![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Department of Biochemistry and Molecular Biology, Faculty of Science, Saitama University, Saitama 338-8570, Japan
Received 3 June 2002/ Accepted 9 July 2002
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
K) lacking the COOH-terminal lysine residue (required for covalent linking to peptidoglycan) did not cause lysis and allowed for the clear separation of the outer and inner membranes. We propose that in pgsA mutants LppK could not be modified by the addition of a diacylglyceryl moiety normally provided by phosphatidylglycerol and that this defect caused unmodified LppK to accumulate in the inner membrane. Although LppK accumulation did not lead to lysis, the accumulation of unmodified wild-type Lpp apparently led to the covalent linking to peptidoglycan, causing the inner membrane to be anomalously anchored to peptidoglycan and eventually leading to lysis. We suggest that this anomalous anchoring largely explains a major portion of the nonviable phenotypes of pgsA null mutants. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Phosphatidylglycerol was, however, found to be dispensable in E. coli, since the null pgsA30::kan allele (30) can be introduced by P1 transduction into lpp-2 strains, which lack the major outer membrane lipoprotein (Braun's lipoprotein, Lpp) (11, 14). The synthesis of Lpp consumes an equimolecular amount of phosphatidylglycerol: the latter supplies a diacylglyceryl moiety that modifies prolipoprotein, the primary gene product of lpp (23). This modification is a prerequisite for the cleavage of the Lpp signal peptide by signal peptidase II, and the modified cysteine residue of Lpp is considered to constitute a part of the recognition site for the unique signal peptidase (10, 28, 29) (Fig. 1). Thus, the absence of prolipoprotein may relieve a drain on the limited pool of phosphatidylglycerol in leaky pgsA mutants; however, the apparently dispensable nature of phosphatidylglycerol in pgsA null mutants poses a question as to why the lack of phosphatidylglycerol is lethal in lpp+ cells.
|
To test this assumption, Lpp and an Lpp derivative lacking the COOH-terminal lysine residue were expressed in a pgsA lpp mutant and their effects on growth and on the separation of the outer and inner membranes were examined.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
|
A plasmid with a mutant lpp allele lacking the COOH-terminal lysine codon (AAG) was constructed as follows. Antisense primer (with an NdeI site) starting 17 nt downstream of the stop codon of lpp contained an alteration of the terminal lysine codon AAG to the stop codon TAG (LpNdeF [5'-TCACTTCAtAtGTACTATTACTAGCGG], which introduces an NdeI site [underlined] with two mismatches [lowercase] and the new stop codon [italicized]). The antisense primer for the corresponding wild-type allele was LpNdeB (5'-TCACTTCAtAtGTACTATTAC), which introduces an NdeI site like that introduced by LpNdeF. The sense primer UBAD (5'-TGCTGGCGGAAAAGATGTGAC) recognizes the sequence starting 997 nt upstream of the initiation codon of lpp in pLPW. These primers were used to amplify the respective lpp alleles, and amplified products were digested with NdeI and BssHII (whose recognition sequence is located upstream of lpp). The downstream region of lpp of pLPW was amplified with the antisense primer DBAD (5'-CTTCATTCAGCTCCGGTTCCC), which recognizes the sequence starting 1,185 nt downstream of the stop codon, and the sense primer LpNde (5'-ACCGCTAGTAATAGTACAtAtGAAG), which introduces an NdeI site like that introduced by LpNdeF and recognizes sequences starting 8 nt upstream of the stop codon TAA. The fragment obtained was digested with NdeI and ApaLI (whose recognition sequence is located in bla). These fragments were then ligated to the large BssHII-ApaLI fragment of pSK6, and the resulting plasmids containing the wild-type or the mutant lpp allele lacking the COOH-terminal lysine residue were designated pLPB and pLPF, respectively.
Derivatives of pMS6 containing lpp alleles were constructed by replacing the pSC101 ori region (the 2.8-kbp PvuI-EcoT22I fragment) of pLPB and pLPF with the ori-rop region of pBR322 (the 2.1-kbp PvuI-EcoT22I fragment). The resulting plasmids were designated pLPB2 and pLPF2, respectively, and their gene products were Lpp and LppK, respectively. A suppressor allele, lpp-12, was amplified from SD12CLR (18) and processed to construct pLPSD2 as above.
PCR was conducted with the Expand High Fidelity PCR system (Boehringer Mannheim Biochemicals, Indianapolis, Ind.). Other DNA manipulations were performed as previously described (22). The DNA sequences of wild-type and mutant alleles of lpp in plasmid constructs were confirmed by using an ABI Prism 310 genetic analyzer (Perkin-Elmer Biosystems).
Media and growth conditions. Luria-Bertani (LB) medium containing 1% tryptone (Difco Laboratories, Detroit, Mich.), 0.5% yeast extract (Difco), and 1% NaCl (15) was supplemented with 0.1 M Na-phosphate buffer (pH 7.2). For solid medium, 1.5% agar was included and the top agar contained 0.6% agar. When required, ampicillin (Sigma Chemical Co., St. Louis, Mo.) was added at a final concentration of 20, 50, or 100 mg per liter. To express lpp and its mutant derivatives under the control of the araBAD promoter, L-arabinose (Wako Pure Chemicals, Tokyo, Japan) was added at 0.2%. Cells were grown at 37°C, and turbidity was monitored with a Klett-Summerson photoelectric colorimeter (no. 54 filter).
Phospholipid analysis. After addition of L-arabinose, cells were labeled with 25 µCi of 32Pi/ml (carrier free; ICN Pharmaceuticals Inc.) for 4.5 h in LB medium supplemented with 20 mM Tris-HCl (pH 7.2). Lipids were extracted by the method of Ames (1) and separated by two-dimensional thin layer chromatography on silica gel plates (silica gel no. 60; E. Merck AG, Darmstadt, Germany) as described previously (25).
Membrane localization of Lpp and its derivatives. The localization of Lpp and its derivatives in the cell envelope was determined as described previously by Yakushi et al. (31). S330 cells containing pLPB2, pLPF2, or pMS6 were grown in LB medium supplemented with 0.1 M Na-phosphate buffer (pH 7.2) at 37°C. At an early exponential growth phase, 0.2% L-arabinose was added and cells were incubated for 1.5 or 3.5 h to express lpp or lpp mutant allele, respectively. The cells were harvested from 100-ml cultures by centrifugation and resuspended in 10 ml of 10 mM Tris-HCl buffer (pH 7.8) containing 0.75 M sucrose, and they were converted into spheroplasts by incubation with 100 µg of lysozyme/ml (Sigma) for 2 min on ice followed by the addition of 20 ml of cold 1.5 mM EDTA (pH 7.5) (19). The spheroplasts were disrupted by sonication, and total membrane fractions were obtained by ultracentrifugation (100,000 x g, 30 min) after unbroken cells were removed by low-speed centrifugation. The total membrane fractions were resuspended in 0.5 ml of 50 mM Tris-HCl (pH 7.5) containing 1 mM EDTA and were layered on 30 to 55% (wt/wt) sucrose gradients (11 ml) containing 5 mM EDTA (pH 7.5). The gradients were centrifuged at 80,000 x g for 12 h at 4°C and were fractionated from the bottom. The density of sucrose (percent) was determined with a refractometer (Atago, Tokyo, Japan). The fractions were separated by Tricine-sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (23), and Lpp and OmpA were detected with their respective antibodies by using an HRP-1000 immunostain kit (Konica, Tokyo, Japan). Antisera against Lpp and OmpA were kindly provided by Henry C. Wu and Yasuhiro Anraku, respectively. The densities of bands detected in the immunoblots were quantified with a densitograph system (Atto, Tokyo, Japan). Succinate dehydrogenase activity was determined as described previously (19).
Mass spectrometry analysis.
Peptidoglycan-linked form (SDS pellet) and free-form (SDS supernatant) fractions of Lpp molecules for mass spectrometric analysis were prepared as described previously (31). After the total membrane was boiled for 20 min with 1% SDS, the supernatant (SDS supernatant fraction) and pellet were obtained by centrifugation at 100,000 x g for 1 h. The pellet was resuspended and treated with lysozyme for 16 h at 37°C (SDS pellet fraction). These fractions were adjusted to 0.1% trifluoroacetic acid and 50% ethanol in water (vol/vol/vol) and mixed with 
-cyna-4-hydroxycinnamic acid (2 mg/200 µl). Matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectra were acquired on a Kratos Kompact MALDI 4 mass spectrometer (Kratos, Manchester, United Kingdom) equipped with a pulsed nitrogen laser (337 nm) in linear mode by using an acceleration voltage of 20 kV.
|
RESULTS AND DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Another interesting characteristic of the pgsA null mutant is its inability to grow in acidic media. Addition of either glucose or L-arabinose (0.2% final concentration) to a culture growing in LB liquid medium (adjusted to pH 7.2 with NaOH) caused lysis within 4 h, in concert with a rapid decrease in pH (below 6). Supplementation with 0.1 M sodium phosphate buffer (pH 7.2) remedied the lysis. Accordingly, we used LB medium supplemented with 0.1 M sodium phosphate buffer and 20 µg of ampicillin per ml in the following experiments.
Effects of the expression of Lpp and its COOH-terminal lysine-lacking derivative on the growth of pgsA null cells. Expression of Lpp in the pgsA+ strain S301 (lpp-2) harboring pLPB2 had little effect on its growth (data not shown). However, when Lpp was expressed in the pgsA null cells (S330 harboring pLPB2), the cells started to lyse 3 h after the addition of 0.2% L-arabinose (Fig. 2b). Viable cell counts decreased much earlier (as early as 2 h after the addition of the inducer) and dropped to a level lower than 10-3 of the maximal value (Fig. 2e). The level of Lpp expressed from pLPB2 (a pBR322 derivative) was about 30% of that of chromosomally expressed Lpp (data not shown). When Lpp was expressed from a low-copy-number plasmid (pLPB, a pSC101 derivative), growth was almost uninhibited.
|
K) which lacks the COOH-terminal lysine residue required for the linking of Lpp to peptidoglycan (29, 31). The expression of LppK from plasmid pLPF2 (Table 1) did not cause cell lysis, although it slightly inhibited an increase in viable cell count (Fig. 2c and f). These results indicate that the deletion of the COOH-terminal lysine residue relieves the toxicity of wild-type Lpp in cells that lack phosphatidylglycerol.
Localization of LppK in the inner membrane of pgsA null cells.
Total membranes of cells containing either Lpp or LppK harvested 1.5 h (before the loss in viability) or 3.5 h after the addition of L-arabinose, respectively, were fractionated by sucrose density gradient centrifugation, and the localization of Lpp or LppK in membranes was examined by SDS-polyacrylamide gel electrophoresis (Fig. 3 and 4). Membranes from pgsA+ cells expressing Lpp were separated into two distinct peaks, and Lpp was detected in the outer membrane fraction as multiple bands (Fig. 3a and 4a). In contrast, a single band was observed in the outer membrane fraction of pgsA+ cells expressing LppK (Fig. 3b and 4b), indicating that the appearance of the multiple bands depended on the COOH-terminal lysine residue needed for the covalent linking of Lpp to peptidoglycan. Thus, the slower-migrating bands of wild-type Lpp were the lysozyme-digested products of the peptidoglycan-bound form, which were complexes linked via the COOH-terminal lysine residue to one, two, or more disaccharide units derived from peptidoglycan (31) (in the absence of lysozyme, only the free-form Lpp, unlinked to peptidoglycan, was detectable [Fig. 5A, band 1 in lanes a and c]).
|
|
|
K in the envelope of cells lacking phosphatidylglycerol was examined. The inner and outer membranes were separated, and the majority of LppK was recovered in the inner membrane fraction and a minority was found in the outer membrane. In the inner membrane fraction, two slower-migrating LppK bands (Fig. 3c and 4c, bands 2 and 3) were observed, and in the outer membrane, the mature form was observed (Fig. 4c, band 1). The appearance of bands 2 and 3 depended on the absence of phosphatidylglycerol, since they were not detected in pgsA+ cells (Fig. 4b). In an in vitro experiment, phosphatidylglycerol was shown to be required for the diacylglyceryl modification (23), which is a prerequisite for the cleavage of the signal peptide by signal peptidase II (10, 28). The cleavage of the signal peptide may therefore be impaired in cells lacking phosphatidylglycerol. Bands 2 and 3 were thus expected to be incompletely processed LppK molecules.
Unmodified and diacylglyceryl-modified forms of LppK accumulate in the inner membrane of pgsA null cells.
To examine the nature of the slower-migrating bands (Fig. 4c, bands 2 and 3) of LppK in the inner membrane, cells were disrupted with sonication, either treated with lysozyme or not treated, and subjected to SDS-polyacrylamide gel electrophoresis. In cells lacking phosphatidylglycerol, bands 2 and 3 were present in samples derived from cells not treated with lysozyme (Fig. 5A, compare lanes i and j), demonstrating that these are not bound to peptidoglycan (free forms). Bands 2 and 3 in lane i of Fig. 5A correspond to those observed in the inner membrane fraction (bands 2 and 3 in lane k). The relative mobilities of these bands suggest that bands 2 and 3 are UPLP and diacylglyceryl prolipoprotein (DGPLP) forms (23) of LppK, respectively.
To identify these bands as UPLP and DGPLP in cells lacking phosphatidylglycerol, the known donor of the diacylglyceryl moiety (23), we treated these cells with globomycin, an inhibitor of signal peptidase II (10). After addition of globomycin (100 µg/ml) to S301 cultures, the mature form disappeared, with the concomitant appearance of DGPLP and UPLP (Fig. 5B). As S330 cells were hypersusceptible to globomycin, the drug was added to S330 cultures at a concentration of 20 µg/ml. After the addition of globomycin, the amount of the mature form decreased, with a concomitant increase in the intensity of bands 2 and 3. These results indicate that bands 2 and 3 are the UPLP and DGPLP forms of LppK, respectively. The mobilities of the bands found in the inner membrane fraction of S330 cells (Fig. 5B, lane IM) were the same as those of the bands detected in globomycin-treated S301 and S330 cells.
To further confirm the identification, we determined the molecular masses of LppK molecules by mass spectrometry. Mass spectra of the SDS supernatant fraction (31), with no lysozyme treatment, of the membrane from S330 cells expressing LppK were taken. The area with a mass/charge ratio of 7,000 to 9,000 contained mass spectral peaks corresponding to bands 1, 2, and 3. A cluster of peaks corresponding to the masses calculated for mature forms of LppK with total acyl chain lengths ranging from 48 to 54 carbon atoms and containing zero to three unsaturated bonds was observed (peaks corresponding to 7,044, 7,068, 7,096, 7,100, 7,102, 7,124, and 7,128 Da were significant). The composition of the acyl chain lengths reflected the previously established composition (17). These masses were consistent with the apparent mass estimated from the mobility of band 1. Two peaks corresponding to the masses calculated for DGPLP forms of LppK having the diacylglyceryl moiety with total acyl chain lengths of 32 and 34 carbon atoms and with one unsaturated bond (8,744 and 8,772 Da) were observed; the masses were consistent with those from the mobility of band 3. A peak corresponding to UPLP of LppK (8,195 Da) was consistent with the apparent mass of band 2.
The presence of LPP and DGPLP indicates that prolipoprotein is modified by a diacylglyceryl moiety from some other phospholipid(s) in pgsA null cells, although we had at first assumed that prolipoprotein could not be modified in the absence of phosphatidylglycerol. For this modification, phosphatidic acid and CDP-diacylglycerol, which are precursors for phosphatidylglycerol synthesis and which accumulate to a level that is a few percent of the total lipid in pgsA null cells (11), were probably substituted for phosphatidylglycerol, since these biosynthetic intermediates can serve as substrates for modification in an in vitro system (23). The mature form of LppK was indeed detectable in the outer membrane (Fig. 3c and 4c, band 1), indicating that a fraction was modified, processed, and localized properly. This may also be the case for Lpp in pgsA null cells; bands corresponding to the mature form were detectable, as seen in Fig. 4d, band 1, and 5A, band 1 in lanes g and h.
The accumulation of DGPLP suggested that cleavage by signal peptidase II was delayed in pgsA null cells. The cleavage step may require phosphatidylglycerol and/or cardiolipin, although in vitro experiments suggested that signal peptidase II does not require phospholipids for activity (28). It should also be noted that the band representing mature form LppK (Fig. 5A, band 1 in lanes i and j) in samples from cells lacking phosphatidylglycerol was thicker than that of Lpp (band 1 in lanes g and h). The delay in cleavage might be partly overcome for LppK by a mechanism we cannot account for at present.
Inner and outer membranes cannot be separated in pgsA null cells expressing Lpp: a cause of lethality.
When Lpp was expressed in pgsA null cells, the two membranes could not be separated and were recovered in a single broad peak (Fig. 3d). Multiple Lpp bands, including high-molecular-weight forms, were detected in fractions in which the inner and the outer membranes were not separated (Fig. 4d, bands 1 to 5 in fractions 6 to 14). To clarify the nature of these multiple bands, disrupted cell preparations were treated with lysozyme or left untreated and were directly subjected to SDS-polyacrylamide gel electrophoresis. Bands 2 and 3 in lanes g and h of Fig. 5A were detected independent of lysozyme treatment and corresponded to UPLP and DGPLP. Bands 4 and 5 in lane h appeared only after lysozyme digestion and were peptidoglycan-linked molecules, an identification that was also supported by their absence from LppK preparations.
To confirm the linkage of UPLP and DGPLP to peptidoglycan, mass spectra of the SDS pellet fraction of Lpp molecules, which was digested with lysozyme (31), were taken. The area with a mass/charge ratio of 9,000 to 10,000 contained mass spectral peaks corresponding to the UPLP and DGPLP molecules linked to one muropeptide unit. These masses were consistent with the apparent masses estimated from the mobility of bands 4 and 5 (Fig. 5A, lane h). Mass spectral peaks corresponding to the UPLP and DGPLP molecules linked to two units of muropeptide were detectable in the area with a mass/charge ratio of 10,000 to 11,000. Those corresponding to the molecules linked to three units were in the area with a mass/charge ratio of 11,000 to 12,000. Peaks corresponding to mature Lpp molecules linked to one, two, and three muropeptide units were also detected.
Taken together, these results indicate that muropeptide-linked forms of DGPLP and UPLP comprised the major portion of the high-molecular-weight bands found in the membranes of cells lacking phosphatidylglycerol. These molecules should be localized in the inner membrane, as was the case for LppK. The linkage of such molecules to peptidoglycan should result in the anchoring of the inner membrane to peptidoglycan and should therefore promote tight cohesion between the inner and outer membranes. Such a structural disorder of the envelope is probably a major reason for the lethality of Lpp in cells lacking phosphatidylglycerol.
The pgsA lpp null strain is temperature sensitive and lyses when incubated for 2.5 h at 42°C (11). E. coli is predicted to have more than 80 other lipoproteins (27, 29), and these lipoproteins may require phosphatidylglycerol (or other acidic phospholipids) for modification. If some are expressed in larger amounts at high temperatures, then a similar cohesion of the inner and outer membranes provoked by unmodified forms of lipoproteins might occur and lead to cell lysis.
A suppressor allele identified in the strain SD12, lpp-12, the product of which has an internal deletion of 21 amino acid residues (32), suppresses the lethality conferred by the reduced levels of phosphatidylglycerol in the pgsA3 mutant (2). Expression of lpp-12 in S330 cells, however, inhibited growth, and the cells subsequently lysed and lost viability. The inner and outer membranes could not be separated after sucrose gradient centrifugation (data not shown). This suggests that UPLP and DGPLP forms of this shorter version of Lpp can link, although inefficiently, the inner membrane to peptidoglycan. The notion that the linking of the inner membrane to peptidoglycan through the COOH-terminal lysine residue of Lpp causes cohesion between the inner and outer membranes might also apply to the lpp-12 product.
Decrease in the content of acidic phospholipids and growth arrest.
Expression of LppK did not cause lysis, and the envelope of S330 cells was clearly separable into outer and inner membrane fractions. The absence of the COOH-terminal lysine residue, however, did not fully suppress the growth defect of S330 cells, and growth was arrested (Fig. 2c and f). In growth-arrested cells, the level of the acidic biosynthetic intermediates, phosphatidic acid and CDP-diacylglycerol, decreased to 5.4% of total phospholipid, compared to 8.4% in cells that did not express LppK. The number of Lpp molecules corresponds to as much as 3.0% of total phospholipid molecules in wild-type cells (24), and the level of LppK expressed from the plasmid-borne lpp allele was about 30% of that of the chromosomally expressed Lpp in wild-type cells; thus, the number of LppK molecules represents about 1% of total phospholipid molecules. Since an appreciable proportion of LppK was lipidated in the cells lacking phosphatidylglycerol (Fig. 4 and 5), this decrease in acidic intermediates suggests that at least a part of the small pool is consumed in the modification of LppK. This decrease may be the major cause of the growth arrest.
The growth arrest brought about by the acidic phospholipid deficiency in HDL1001 (Plac-pgsA+ lpp+) cells is suppressed by an rnhA mutation. This has been explained by the defective initiation of DNA replication at oriC (30, 33) that results from a failure to regenerate the active ATP-bound form from the inactive ADP-bound form of the initiator protein DnaA, a process that is known to rely on acidic phospholipids in vitro (4, 6). In HDL1001 cells, a small amount of phosphatidylglycerol is produced by basal-level expression of pgsA even in the absence of IPTG (isopropyl-ß-D-thiogalactopyranoside) (9). Accordingly, Lpp is probably properly modified by phosphatidylglycerol and localizes to the outer membrane without provoking disorder in the envelope. The lower limit of total acidic phospholipids (including biosynthetic intermediates) required for the growth of HDL1001 cells has been shown to be ca. 7% (9). Thus, in both HDL1001 (Plac-pgsA+ lpp+) cells and S330 (pgsA30::kan lpp-2) cells expressing LppK, the major cause for growth arrest may be the consumption of the limited pool of acidic phospholipids, including biosynthetic intermediates, required for modification. In pgsA null cells expressing LppK, a shortage of the acidic biosynthetic intermediates may cause defects in essential processes involving peripheral membrane proteins such as DnaA (4, 7), SecA (3, 12, 26), FtsY (5), and PssA (13, 20, 21).
|
ACKNOWLEDGMENTS |
|---|
This work was supported in part by grants-in-aid for scientific research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Appl. Environ. Microbiol. | Infect. Immun. | Eukaryot. Cell |
|---|---|---|
| Mol. Cell. Biol. | J. Virol. | Microbiol. Mol. Biol. Rev. |
| ALL ASM JOURNALS |
). At the time indicated by the arrows, L-arabinose (0.2%) was added (•) and Klett units were measured (a to c). The numbers of viable cells were determined by plating aliquots of each culture on LB agar plates supplemented with 0.2% glucose (d to f).
), a representative inner membrane protein, were determined by a specific enzyme assay, and those of OmpA (
), a representative outer membrane protein, were determined with anti-OmpA antiserum. Sucrose density (percent) is indicated by the dashed line.
