Lipopolysaccharide Transport to the Bacterial Outer Membrane in Spheroplasts
2004; Elsevier BV; Volume: 280; Issue: 6 Linguagem: Inglês
10.1074/jbc.m409259200
ISSN1083-351X
AutoresBoris Tefsen, Jeroen Geurtsen, Frank Beckers, Jan Tommassen, Hans de Cock,
Tópico(s)Escherichia coli research studies
ResumoThe mechanism of lipopolysaccharide (LPS) transport in Gram-negative bacteria from the inner membrane to the outer membrane is largely unknown. Here, we investigated the possibility that LPS transport proceeds via a soluble intermediate associated with a periplasmic chaperone analogous to the Lol-dependent transport mechanism of lipoproteins. Whereas newly synthesized lipoproteins could be released from spheroplasts of Escherichia coli upon addition of a periplasmic extract containing LolA, de novo synthesized LPS was not released. We demonstrate that LPS synthesized de novo in spheroplasts co-fractionated with the outer membranes and that this co-fractionation was dependent on the presence in the spheroplasts of a functional MsbA protein, the protein responsible for the flip-flop of LPS across the inner membrane. The outer membrane localization of the LPS was confirmed by its modification by the outer membrane enzyme CrcA (PagP). We conclude that a substantial amount of LPS was translocated to the outer membrane in spheroplasts, suggesting that transport proceeds via contact sites between the two membranes. In contrast to LPS, de novo synthesized phospholipids were not transported to the outer membrane in spheroplasts. Apparently, LPS and phospholipids have different requirements for their transport to the outer membrane. The mechanism of lipopolysaccharide (LPS) transport in Gram-negative bacteria from the inner membrane to the outer membrane is largely unknown. Here, we investigated the possibility that LPS transport proceeds via a soluble intermediate associated with a periplasmic chaperone analogous to the Lol-dependent transport mechanism of lipoproteins. Whereas newly synthesized lipoproteins could be released from spheroplasts of Escherichia coli upon addition of a periplasmic extract containing LolA, de novo synthesized LPS was not released. We demonstrate that LPS synthesized de novo in spheroplasts co-fractionated with the outer membranes and that this co-fractionation was dependent on the presence in the spheroplasts of a functional MsbA protein, the protein responsible for the flip-flop of LPS across the inner membrane. The outer membrane localization of the LPS was confirmed by its modification by the outer membrane enzyme CrcA (PagP). We conclude that a substantial amount of LPS was translocated to the outer membrane in spheroplasts, suggesting that transport proceeds via contact sites between the two membranes. In contrast to LPS, de novo synthesized phospholipids were not transported to the outer membrane in spheroplasts. Apparently, LPS and phospholipids have different requirements for their transport to the outer membrane. The cell envelope of Gram-negative bacteria consists of an inner (IM) 1The abbreviations used are: IM, inner membrane; OM, outer membrane; OMP, OM protein; LPS, lipopolysaccharide; Kdo, 3-deoxy-d-manno-octulosonic acid; ts, temperature-sensitive; SV, synthetic minimal medium.1The abbreviations used are: IM, inner membrane; OM, outer membrane; OMP, OM protein; LPS, lipopolysaccharide; Kdo, 3-deoxy-d-manno-octulosonic acid; ts, temperature-sensitive; SV, synthetic minimal medium. and an outer membrane (OM) separated by the peptidoglycan-containing periplasm. The lipidic components of the OM, phospholipids, lipopolysaccharide (LPS), and lipoproteins, are all synthesized in the IM. Because of their hydrophobic lipid moiety, transport directly through the aqueous periplasm is thermodynamically unfavorable. Therefore, the transport process is expected to require a special machinery, including, for example, a soluble periplasmic chaperone that shields the lipid moiety. Alternatively, transport might proceed directly from IM to OM via a macromolecular complex bridging the periplasm. The existence of contacts between IM and OM has been reported previously (1Bayer M.E. J. Gen. Microbiol. 1968; 53: 395-404Crossref PubMed Scopus (180) Google Scholar). LPS can be divided in a hydrophobic lipid A part and a hydrophilic part, consisting of the oligosaccharide core and an O-antigen. In Escherichia coli K-12, lipid A consists of a glucosamine dimer substituted with, in general, six acyl chains. The core of the LPS consists of 3-deoxy-d-manno-octulosonic acid (Kdo) residues, which are attached to lipid A, and other sugars (2Raetz C.R. Whitfield C. Annu. Rev. Biochem. 2002; 71: 635-700Crossref PubMed Scopus (3358) Google Scholar). The biosynthesis is well described (2Raetz C.R. Whitfield C. Annu. Rev. Biochem. 2002; 71: 635-700Crossref PubMed Scopus (3358) Google Scholar), and, recently, considerable progress has been made in understanding the mechanism of its transport from the site of synthesis to its destination. The translocation of the lipid A core moiety of LPS across the IM requires the MsbA protein (3Zhou Z. White K.A. Polissi A. Georgopoulos C. Raetz C.R. J. Biol. Chem. 1998; 273: 12466-12475Abstract Full Text Full Text PDF PubMed Scopus (281) Google Scholar, 4Doerrler W.T. Reedy M.C. Raetz C.R. J. Biol. Chem. 2001; 276: 11461-11464Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar, 5Doerrler W.T. Gibbons H.S. Raetz C.R. J. Biol. Chem. 2004; 279: 45102-45109Abstract Full Text Full Text PDF PubMed Scopus (196) Google Scholar). When a temperature-sensitive (ts) msbA mutant was shifted to the restrictive temperature, LPS was found to accumulate in the IM (4Doerrler W.T. Reedy M.C. Raetz C.R. J. Biol. Chem. 2001; 276: 11461-11464Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar). MsbA is a transporter of the ATP-binding cassette protein family, and the presence of (Kdo)2-lipid A, a precursor of LPS, increased the ATPase activity of MsbA reconstituted in liposomes (6Doerrler W.T. Raetz C.R. J. Biol. Chem. 2002; 277: 36697-36705Abstract Full Text Full Text PDF PubMed Scopus (174) Google Scholar). The crystal structure of MsbA has been solved (7Chang G. Roth C.B. Science. 2001; 293: 1793-1800Crossref PubMed Scopus (584) Google Scholar). Additionally, it was recently shown that an OM protein designated Imp is required for the correct localization of LPS at the cell surface of Neisseria meningitidis. This protein is likely involved in a very late step in LPS transport, presumably the flip-flop from the inner to the outer leaflet of the OM (8Bos M.P. Tefsen B. Geurtsen J. Tommassen J. Proc. Natl. Acad. Sci. U. S. A. 2004; 101: 9417-9422Crossref PubMed Scopus (203) Google Scholar). However, the transport of LPS through the aqueous periplasm remains enigmatic. Potentially, LPS could be transported across the periplasm in a way analogous to lipoprotein transport. The transport of lipoproteins, such as Braun's lipoprotein, also known as Lpp, to the OM is reasonably well understood nowadays. Lipoproteins are N-terminally lipidated (9Gupta S.D. Wu H.C. FEMS Microbiol. Lett. 1991; 62: 37-41Crossref PubMed Scopus (46) Google Scholar, 10Sankaran K. Wu H.C. J. Biol. Chem. 1994; 269: 19701-19706Abstract Full Text PDF PubMed Google Scholar) after translocation across the IM via the Sec machinery and then recruited by the IM-located LolCDE complex, which, like MsbA, belongs to the ATP-binding cassette transporter superfamily (11Yakushi T. Masuda K. Narita S. Matsuyama S. Tokuda H. Nat. Cell. Biol. 2000; 2: 212-218Crossref PubMed Scopus (204) Google Scholar). This complex utilizes ATP to enable the periplasmic chaperone LolA to form a complex with the lipoprotein (11Yakushi T. Masuda K. Narita S. Matsuyama S. Tokuda H. Nat. Cell. Biol. 2000; 2: 212-218Crossref PubMed Scopus (204) Google Scholar, 12Matsuyama S. Tajima T. Tokuda H. EMBO J. 1995; 14: 3365-3372Crossref PubMed Scopus (164) Google Scholar). After transport across the periplasm, lipoproteins are transferred from the LolA-lipoprotein complex to the receptor LolB in the OM (13Matsuyama S. Yokota N. Tokuda H. EMBO J. 1997; 16: 6947-6955Crossref PubMed Scopus (171) Google Scholar). In this study, we have investigated whether LPS can be released from spheroplasts by periplasmic components. However, we found that LPS is transported to the OM in spheroplasts, suggesting that transport proceeds via contact sites between the membranes. Strains and Media—E. coli strains used in this study are listed in Table I. Cells were grown in LB medium (14Tommassen J. van Tol H. Lugtenberg B. EMBO J. 1983; 2: 1275-1279Crossref PubMed Scopus (117) Google Scholar), synthetic minimal medium (SV) (15Lugtenberg B. Peters R. Bernheimer H. Berendsen W. Mol. Gen. Genet. 1976; 147: 251-262Crossref PubMed Scopus (175) Google Scholar) supplemented with 0.2% glucose, and 10 μg/ml thiamine or on LB agar plates. Ampicillin was supplied, where necessary, at a final concentration of 100 μg/ml. Experiments with genetically modified organisms were performed under license number GGO 99-139/2.Table IE. coli strains and plasmidsStrain or plasmidRelevant characteristicsaKmr, kanamycin-resistant; Apr, ampicillin-resistant.Source or Ref.Strains Top10 F′hsdRMS recA1Invitrogen DH5ahsdR17 recA135Hanahan D. J. Mol. Biol. 1983; 166: 557-580Crossref PubMed Scopus (8170) Google Scholar MC4100K-12 laboratory strain36Casadaban M.J. J. Mol. Biol. 1976; 104: 541-555Crossref PubMed Scopus (1301) Google Scholar O111::B4Smooth laboratory straincContains LPS molecules with an O-antigen.37Coleman Jr., W.G. Goebel P.J. Leive L. J. Bacteriol. 1977; 130: 656-660Crossref PubMed Google Scholar W3110K-12 laboratory strain4Doerrler W.T. Reedy M.C. Raetz C.R. J. Biol. Chem. 2001; 276: 11461-11464Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar WD2ts msbA mutant derivative of W31104Doerrler W.T. Reedy M.C. Raetz C.R. J. Biol. Chem. 2001; 276: 11461-11464Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar WD201ts msbA pmrA mutant derivative of W31105Doerrler W.T. Gibbons H.S. Raetz C.R. J. Biol. Chem. 2004; 279: 45102-45109Abstract Full Text Full Text PDF PubMed Scopus (196) Google Scholar CAG12077crcA::Tn10CGSCbE. coli Genetic Stock Center.Plasmids pCRII-TOPOCloning vector, Kmr, AprInvitrogen pCRII-TOPO-crcApCRII-TOPO containing crcAThis study pET11aExpression vector, AprNovagen pET11a-crcApET11a containing crcAThis study pMMB67EHExpression vector, Apr38Furste J.P. Pansegrau W. Frank R. Blocker H. Scholz P. Bagdasarian M. Lanka E. Gene (Amst.). 1986; 48: 119-131Crossref PubMed Scopus (813) Google Scholar pMMB67EH-crcApMMB67EH containing crcAThis studya Kmr, kanamycin-resistant; Apr, ampicillin-resistant.b E. coli Genetic Stock Center.c Contains LPS molecules with an O-antigen. Open table in a new tab Plasmid Constructions—The plasmids used in this study are listed in Table I. To clone the crcA gene, a DNA fragment containing this gene was amplified from genomic DNA obtained from E. coli K-12 strain MC4100 by PCR performed in a volume of 50 μl with 1.75 units of Expand High Fidelity enzyme mix (Roche Applied Science), 0.2 mm dNTPs, 1.5% dimethyl sulfoxide and 25 pmol of primers (Isogen) with the sequences AACATATGAACGTGAGTAAATATGTCG (NdeI restriction site is underlined) and AAGGATCCTCAAAACTGAAAGCGCATC (BamHI site is underlined). After initial denaturation for 3 min at 94 °C, 30 cycles of 1 min at 94 °C, 1 min at 55 °C, and 90 s at 72 °C were performed, followed by 10 min at 72 °C. The PCR product was cloned into pCRII-TOPO according to manufacturer's protocol (Invitrogen). From the resulting plasmid, pCRII-TOPO-crcA, an NdeI-BamHI fragment was cloned into pET11a, and from the resulting construct the sequence of crcA was confirmed. An XbaI-HindIII fragment from pET11a-crcA, including the ribosome-binding site of pET11a, was cloned into pMMB67EH, yielding pMMB67EH-crcA. Preparation of Spheroplasts—Cultures, pregrown overnight at 37 °C in either SV or LB medium for subsequent protein or LPS labeling, respectively, were diluted 10-fold in the same medium and grown to an optical density at 660 nm (A660) between 0.4 and 0.7. Cells from a 10-ml culture were converted to spheroplasts as described (16Osborn M.J. Gander J.E. Parisi E. J. Biol. Chem. 1972; 247: 3973-3986Abstract Full Text PDF PubMed Google Scholar). Briefly, the cells were pelleted and washed in 2 ml of 10 mm Tris-HCl (pH 7.5), 0.75 m sucrose and resuspended in 500 μl of the same buffer. Subsequently, 25 μl of lysozyme (2 mg/ml) and 1050 μl of 1.5 mm EDTA were added, and the mixture was incubated on ice for 10 to 20 min. After this procedure, more than 99% of the cells were converted to spheroplasts as evaluated under a light microscope. Isolation of Periplasmic Fractions—MC4100 cells from five 1-liter cultures grown in LB at 37 °C to an A660 of ∼0.6 were pelleted and converted to spheroplasts as described above in a total volume of 1.2 liters. After removal of the spheroplasts by centrifugation in an SLA1500 rotor (Beckman Coulter) at 10,000 rpm for 10 min at 4 °C and of residual membranes by ultracentrifugation in a 60 Ti rotor (Beckman Coulter) at 38,000 rpm for 30 min at 4 °C, the supernatant was concentrated to ∼15 ml in an Amicon 8400 concentrator using a 10-kDa cut-off polyethersulfone filter and nitrogen gas. The periplasmic fraction was finally dialyzed against 20 mm Tris-HCl (pH 8.0) at 4 °C. Aliquots were immediately frozen in liquid nitrogen, stored at –80 °C and used only once. Quantitative Western Blotting—A dilution series of purified LolA with known protein concentration was loaded as a standard next to a dilution series of periplasmic extracts on an SDS-polyacrylamide gel, and the proteins were blotted onto a nitrocellulose membrane (17Sambrook J. Fritsch E.F. Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY1989Google Scholar). After immunodetection with antiserum directed against LolA, peroxidase-conjugated secondary antibody, H2O2, and 4-chloro-1-naphthol (18Foster R.G. Provencio I. Bovee-Geurts P.H. DeGrip W.J. J. Neuroendocrinol. 2003; 15: 355-363Crossref PubMed Scopus (13) Google Scholar), the concentration of LolA in the periplasmic extracts was estimated by comparing the intensity of the bands with the standard. Spheroplast Release Assay—One-tenth of a spheroplast suspension was added to 375 μl of SV supplemented with 0.25 m sucrose. A concentrated periplasmic fraction (usually 50 μl), purified LolA (∼4 μg), or Tris-HCl (pH 8.0) was added to the mixture. To label de novo synthesized proteins, the spheroplasts were incubated with 5 μCi [35S] cysteine/methionine (Promix label) for 3 min at 30 °C followed by a chase for 4 min with 2 μl of 0.1 m unlabeled cysteine/methionine. LPS and phospholipids were labeled by incubating the spheroplasts for 4 min at 37 °C with 2 μCi of [1-14C]sodium acetate, followed by a chase with 10 μl of 0.48 m unlabeled sodium acetate for 4 min. Alternatively, LPS was labeled by incubation of the spheroplasts with 2 μCi N-acetyl-d-[1-14C]glucosamine for 25 min. All radioactive compounds were purchased from Amersham Biosciences. After labeling, spheroplasts were chilled on ice for at least 1 min and centrifuged at 16,400 × g in an Eppendorf centrifuge to separate the spheroplasts from the medium. Spheroplasts were resuspended in water, and hereafter, the proteins present in the spheroplast and supernatant fractions were precipitated by adding 0.10 volume of 70% (w/w) trichloroacetic acid. Precipitated proteins were resuspended in SDS buffer followed by immunoprecipitation as described (19Bosch D. de Boer P. Bitter W. Tommassen J. Biochim. Biophys. Acta. 1989; 979: 69-76Crossref PubMed Scopus (34) Google Scholar) with polyclonal rabbit antisera directed against Lpp or OmpC and analysis of the precipitated proteins by SDS-PAGE (20Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207018) Google Scholar) and subsequent autoradiography. Notably, spheroplasts lost the ability to release Lpp in the presence of LolA and of OM proteins (OMPs) into the medium if they were collected by centrifugation before the labeling procedure (data not shown). Purification of Radiolabeled LPS—E. coli O111::B4 cells were labeled overnight in 10 ml LB medium with 2 μCi of [1-14C]sodium acetate. Cells were collected by centrifugation for 10 min at 5000 rpm in an Eppendorf centrifuge, and LPS was extracted with the hot-phenol method (21Jann K. Jann B. Orskov F. Orskov I. Westphal O. Biochem. Z. 1965; 342: 1-22PubMed Google Scholar). The LPS concentration was determined by measuring the Kdo content of the samples (22van Alphen L. Verkleij A. Leunissen-Bijvelt J. Lugtenberg B. J. Bacteriol. 1978; 134: 1089-1098Crossref PubMed Google Scholar). Isopycnic Sucrose Gradient Centrifugation—After labeling for 1 min with 2 μCi of [1-14C]sodium acetate and a 30-min chase period with 15 μl of 0.48 m unlabeled acetate, cells or spheroplasts were resuspended in 1 ml of 10 mm Tris-HCl, 1 mm EDTA (pH 8.0). Alternatively, 10 ml of mid-log cells were converted into spheroplasts as described and resuspended into LB containing 0.25 m sucrose (end volume of 22 ml). Spheroplasts or intact cells were labeled for 25 min with 80 μCi of [1-14C]sodium acetate. The cells or spheroplasts were broken by two 40-watt pulses of 10 s each using a microtip on a Sonifier B-12 (Branson Sonic Power Co.) and cooling on ice between the pulses. The sonicated samples were applied to the top of a discontinuous sucrose gradient formed by layering sucrose solutions (buffered with 10 mm Tris-HCl, 1 mm EDTA (pH 7.5)) into a centrifuge tube as follows: 2 ml of 56% sucrose; 4.75 ml of 42% sucrose and 4.25 ml of 25% sucrose (all w/w). After 22.5-h centrifugation at 38,000 rpm in a Beckman SW41 rotor at 4 °C, 1-ml fractions were collected starting from the top (23Thom J.R. Randall L.L. J. Bacteriol. 1988; 170: 5654-5661Crossref PubMed Google Scholar). Lactate dehydrogenase activity (24Osborn M.J. Gander J.E. Parisi E. Carson J. J. Biol. Chem. 1972; 247: 3962-3972Abstract Full Text PDF PubMed Google Scholar) was used as an IM marker. The porins, visualized by Coomassie staining after SDS-PAGE (20Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207018) Google Scholar), were used as a marker for the OM. LPS Analysis—LPS was routinely analyzed on SDS-polyacrylamide gels containing 14% acrylamide and subsequently visualized by silver staining (25Tsai C.M. Frasch C.E. Anal. Biochem. 1982; 119: 115-119Crossref PubMed Scopus (2309) Google Scholar) and/or autoradiography. LPS from sucrose fractions was precipitated by adding 0.10 volume of 70% trichloroacetic acid. Pellets were resuspended in 160 μl of phosphate-buffered saline (pH 7.4), and lipid A was subsequently isolated as described (26Zhou Z. Ribeiro A.A. Lin S. Cotter R.J. Miller S.I. Raetz C.R. J. Biol. Chem. 2001; 276: 43111-43121Abstract Full Text Full Text PDF PubMed Scopus (182) Google Scholar). The total amount of radioactive lipid A in each fraction was determined by scintillation counting in 5 ml of Lipoluma Plus (LUMAC) using a 1209 Rackbeta LSC (LKB Wallac) and/or analyzed by TLC as described (26Zhou Z. Ribeiro A.A. Lin S. Cotter R.J. Miller S.I. Raetz C.R. J. Biol. Chem. 2001; 276: 43111-43121Abstract Full Text Full Text PDF PubMed Scopus (182) Google Scholar). Analysis of CrcA/PagP-mediated LPS Modifications—Cells containing pMMB67EH-crcA or the empty vector pMMB67EH were pregrown in LB medium at 37 °C, diluted 10-fold in 5 ml of LB containing 1 mm isopropyl β-d-thiogalactopyranoside to induce crcA expression, and grown for 80 min when the A660 reached ∼0.5. Hereafter, the cells were converted to spheroplasts. These spheroplasts or an equivalent amount of intact cells were resuspended into 3.75 ml of LB medium containing 0.25 m sucrose and labeled for 25 min with 20 μCi of [1-14C]sodium acetate at 37 °C. Pellets obtained from labeled cells or spheroplasts were resuspended in 160 μl phosphate-buffered saline (pH 7.4). Alternatively, pellets were resuspended, sonicated, and fractionated as described in isopycnic sucrose gradient centrifugation section. Phospholipid Analysis—Phospholipids were isolated from resuspended spheroplasts or medium fractions obtained after spheroplast release assays or from 200 μl of 1-ml sucrose fractions using a two-phase lipid extraction method (27Bligh E.G. Dyer W.J. Can. J. Med. Sci. 1959; 37: 911-917Google Scholar). The total amount of radioactive phospholipids was determined by scintillation counting in 5 ml of Lipoluma Plus (LUMAC) using a 1209 Rackbeta LSC (LKB Wallac). Phospholipids were routinely separated by TLC. The plates (silica gel 60, 20 × 10 cm, Merck) were developed with chloroform/methanol/acetic acid at a ratio of 65:25:10 and subjected to autoradiography. Release of Lipoproteins from Spheroplasts—Previously, it has been demonstrated that de novo synthesized lipoproteins can be released from spheroplasts upon addition of concentrated periplasmic extracts containing the chaperone LolA (12Matsuyama S. Tajima T. Tokuda H. EMBO J. 1995; 14: 3365-3372Crossref PubMed Scopus (164) Google Scholar). Here, we wished to determine whether such extracts contain a chaperone for the release of LPS as well. Thus, concentrated periplasmic extracts were prepared as described and first tested for their activity in the release of Lpp. Indeed, de novo synthesized Lpp was released from the spheroplasts into the medium, specifically and efficiently, upon addition of either purified LolA or a concentrated periplasmic fraction (Fig. 1). With different batches of periplasmic extracts, the efficiency of Lpp release was variable (ranging from 30 to 80%). Quantification by Western blotting revealed that the amount of LolA in these periplasmic fractions correlated with their efficiency in releasing Lpp (data not shown). Thus, de novo synthesized Lpp was released from spheroplasts in a LolA-dependent manner. Failure to Release LPS from Spheroplasts—We subsequently investigated the possibility to release newly synthesized LPS from spheroplasts by addition of periplasmic extracts. Spheroplasts were incubated with [1-14C]sodium acetate to label, unspecifically, LPS. After labeling, the spheroplasts were separated from the medium by centrifugation and both fractions were analyzed by SDS-PAGE and autoradiography. All newly synthesized LPS remained associated with the spheroplasts both in the absence and in the presence of concentrated periplasmic fractions (Fig. 2). Even much further concentrated periplasmic fractions did not stimulate LPS release (data not shown). Similar results were obtained when LPS was specifically labeled with N-acetyl-d-[1-14C]glucosamine, a precursor of lipid A (data not shown). Since acetate enters the general metabolic pathway, also proteins were labeled with [14C]acetate. Several de novo synthesized proteins were released from spheroplasts even in the absence of a periplasmic extract (Fig. 2), consistent with the previously reported release of periplasmic proteins and OMPs from spheroplasts (28Sen K. Nikaido H. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 743-747Crossref PubMed Scopus (65) Google Scholar). By immunoprecipitation, one of the released proteins could be identified as porin OmpC (data not shown). At least two proteins, indicated with an asterisk in Fig. 2, were released only in the presence of periplasmic extract. Most likely, they represent lipoproteins. The results of these experiments suggest that LPS cannot be released (in a soluble form) from spheroplasts by periplasmic components. An alternative explanation could be that LPS after being released from spheroplasts forms aggregates that co-fractionate with the spheroplasts. To test this possibility, different amounts of purified 14C-labeled LPS were added to a spheroplasts suspension, and we investigated whether this LPS formed aggregates that co-fractionated with spheroplasts. However, after centrifugation, more than 98% of the labeled LPS molecules remained in the supernatant fractions. Therefore, we conclude that de novo synthesized LPS molecules remain associated with spheroplasts in the absence and presence of periplasmic components. Localization of de Novo Synthesized LPS in Spheroplasts—It has been described that part of the OM remains attached to the IM in spheroplasts (29Birdsell D.C. Cota-Robles E.H. J. Bacteriol. 1967; 93: 427-437Crossref PubMed Google Scholar). Therefore, we considered the possibility that LPS, synthesized de novo in spheroplasts, might still be transported to the OM. To investigate this possibility, pulse-labeled spheroplasts generated from strain MC4100 were disrupted by sonication and fractionated by isopycnic sucrose gradient centrifugation. More then 90% of the IM marker lactate dehydrogenase was localized in the low density membrane fractions, whereas the porins were mainly present in the high density membrane fractions (data not shown). Also most LPS molecules, which were synthesized prior to the conversion of the cells to spheroplasts, were found in the high density membrane fractions when analyzed on a silver-stained gel (data not shown). Interestingly, the radiolabeled LPS molecules were equally distributed among the low density (inner) membrane fractions and the high density (outer) membrane fractions (Fig. 3). Apparently, a substantial proportion of the de novo synthesized LPS molecules was translocated to the (remnants of the) OM in the spheroplasts. The addition of concentrated periplasmic extracts during the labeling did not affect the localization of the newly synthesized LPS (data not shown). Inactivation of MsbA Prevents the OM Localization of de Novo Synthesized LPS in Spheroplasts—To exclude any artifacts in the procedure described in the previous section, we determined whether the OM localization of LPS in spheroplasts was dependent on the MsbA protein, which is required for the transport of LPS across the IM. When spheroplasts from the ts msbA mutant strain WD201 were labeled with [14C]acetate at the permissive temperature, the newly synthesized LPS fractionated mainly in the high density OM fractions in a sucrose gradient (Fig. 4), similarly as observed in spheroplasts from MC4100 (as described above). However, when the spheroplasts were labeled after 30-min incubation at the restrictive temperature (44 °C), the de novo synthesized LPS accumulated in the low density membrane fractions corresponding to the IM (Fig. 4). Newly synthesized LPS in spheroplasts generated from the isogenic wild-type strain W3110 that were similarly treated and labeled at 44 °C fractionated normally in the OM fractions of a sucrose gradient (data not shown). This demonstrates that LPS transport to the OM in spheroplasts is dependent on a functional MsbA protein as described previously and was confirmed in the present study (data not shown) for intact cells (4Doerrler W.T. Reedy M.C. Raetz C.R. J. Biol. Chem. 2001; 276: 11461-11464Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar). De Novo Synthesized LPS Molecules Are Modified by CrcA in Spheroplasts—To confirm that LPS can indeed be transported to the OM in spheroplasts, we determined whether the OM protein CrcA could modify these LPS molecules. CrcA, also known as PagP, transfers a palmitate residue from the sn-1 position of a phospholipid to the N-linked hydroxymyristate on the proximal unit of lipid A (30Bishop R.E. Gibbons H.S. Guina T. Trent M.S. Miller S.I. Raetz C.R. EMBO J. 2000; 19: 5071-5080Crossref PubMed Scopus (280) Google Scholar). From its recently solved structure, CrcA was predicted to have its active site on the edge of the outer leaflet of the OM (31Hwang P.M. Choy W.Y. Lo E.I. Chen L. Forman-Kay J.D. Raetz C.R. Prive G.G. Bishop R.E. Kay L.E. Proc. Natl. Acad. Sci. U. S. A. 2002; 99: 13560-13565Crossref PubMed Scopus (278) Google Scholar). To apply this approach, cells producing CrcA (CAG12077-pMMB67EH-crcA) or not (CAG12077-pMMB67EH) were converted to spheroplasts and labeled with [14C]acetate. Subsequently, their lipid A species were isolated and analyzed by TLC. Indeed, hepta-acylated lipid A species were produced abundantly in spheroplasts from cells expressing CrcA (indicated by arrows in Fig. 5), and these species were absent in those of the crcA disruption mutant CAG12077 carrying the empty vector (Fig. 5). The CrcA-modified lipid A species that were observed in the spheroplasts migrated to the same position as the lipid A species produced in labeled intact cells (Fig. 5). When these labeled spheroplasts were fractionated on a sucrose gradient, and the different fractions were analyzed for their lipid A species, the majority (>90%) of the CrcA-modified lipid A species was present in the high density OM fractions (data not shown), whereas the lipid A species in the IM fractions were not modified by CrcA. In conclusion, the results show that CrcA indeed modifies newly synthesized LPS molecules produced in spheroplasts. Since these modifications occur in the OM, these results confirm that LPS is indeed transported to the OM in spheroplasts. Transport of de Novo Synthesized Phospholipids Is Defective in Spheroplasts—In the previous sections, we demonstrated that LPS is transported to the OM in spheroplasts. We subsequently investigated the fate of newly synthesized phospholipids in spheroplasts. When intact cells of the msbA mutant WD2 were labeled at the permissive temperature of 30 °C with [14C]acetate, and the membranes were separated by isopycnic sucrose gradient centrifugation, de novo synthesized phospholipid molecules were found in both the low density (IM) and high density membrane (OM) fractions (Fig. 6). In contrast, in spheroplasts, the vast majority of de novo synthesized phospholipids localized in the low density membrane fractions (Fig. 6). Similar results were found when the wild-type strains W3110 and MC4100 were used (data not shown). These results demonstrate that the transport of newly synthesized phospholipids is defective in spheroplasts. Possibly, the transport of phospholipids to the OM requires periplasmic components analogous to LolA that were removed by spheroplast formation. To evaluate this possibility, we investigated whether phospholipids were released when spheroplasts were labeled in the presence of a periplasmic extract. However, these extracts did not enhance the release of phospholipids from the spheroplasts (data not shown). In addition, periplasmic extracts did not change the distribution of newly synthesized phospholipids between IM and OM of the spheroplasts (data not shown). Taken together, we conclude that phospholipid transport to the OM is defective in spheroplasts, where at least a significant proportion of the LPS molecules is still transported to the OM. Thus, LPS and phospholipids appear to have different requirements for their transport through the periplasm. The transport of lipoproteins from the periplasmic side of the IM to the OM via the Lol system has been characterized in considerable detail (12Matsuyama S. Tajima T. Tokuda H. EMBO J. 1995; 14: 3365-3372Crossref PubMed Scopus (164) Google Scholar, 13Matsuyama S. Yokota N. Tokuda H. EMBO J. 1997; 16: 6947-6955Crossref PubMed Scopus (171) Google Scholar, 32Yokota N. Kuroda T. Matsuyama S. Tokuda H. J. Biol. Chem. 1999; 274: 30995-30999Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar). In this study, we have investigated the possibility that LPS is transported to the OM via a similar mechanism, a possibility also suggested by others (see, e.g. reference (33Reuter G. Janvilisri T. Venter H. Shahi S. Balakrishnan L. van Veen H.W. J. Biol. Chem. 2003; 278: 35193-35198Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar). A putative periplasmic chaperone for LPS might shield the hydrophobic acyl chains of the LPS molecules from the aqueous periplasm by forming a complex that could traverse the periplasm and target to the OM. Our results demonstrate that LPS, in contrast to Lpp, could not be released in a soluble form from spheroplasts by periplasmic components. We expect that the periplasmic fractions were concentrated enough to contain sufficient amounts of this putative component. Since the OM contains much more LPS molecules than lipoproteins, one would expect this putative LPS chaperone to be at least as abundant as LolA. The latter was present in sufficient amounts in the isolated periplasmic fractions to release de novo synthesized Lpp efficiently. OmpA and the porins were released from the spheroplasts independent of the presence of a periplasmic extract. If LPS were released from spheroplasts, independent of a periplasmic chaperone, it would possibly form aggregates that could co-fractionate with the spheroplasts. However, this possibility is unlikely since purified LPS, added to spheroplasts, remained in the culture supernatant. Thus, it appears that LPS is not released from spheroplasts, neither in the presence nor in the absence of periplasmic extracts. This indicates that LPS is transported via a route different from that of lipoproteins as wells as from that of integral OMPs. Previously, electron microscopy studies showed that large parts of the OM remain attached to the IM of spheroplasts (29Birdsell D.C. Cota-Robles E.H. J. Bacteriol. 1967; 93: 427-437Crossref PubMed Google Scholar), presumably via some sort of contact sites. When we localized the newly synthesized LPS produced by the spheroplasts, we observed that a large proportion of these LPS molecules co-fractionated with (remnants of) the OM. The distribution of the LPS molecules was not affected by the addition of periplasmic components. To demonstrate that the apparent OM localization of the LPS was not due to aberrant fractionation behavior of LPS-containing IM vesicles, we demonstrated that the transport of de novo synthesized LPS molecules to the OM in spheroplasts was dependent on a functional MsbA protein. Thus, it appears that transport of LPS in spheroplasts proceeds via the pathway that is normally used in intact cells. Furthermore, we confirmed the correct localization of newly synthesized LPS molecules to the OM of spheroplasts by demonstrating that these molecules were modified by CrcA in the OM. Together, this shows that transport of LPS is possible in spheroplasts, presumably via the contact sites between the IM and OM. Interestingly, a role for the zones of adhesion between IM and OM in the transport of LPS has been suggested previously, based on the observation that newly synthesized LPS molecules appeared at specific connection sites between IM and OM in freeze-etched Salmonella typhimurium (34Mühlradt P.F. Menzel J. Golecki J.R. Speth V. Eur. J. Biochem. 1973; 35: 471-481Crossref PubMed Scopus (101) Google Scholar). Although the nature of such contact sites is unknown, they might be formed by proteins. Parts of the OM are known to remain attached to spheroplasts (29Birdsell D.C. Cota-Robles E.H. J. Bacteriol. 1967; 93: 427-437Crossref PubMed Google Scholar), probably via such contact sites. Therefore, our results are consistent with the transport of LPS to the OM via the contact sites. In contrast to LPS, newly synthesized phospholipids did not fractionate with the high density OM in sucrose gradients, indicating that their transport is abolished in spheroplasts. It was proposed previously (4Doerrler W.T. Reedy M.C. Raetz C.R. J. Biol. Chem. 2001; 276: 11461-11464Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar) that the transport of LPS and phospholipids share a common component at the level of the IM, MsbA. If phospholipids would subsequently continue on the same route of transport as LPS, we would have expected also phospholipid transport to occur in spheroplasts. Since this was not the case, the mechanism of phospholipid transport across the periplasm must be different from the mechanism of LPS transport. We also investigated the possibility that phospholipid transport occurs via a periplasmic chaperone, analogous to the lipoprotein chaperone LolA. However, we were not able to find evidence for a periplasmic chaperone that releases de novo synthesized phospholipids from the IM. Thus, the transport of phospholipids appears to have different requirements than that of LPS and lipoproteins. Future research must focus on elucidating the components involved in the transport to the OM of both LPS and phospholipids. We thank Dr. Hajime Tokuda for providing purified LolA and LolA antiserum, Dr. William Doerrler and Dr. Christian Raetz for providing strains WD2 and WD201, and the E. coli Genetic Stock Center for providing strain CAG12077.
Referência(s)