INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Protein translocation across the cytoplasmic membrane of Escherichia coli is mediated by a machinery comprising six Sec factors (A, D, E, F, G, and Y) with the help of a secretion-specific molecular chaperone, SecB (4, 6, 14, 20, 29, 38). The SecA cycle of membrane insertion and deinsertion coupled to ATP binding and hydrolysis, respectively, has been thought to drive protein translocation (8, 9). SecG stimulates the SecA cycle by undergoing a membrane topology inversion cycle, leading to efficient protein translocation (26). A proton motive force, which also enables efficient protein translocation, was recently found to accelerate the SecA cycle (23). Moreover, a number of Sec mutants affecting protein translocation have been shown to modulate the SecA cycle (7, 8, 21, 22, 36). These observations, taken together, indicate that the SecA cycle is a critical step for protein translocation, although the details of molecular events involved in the SecA cycle remain to be clarified.

A secG null mutant exhibits cold-sensitive growth in a strain-specific manner (1, 11), whereas protein translocation in the absence of SecG is defective irrespective of the E. coli strain (1, 11, 24, 37). Overexpression of pgsA, encoding phosphatidylglycerophosphate synthase (12, 39), and of gpsA, encoding a biosynthetic sn-glycerol-3-phosphate dehydrogenase (3, 30), suppresses the cold-sensitive phenotype of the secG null mutant (16, 34, 36, 37). Depletion of glycerol-3-phosphate due to the glpR mutation was proposed to be responsible for the cold-sensitive growth in the absence of SecG (10). Since both PgsA and GpsA are involved in phospholipid synthesis, these results indicate that the absence of the SecG function is compensated for by manipulation of the phospholipid composition in membranes. Indeed, an increase in the acidic-phospholipid content on pgsA overexpression was found to stimulate the SecA cycle in the absence of SecG (37).

Not only acidic phospholipids (5, 18, 40) but also nonbilayer lipids (31, 40) have been reported to be important for protein translocation. On the other hand, the effects of the fatty acid composition on protein translocation have not been reported. The synthesis of unsaturated fatty acids, which are essential for the growth of E. coli, requires FabA and FabB (19). FabA catalyzes the dehydration of the -hydroxydecanoyl-acyl carrier protein (ACP) to a mixture of trans-2-decanoyl-ACP and cis-3-decanoyl-ACP, thereby introducing a double bond into the growing fatty acid chain. The trans-2 isomer is then reduced to a saturated fatty acid while the double bond in the cis-3 isomer is preserved throughout the elongation to form the unsaturated fatty acid. E. coli possesses only three major fatty acids, palmitic acid (16:0), palmitoleic acid (16:1), and cis-vaccenic acid (18:1). The fabA6 mutant is a temperature-sensitive unsaturated fatty acid auxotroph since the mutated enzyme is unstable at high temperature (32). The sfa gene has been reported to be a multicopy suppressor for the fabA6 mutation (32). Overexpression of sfa was shown to cause the overproduction of unsaturated fatty acids in the wild-type strain (32).

We report here the isolation of a new multicopy suppressor of the cold-sensitive phenotype of the secG null mutation. The suppressor also corrected the temperature-sensitive phenotype of the fabA6 mutant. We therefore investigated the temperature-sensitive and cold-sensitive suppression in more detail and found that a single gene, which is different from sfa, is responsible for the suppression of both the temperature-sensitive and cold-sensitive phenotypes. The overexpression of the suppressor gene caused a significant increase in the unsaturated fatty acid content and corrected the defective protein translocation in the secG null mutant.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. E. coli JT2602 (fabAts6 zdf::Tn10 leuB6 thr-1 lacY1 thi-1 supE44 tonA33 F) (32), kindly supplied by Charles O. Rock, FS1576 (C600 recD1009) (27, 35), KN370 (FS1576 secG::kan) (24), K003 (HfrH pnp-13 tyr met RNase I Lpp uncB-C::Tn10) (42), and KN553 (K003 secG::kan) (26) were used. K003 and KN553 were grown on the previously reported medium (42). The others were grown on Luria-Bertani (LB) medium.

Materials. Restriction enzymes and DNA-modifying enzymes were obtained from Takara Shuzo Co. Anti-GnsA/GnsB antibodies were raised in rabbits against a synthetic peptide corresponding to the Thr³⁹-Lys⁵³ region of GnsA (identical to the Thr⁴⁰-Lys⁵⁴ region of GnsB). [³²P]orthophosphoric acid was from NEN Life Science Products.

Cloning of a multicopy suppressor of the secG null mutation. Chromosomal DNA prepared from secG null mutant cells was digested with various restriction enzymes and cloned into appropriate sites of pBR322. The secG null mutant, KN370, was transformed with these plasmids and grown on LB plates for 3 days at 20°C. Colonies on the plates were pooled, and their plasmids were analyzed. Among these plasmids, pSG1, containing an 1-kbp BamHI-EcoRI fragment (Fig. 1A), suppressed the cold-sensitive growth defect of KN370. This fragment was inserted into the BamHI-EcoRI sites of pUC19 to construct pSG2. To truncate the fragment from the EcoRI site, pSG2 was cut with BbeI and EcoRI. For truncation from the BamHI site, pSG2 was cut with PstI and BamHI. After partial digestion with exonuclease III, the truncated derivatives of pSG2 were successively treated with mung bean nuclease and Klenow enzyme to make the ends blunt, followed by self-ligation. Plasmids pSG70, pSG63, and pSG86 were constructed thus. To construct pSG5, the EcoRI-KpnI region of pSG2 was deleted, followed by self-ligation. Truncation of the fragments was confirmed by DNA sequencing.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Multicopy suppressors of the secG null mutation. (A) An 1-kbp EcoRI-BamHI fragment and its truncated derivatives are shown with their abilities to suppress the cold-sensitive (cs) secG null mutation. A possible promoter for yccL found in sfa is shown above pSG1. (B) Amino acid sequences deduced from the nucleotide sequences of yccL (gnsA) and ydfY (gnsB). Anti-GnsA and -GnsB antibodies were raised against the indicated identical sequence.

Construction of pSRA and pSRB. A DNA fragment containing the gnsA or gnsB gene with an ideal Shine-Dalgarno sequence was amplified by PCR with a pair of primers, i.e., for gnsA, 5'-TTGGATCCTAGGAGGTTTAAATTTATGAATATTGAAGAGTTAAAAAAACAAGCC-3' and 5'-TTAGATCTTCACAAAATTACATTATTTTGATTTTGACATCATAA-3', and for gnsB, 5'-TTGGA TCCTAGGAGGTTTAAATTTATGATGAATATTGAAAACTTAAAAACA AAAGCAGAAGCA-3' and 5'-TTAGATCTTTGAATACATTAGATTAAATTAATCTTGACATCATAG-3' (italicized letters represent created restriction sites and initiation codons). The amplified fragment was subcloned into the pGEM-T Easy vector (Promega) and confirmed by sequencing. These plasmids were digested with BamHI and BglII, whose sites were created upstream and downstream of the PCR fragments, respectively. The BamHI-BglII fragment of 200 bp was then inserted into the BglII site of pKQ2 (24) to construct pSRA and pSRB, encoding gnsA and gnsB, respectively, under the control of the ara regulon.

Construction of gnsA and gnsB null mutants. DNA fragments used to disrupt gnsA and gnsB were amplified by PCR using the specified oligonucleotides having a unique restriction site (Table 1) and then cloned into pGEM-T Easy vector. To amplify the upstream (2.7-kbp) and downstream (2.8-kbp) regions of gnsA, primers A up 5'/A up 3' and A down 5'/A down 3', respectively, were used with chromosomal DNA prepared using FS1576 as a template. The cat gene was amplified using primers cat 5'/cat 3' with pHSG399 (Takara Shuzo Co.) as a template. The upstream and downstream regions of gnsA and the cat gene were excised from the cloning vector by digesting the respective unique sites and then cloned together into pBR322, which had been cut with Eam1105I and ClaI, to construct the gnsA::cat allele. This plasmid (pSR102) was linearized by ClaI and NheI and then transformed into FS1576 harboring pSRA. Chloramphenicol-resistant colonies were then isolated in the presence of arabinose (0.2%) and examined for the loss of pSR102. The gnsA::cat allele was confirmed by PCR with one of the chloramphenicol-resistant strains, RS1110. To disrupt gnsB, the upstream (2.7-kbp) and downstream (2.1-kbp) regions of gnsB were amplified using primers B up 5'/B up 3' and B down 5'/B down 3', respectively. The spc gene was amplified using primers spc 5'/spc 3' with pHP45 (28) as a template. These DNA fragments were then cloned together into pBR322, which had been digested with Eam1105I and NdeI, to construct pSR103 carrying the gnsB::spc allele. After linearization by Eam1105I and NdeI, pSR103 was introduced into FS1576 harboring pSRB. RS1103, the spectinomycin-resistant strain thus isolated, carried the gnsB::spc allele. Both RS1102 (gnsA::cat) harboring pSRA and RS1103 (gnsB::spc) harboring pSRB grew normally in the absence of arabinose.

gnsA

gnsB

SDS-PAGE and immunoblotting. GnsA and GnsB were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a gel composed of 12.5% acrylamide and 0.27% N,N'-methylenebisacrylamide as previously described (13). For the analysis of OmpA and proOmpA, a gel comprising 13.5% acrylamide and 0.36% N,N'-methylenebisacrylamide was used as previously described (17). Immunoblotting was performed as described previously (25).

Determination of fatty acid compositions. E. coli cells grown on KYG medium (42) were harvested at the mid-exponential-growth phase (optical density at 660 nm [OD₆₆₀], 0.8) and then washed with and suspended in distilled water. The cells were disrupted by passage through a French pressure cell (15,000 lb/in²). After removal of the unbroken cells by centrifugation (10,000 × g, 4°C), the total membrane fraction was obtained by centrifugation at 200,000 × g for 2 h at 4°C. Fatty acid methyl esters were prepared from the membrane fraction as described previously (2) and then analyzed by gas chromatography.

Phospholipid compositions. E. coli cells were labeled with [³²P]orthophosphate (10 µCi/ml) for 2 h at 37°C. Where indicated, labeling was continued for 3 h at 20°C. Lipids were extracted with chloroform-methanol (1:2) and then analyzed by thin-layer chromatography on Silicagel 60 (Merck) with chloroform, methanol, H₂O, and 30% NH₄OH (120:75:6:2) as the developing solvent, as described previously (37). Phospholipids were identified on the chromatogram by autoradiography. The phospholipid spots were scraped off to determine radioactivity.

In vitro protein translocation. E. coli K003, KN553/pKQ2, and KN553/pSRA were grown at 37°C in the presence of 0.2% arabinose. Inverted membrane vesicles (IMVs) were prepared from these cells as described previously (41). Translocation of proOmpA into the IMVs was examined at 20°C in a reaction mixture comprising IMVs (0.1 mg/ml), SecA (60 µg/ml), SecB (50 µg/ml), 1 mM ATP, 5 mM succinate, 1 mM MgSO₄, 10 mM dithiothreitol, an ATP-generating system (10 µg of creatine kinase per ml plus 5 mM creatine phosphate), and 50 mM potassium phosphate (pH 7.5). The reaction was initiated by the addition of prewarmed [³⁵S]proOmpA (1.3 × 10⁶ cpm/ml). Aliquots (25 µl) of the reaction mixture were withdrawn at various times and mixed with proteinase K (1 mg/ml) to terminate the reaction. After proteinase K digestion on ice, the OmpA materials were recovered by trichloroacetic acid precipitation (final concentration, 10%), successively washed with acetone and ether, and then analyzed by SDS-PAGE and fluorography. The translocation activity was determined by densitometric quantitation of the OmpA materials and expressed as a percentage of the input proOmpA.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Overexpression of yccL suppresses the cold-sensitive secG null mutation. Plasmid pSG1, which carries an 1-kbp BamHI-EcoRI fragment at BamHI-EcoRI sites of pBR322, was obtained as a multicopy suppressor of the cold-sensitive secG null mutation of KN370. Restriction mapping and hybridization analyses with the ordered library of the E. coli chromosome revealed that the fragment corresponded to the 22-min region, and carried cspG, sfa, yccL, and yccM, although cspG and yccM were truncated (Fig. 1A). Complete deletion of cspG (pSG70) and further truncation of yccM (pSG63) had no effect on the suppressor activity, whereas truncation of yccL (pSG86) abolished the suppressor activity, suggesting that yccL but not sfa is responsible for the suppressor activity. Deletion of the upstream region of yccL (pSG5) also abolished the suppressor activity. We found a possible promoter sequence for yccL in the deleted region of sfa (Fig. 1A). We then constructed pSRA carrying yccL under the control of the ara regulon. This plasmid suppressed the cold-sensitive phenotype of KN370 in the presence of 0.2% arabinose (Table 2). Hereafter, yccL encoding a small protein of 6.6 kDa (Fig. 1B) is renamed gnsA (secG null mutant suppressor A).

Both gnsA and gnsB are multicopy suppressors of the temperature-sensitive fabA6 mutation. The sfa gene present in pSG1 was reported to be responsible for suppression of the temperature-sensitive fabA6 mutation (32). Since FabA is an essential enzyme for unsaturated fatty acid synthesis, the sfa gene was assumed to play a role in unsaturated fatty acid synthesis (32). Indeed, unsaturated fatty acids increased when wild-type E. coli harbored a multicopy plasmid carrying sfa (32). The sfa gene and its downstream gnsA genes overlap by 11 bp. However, the reported shortest suppressor fragment carried not only sfa but also gnsA (32). We therefore examined the suppression of the fabA6 mutation of JT2602 by the plasmids constructed in this study (Table 2). To our surprise, both gnsA and gnsB suppressed not only the cold-sensitive secG null mutation but also the temperature-sensitive fabA6 mutation of JT2602 when they were induced by 2% arabinose (Table 2). In contrast, pSG86 carrying sfa but not gnsA (Fig. 1A) did not suppress the fabA6 mutation, indicating that the suppressor of the fabA6 mutation is gnsA but not sfa.

Overexpression of gnsA or gnsB stimulates protein translocation. The secG null mutation causes a cold-sensitive growth defect in a strain-specific manner (1, 11), whereas protein translocation at low temperature was defective in all secG null strains examined (1, 11, 24, 37). Accumulation of proOmpA was examined in two secG null mutants, KN370 and KN553, the latter of which did not exhibit cold-sensitive growth. Both strains harboring a vector, pBR322 or pKQ2, accumulated proOmpA at 20°C (Fig. 2). On the other hand, proOmpA accumulation was hardly detectable in KN370 harboring pSG1 (Fig. 2A). Furthermore, when gnsA was expressed from pSRA with 0.2% arabinose, proOmpA was almost undetectable in both KN370 (Fig. 2A) and KN553 (Fig. 2B). Expression of gnsB from pSRB on the addition of 0.2% arabinose had no effect on the level of proOmpA. However, the level of proOmpA in both strains significantly decreased when gnsB was expressed with 2% arabinose (Fig. 2). Taken together, these results indicate that both gnsA and gnsB correct the defective protein translocation of the secG null mutant upon overproduction.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Overexpression of gnsA and gnsB stimulates proOmpA translocation in the secG null mutant. KN370 (A) and KN553 (B) harboring the indicated plasmids were grown at 37°C. Where specified, the medium contained the indicated concentrations of arabinose. When the turbidity at 660 nm reached 0.8, the cultures were transferred to 20°C and further incubated for 3 h. Cellular proteins (5 µg) were analyzed by SDS-PAGE and immunoblotting with anti-OmpA antibodies.

View larger version (39K): [in this window] [in a new window]   FIG. 3.   Overproduction of GnsA and GnsB. KN370 (A) and KN553 (B) harboring the indicated plasmids were grown at 37°C in the presence of the indicated concentrations of arabinose. At an OD660 of 1.0, cellular proteins (10 µg for panel A and 5 µg for panel B) were analyzed by SDS-PAGE and immunoblotting with anti-GnsA and -GnsB antibodies. As a control, 5 µg of cellular proteins of K003 was analyzed in the left lane of panel B. The migration position of a 6.2-kDa marker protein is indicated at the left.

Unsaturated fatty acid compositions. GnsA and GnsB were each overproduced in KN553 cells, and then their major fatty acid compositions at 20°C were analyzed (Table 3). Overproduction of GnsA with 0.2% arabinose caused a significant increase in cis-vaccenic acid (18:1) at the expense of palmitic acid (16:0). The total unsaturated fatty acid content increased from about 50 to near 75% upon GnsA overproduction. Overproduction of GnsB also increased the cis-vaccenic acid and total unsaturated fatty acid contents, albeit to lesser extents. When the arabinose concentration was increased from 0.2 to 2%, the suppressor activity of GnsB markedly increased, whereas the increase in unsaturated fatty acids was relatively small (Tables 2 and 3).

View larger version (19K): [in this window] [in a new window]   FIG. 4.   GnsA does not exhibit a cold shock response. KN370 harboring pSG1 was grown at 37°C until the OD660 reached 0.8 and then was transferred to 20 or 10°C. The cultures were then kept at 20°C for 3 h or at 10°C for the specified time. Cellular proteins (10 µg) were then analyzed as for Fig. 3.

An increase in unsaturated fatty acids stimulates in vitro protein translocation. K003, KN553/pKQ2, and KN553/pSRA were grown at 37°C in the presence of 0.2% arabinose. IMVs were prepared from these cells and then subjected to the proOmpA translocation assay at 20°C (Fig. 5). The translocation activity was significantly retarded when IMVs lacked SecG (Fig. 5, compare open circles and closed triangles). On the other hand, the translocation activity was near normal when the unsaturated fatty acid content was increased by the overexpression of gnsA (Fig. 5, closed circles). These results indicate that the absence of the SecG function is compensated for by an increase in unsaturated fatty acids.

View larger version (14K): [in this window] [in a new window]   FIG. 5.   Stimulation of in vitro proOmpA translocation by an increase in the unsaturated fatty acid content. IMVs were prepared from K003 (open circles), KN553/pKQ2 (triangles), and KN553/pSRA (closed circles) and then subjected to the proOmpA translocation assay at 20°C as described in Materials and Methods.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We showed here that, in addition to pgsA (16, 36, 37) and gpsA (34), gnsA and gnsB are multicopy suppressors of the secG null mutation. Strikingly, all these genes participate in the synthesis of phospholipids, suggesting the functional relationship between SecG and membrane phospholipids. Defective protein translocation due to SecG depletion was corrected both in vivo and in vitro by the overexpression of gnsA. An increase in the acidic-phospholipid content upon the pgsA overexpression was previously shown to enhance the translocation-coupled ATPase activity of SecA when IMVs lacking SecG or SecAR11(Cs) derivative were used (37). However, gnsA overexpression did not suppress the secAR11(Cs) mutation, suggesting different mechanisms for the gnsA- and pgsA-dependent suppression. An increase in membrane fluidity due to the increase in unsaturated fatty acids seems to be most likely responsible for the gnsA-dependent suppression, since the SecG function is especially important at low temperature, under which condition the membrane fluidity is low. Furthermore, the conformation and activity of SecA are also dependent on temperature (33). With these facts, it is of great interest to determine whether the increase in membrane fluidity has any effect on the efficiency of the SecA cycle or the ATPase activity of SecA.

Overexpression of gnsA and gnsB caused a remarkable increase in the unsaturated fatty acid content. However, the gnsA-gnsB double null mutant exhibited no defect. Moreover, the two proteins were undetectable unless they were overproduced. Therefore, their physiological functions under normal conditions remain to be clarified, although both proteins were predicted to possess a helix-turn-helix structure. Since gnsA overexpression increased the unsaturated fatty acid content in the secG null mutant, which carries the wild-type fabA gene, stabilization of the FabA enzyme by overproduced GnsA seems to be unlikely. On the other hand, overexpression of sfa, which is now called gnsA, was reported to suppress the temperature-sensitive phenotype of the fabA6 mutant but not that of the fabA2 mutant (32). The allele-specific suppression seems to be due to the phenotypic difference between the two Fab mutants.

We searched for GnsA/B homologs in eubacteria and found that only Salmonella strains carry a gene encoding a homolog. The Salmonella gene is also located near a cold shock gene, cspB, suggesting that the gnsA(B) gene is a part of a cold shock gene cluster in both E. coli and Salmonella.