J Biol Chem, Vol. 274, Issue 35, 24567-24574, August 27, 1999

Skp, a Molecular Chaperone of Gram-negative Bacteria, Is Required for the Formation of Soluble Periplasmic Intermediates of Outer Membrane Proteins^*

Ute Schäfer^§, Konstanze Beck^§, and Matthias Müller^¶

From the  Institut für Biochemie und Molekularbiologie and ^§ Fakultät für Biologie, Universität Freiburg, Hermann-Herderstrasse 7, D-79104 Freiburg, Germany

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Using a cross-linking approach, we have analyzed the function of Skp, a presumed molecular chaperone of the periplasmic space of Escherichia coli, during the biogenesis of an outer membrane protein (OmpA). Following its transmembrane translocation, OmpA interacts with Skp in close vicinity to the plasma membrane. In vitro, Skp was also found to bind strongly and specifically to pOmpA nascent chains after their release from the ribosome suggesting the ability of Skp to recognize early folding intermediates of outer membrane proteins. Pulse labeling of OmpA in spheroplasts prepared from an skp null mutant revealed a specific requirement of Skp for the release of newly translocated outer membrane proteins from the plasma membrane. skp mutant cells are viable and show only slight changes in the physiology of their outer membranes. In contrast, double mutants deficient both in Skp and the periplasmic protease DegP (HtrA) do not grow at 37 °C in rich medium. We show that in the absence of an active DegP, a lack of Skp leads to the accumulation of protein aggregates in the periplasm. Collectively, our data demonstrate that Skp is a molecular chaperone involved in generating and maintaining the solubility of early folding intermediates of outer membrane proteins in the periplasmic space of Gram-negative bacteria.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Gram-negative bacteria are characterized by the existence of an outer membrane, which together with the inner membrane (also termed plasma membrane) delimits the intervening periplasmic space. Outer membranes harbor two unique features, (i) the outer leaflet of the lipid bilayer is composed of a specific glycolipid, called lipopolysaccharide (LPS)¹, which plays a crucial role in maintaining the permeability barrier of a Gram-negative cell (1); (ii) outer membrane proteins, many of which function as porins, usually form trimers that are inserted into the lipid bilayer as antiparallel -barrels (2). Like the soluble periplasmic proteins, outer membrane proteins are synthesized as preproteins in the cytoplasm. They are translocated in a SecA- as well as SecB-dependent manner (3) across the inner membrane into the periplasm where an ill-defined sequence of folding events must occur, leading to soluble intermediates, to the conversion of monomers into trimers, and usually to the association with LPS allowing insertion into the outer membrane (4).

A picture of how the distinct folding events are catalyzed by individual chaperones is only now emerging (recently reviewed in Ref. 5). Most of the periplasmic chaperones thus far identified belong to two major groups, the Dsb proteins catalyzing thiol-disulfide exchange reactions, and peptidyl prolyl isomerases (PPIases) catalyzing the cis-trans isomerization around Xaa-Pro peptidyl bonds. Representatives of all major families of PPIases have been detected in the periplasm of Gram-negative bacteria. RotA (PpiA) is a cyclophilin-type PPIase; FkpA, a FK506-binding protein type PPIase; and SurA and PpiD belong to the parvulin type. The latter two have recently been analyzed in more detail as to an involvement in the biogenesis of outer membrane proteins (6-9).

In addition, Skp has been suggested to function as a periplasmic chaperone. This 16-kDa, basic protein was first purified 20 years ago as a LPS-associated protein of Salmonella minnesota (10). In Escherichia coli, however, it was initially ascribed the function of a histone-like protein (HLP1; hlpA) (11) and for some time was mistaken as the product of the downstream firA gene (12). Kleppe and co-workers (13, 14) purified it as a 17-kDa, basic DNA-binding protein from E. coli, cloned its gene, and coined the name Skp for "17-kilodalton protein." Vaara and co-workers (15, 16) identified a homologue in S. typhimurium, which they characterized as outer membrane protein OmpH. Thome et al. (17) reported a SecA-like activity of E. coli Skp, which was used to purify the protein to homogeneity. SecA is a cytosolic translocation ATPase that targets unfolded preproteins to, and energizes their transport across, the bacterial plasma membrane (3). Since Skp is a periplasmic protein, its SecA-like activity was interpreted as that of a molecular chaperone preventing premature folding of preproteins in vitro and thereby alleviating the need for SecA (18). Notably, the in vitro substrates of Skp thus identified were the outer membrane proteins LamB and OmpA. A specific interaction of Skp with outer membrane proteins was later confirmed (19, 20). A skp null mutation leads to a moderate reduction in outer membrane proteins (19), suggesting that Skp is involved in the assembly process of outer membrane proteins. Lately, Skp was also identified as a periplasmic protein improving folding of recombinant proteins (21). By analyzing early steps of the biogenesis of OmpA in vitro in combination with characterizing a skp null phenotype we now show that Skp in fact is a molecular chaperone whose function is the generation and maintenance of early soluble folding intermediates of outer membrane proteins in the periplasmic space of Gram-negative bacteria.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Bacterial Strains and Plasmids-- If not stated otherwise E. coli K12 strain MC4100 (22) was used. Strains CAG18515 (proAB::Tn10 containing the kanamycin resistance gene) (23), CLC198 (degP::Tn10) (24), ID18 (firA200) (25), and MRE600 (17) have been described. For the lac-dependent expression of firA and skp, plasmids pSLF3 and pTRS7 were used (26). Plasmids pDMB (27) and p717OmpA² were used for T7 promoter-dependent in vitro expression of ompA. Similarly, mtlA encoding mannitol permease of E. coli was subcloned into vector pKSM717 (28) yielding p717MtlA-B.² In plasmid pTompA, ompA excised from pRD87 (29) as a PstI-EcoRI fragment and ligated into pTrc99A (Amersham Pharmacia Biotech) is under trc promoter control. Wild type MalE was expressed from pJF2 (30) and the COOH-terminally truncated MalE601 from the pJF2 derivative pLH13 (31).

In Vitro Synthesis and Translocation-- In vitro transcription/translation using E. coli S-135s (32) and the reconstituted system (27) was performed as described previously (27, 32). Triethanolamine acetate was replaced as buffer by Hepes/NaOH, pH 7.5, in order to avoid quenching of the cross-linker DSS. The oligonucleotide-directed synthesis and isolation of ribosome-associated nascent chains (RANCs) has been described (27). The oligonucleotides were chosen such that they allowed synthesis of a 125-amino acid-long RANC of pOmpA (pOmpA-5) and a 189-amino acid-long RANC of MtlA (MtlA189). The synthesis of RANCs was improved by including an antisense 10Sa RNA oligonucleotide (33). Gradient-purified, inside-out plasma membrane vesicles of E. coli strains MC4100 and MRE600 were prepared and extracted with 1 M potassium acetate as described previously (27).

Cross-linking-- A 25 mM stock solution of DSS (Pierce) was diluted 10-fold into samples and incubated at 25 °C for 30 min. The reaction was stopped by adding Tris-HCl, pH 7.5, to 50 mM and incubating at 25 °C for 15 min.

Construction of a Mutant Carrying an In-frame Deletion in skp-- The skp gene was excised from plasmid pGAH317 (13) by KpnI and BssHII yielding a fragment with about 300 base pairs on either side of skp. This fragment was ligated into pGEM3Z (Promega) cut with KpnI and HincII. By using primers that hybridized to the insert upstream of nucleotide 12 and downstream of nucleotide 376 of skp (cf. Fig. 3A) an inverse PCR product of this plasmid was obtained that lacked 363 nucleotides of skp (5'-primer, C GCG GAT CCC CAC TTT TTC ACA ATA AAC TCC; 3'-primer, CGC GGA TCC TCC GTT GCC AAC AGC CAG GAT ATC). Both primers contained BamHI linkers by which the PCR product was religated. These linker sequences introduced a Gly and a Ser between the remaining flanking regions of skp. The deletion in skp thus generated was confirmed by nucleotide sequence analysis.

The deleted version of skp was introduced into the chromosome of strain MC4100. To this end, it was excised from the pGEM3Z vector by KpnI and PstI and ligated into pMAK705 (34), which is temperature-sensitive for replication and carries the chloramphenicol resistance gene. The plasmid was therefore forced to cointegrate into the chromosome via homologous recombination at the residual skp sequences when cells were grown at 44 °C in the presence of chloramphenicol. Cointegrates were subsequently resolved by growth at 30 °C. Cells thus obtained grew in small and larger colonies. One out of 10 small colonies, as opposed to none of 200 larger colonies, had the intact skp gene on the plasmid and the deleted form on the chromosome. These cells lost their plasmid with a frequency of 1:100 when grown at 44 °C in the absence of chloramphenicol.

To move the deleted version of skp into new MC4100 background by P1 transduction, a Tn10 insertion in proAB, which maps 1.3 min away from skp on the E. coli chromosome, was first transduced into the skp mutant. Subsequently, both the kanamycin resistance of Tn10 and the skp deletion were moved by the same strategy into MC4100.

Isolation of skp,degP Double Mutants-- The tetracycline resistance of a degP::tet mutant was moved into the skp mutant by P1 transduction. P1 lysates were then obtained from tetracycline-resistant cells carrying the skp deletion and used to cotransduce both phenotypes into MC4100.

Spheroplasts-- For pulse labeling of spheroplasts, cells were grown overnight in Luria broth medium at 37 °C in the presence of 0.4% glucose to repress expression of plasmid-encoded genes from lac-type promoters where applicable. Cells were inoculated 1:50 into minimal E-medium (35) supplemented with 18 amino acids (without Met and Cys; 100 µM each) and 0.4% glucose and grown for 4 h at 37 °C. Cells present in 2 ml of cell culture (A₅₈₀ = 1) were collected at 3500 × g, resuspended in 450 µl of 100 mM Tris-HCl, pH 7.5, 0.5 M sucrose and converted to spheroplasts by the addition of 450 µl of 0.1 mg/ml lysozyme dissolved in 8 mM EDTA, pH 8.0, during 20 min on ice. They were pelleted at 6000 × g for 2 min, resuspended in E-medium supplemented with 18 amino acids (without Met and Cys; 100 µM each) and 2% glycerol, 0.25 M sucrose, and 2 mM IPTG. After a 10-min incubation at 37 °C, [³⁵S]methionine/cysteine was added to a final concentration of 50 µCi/ml. The labeling period of 5 min was stopped in ice water. One aliquot was directly treated with 0.3 mg/ml proteinase K for 60 min on ice. Spheroplasts were collected either by high speed (2 min at 16,000 × g) or low speed (2 min at 6 000 × g) centrifugation. For the preparation of PPPs, treatment of cells with EDTA/lysozyme was extended to 30 min followed by the addition of MgCl₂ to 15 mM.

Miscellaneous Methods-- Immunoprecipitation was performed essentially as described previously (36) with some modifications. Four-fold scaled up in vitro reactions were denatured in 1% SDS and applied to 10-15 µl of antiserum preadsorbed to 5-10 mg of protein A-Sepharose for 90 min at 4 °C. Formation of antigen-antibody complexes was allowed to proceed for at least 60 min at 4 °C at an SDS concentration below 0.2%. No proteinase inhibitors were used. SecA (32) and Skp (20) were purified as detailed elsewhere. Immunoblotting was performed as described previously (37). For determining sensitivity toward antibiotics, cells were embedded 1:100 in 1.5% agar plates prepared in E-medium supplemented with 0.8% maltose or LB medium. The agar was overlayered with 5-mm filter discs onto which the antibiotics had been applied in a final volume of 10 µl (0.2 mg of rifampicin dissolved in dimethyl formamide; 1 mg each of vancomycin dissolved in water and novobiocin dissolved in 70% ethanol).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Skp Interacts with Newly Translocated Non-native Outer Membrane Proteins-- To demonstrate interaction of Skp with its substrates, we synthesized outer membrane proteins by coupled transcription/translation and analyzed their translocation into inside-out plasma membrane vesicles (INV). As shown in Fig. 1A, the precursor of the outer membrane protein A (pOmpA) synthesized in vitro was partially converted to its signal sequence-free form (OmpA) by the cotranslational addition of INV (lane 3). Translocation into the lumen of INV is indicated by the accumulation of proteinase K-resistant material (lane 4), which was not obtained in the absence of INV (lane 2). Treatment with the membrane-permeable cross-linker DSS gave rise to an approximately 50-kDa cross-linking product (lane 5, asterisk), which was recognized by antisera raised against OmpA (lane 6) and Skp (lane 7). The 50-kDa material corresponds in size to one molecule of OmpA (35 kDa) and Skp (16 kDa) each. As expected for an OmpA-Skp complex formed after translocation of OmpA into INV, it was resistant to digestion with proteinase K as long as the vesicles were not disrupted with Triton X-100 (lanes 8 and 9). In addition, the intensity of the cross-link decreased when translocation was reduced by a lowered reaction temperature or by blockage of SecA with a nonhydrolyzable analogue of ATP (Fig. 1B, lanes 1-12; see lanes 2, 6, and 10 for the different extents of translocation). The specificity of the interaction of Skp with OmpA thus detected is demonstrated by the absence of this 50-kDa cross-linking product from assays containing INV of a skp null mutant (Fig. 1A, compare lanes 7 and 13). Similar results were obtained when the outer membrane protein LamB was synthesized in vitro (not shown).

View larger version (70K): [in this window] [in a new window]   Fig. 1.   Newly translocated OmpA binds to Skp on the periplasmic side of the membrane. A, pOmpA was synthesized by an S-135 from the skp mutant in the presence of INV as indicated. Wild type INV had been freed of Skp on the cytosolic side by high salt washing (K-INV wt). For cross-linking, samples were incubated post-translationally with DSS. Samples were either directly precipitated with trichloroacetic acid or first treated with proteinase K or immunoprecipitated with antibodies against OmpA and Skp (OmpA, Skp). Shown are radioactively labeled proteins separated by SDS-PAGE and visualized by phosphorimaging. The asterisk marks the cross-link between OmpA and Skp. The electrophoretic mobility of marker proteins is indicated on the right in kilodaltons. B, same as described for A except that INV were added post-translationally and incubated at the indicated temperatures either in the presence of 2.5 mM ATP or 25 mM AMP-PNP.

These results suggest that the used INV, which had been washed with 1 M salt to remove contaminating Skp from the outside of the vesicles, retained Skp sequestered within their lumen. This portion of Skp must tightly be associated with the periplasmic side of the membrane, since it copurified with the INV. Consequently, the interaction between Skp and newly translocated outer membrane proteins demonstrated to occur in these vesicles must reflect an event taking place in close proximity to the outer leaflet of the plasma membrane.

Proteins translocated across the plasma membrane before they acquire stable tertiary structures on the periplasmic side conceivably pass through loosely folded intermediate states which might resemble those of nascent polypeptides following release from the ribosome. To examine whether Skp has the potential ability to interact with non-native proteins, RANCs were synthesized in vitro. The pOmpA-RANCs were isolated by centrifugation and subsequently incubated with purified Skp and SecA and in addition with puromycin to release the ribosomes. SecA has been shown to interact with nascent precursor proteins (27). Cross-linking was then performed with DSS (Fig. 2A). After incubation with SecA alone, a cross-linking product of about 112 kDa appeared that was recognized by antibodies directed against SecA (lanes 2 and 3, asterisk). This cross-link was dependent on the addition of SecA (cf. lanes 6 and 7). It corresponds in size to a complex between one molecule of SecA (102 kDa) and a 125-amino acid-long fragment of pOmpA (pOmpA-5). Instead, when the pOmpA-5 RANCs were incubated with Skp, three comparably strong cross-linking products of approximate sizes of 27, 50, and 72 kDa appeared (lanes 6 and 10, double arrows), which all together were recognized by anti-Skp antibodies (lanes 8 and 12) and are therefore most likely complexes between pOmpA-5 and monomeric and oligomeric forms of Skp.

View larger version (66K): [in this window] [in a new window]   Fig. 2.   Skp efficiently and specifically interacts with nascent chains of an outer membrane protein after release from the ribosome. A, in vitro synthesis of a 125-amino acid-long NH2-terminal fragment of pOmpA (pOmpA5). RANCs were isolated by sucrose gradient centrifugation and subsequently incubated with purified SecA and Skp (approximately 1 µM each) prior to cross-linking with DSS where indicated. Some full-length pOmpA cosynthesized with the nascent chains was usually also found in the ribosomal pellet. Cross-linking products of pOmpA5 with SecA (asterisk) and Skp (double arrows) are indicated. B, same as described for A except that RANCs were incubated with a total cytosolic extract containing Skp. In addition, a 189-amino acid-long nascent chain of mannitol permease (MtlA189) was analyzed. Where indicated, samples had been treated with 0.8 mM puromycin prior to cross-linking.

Binding of Skp to nascent pOmpA required release from the ribosome (Fig. 2B, compare lanes 3 and 6), suggesting either that binding sites on pOmpA-5 are not accessible while bound to the ribosome or that recognition by Skp depends on the prior formation of some secondary structure. A requirement for some structural motif of the interaction with Skp is indicated by the failure of Skp to bind to nascent chains of an inner membrane protein, mannitol permease (MtlA), under identical experimental conditions (Fig. 2B, lanes 7-12). These results demonstrate that Skp interacts with early folding intermediates of outer membrane proteins in a specific manner.

A skp Knock-out Mutant Exhibits an Early Defect in Outer Membrane Protein Assembly Leading to an Altered Structure and Physiology of the Outer Membrane-- To generate a knock-out mutant of skp an in-frame deletion removing 121 of 161 amino acids of Skp was constructed (Fig. 3). The skp mutant showed no polar effect on downstream firA (Fig. 3), which is an essential gene involved in lipid A biosynthesis (38). In contrast, skp was not found to be essential for growth under all conditions tested. The skp mutant grew normally on rich and minimal media with various carbon sources at temperatures between 20 and 42 °C.

View larger version (47K): [in this window] [in a new window]   Fig. 3.   Characterization of a skp null mutant. A depicts the organization of the E. coli chromosome in the 4-min region. Numbers reflect nucleotides beginning with the start codon of skp. The deletion is indicated by the hatched area. Arrows represent the two primers used for the PCR shown in B. B, PCR products separated on agarose gel electrophoresis and stained with ethidium bromide. The right lane contains DNA standards of the indicated sizes in nucleotides. The expected length of the PCR product derived from wild type DNA was 500, that from the skp mutant 136 plus an additional 6 nucleotides of a BamHI cleavage site introduced for ligation of the two skp flanking regions. The two other panels are immunoblots of total cells decorated with polyclonal antibodies directed against Skp and FirA. FirA was also identified by overproduction from plasmid pSLF3 (pfirA).

Chen and Henning (19) reported previously that inactivation of skp results in a reduction of outer membrane proteins. Changes in the profile of outer membrane proteins are usually paralleled by an altered permeability of the outer membrane. skp cells exhibited a higher sensitivity than wild type cells toward antibiotics that cross the lipids of the outer membrane, such as rifampicin, vancomycin, and novobiocin (Fig. 4). The increase in sensitivity was more pronounced when cells were grown on minimal medium and at lower temperature (20 °C). It was reversed by expressing Skp in the skp mutant from an extrachromosomal copy.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Increased sensitivity of skp mutant cells toward hydrophobic antibiotics. Indicated strains (pskp is pTRS7) were grown in agar plates overlayered with antibiotics-soaked filter discs. Zones of clearance were measured after growth at 36 h at 20 °C. The plotted data are the means of at least four independent assays, each with the S.D. included.

Because skp is cotranscribed with firA, whose gene product is involved in lipid A biosynthesis, we compared the LPS content of the skp mutant to that of wild type cells (data not shown). In contrast to the described LPS deficiency of the firA200 mutant (39), deletion of skp resulted in an about 2-fold higher amount of LPS, which became evident only when cells were grown on minimal medium.

Collectively, the phenotypic features of the skp mutant are rather moderate and are more pronounced upon growth in minimal medium and at low temperature. To rule out the possibility that the lack of a strong phenotype of the skp mutant was caused by the acquisition of a second site suppressor mutation the skp deletion was moved by P1 transduction into wild type cells. No change in phenotype was obtained (not shown).

It was conceivable that in whole cells, a lack of Skp remained cryptic because of compensatory periplasmic proteins. We therefore employed spheroplasts as a periplasm-free semi-in vitro model. Wild type and mutant cells each transformed with a plasmid carrying ompA under tac promoter control were converted to spheroplasts which were then induced for OmpA synthesis and labeled with [³⁵S]methionine/cysteine (Fig. 5A). Spheroplasts (S) were separated from proteins secreted during pulse labeling (P, periplasmic fraction) either by centrifugation at 16,000 × g (lanes 1 and 2, 7, and 8) or by low speed centrifugation at 6000 × g. In the latter case, the supernatant was subsequently resolved into soluble (sP) and aggregated (aP) periplasmic material by centrifugation at 16,000 × g (lanes 5 and 6, 11, and 12). Fractions thus obtained were displayed by SDS-PAGE and phosphorimaging. Upon induction with IPTG, OmpA was the major labeled protein in both spheroplasts and the periplasm. In these overproducing conditions, the spheroplasts even retained some precursor of OmpA (pOmpA) as indicated by its resistance toward proteinase K (lanes 3 and 9). Whereas 68% of newly synthesized mature OmpA was secreted from wild type spheroplasts into the medium (lanes 1 and 2), only 13% was released from the skp mutant spheroplasts (lanes 7 and 8). Despite the fact that 87% remained associated with the mutant spheroplasts, this OmpA material had been translocated across the plasma membrane, because it was completely digested by proteinase K (lane 9). Cosedimentation of OmpA with the mutant spheroplasts was not due to aggregates of OmpA formed in the absence of Skp because the same high amount of OmpA was pelletable upon low speed centrifugation (compare lanes 7 and 10). Rather, OmpA was not released from the mutant spheroplasts because of the absence of Skp. Accordingly, induction of Skp synthesis from plasmid pTRS7 during pulse labeling of skp mutant spheroplasts restored secretion of OmpA (Fig. 5B). Attempts to achieve release by adding purified Skp to spheroplasts were, however, unsuccessful (not shown). Possibly, the active conformation of Skp requires its biosynthetic association with the plasma membrane. These results clearly demonstrate an involvement of Skp in the acquirement of a soluble periplasmic conformation of newly translocated outer membrane proteins.

View larger version (55K): [in this window] [in a new window]   Fig. 5.   In the absence of Skp, newly synthesized OmpA remains attached to the plasma membrane. A, E. coli cells of wild type (wt) strain MC 4100 and the MC 4100 skp mutant each transformed with plasmid pTompA containing ompA under tac promoter control were converted to spheroplasts. Spheroplasts were induced with IPTG and pulse-labeled with [35S]methionine/cysteine. PK, proteinase K. In lanes 1 and 2, 7 and 8, spheroplasts (S) were separated from proteins secreted into the periplasm-like medium (P) by high speed centrifugation. In lanes 4-6 and 10-12, a low speed supernatant was first obtained that was subsequently separated into soluble (sP) and aggregated (aP) periplasmic material. Proteins were separated by SDS-PAGE and visualized by phosphorimaging. The band representing OmpA was used for calculating the indicated distribution between various fractions. B, except for the last two lanes, cells expressed ompA only from the chromosome. OmpA secreted into the periplasm (arrow) was identified by its comigration with the protein overproduced by plasmid pTompA. Plasmid pTRS7 encodes skp under trc promoter control. Periplasmic Skp is indicated (arrow). Spheroplasts were collected by high speed centrifugation. The identity of the periplasmic protein slightly larger than Skp is unknown. C, same as A except that wild type and mutant cells had been transformed with plasmids encoding wild type MalE and the COOH-terminally truncated MalE*. Spheroplasts were collected by high speed centrifugation.

A similar experiment was performed with cells expressing either a wild type copy of maltose binding protein or a mutant allele lacking the authentic C terminus (Fig. 5C, MalE and MalE*). With both species, the distribution between spheroplast-associated and soluble, periplasmic fraction was identical for wild type and skp mutant cells, indicating a specificity of Skp for outer membrane proteins as suggested previously (17, 20).

Skp Is a Chaperone Maintaining the Soluble State of Periplasmic Proteins-- DegP is a periplasmic protease whose function is to degrade misfolded proteins in the periplasm (Ref. 40 and references therein). It was conceivable that the effect of an inactivation of Skp was obscured by the proteolytic activity of DegP. We therefore combined the skp deletion with a degP allele, which had been inactivated by Tn10 insertion (24), by transducing tetracycline resistance and deletion of degP into the skp mutant using P1 (Fig. 6A). Two double mutants independently obtained (1 and 2) grew normally at 30 °C (Fig. 6B). In contrast to both single mutants, skp and degP::tet, double mutant cells, however, stopped growing in Luria broth medium about 3 h after shifting them to 37 °C (Fig. 6C). In other words, the near normal growth behavior of cells lacking an active DegP protease is strictly dependent on an active Skp and vice versa, a lack of Skp is tolerated as long as DegP is functional.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   The combined inactivation of skp and degP results in a synthetic temperature sensitivity. A, immunoblot with anti-Skp and anti-DegP antisera of wild type (wt), skp mutant, degP::tet mutant, and two double mutants (1 and 2). B and C, strains indicated were subcultured at a starting A580 of 0.03 in Luria broth medium and grown at 30 and 37 °C, respectively.

The skp,degP::tet double mutant finally allowed us to directly demonstrate a chaperone activity of Skp. In Fig. 7, wild type (lanes 1), skp single mutant (lanes 2), degP::tet single mutant (lanes 3), and skp,degP::tet double mutant (lanes 4 and 5) cells were converted to spheroplasts and periplasmic fractions were obtained. These were subsequently divided by high speed centrifugation into soluble and pelletable, i.e. aggregated PPPs. When grown at the permissive temperature of 30 °C, all strains exhibited the same pattern of soluble PPPs (A) as revealed by SDS-PAGE and staining with Coomassie Blue. The same held true for aggregated PPPs with the exception that this fraction was consistently found diminished in the degP::tet single mutant (B). In contrast, when cells had been shifted to 37 °C for 3 h before preparing PPPs, significantly more aggregated proteins of all sizes accumulated specifically in the periplasm of the double mutant (D). This is corroborated by a Western blot of these fractions using anti-OmpA antibodies (inset of D). Aggregation was clearly a phenotypic feature of the double mutant, because it developed gradually during growth at 37 °C (G). The periplasm of the double mutant also contained more soluble proteins (C), this phenomenon too being dependent on growth at 37 °C (F). This, however, was not due to an enhanced fragility of spheroplasts obtained from the double mutant as shown by the almost complete absence of a cytosolic protein (P48 or Ffh) from all periplasmic protein fractions prepared (E), nor was it caused by an increased synthesis of outer membrane proteins (not shown). Thus, the absence of Skp leads to an accumulation of denatured proteins in the periplasm if cells lack DegP to degrade them. These results strongly support the view that Skp functions as a chaperone maintaining periplasmic intermediates of outer membrane proteins in a soluble state.

View larger version (72K): [in this window] [in a new window]   Fig. 7.   The simultaneous lack of Skp and DegP leads to a temperature-dependent accumulation of periplasmic proteins. Overnight cultures of the strains indicated grown in Luria broth medium at 20 °C were subcultured 1:100 in fresh Luria broth medium for an additional 3 h at either 30 or 37 °C. In F and G growth was stopped at the specified times. PPPs were separated from spheroplasts by low speed centrifugation and then resolved into soluble and aggregated PPPs by centrifugation for 10 min at 16,000 × g, precipitated with trichloroacetic acid, resolved by SDS-PAGE, and stained with Coomassie Blue. D (inset) and E, immunoblots with antisera against OmpA and the cytoplasmic protein P48 (Ffh). E serves as control for the integrity of spheroplasts and the equal sizes of the samples withdrawn from each strain.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Folding of Outer Membrane Proteins in the Periplasm-- After their translocation across the plasma membrane, outer membrane proteins of Gram-negative bacteria sequentially pass through several intermediate stages characterized as unfolded monomers, folded monomers, metastable trimers, and stable trimers (summarized in Ref. 4). The data now available suggest that Skp participates as molecular chaperone early in these folding events, because first, Skp physically associates with the plasma membrane, and second, it recognizes an NH₂-terminal part of OmpA during or immediately after folding is initiated.

Skp Is Firmly Associated with the Plasma Membrane-- Purified Skp penetrates into phospholipid bilayers and acquires a partial resistance toward proteolysis by the addition of phospholipids (20). Here, we demonstrate that during purification involving sucrose gradient centrifugation, plasma membrane vesicles retain Skp bound on their lumenal membrane leaflet. These findings suggest that Skp is tightly associated with the periplasmic side of the plasma membrane also in vivo. DegP also behaves like a peripheral membrane protein, which cannot be released by osmotic shock (40) and which interacts with phospholipids (41), whereas the PPIase SurA, which is specifically involved in the conversion of unfolded to folded monomers of outer membrane proteins (8), is released from osmotically shocked cells (6).

Skp Binds to Non-native Outer Membrane Proteins-- Purified Skp forms strong cross-links with nascent pOmpA chains after release from the ribosomes. This suggests that Skp has the capacity to bind to non-native structures of outer membrane proteins such as they might exist following release from the translocon in the plasma membrane. We have detected cross-links presumably with oligomeric forms of Skp indicating that Skp might function in its active form as oligomer as suggested previously (14). For complex formation between Skp and nascent chains to occur, an NH₂-terminal stretch of OmpA encompassing the first five -strands is sufficient. On the other hand, Skp does not bind to the COOH-terminal part of OmpA (19). These findings suggest that Skp recognizes its substrates soon after or even during their appearance in the periplasm.

Skp Is Required for the Solubility of Outer Membrane Proteins in the Periplasm-- The latter notion is also consistent with the involvement of Skp in the release of newly synthesized outer membrane proteins from the plasma membrane into the periplasm as demonstrated here. Because a lack of Skp did not cause a reduction in the steady-state level of periplasmic proteins, its influence on the release of newly translocated OmpA can only be of kinetic nature. By combining a defect in skp with one in degP we demonstrate that Skp in addition is involved in the maintenance of the solubility of periplasmic proteins. This is the as yet strongest biochemical indication that Skp functions as a molecular chaperone. Other facts support this idea. (i) As shown for the inactivation of the other periplasmic chaperones SurA (7, 8), PpiD (9), Dsb proteins (7), and FkpA (7), a loss of Skp induces the ^E-response (42), which leads to an increase in the transcription of degP to enhance cleavage of transiently or globally denatured periplasmic proteins (summarized in Ref. 40). (ii) Upstream of the skp gene, a CpxR-binding consensus element has been disclosed (43). CpxR is the response regulator of a two-component system involved in the regulation of the ^E-factor (43, 44), suggesting that skp itself might be regulated by the ^E-factor, much like fkpA also (45). (iii) A lack of Skp conveys on the mutant cell an outer membrane phenotype, which is characterized by a decreased level of outer membrane proteins concomitant with an enhanced permeability toward hydrophobic substances. The same, yet quantitatively much stronger, phenotypic features are associated with mutations in surA, ppiD, and fkpA (7-9). A reduced amount of outer membrane proteins is in fact expected if periplasmic folding catalysts are compromised.

Toward an Elucidation of Skp's Chaperone Activity-- In summary, the function of Skp as periplasmic folding catalyst is indicated by several criteria: (i) a lack of Skp enhances the ^E-response and skp itself appears to be a member of the ^E-regulon; (ii) a lack of Skp is associated with fewer outer membrane proteins and an elevated outer membrane permeability much like described for other periplasmic chaperones; (iii) Skp is required for the efficient release of newly translocated outer membrane proteins from the plasma membrane; (iv) Skp is needed to maintain periplasmic proteins in solution; (v) Skp binds to non-native outer membrane proteins, a tendency that might be facilitated by its association with the plasma membrane. These findings suggest a time point of interaction between Skp and outer membrane proteins early during their folding cascade and soon after translocation into the periplasm. It had been hypothesized that Skp, due to its genomic localization in close proximity to LPS biosynthetic genes and its especially basic nature, binds LPS species and serves as an exchange factor removing LPS molecules from folded monomers of outer membrane proteins (7). However, Skp does not appear to bind to folded monomers (20) but as shown here rather to early folding stages like the unfolded monomer. This notion would be in full agreement with the fact that like Skp, early folding intermediates of outer membrane proteins are still membrane-associated (46).

	ACKNOWLEDGEMENTS

We gratefully acknowledge Dr. Michael Ehrmann for providing strains CLC198 and CAG18515 and anti-DegP antiserum and for many invaluable comments; Drs. August Böck, Thomas Maier, and Markus Pajatsch for pMAK705 and a P1 lysate and for their initial help with the preparation of the skp mutant; Dr. Karl Ludwig Schimz for antisera against SecA and OmpA; Dr. Ira Dicker for strain ID18 and anti-FirA antiserum; Dr. Howard Shuman for plasmid pLH13; Drs. Genny Gallagher and Tracy Topping for pJF2; Dr. Long-Fei Wu for many helpful discussions; and Dr. Hans Georg Koch for a critical reading of the manuscript.

	FOOTNOTES

^* This work was supported by grants from the Sonderforschungsbereiche 184 (Munich) and 388 (Freiburg) and from the Fonds der Chemischen Industrie.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ To whom correspondence should be addressed. Tel.: 49-761-203-5265; Fax: 49-761-203-5274; E-mail: mumatthi@ruf.uni-freiburg.de.

² K. Beck and M. Müller, manuscript in preparation.

	ABBREVIATIONS

The abbreviations used are: LPS, lipopolysaccharide; PPIase, peptidyl prolyl isomerase; DSS, disuccinimidyl suberate; RANC, ribosome-associated, nascent chain; INV, inside-out plasma membrane vesicle(s); PPPs, periplasmic proteins; IPTG, isopropyl--D-thiogalactopyranoside; PCR, polymerase chain reaction; PAGE, polyacrylamide gel electrophoresis; AMP-PNP, adenosine 5'-(,-imino)triphosphate.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

1.	Nikaido, H. (1996) in Escherichia coli and Salmonella: Cellular and Molecular Biology (Neidhardt, F. C., ed) , pp. 29-47, ASM Press, Washington, D. C.
2.	Weiss, M. S., Wacker, T., Weckesser, J., Welte, W., and Schulz, G. E. (1990) FEBS Lett. 267, 268-272[CrossRef][Medline] [Order article via Infotrieve]
3.	Duong, F., Eichler, J., Price, A., Leonard, M. R., and Wickner, W. (1997) Cell 91, 567-573[CrossRef][Medline] [Order article via Infotrieve]
4.	van Gelder, P., de Cock, H., and Tommassen, J. (1997) in Membrane Protein Assembly (von Heijne, G., ed) , pp. 63-82, R. G. Landes Co.
5.	Missiakas, D., and Raina, S. (1997) J. Bacteriol. 179, 2465-2471[Free Full Text]
6.	Lazar, S. W., and Kolter, R. (1996) J. Bacteriol. 178, 1770-1773[Abstract/Free Full Text]
7.	Missiakas, D., Betton, J. M., and Raina, S. (1996) Mol. Microbiol. 21, 871-884[CrossRef][Medline] [Order article via Infotrieve]
8.	Rouvière, P. E., and Gross, C. A. (1996) Genes Dev. 10, 3170-3182[Abstract/Free Full Text]
9.	Dartigalongue, C., and Raina, S. (1998) EMBO J. 17, 3968-3980[CrossRef][Medline] [Order article via Infotrieve]
10.	Geyer, R., Galanos, C., Westphal, O., and Golecki, J. R. (1979) Eur. J. Biochem. 98, 27-38[Medline] [Order article via Infotrieve]
11.	Lathe, R., Buc, H., Lecocq, J. P., and Bautz, E. K. F. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 3548-3552[Abstract/Free Full Text]
12.	Aasland, R., Coleman, J., Holck, A. L., Smith, C. L., Raetz, C. R. H., and Kleppe, K. (1988) J. Bacteriol. 170, 5916-5918[Abstract/Free Full Text]
13.	Holck, A., and Kleppe, K. (1988) Gene (Amst.) 67, 117-124[CrossRef][Medline] [Order article via Infotrieve]
14.	Holck, A., Lossius, I., Aasland, R., and Kleppe, K. (1987) Biochim. Biophys. Acta 914, 49-54[CrossRef][Medline] [Order article via Infotrieve]
15.	Koski, P., Hirvas, L., and Vaara, M. (1990) Gene (Amst.) 88, 117-120[CrossRef][Medline] [Order article via Infotrieve]
16.	Koski, P., Rhen, M., Kantele, J., and Vaara, M. (1989) J. Biol. Chem. 264, 18973-18980[Abstract/Free Full Text]
17.	Thome, B. M., Hoffschulte, H. K., Schiltz, E., and Müller, M. (1990) FEBS Lett. 269, 113-116[CrossRef][Medline] [Order article via Infotrieve]
18.	Thome, B. M., and Müller, M. (1991) Mol. Microbiol. 5, 2815-2821[CrossRef][Medline] [Order article via Infotrieve]; Correction (1992) Mol. Microbiol. 6, 1077
19.	Chen, R., and Henning, U. (1996) Mol. Microbiol. 19, 1287-1294[Medline] [Order article via Infotrieve]
20.	de Cock, H., Schäfer, U., Potgeter, M., Demel, R., Müller, M., and Tommassen, J. (1999) Eur. J. Biochem. 259, 96-103[Medline] [Order article via Infotrieve]
21.	Bothmann, H., and Plückthun, A. (1998) Nat. Biotech. 16, 376-380 [CrossRef][Medline] [Order article via Infotrieve]
22.	Oliver, D. B., and Beckwith, J. (1982) Cell 30, 311-319[CrossRef][Medline] [Order article via Infotrieve]
23.	Singer, M., Baker, T. A., Schnitzler, G., Deischel, S. M., Goel, M., Dove, W., Jaacks, K. J., Grossman, A. D., Erickson, J. W., and Gross, C. A. (1989) Microbiol. Rev. 53, 1-24[Abstract/Free Full Text]
24.	Snyder, W. B., Davis, L. J. B., Danese, P. N., Cosma, C. L., and Silhavy, T. J. (1995) J. Bacteriol. 177, 4216-4223[Abstract/Free Full Text]
25.	Dicker, I. B., and Seetharam, S. (1991) J. Bacteriol. 173, 334-344[Abstract/Free Full Text]
26.	Potgeter, M. (1994) Biochemical and Molecular-Biological Characterization of Skp from Escherichia coli.Doctoral thesis , University of Freiburg, Germany
27.	Behrmann, M., Koch, H. G., Hengelage, T., Wieseler, B., Hoffschulte, H. K., and Müller, M. (1998) J. Biol. Chem. 273, 13898-13904[Abstract/Free Full Text]
28.	Maneewannakul, S., Maneewannakul, K., and Ippen-Ihler, K. (1994) Plasmid 31, 300-307[CrossRef][Medline] [Order article via Infotrieve]
29.	Freudl, R., Schwarz, H., Klose, M., Movva, N. R., and Henning, U. (1985) EMBO J. 4, 3593-3598[Medline] [Order article via Infotrieve]
30.	Fikes, J. D., and Bassford, P. J. (1987) J. Bacteriol. 169, 2352-2359[Abstract/Free Full Text]
31.	Treptow, N. A., and Shuman, H. A. (1988) J. Mol. Biol. 202, 809-822[CrossRef][Medline] [Order article via Infotrieve]
32.	Helde, R., Wieseler, B., Wachter, E., Neubüser, A., Hoffschulte, H. K., Hengelage, T., Schimz, K. L., Stuart, R. A., and Müller, M. (1997) J. Bacteriol. 179, 4003-4012[Abstract/Free Full Text]
33.	Hanes, J., and Plückthun, A. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 4937-4942[Abstract/Free Full Text]
34.	Hamilton, C. M., Aldea, M., Washburn, B. K., Babitzke, P., and Kushner, S. R. (1989) J. Bacteriol. 171, 4617-4622[Abstract/Free Full Text]
35.	Vogel, H. J., and Bonner, D. M. (1956) J. Biol. Chem. 218, 97-106[Free Full Text]
36.	Ahrem, B., Hoffschulte, H. K., and Müller, M. (1989) J. Cell Biol. 108, 1637-1646[Abstract/Free Full Text]
37.	Wieseler, B., and Müller, M. (1993) Mol. Microbiol. 7, 167-176[CrossRef][Medline] [Order article via Infotrieve]
38.	Kelly, T. M., Stachula, S. A., Raetz, C. R. H., and Anderson, M. S. (1993) J. Biol. Chem. 268, 19866-19874[Abstract/Free Full Text]
39.	Roy, A. M., and Coleman, J. (1994) J. Bacteriol. 176, 1639-1646[Abstract/Free Full Text]
40.	Pallen, M. J., and Wren, B. W. (1997) Mol. Microbiol. 26, 209-221[CrossRef][Medline] [Order article via Infotrieve]
41.	Skórko-Glonek, J., Lipiska, B., Krzewski, K., Zolese, G., Bertoli, E., and Tanfani, F. (1997) J. Biol. Chem. 272, 8974-8982[Abstract/Free Full Text]
42.	Missiakas, D., and Raina, S. (1998) Mol. Microbiol. 28, 1059-1066[CrossRef][Medline] [Order article via Infotrieve]
43.	Pogliano, J., Lynch, A. S., Belin, D., Lin, E. C. C., and Beckwith, J. (1997) Genes Dev. 11, 1169-1182[Abstract/Free Full Text]
44.	Danese, P. N., Snyder, W. B., Cosma, C. L., Davis, L. J. B., and Silhavy, T. J. (1995) Genes Dev. 9, 387-398[Abstract/Free Full Text]
45.	Danese, P. N., and Silhavy, T. J. (1997) Genes Dev. 11, 1183-1193[Abstract/Free Full Text]
46.	Freudl, R., Schwarz, H., Stierhof, Y. D., Gamon, K., Hindennach, I., and Henning, U. (1986) J. Biol. Chem. 261, 11355-11361[Abstract/Free Full Text]