skp acts very early in the biogenesis of outer membrane proteins. The model takes into account the important observation that Skp interacts specifically with non-native outer membrane proteins and has affinity for phospholipids.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

H.d.C is a recipient of a fellowship from the Royal Netherlands Academy of Arts and Sciences. The work of M.M. was supported by the Sonderforschungsbereich 184. We would like to thank the anonymous referee who suggested that we look at protease sensitivity and release of Skp from permeabilized cells.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References