A Novel Class of secA Alleles That Exert a Signal-Sequence-Dependent Effect on Protein Export in Escherichia coli

Corresponding author: Dominique Belin, CMU 1, rue Michel-Servet, CH-1211 Geneva 4, Switzerland., dominique.belin{at}medecine.unige.ch (E-mail)

Communicating editor: G. R. SMITH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The murine plasminogen activator inhibitor 2 (PAI2) signal sequence inefficiently promotes the export of E. coli alkaline phosphatase (AP). High-level expression of PAI2::AP chimeric proteins from the arabinose P_BAD promoter is toxic and confers an Ara^S phenotype. Most Ara^R suppressors map to secA, as determined by sequencing 21 independent alleles. Mutations occur throughout the gene, including both nucleotide binding domains (NBDI and NBDII) and the putative signal sequence binding domain (SSBD). Using malE and phoA signal sequence mutants, we showed that the vast majority of these secA suppressors exhibit weak Sec phenotypes. Eight of these secA mutations were further characterized in detail. Phenotypically, these eight suppressors can be divided into three groups, each localized to one domain of SecA. Most mutations allow near-normal levels of wild-type preprotein export, but they enhance the secretion defect conferred by signal sequence mutations. Interestingly, one group exerts a selective effect on the export of PAI2::AP when compared to that of AP. In conclusion, this novel class of secA mutations, selected as suppressors of a toxic signal sequence, differs from the classical secA (prlD) mutations, selected as suppressors of defective signal sequences, although both types of mutations affect signal sequence recognition.

SINCE the development of the signal sequence model (BLOBEL and DOBBERSTEIN 1975 ), protein translocation across the plasma membrane has been characterized both genetically and biochemically in Escherichia coli (SCHATZ and BECKWITH 1990 ; DANESE and SILHAVY 1998 ; DRIESSEN et al. 1998 , DRIESSEN et al. 2001 ). In prokaryotes, periplasmic and outer membrane proteins are synthesized in the cytoplasm and translocated post-translationally. The signal sequence of secreted proteins functions both as a targeting signal and as a recognition signal during export. Signal sequences are composed of three domains: a positively charged N-terminal region, a hydrophobic core, and a polar C-terminal region containing the leader peptidase cleavage site. The precursors of exported proteins are maintained in a translocation-competent conformation by chaperones such as SecB (COLLIER et al. 1988 ), GroE (LECKER et al. 1989 ), or DnaK (WILD et al. 1996 ). At the membrane, they bind the homodimeric ATPase SecA (PrlD) and the heterotrimeric SecYEG (PrlAGH), the integral membrane complex that constitutes the core of the translocase. SecYEG also promotes the insertion and determines the topology of integral membrane proteins (PRINZ et al. 1998 ).

Precursor translocation across the membrane is an energy-requiring process driven by ATP hydrolysis and by the proton motive force (CHEN and TAI 1986 ; GELLER et al. 1986 ; SCHIEBEL et al. 1991 ). Translocation is initiated by the ATP-dependent co-insertion of SecA and the preprotein at SecYEG. Interaction of SecA and SecYEG stimulates the ATPase activity of SecA, which leads to the dissociation of SecA from the preprotein. During this process, SecA undergoes large conformational changes (ECONOMOU and WICKNER 1994 ; KIM et al. 1994 ). After this initial stage, the proton motive force can drive preprotein translocation to completion (DRIESSEN and WICKNER 1991 ). Electron microscopic analysis of both the Bacillus subtilis and E. coli SecYE complexes indicates that they form a large channel with a quasi-pentameric structure, at least in octyl-glucoside detergent (MEYER et al. 1999 ; MANTING et al. 2000 ). However, two other studies suggest that the functional state of SecYEG is either monomeric (YAHR and WICKNER 2000 ) or dimeric (BESSONNEAU et al. 2002 ). SecY and SecE are encoded by essential genes and promote both SecA-dependent protein translocation into proteoliposomes (AKIMARU et al. 1991 ) and integral membrane protein insertion mediated by the signal recognition particle (NEWITT and BERNSTEIN 1998 ). The latter process is also SecA dependent (QI and BERNSTEIN 1999 ), although initial membrane insertion can occur in vitro in the absence of SecA (SCOTTI et al. 1999 ). SecG is a nonessential component of the translocase (NISHIYAMA et al. 1994 ), which stimulates translocation both in vivo and in vitro (NISHIYAMA et al. 1993 ; BOST and BELIN 1995 ; HANADA et al. 1996 ). SecG undergoes an inversion of its membrane topology coupled with SecA insertion-deinsertion cycles (NISHIYAMA et al. 1996 ). Two other integral membrane proteins, SecD and SecF, are required for efficient translocation (ARKOWITZ and WICKNER 1994 ; POGLIANO and BECKWITH 1994 ). However, despite the fact that these two proteins can be isolated together with SecYEG (DUONG and WICKNER 1997 ), their precise function during translocation remains poorly understood.

Early genetic selections led to the identification of most sec genes and gave rise to two kinds of mutations in most of these genes (SCHATZ and BECKWITH 1990 ). Mutations conferring a Sec phenotype have a generalized secretion defect. Suppressors with a Prl phenotype are more subtle mutants that have lost the ability to discriminate against inefficient signal sequences and that even allow export of proteins lacking a recognizable signal sequence (DERMAN et al. 1993 ). These observations led to the hypothesis that the translocase has a signal sequence proofreading activity (OSBORNE and SILHAVY 1993 ). Prl mutations have been isolated in secA, secY, secE, and secG (SCHATZ and BECKWITH 1990 ; BOST and BELIN 1997 ), suggesting that each protein contributes somehow to the signal sequence recognition function of the translocase.

In this study, we describe the isolation and characterization of a novel class of secA alleles. These mutants were isolated as suppressors of the murine PAI2 signal sequence, which is toxic when expressed at high levels. Most suppressors map to secA, secY, and secG, supporting the notion that the toxicity of PAI2::AP fusion proteins is due to a defective interaction between the PAI2 signal sequence and the translocase (BOST and BELIN 1995 ; BOST et al. 2000 ). Since many suppressors map to secA, we reasoned that their molecular defect could help us understand the function of SecA during an early step of the translocation process. Unlike secG suppressors, which are mostly localized to three contiguous codons (the central T₄₁L₄₂F₄₃ domain), secA mutations are distributed throughout the first two-thirds of the gene. Like secG suppressors, all secA mutants exhibit a weak Sec phenotype. Most of these suppressors have a different effect on the activity of several wild-type and mutated signal sequences, and one group has a selective effect on PAI2::AP export. Thus, these mutants, unlike classical sec mutants, do not confer a strong and generalized secretion defect. In conclusion, this class of suppressors defines a weak Sec phenotype that appears opposite to that of prl mutants in the context of mutated or inefficient signal sequences, suggesting that they are affected mainly in the signal sequence recognition activity of SecA.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Reagents:
Liquid and solid media were prepared as described (MILLER 1992 ). Antibiotics were used at the following concentrations: kanamycin, 40 µg/ml; ampicillin, 200 µg/ml; and tetracycline, 15 µg/ml. 5-Bromo-4-chloro-3 indolyl-phosphate (XP) was purchased from DCL (Oxford, CT) and used at 40 µg/ml. Rabbit antisera against MalE (MBP)-DegP, RbsB (RBP), OmpA, and SecA were kind gifts from J. Beckwith, M. Ehrmann, and. W. Wickner. Anti-MalE serum was purchased from New England Biolabs (Beverly, MA).

Bacterial strains:
All strains are described in Table 1 or were constructed by phage P1-mediated transduction or by transformation (MILLER 1992 ); the secA gene cotransduces with the leu::Tn10 transposon of LMG194 with a frequency of 40%. All studies have been performed with chromosomal secA alleles. The secA (prlD) strains were kindly provided by T. Silhavy. In the absence of registered sec allele numbers, they are referred to as secA (prlD2), secA (prlD5), and secA (prlD21) in the text and as prlD2, prlD5, and prlD21 in the figures, using the published prl numbers (FIKES and BASSFORD 1989 ; HUIE and SILHAVY 1995 ).

Plasmid construction:
Plasmids pBAD72K, pBADhBK, and pBADhBS are pBAD24 derivatives (BOST and BELIN 1995 ; GUZMAN et al. 1995 ; BOST et al. 2000 ). psecA⁺ is a derivative of pBE2 (KIM et al. 1994 ) lacking both the promoter of the secA operon and most of the gene X (secM) coding sequence; secA expression is driven from the tetracycline promoter of pACYC184. Briefly, a 1.6-kbp MluI/Klenow fragment from pBE2 encoding the 3' part of secA was first subcloned in the SmaI site of pBS-KS. The resulting vector was digested with BssHII and EcoRI and ligated to a 1.1-kbp BssHII fragment of pBE2 encoding the promotorless 5' terminus of secA. At that step, one of the BssHII sites of the insert was not ligated to the EcoRI site of the vector. The ends of the linear intermediate were filled in with the Klenow fragment of DNA PolI and ligated prior to transformation. From the resulting plasmid, a XhoI/XbaI/Klenow fragment containing the entire secA coding sequence was cloned into pACYC184 digested with SalI/Klenow and EcoRV to generate psecA⁺. SecA expression was confirmed by the fact that psecA⁺, like pBE2, confers a dominant Ara^S phenotype to all the Ara^R secA strains. Plasmids containing the mutant secA alleles were constructed with the primers described in Table 2 as follows: psecA303, a PCR fragment (primers A3UP/A4DON) was digested with PvuI/SphI and cloned in psecA⁺; psecA307–310, a PCR fragment (primers A5UP/A8DON) was digested with SphI/BssHII and cloned in psecA⁺; psecA311–312, a PCR fragment (primers A7UP/A10DON) was digested with BssHII/BglII and cloned in psecA⁺. psecA314, a PCR fragment (primers A9UP/A12DON) was digested with BglII/KpnI and cloned in psecA⁺. psecA315–319, a PCR fragment (primers A11UP/A16DON) was digested with KpnI/SnaBI and cloned in psecA⁺. All psecA mutant plasmids confer an Ara^R phenotype to the KK1 strain.

Measurements of doubling times:
Saturated cultures grown overnight from single colonies in Luria broth (LB) medium were diluted 1/50 in M63 medium containing 0.2% glucose and 40 µg/ml leucine and grown overnight. Cells were collected by centrifugation, washed once in M63 medium, and diluted 1/50 in M63 medium supplemented with 1% LB medium and either 0.2% glucose or 0.2% maltose; under these conditions, growth of secA⁺ cells is sufficiently rapid to measure the effect of secA alleles exhibiting a mild-to-strong Sec phenotype. For the comparison of secA⁺ and secA (prlD) strains, LB was not added to the minimal media. Growth was followed by measuring the absorbance at 600 nm with an open chamber spectrophotometer (S250, Secomam).

Isolation of suppressors of PAI2::AP toxicity:
Cells expressing the PAI2::AP or the hB::AP chimeric proteins from the arabinose P_BAD promoter are unable to form colonies on plates containing arabinose. Spontaneous and UV-induced suppressors occur at a frequency of 10^-7 and 10^-6, respectively (BOST and BELIN 1995 ; BOST et al. 2000 ). To ensure that suppressor strains retain the capacity to export the chimeric protein, the alkaline phosphatase (AP) substrate XP was included in the plates, and only dark blue colonies were further characterized; white or pale blue colonies represent about one-half of the total suppressors. Of a total of 205 mutants (105 spontaneous), 59 carried plasmid-encoded mutations, and 67 had chromosomally encoded pcnB mutations or synthesized reduced levels of the chimeric protein. The remaining 79 were mapped by P1 transduction, and 35 suppressors mapped near or in secA (11 spontaneous and 24 UV induced); the localization of 8 secY and 7 secG suppressors was confirmed by sequencing (BOST and BELIN 1995 ; BOST et al. 2000 ). The remaining mutations map to rpoA (5), ydeA (2), and other yet-uncharacterized loci.

Alkaline phosphatase assay:
AP activity was measured by determining the rate of p-nitro-phenyl-phosphate (Sigma, St. Louis) hydrolysis and was normalized to the A_600nm of the cell suspension (BOST and BELIN 1995 ).

Pulse-labeling and immunoprecipitation:
Cell cultures were grown and pulse labeled with [³⁵S]methionine (IS-103, Hartmann, Braunschweig, Germany) at 37° as described previously (BOST and BELIN 1995 ). For the analysis of MalE, cultures were induced with 0.2% maltose for 60 min before the labeling. Quantification was performed by scanning the gels with a Molecular Dynamics (Sunnyvale, CA) phosphorimager, using the ImageQuant version 3.22 software.

Single-strand conformation analysis and sequencing:
Primers were designed to cover all of the secA coding sequence with eight overlapping PCR fragments of 400 bp (Table 2). The SecA primers were labeled with [-³²P]ATP and T4-polynucleotide kinase for 30 min at 37°. Reactions were stopped by heating for 5 min at 95°. Labeled and unlabeled primers were mixed at a ratio of 1:5 and used to amplify the chromosomal secA sequences (RUSSO et al. 1993 ). PCR products were boiled for 20 min, chilled on ice, and loaded on mutation detection enhancement 10% glycerol gels (AT Biochem, Malvern, PA). Gels were run for 18 hr at 40 W at 4° and at room temperature for all tested PCR fragments, dried, and exposed to XAR films (Kodak). PCR-amplified products were sequenced with a PCR sequencing kit (Amersham, Buckinghamshire, UK).

Western blot and quantification:
Triplicate cultures were grown for 2 hr at 37°. Cells (1 ml, A_600nm = 0.3) were centrifuged for 2 min at 13,000 rpm, resuspended in 100 µl of SDS sample buffer, diluted 1/100, and boiled for 5 min. Lysates (20 µl) were loaded on 10% SDS-polyacrylamide gels. After electrophoresis, proteins were transferred onto a nitrocellulose membrane (Protran BA85, Schleicher & Schuell, Keene, NH), with a semidry transfer apparatus (Biorad, Richmond, CA). SecA was detected with an enhanced chemiluminescence kit (Amersham) using a rabbit anti-SecA serum followed by a goat anti-rabbit IgG coupled to peroxidase. Membranes were exposed to XAR films. Quantification was performed by scanning nonsaturated films with an Arcuss II scanner (Agfa), using the ImageQuant version 3.22 software. For calibration, each gel was loaded with triplicates of four different concentrations of the secA⁺ sample.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Rationale of the genetic selection:
The selection used to isolate mutants in the translocation machinery genes is described in MATERIALS AND METHODS. Briefly, two different PAI2::AP fusions were cloned in derivatives of pBAD24, an inducible expression vector containing the promoter of the araBAD operon (GUZMAN et al. 1995 ). The eukaryotic signal sequence promotes the export of these chimeric proteins, albeit inefficiently. Upon arabinose induction, high-level expression of the PAI2::AP fusions is toxic probably because of a defective and prolonged interaction between the eukaryotic signal sequence and the translocase. Thus, wild-type cells expressing these fusions are unable to form colonies on arabinose-containing plates (Ara^S phenotype). Spontaneous and UV-induced Ara^R suppressors have been isolated, and most mutations suppress both toxic fusions (BOST and BELIN 1995 ; BOST et al. 2000 ). Approximately one-half of the Ara^R suppressors formed blue colonies in the presence of the AP substrate XP and therefore retain the capacity to export PAI2::AP, at least to a certain extent. This second condition was imposed in the selection of Ara^R suppressors to avoid the isolation of sec mutants with a generalized secretion defect. Most Ara^R suppressors mapped in or near secA, while several others mapped to secY and secG. So far no mutations have been detected in secD or secF, suggesting that the chimeric proteins are toxic because they affect an early step of the translocation process. The secA alleles were recessive to secA⁺ expressed from a multicopy plasmid (KIM et al. 1994 ; data not shown). The high prevalence of secA suppressors could be explained by a combination of the following facts: (i) SecA plays a key role in the export step blocked by the PAI2::AP fusion; (ii) SecA acts upstream of the affected step; (iii) secA is the largest sec gene; or (iv) one or more mutational hot spots are in the secA gene. We thus characterized these secA alleles.

DNA sequence analysis:
Since secA is a large gene, we first mapped the mutations by single-strand conformation analysis, a highly sensitive technique that can detect most single-base substitutions (ORITA et al. 1989 ). For each mutant, eight overlapping PCR fragments were analyzed on nondenaturing gels (data not shown). Among the 26 mutants analyzed, we detected 18 single and one double mutation; only 2 of the single mutations occurred twice (Fig 1). Five mutations could not be mapped by this technique and were not further characterized. For the purpose of clarity, the sequenced alleles were numbered from secA301 to secA319 according to the position of the mutations along the gene (Fig 1). None of the sequenced mutations have been described previously and they all affect well-conserved amino acids in bacterial and chloroplastic secA homologs, as determined with the ProDom database (http:://protein.inra.fr/prodom.html). To ensure that the detected mutations are responsible for the Ara^R phenotype, we constructed plasmids expressing several of these secA alleles. All tested plasmids (psecA303, -307–312, and -314–319) were dominant over chromosomal secA⁺, conferring the expected Ara^R phenotype (data not shown).

View larger version (18K): In this window In a new window Download PPT slide   Figure 1. Localization of the AraR SecA substitutions. The entire SecA protein is shown in the middle line, where three known interacting domains are indicated: NBDI, SSBD, and NBDII. These domains and flanking regions are enlarged in the top and bottom lines. Previously identified Sec (SCHMIDT et al. 1988 ) and Prl (FIKES and BASSFORD 1989 ; HUIE and SILHAVY 1995 ) substitutions in SecA are indicated above the line in the one-letter code with the position of the affected amino acids. Mutated residues identified in this work are shown below the lines; the eight alleles studied in detail are indicated by boldface type. secA312 and secA317 have been isolated twice. SecA315 is a double mutant, and the S458F substitution is not indicated because the L588Q substitution expressed on a plasmid alone was sufficient to confer the AraR phenotype.

Five mutations are located close to or within a region that encodes the domain of SecA defined by crosslinking as a putative signal-sequence-binding domain (SSBD; KIMURA et al. 1991 ). To our knowledge, only two chromosomal mutations, secA (prlD5) (A373V) and secA283 (H309Y), have been previously described near this region (FIKES and BASSFORD 1989 ; MATSUMOTO et al. 2000 ). secA site-directed mutations encoding substitutions at Y326 provided evidence that this region is critical in controlling SecA-preprotein interaction (KOURTZ and OLIVER 2000 ). Five mutations are located in a region encoding the nucleotide binding domain I (NBDI) domain, where most sec mutations map. Four single mutations and one double mutation are located in and around the region encoding the NBDII domain. The remaining mutations encode two substitutions in the N-terminal part of SecA and two between SSBD and NBDII.

The effect of secA suppressors on maltose binding protein (malE) signal sequence mutations defines three classes of mutants:
Two types of mutations have been described in sec genes: Sec mutations, with a generalized defect that decreases protein export, and Prl mutations, which improve export mediated by defective signal sequences. Some Prl mutations have a selective effect depending on the signal sequence mutation, and mutations like secE (prlG1) and secA (prlD21) also reduce the kinetics of export of wild-type preproteins (STADER et al. 1989 ; HUIE and SILHAVY 1995 ). Since all of the secG suppressors isolated in the same selection have a weak Sec phenotype (BOST and BELIN 1995 ; BOST et al. 2000 ), it was interesting to determine whether the secA suppressors are also Sec mutations. The phenotype of the secA alleles was first investigated in the malE14-1 strain, which carries a signal sequence mutation in preMalE that interferes with its export (BEDOUELLE et al. 1980 ). The Mal phenotype of this strain can be scored on McConkey maltose indicator plates, where improved (Prl phenotype) or decreased (Sec phenotype) MalE export can be easily monitored. All secA suppressors exhibited a Sec phenotype and none showed a Prl phenotype (data not shown). In addition, the phenotype of the 26 secA alleles in this strain showed that we could qualitatively differentiate them according to the strength of their secretion defect.

To further characterize the secA suppressors, we decided to concentrate our efforts on a set of suppressors in which all mutated regions of secA are represented. secA301 and secA302 affect the N-terminal region, secA307 affects the Walker B motif in NBDI, secA311 affects the SSBD domain and showed the strongest Sec phenotype, secA313 affects a residue located between SSBD and NBDII, and secA315, -316, and -318 affect NBDII. We asked whether a similar Sec phenotype could be observed in other malE signal sequence mutant strains and whether the strength of this phenotype could help classify the secA alleles. Three other malE signal sequence mutations (malE10-1, malE18-1, and malE19-1) were used in addition to malE14-1; the mutations are described in Table 3. All of these signal sequence mutations exert a detectable effect on MalE export that ranges from a 70% (malE10-1) to a 96% (malE18-1) export defect, as defined by pulse-labeling experiments (BEDOUELLE et al. 1980 and our unpublished results). The Sec phenotypes of the selected secA alleles in the four different malE strains are shown in Table 3. The amino acid substitutions in SecA can be separated into three classes: (i) Those that confer a very weak Sec phenotype are located in the N-terminal region and in NBDI; (ii) those with an intermediate Sec phenotype localize in and around NBDII; and (iii) the strongest Sec phenotype observed among all tested secA alleles was that of secA311, the only suppressor isolated so far that alters the SSBD domain. It appears striking that the three classes of mutations localize to defined regions of SecA and are not randomly distributed along the protein.

The analysis shown in Table 3 is based on color or size of colonies and is therefore only qualitative. To provide a more quantitative assay, we measured the growth rates in maltose or glucose minimal medium of malE14-1 strains carrying each of the secA alleles. Indeed, in strains carrying this malE signal sequence mutation, growth in maltose as the only carbon source is dependent on the amount of periplasmic MalE (BEDOUELLE et al. 1980 ). Thus, a Sec mutation should confer a slower growth rate and a Prl mutation should confer a faster growth rate. The different growth rates, expressed as the doubling times, are shown in Fig 2. The doubling times in glucose minimal medium were equivalent to those of secA⁺ for all secA mutant strains, indicating that growth is not significantly affected by these secA mutations. In contrast, growth in maltose minimal medium showed the same trend as that observed on maltose indicator plates (Fig 2A). The secA alleles could be grouped in three categories: secA311 had the strongest effect, with an almost threefold increase in doubling time (scored as "-" in Table 3); secA313, -315, and -318 conferred an intermediate but significant increase in doubling time (scored as "+" in Table 3); and the four remaining alleles had only a marginal effect on doubling time (scored "++" to "++++" in Table 3). The McConkey maltose indicator plates therefore appear to be the most sensitive assay to detect subtle differences in the amount of periplasmic MalE.

View larger version (58K): In this window In a new window Download PPT slide   Figure 2. Doubling time of strains carrying different secA alleles in minimal media with either maltose or glucose as the carbon source. Cells were grown at 37° in minimal media with either glucose or maltose as described in MATERIALS AND METHODS. Growth rates, which depend on sugar import, are expressed as doubling times. All alleles were transduced into the KK3 strain, which contains the malE14-1 signal sequence mutation. Each bar graph represents the average result from at least three independent cultures; error bars indicate the SD. (A) AraR secA mutants are compared to the secA+ isogenic strain. The significance of the results was evaluated with the Mann-Whitney nonparametric test. For growth in maltose minimal medium, the P value was 0.004 for secA313, -315, and -318. (B) secA (prlD) mutants were compared to the secA+ isogenic strain; for growth in maltose minimal medium, the P value was 0.03 for the three secA (prlD) strains.

We have also determined growth rates with three Prl mutations in secA: secA (prlD2), secA (prlD5) (FIKES and BASSFORD 1989 ), and secA (prlD21) (HUIE and SILHAVY 1995 ). The first two secA (prlD) mutations were selected as suppressors of malE14-1 on maltose indicator plates. As expected, growth rate in maltose minimal medium was faster with the secA (prlD) strains than with the secA⁺ strain (Fig 2B).

The two assays described above, fermentation on maltose indicator plates and growth in maltose minimal medium, both reflect the steady-state level of periplasmic MalE. To directly assay preMalE export, we measured the efficiency of signal sequence cleavage in pulse-chase labeling experiments (Fig 3). These experiments were performed with strains carrying the malE14-1 mutation, since the amount of mature MalE produced in a secA⁺ strain is low but sufficient to accurately measure an export defect conferred by the secA alleles. In the first set of experiments, cells were labeled for 30 sec and chased for 12 min. A small but significant decrease in the relative amount of mature MalE was observed with secA311 and secA318, which were scored as - and + in Table 3; similar results were observed with secA313 and -315, which are similar to secA318 (data not shown). No significant difference was observed with the other four strains (scored as ++ to ++++ in Table 3). In contrast, the amount of mature MalE was increased more than twofold by the secA (prlD5) mutation.

View larger version (42K): In this window In a new window Download PPT slide   Figure 3. Effect of the secA alleles on MalE export, determined by the efficiency of signal sequence cleavage. All secA alleles were transduced into the KK3 strain bearing the malE14-1 mutation. (A) Cells were pulse labeled for 30 sec and chased for 12 min. Lysates were immunoprecipitated with anti-MalE antibodies. The amounts of precursor and mature MalE were measured, and the relative amount of mature protein is indicated. Each bar represents the average of three independent cultures; error bars indicate the SD. (B) Time course of preMalE export. Cells were pulse labeled for 30 sec and chased for the indicated times. p, preMalE; m, mature MalE.

We also performed a series of pulse-chase experiments to determine whether export of the MalE14-1 protein occurs with similar kinetics in secA⁺ and secA mutant strains (Fig 3B). In secA⁺ cells, 5–6% of the protein synthesized during the 30-sec pulse was converted into mature MalE, and there was no significant increase during the chase period. The difference between this value and that shown in Fig 3A reflects the instability of cytoplasmic preMalE (POGLIANO and BECKWITH 1993 ). The strongest difference was observed with secA311, where only 1–2% of mature MalE was produced during the pulse, with no detectable increase during the 15-min chase. In addition, we consistently observed that the total amount of MalE synthesized in a strain carrying the secA311 allele was reduced two- to fivefold. The most likely explanation of this observation is that the added effects of the secA311 and malE14-1 mutations result in too little MalE in the periplasm to fully induce the maltose regulon. This difference probably also explains the drastic effect of secA311 on maltose fermentation (Table 3) and on growth in minimal maltose medium (Fig 2). With secA307 and -315, there was an approximately twofold reduction in the amount of mature MalE. Surprisingly, the effect of secA316 on preMalE signal sequence cleavage was smaller than that of secA307 and -315, although its effect on indicator plates was more severe than that of secA307. It is possible that these secA alleles also affect other components of the maltose import system, such as integration of MalF and MalG into the membrane. If this is the case, the good correlation observed between phenotypes on maltose indicator plates and growth rates in minimal maltose medium may not completely extend to the export of the MalE14-1 protein.

In conclusion, these experiments showed that none of the secA suppressors confers a Prl phenotype and that secA311, the only mutation affecting the SSBD domain, showed the strongest effect with all tested malE signal sequence mutations.

The effect of secA suppressors on an AP signal sequence mutation:
To provide an independent quantitative assay of protein export in strains carrying the Ara^R secA alleles, we used the Mph56 strain that carries the phoA73 mutation altering the signal sequence of AP. The phoA gene is constitutively expressed in this strain and the L14Q substitution results in a 70% decrease in AP export (MICHAELIS et al. 1986 ). On plates containing the PhoA substrate XP, the secretion phenotype can be scored by the blue color intensity of the colonies. When compared to the secA⁺ strain, all strains carrying the Ara^R secA alleles formed more pale colonies, indicative of a Sec phenotype. In contrast, a strain carrying the secA (prlD5) allele formed darker colonies, as expected (data not shown). To quantify the secretion defect conferred by the Ara^R secA alleles, we measured the steady-state level of AP activity in derivatives of the Mph56 strain (Fig 4A). The same general trend described in Table 3 was also observed with the phoA73 signal sequence mutation. The secA311 allele had a strong defect resulting in a 10- to 20-fold reduction in AP activity. The secA313, -315, and -316 alleles had an intermediate defect, resulting in a 3- to 5-fold reduction in AP activity. The remaining mutants had a weaker defect, resulting in a 2-fold reduction in AP activity. It should be noted that secA318, which had an effect similar to that of secA301, -302, and -307 with phoA73, exhibited the most variable effects with the different malE signal sequence mutations (Table 3). In contrast, the three secA (prlD) alleles tested improved export mediated by the phoA73 signal sequence mutation (Fig 4D).

View larger version (42K): In this window In a new window Download PPT slide   Figure 4. Effect of secA alleles on the export of alkaline phosphatase derivatives. Cells were grown at 37° to an A600nm of 0.2 in LB medium and induced with 0.2% arabinose for 10 min. Cells were assayed for AP activity. Each bar represents the average from three independent cultures; error bars indicate the SD. In A–F, the activity determined with the secA+ strain was taken as 100% and used to normalize the values measured with the secA mutant strains. (A and D) Effect of secA alleles on export of the phoA73 mutant; all secA alleles were transduced into the Mph56 strain; cultures were not induced with arabinose. (B and C) The strains are derivatives of the DHB3 (B) or MM2 (C) strains, carry the pDB3 plasmid, and express wild-type AP. (E and F) The strains are derivatives of the KK1 strain, carry the pBADhBK plasmid, and express a PAI2::AP fusion. The significance of the results was evaluated with the Mann-Whitney nonparametric test. (A) For PhoA73, P values were 0.08 for all secA alleles. (D) P values were 0.08 for all secA (prlD) alleles. (B) For wild-type AP, the P value was 0.08 for secA302, secA311, and secA315, when compared to secA+. (C) The results were not significantly different, except for secA (prlD21) (P value was 0.08). (E) For PAI2::AP, P values were 0.08 for all mutant alleles, when compared to secA+. (F) The P value was 0.08 for secA (prlD5) and secA (prlD21). When PAI2::AP and wild-type AP were compared in B and E, the P value was 0.08 for all mutant alleles except secA302 and secA311.

Differential effect on PAI2::AP and wild-type AP export:
Most secA suppressors conferred only a weak secretion defect on the export of mutated MalE signal sequences (Table 3 and Fig 3). This is probably due to the fact that the Ara^R suppressors were screened for forming blue colonies on selective plates containing XP and thus are still able to export the PAI2::AP fusions. The mechanism of suppression by secG mutations was proposed to result from a selective reduction in the export kinetics of PAI2::AP, without affecting that of wild-type AP (BOST and BELIN 1995 ; BOST et al. 2000 ). PAI2, like ovalbumin, lacks a leader peptidase cleavage site at the end of its signal sequence. Thus, we cannot analyze the export of the chimeric proteins by pulse-chase labeling and determination of cleavage efficiency.

We therefore determined the effect of the secA alleles on the export of wild-type AP (Fig 4B) and PAI2::AP (Fig 4E) by measuring the amount of AP activity after 10 min of induction with arabinose. For most secA suppressors, the accumulation of active PAI2::AP at this early time after induction was less than that measured in the secA⁺ strain (Fig 4E). Interestingly, the PAI2::AP export defect was again stronger with secA311 than with the other alleles. In contrast, accumulation of active AP was slightly higher [secA (prlD5)] or equal in strains carrying a secA (prlD) mutation when compared to the secA⁺ isogenic strain (Fig 4F). The situation was somewhat different for wild-type AP export: Five of the eight secA alleles tested conferred only a slight reduction in export, and only secA311 showed a twofold reduction in wild-type AP export (Fig 4B). A similar slight reduction in wild-type AP export was also detected with the secA (prlD) alleles (Fig 4C).

When PAI2::AP and wild-type AP export were compared, the secA alleles showed three different behaviors. Surprisingly, secA301 and secA307 had a slightly stronger effect on wild-type AP export than on PAI2::AP export (P value = 0.08). Mutants secA302 and secA311 have the same effect on both proteins. Finally, the four mutants that localize to NBDII had a weaker effect on wild-type AP (P value = 0.08) and thus appear selective for the chimeric protein. Thus, the secA suppressors of the toxic murine signal sequence appear to have a more complex phenotype than that of the secG suppressors (BOST and BELIN 1995 ; BOST et al. 2000 ), and suppression is not necessarily achieved by a selective reduction in PAI2::AP export.

Export of wild-type proteins is not drastically affected by the secA suppressors:
Since the Sec phenotype of the secA suppressors was defined in strains carrying malE and phoA signal sequence mutations (BEDOUELLE et al. 1980 ), the observed phenotypes could reflect interactions that are particularly defective with mutated signal sequences. Indeed, seven secA suppressors exerted only a minor effect on wild-type AP export (Fig 4B). It was therefore interesting to determine whether a secretion defect could be observed with other wild-type preproteins. Short pulse-labelings and immunoprecipitations were performed to quantify the export of RbsB, DegP, OmpA, and MalE; export efficiency was determined by measuring the extent of signal sequence cleavage calculated from the relative amounts of precursor and mature proteins (Fig 5). The strongest secretion defect was observed with secA311, although a more than twofold increase of the amount of uncleaved precursor was observed only with RbsB and DegP. A similar effect was also observed for RbsB with secA302, secA315, and secA316 and for DegP and OmpA with secA316. Finally, practically no export defect was observed for the four remaining mutants. The effects observed with wild-type MalE are too small to produce an altered phenotype on maltose indicator plates.

View larger version (56K): In this window In a new window Download PPT slide   Figure 5. Effect of the secA alleles on wild-type preprotein export, determined by the efficiency of signal sequence cleavage. All secA alleles were transduced into the KK1 strain. Cells were pulse labeled for 30 sec at 37°. Lysates were immunoprecipitated with anti-RbsB, anti-MalE-DegP, and anti-OmpA antibodies. For each protein analyzed, the precursor and the mature form were measured, and the relative amount of precursor is indicated. Each bar graph represents the average from three independent cultures; error bars indicate the SD.

Some Prl mutations decrease the kinetics of export of wild-type proteins. For instance, the secA (prlD21) allele confers a slight secretion defect on MalE and LamB (HUIE and SILHAVY 1995 ). A similar effect was observed with RbsB and DegP, but not with OmpA (Fig 5). The two other secA (prlD) alleles did not significantly affect the export of these three wild-type proteins.

SecA expression is translationally autoregulated (NAKATOGAWA and ITO 2001 ), and its expression can reach up to 20-fold its normal level in case of severe secretion defects caused (i) by the jamming of the Sec machinery with LacZ fusions or (ii) by Sec mutations in secA, secD, secE, or secY (RIGGS et al. 1988 ; ROLLO and OLIVER 1988 ). Thus, the SecA steady-state level provides an indirect way to assess the overall export capacity. SecA levels in exponentially growing cells were quantified by Western blot for each suppressor strain. The data in Fig 6A show a less than twofold increase in SecA expression in seven of the strains, which therefore exhibit only a very weak overall secretion defect. It is interesting to note that secA313 showed a nearly twofold increase in SecA level, although this allele had little effect on the export of five wild-type proteins (Fig 4B and Fig 5). In the secA307 strain, we detected a 50% increase in SecA level, even though this allele slightly accelerated DegP and MalE export (Fig 5). With secA311, which had the strongest defect in all previous assays, the level of SecA expression was increased threefold. As previously reported (HUIE and SILHAVY 1995 ), secA (prlD21) showed a threefold increase in SecA level and the other secA (prlD) alleles had only a marginal effect (Fig 6B).

View larger version (41K): In this window In a new window Download PPT slide   Figure 6. Level of SecA expression in the different suppressor strains. (A) AraR suppressors. (B) Prl mutants. All secA alleles were transduced into the KK1 strain. Exponentially growing cells were lysed and extracts were subjected to 10% SDS-PAGE. Samples were diluted according to the A600nm of the cultures and each lane contained the same amount of cell equivalents (0.3 mA600nm). SecA was detected by chemiluminescence and quantification was performed by scanning nonsaturated films with an Arcuss II scanner (Agfa) and by using the ImageQuant version 3.22 software. Each bar represents the mean of three independent cultures; error bars indicate the SD.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The secA suppressors affect three different domains of the protein:
We describe here the characterization of a novel class of secA alleles. Since only 2 of 21 independent mutations affect the same site, this collection is not yet saturated for all possible suppressor sites. The 19 substitutions (Fig 1) span a large part of SecA, from the N-terminal region to NBDII, but none was found in the C-terminal 240 residues. A large collection of secA suppressors of the cold-sensitive secY205 allele has been described (MATSUMOTO et al. 2000 ). Since the 51 mutations map to 40 different sites, this collection is saturated only slightly more than our collection, and the total number of sites that could be identified by the two selections is of the same order. Nevertheless, there is no overlap between these two sets of secA alleles, although the Y117H substitution reported here is adjacent to the A116V and L118Q ones. In contrast, two of the secY205 suppressors are identical to secA (prlD3) and secA (prlD5), and four are different substitutions at residues affected by secA (prlD) mutations.

Using malE and phoA signal sequence mutations, we observed that all of our secA alleles enhance their export defect and therefore confer a weak Sec phenotype. The relative strengths of their phenotypes appear to define three classes of alleles: the N-terminal and NBDI mutants (class I), the SSBD mutant (class II), and the last four mutants (class III). This distribution is purely operational at this stage, since secA318 behaves as a class III allele in three malE strains, but as a class I allele in strains carrying the malE18 or phoA73 mutations. With the help of J. Hunt (Columbia University, New York), we could calculate the spatial distribution of these mutant sites in the B. subtilis SecA crystal structure. Our initial conclusion was confirmed by the spatial distribution of the sites. Indeed, all 19 substituted residues identified in Fig 1 are clustered into three folded domains of SecA: secA301-307 and secA319 in the first domain, secA308–312 in the second domain, and secA313–318 in the third domain (J. F. HUNT, S. WEINKAUF, D. B. OLIVER and J. DEISENHOFER, personal communication).

With mutated signal sequences, the secA Ara^R suppressors have a phenotype opposite to that of secA (prlD) mutants:
Two types of mutations have been isolated in most genes encoding components of the protein translocation machinery: Sec mutations and Prl mutations. Suppressors of the toxic PAI2 signal sequence represent a third class of secA mutations. Since Ara^R and Prl mutations suppress inefficient signal sequences, they were systematically compared.

In contrast with the small increase observed with secA (prlD5), the Ara^R alleles do not improve the inefficient export mediated by this mammalian signal sequence. The three secA (prlD) alleles tested confer the same Ara^S phenotype as the secA⁺ strain. Moreover, the irreversibility of this lethal phenotype, measured after induction of the toxic chimeric protein, was more pronounced with the secA (prlD) alleles when compared to a secA⁺ strain (data not shown). Most importantly, the secA Ara^R suppressors decrease export mediated by mutated malE and phoA signal sequences, which are suppressed to various extents by secA (prlD) alleles. Although the strength of the enhanced Sec phenotype conferred by our secA alleles depends on the signal sequence tested, export was more drastically affected with mutated signal sequences than with their wild-type counterparts. Thus, these secA alleles appear to have an increased specificity that preferentially excludes weak signal sequences, while the secA (prlD) alleles improve their export activity.

With wild-type preproteins, the situation is more complex. A weak Sec phenotype can be observed with conditional-lethal sec mutants grown at the permissive temperature (RIGGS et al. 1988 ), with some sec (prl) alleles, including secA (prlD21) (SAKO and IINO 1988 ; STADER et al. 1989 ; HUIE and SILHAVY 1995 ), and with several of our suppressors. While this is probably often the result of a reduced activity of the mutated protein, an enhanced discrimination of signal sequences would have the same effect. If the signal sequence discrimination function is altered, we may also identify wild-type signal sequences that are more rapidly exported in the secA suppressor strains. This is indeed what was observed with secA307 and wild-type DegP and MalE.

Taken together, our results suggest that the secA Ara^R suppressors alter the same function as the secA (prlD) mutations, i.e., the interaction of SecA with signal sequences. With mutated or inefficient signal sequences, these alleles display an opposite phenotype. The variability observed with different signal sequences appears to be an inherent property of the export process, considering that the translocase must process a highly degenerate set of signal sequences and that SecA is a multifunctional protein that participates in all known steps of export. At this stage, it is difficult to make predictions concerning the type of interaction(s) between the signal sequence and different domains of SecA.

Mechanism of suppression of the PAI2 signal sequence:
We observed three different behaviors for the export of PAI2::AP and wild-type AP in the secA strains, suggesting that there are at least three different ways to suppress the toxicity of PAI2::AP. First, mutations affecting the putative SSBD motif are predicted to generally decrease the affinity of SecA for signal sequences. Although we did not directly measure the affinity of SecA311 (I327S) for signal sequences, the phenotypes of the secA311 mutation are fully compatible with this interpretation. Furthermore, substitutions at Y326, which affects the adjacent residue, strongly affect the SecA-preprotein interaction (KOURTZ and OLIVER 2000 ). By reducing protein export mediated by nearly all the signal sequences tested, including the chimeric protein, this type of mutation probably prevents translocation jamming by PAI2::AP. Second, mutations like secA313, sec3A15, secA316, and secA318 show a behavior similar to that of secG suppressors, since they selectively slow down PAI2::AP export when compared to that of wild-type AP. Third, mutations affecting the N-terminal region are somewhat surprising, since they show a stronger defect in the export of wild-type AP when compared to that of PAI2::AP and have only a very weak Sec phenotype. These mutations could make the translocase less susceptible to blockage by PAI2::AP, perhaps by altering the dynamic interaction(s) of SecYEG with SecA (ECONOMOU and WICKNER 1994 ; NISHIYAMA et al. 1996 ).

Are the nucleotide binding domains of SecA involved in signal sequence recognition?
It has been proposed that the main function affected in Prl mutants is signal sequence recognition, suggesting that this recognition provides a proofreading activity to the translocase (OSBORNE and SILHAVY 1993 ). In the case of the SecYEG complex, the Prl phenotype is associated with structural changes affecting the dynamic interaction of the translocase subunits. Indeed, Prl mutations in secY and secE (i) cause a "general relaxation" of the translocase and (ii) stabilize SecA at the SecYEG complex (NOUWEN et al. 1996 ; VAN DER WOLK et al. 1998 ; DUONG and WICKNER 1999 ). The abolition of secA (prlD) suppression in a secG null strain extends the notion of a dynamic interaction to SecA (RAMAMURTHY et al. 1998 ). The distribution of the secA Ara^R and secA (prlD) mutations shows that the SSBD and both NBDI and NBDII motifs contribute to the recognition of signal sequences. The fact that some mutations affect SSBD was expected, since this domain crosslinks to the OmpA signal sequence (KIMURA et al. 1991 ). However, most of our mutations affect either NBDI or NBDII. This may be surprising since these regions are thought to be involved in the binding and hydrolysis of ATP and therefore in the catalytic activity of SecA during translocation. Our results suggest that the nucleotide binding domains of SecA also function during signal sequence recognition and thus contribute to the postulated proofreading activity of the translocase. A similar conclusion was reached by HUIE and SILHAVY 1995 on the basis of the phenotypes of secA (prlD) alleles.

The concept of proofreading requires a "stop and go" energy-dependent process, as is the case for protein synthesis (THOMPSON et al. 1986 ). After ATP binding and insertion of the preprotein in the membrane, a SecA-SecYEG bound state, the stop and go process could depend on the ATPase activity of SecA. The interaction of SecYEG with the preprotein and with SecA would set the timer for ATP hydrolysis and release of SecA. A longer interaction would be expected to favor the release of a low-affinity signal sequence from the translocase. In contrast, a shorter interaction would be less discriminative. Signal peptides modulate the ATPase activity of SecA in a complex manner (WANG et al. 2000 ; TRIPLETT et al. 2001 ). The observation that several SecA (PrlD) proteins have an increased membrane-associated ATPase activity strongly supports this model (SCHMIDT et al. 2000 ). A faster recognition step could also lead to overall faster translocation. Indeed, the secY (prlA4) mutation influences the rate of translocation (NOUWEN et al. 1996 ; VAN DER WOLK et al. 1998 ) probably by accelerating SecA deinsertion (NISHIYAMA et al. 1999 ). The secA mutations described here, together with secA (prlD) mutations, provide useful tools for dissecting the molecular mechanism of signal sequence recognition during protein export.

	ACKNOWLEDGMENTS

We thank D. Oliver, A. Economou, C. Georgopoulos, M. Ehrmann, and J. Hunt for helpful discussions; C. Georgopoulos for critical reading of the manuscript; S. Bost, C. Rudaz, and F. Silva for the isolation of the suppressor strains; F. Silva for excellent technical assistance; and Dr. M. Berlyn for help with the registration of secA alleles. This work was supported by the "Fonds National Suisse de la Recherche Scientifique" and the State of Geneva.

Manuscript received April 16, 2002; Accepted for publication August 2, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AKIMARU, J., S. MATSUYAMA, H. TOKUDA, and S. MIZUSHIMA, 1991  Reconstitution of a protein translocation system containing purified SecY, SecE, and SecA from Escherichia coli.. Proc. Natl. Acad. Sci. USA 88:6545-6549.[Abstract/Free Full Text]

ARKOWITZ, R. A. and W. WICKNER, 1994  SecD and SecF are required for the proton electrochemical gradient stimulation of preprotein translocation. EMBO J. 13:954-963.[Medline]

BEDOUELLE, H., P. J. BASSFORD, JR., A. V. FOWLER, I. ZABIN, and J. BECKWITH et al., 1980  Mutations which alter the function of the signal sequence of the maltose binding protein of Escherichia coli.. Nature 285:78-81.[Medline]

BESSONNEAU, P., V. BESSON, I. COLLINSON, and F. DUONG, 2002  The SecYEG preprotein translocation channel is a conformationally dynamic and dimeric structure. EMBO J. 21:995-1003.[Medline]

BLOBEL, G. and B. DOBBERSTEIN, 1975  Transfer of proteins across membranes. I. Presence of proteolytically processed and unprocessed nascent immunoglobulin light chains. J. Cell Biol. 67:835-851.[Abstract/Free Full Text]

BOST, S. and D. BELIN, 1995  A new genetic selection identifies essential residues in SecG, a component of the Escherichia coli protein export machinery. EMBO J. 14:4412-4421.[Medline]

BOST, S. and D. BELIN, 1997  prl mutations in the Escherichia coli secG gene. J. Biol. Chem. 272:4087-4093.[Abstract/Free Full Text]

BOST, S., F. SILVA, C. RUDAZ, and D. BELIN, 2000  Both transmembrane domains of SecG contribute to signal sequence recognition by the Escherichia coli protein export machinery. Mol. Microbiol. 38:575-587.[Medline]

BOYD, D., C. MANOIL, and J. BECKWITH, 1987  Determinants of membrane protein topology. Proc. Natl. Acad. Sci. USA 84:8525-8529.[Abstract/Free Full Text]

CASADABAN, M. J., 1976  Transposition and fusion of the lac genes to selected promoters in Escherichia coli using bacteriophage lambda and mu. J. Mol. Biol. 104:541-555.[Medline]

CASADABAN, M. J. and S. N. COHEN, 1980  Analysis of gene control signals by DNA fusion and cloning in Escherichia coli.. J. Mol. Biol. 138:179-207.[Medline]

CHEN, L. L. and P. C. TAI, 1986  Roles of H⁺-ATPase and proton motive force in ATP-dependent protein translocation in vitro.. J. Bacteriol. 167:389-392.[Abstract/Free Full Text]

COLLIER, D. N., V. A. BANKAITIS, J. B. WEISS, and P. J. BASSFORD, JR., 1988  The antifolding activity of SecB promotes the export of the E. coli maltose-binding protein. Cell 53:273-283.[Medline]

DANESE, P. N. and T. J. SILHAVY, 1998  Targeting and assembly of periplasmic and outer-membrane proteins in Escherichia coli.. Annu. Rev. Genet. 32:59-94.[Medline]

DERMAN, A. I., J. W. PUZISS, P. J. BASSFORD, JR., and J. BECKWITH, 1993  A signal sequence is not required for protein export in prlA mutants of Escherichia coli.. EMBO J. 12:879-888.[Medline]

DRIESSEN, A. J. and W. WICKNER, 1991  Proton transfer is rate-limiting for translocation of precursor proteins by the Escherichia coli translocase. Proc. Natl. Acad. Sci. USA 88:2471-2475.[Abstract/Free Full Text]

DRIESSEN, A. J., P. FEKKES, and J. P. VAN DER WOLK, 1998  The Sec system. Curr. Opin. Microbiol. 1:216-222.[Medline]

DRIESSEN, A. J., E. H. MANTING, and C. VAN DER DOES, 2001  The structural basis of protein targeting and translocation in bacteria. Nat. Struct. Biol. 8:492-498.[Medline]

DUONG, F. and W. WICKNER, 1997  Distinct catalytic roles of the SecYE, SecG and SecDFyajC subunits of preprotein translocase holoenzyme. EMBO J. 16:2756-2768.[Medline]

DUONG, F. and W. WICKNER, 1999  The PrlA and PrlG phenotypes are caused by a loosened association among the translocase SecYEG subunits. EMBO J. 18:3263-3270.[Medline]

ECONOMOU, A. and W. WICKNER, 1994  SecA promotes preprotein translocation by undergoing ATP-driven cycles of membrane insertion and deinsertion. Cell 78:835-843.[Medline]

FIKES, J. D. and P. J. BASSFORD, JR., 1989  Novel secA alleles improve export of maltose-binding protein synthesized with a defective signal peptide. J. Bacteriol. 171:402-409.[Abstract/Free Full Text]

GELLER, B. L., N. R. MOVVA, and W. WICKNER, 1986  Both ATP and the electrochemical potential are required for optimal assembly of pro-OmpA into Escherichia coli inner membrane vesicles. Proc. Natl. Acad. Sci. USA 83:4219-4222.[Abstract/Free Full Text]

GUZMAN, L. M., D. BELIN, M. J. CARSON, and J. BECKWITH, 1995  Tight regulation, modulation, and high-level expression by vectors containing the arabinose P_BAD promoter. J. Bacteriol. 177:4121-4130.[Abstract/Free Full Text]

HANADA, M., K. NISHIYAMA, and H. TOKUDA, 1996  SecG plays a critical role in protein translocation in the absence of the proton motive force as well as at low temperature. FEBS Lett. 381:25-28.[Medline]

HUIE, J. L. and T. J. SILHAVY, 1995  Suppression of signal sequence defects and azide resistance in Escherichia coli commonly result from the same mutations in secA.. J. Bacteriol. 177:3518-3526.[Abstract/Free Full Text]

KIM, Y. J., T. RAJAPANDI, and D. OLIVER, 1994  SecA protein is exposed to the periplasmic surface of the E. coli inner membrane in its active state. Cell 78:845-853.[Medline]

KIMURA, E., M. AKITA, S. MATSUYAMA, and S. MIZUSHIMA, 1991  Determination of a region in SecA that interacts with presecretory proteins in Escherichia coli.. J. Biol. Chem. 266:6600-6606.[Abstract/Free Full Text]

KOURTZ, L. and D. OLIVER, 2000  Tyr-326 plays a critical role in controlling SecA-preprotein interaction. Mol. Microbiol. 37:1342-1356.[Medline]

LECKER, S., R. LILL, T. ZIEGELHOFFER, C. GEORGOPOULOS, and P. J. BASSFORD, JR. et al., 1989  Three pure chaperone proteins of Escherichia coli—SecB, trigger factor and GroEL—form soluble complexes with precursor proteins in vitro.. EMBO J. 8:2703-2709.[Medline]

MANTING, E. H., C. VAN DER DOES, H. REMIGY, A. ENGEL, and A. J. DRIESSEN, 2000  SecYEG assembles into a tetramer to form the active protein translocation channel. EMBO J. 19:852-861.[Medline]

MATSUMOTO, G., H. NAKATOGAWA, H. MORI, and K. ITO, 2000  Genetic dissection of SecA: suppressor mutations against the secY205 translocase defect. Genes Cells 5:991-999.[Abstract]

MEYER, T. H., J. F. MENETRET, R. BREITLING, K. R. MILLER, and C. W. AKEY et al., 1999  The bacterial SecY/E translocation complex forms channel-like structures similar to those of the eukaryotic Sec61p complex. J. Mol. Biol. 285:1789-1800.[Medline]

MICHAELIS, S., J. F. HUNT, and J. BECKWITH, 1986  Effects of signal sequence mutations on the kinetics of alkaline phosphatase export to the periplasm in Escherichia coli.. J. Bacteriol. 167:160-167.[Abstract/Free Full Text]

MILLER, J. H., 1992 A Short Course in Bacterial Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

NAKATOGAWA, H. and K. ITO, 2001  Secretion monitor, SecM, undergoes self-translation arrest in the cytosol. Mol. Cell 7:185-192.[Medline]

NEWITT, J. A. and H. D. BERNSTEIN, 1998  A mutation in the Escherichia coli secY gene that produces distinct effects on inner membrane protein insertion and protein export. J. Biol. Chem. 273:12451-12456.[Abstract/Free Full Text]

NISHIYAMA, K., S. MIZUSHIMA, and H. TOKUDA, 1993  A novel membrane protein involved in protein translocation across the cytoplasmic membrane of Escherichia coli.. EMBO J. 12:3409-3415.[Medline]

NISHIYAMA, K., M. HANADA, and H. TOKUDA, 1994  Disruption of the gene encoding p12 (SecG) reveals the direct involvement and important function of SecG in the protein translocation of Escherichia coli at low temperature. EMBO J. 13:3272-3277.[Medline]

NISHIYAMA, K., T. SUZUKI, and H. TOKUDA, 1996  Inversion of the membrane topology of SecG coupled with SecA-dependent preprotein translocation. Cell 85:71-81.[Medline]

NISHIYAMA, K, A. FUKUDA, K. MORITA, and H. TOKUDA, 1999  Membrane deinsertion of SecA underlying proton motive force-dependent stimulation of protein translocation. EMBO J. 18:1049-1058.[Medline]

NOUWEN, N., B. DE KRUIJFF, and J. TOMMASSEN, 1996  PrlA suppressors in Escherichia coli relieve the proton electrochemical gradient dependency of translocation of wild-type precursors. Proc. Natl. Acad. Sci. USA 93:5953-5957.[Abstract/Free Full Text]

OLIVER, D. B. and J. BECKWITH, 1981  E. coli mutant pleiotropically defective in the export of secreted proteins. Cell 25:765-772.[Medline]

ORITA, M., H. IWAHANA, H. KANAZAWA, K. HAYASHI, and T. SEKIYA, 1989  Detection of polymorphisms of human DNA by gel electrophoresis as single-strand conformation polymorphisms. Proc. Natl. Acad. Sci. USA 86:2766-2770.[Abstract/Free Full Text]

OSBORNE, R. S. and T. J. SILHAVY, 1993  PrlA suppressor mutations cluster in regions corresponding to three distinct topological domains. EMBO J. 12:3391-3398.[Medline]

POGLIANO, J. A. and J. BECKWITH, 1994  SecD and SecF facilitate protein export in Escherichia coli.. EMBO J. 13:554-561.[Medline]

POGLIANO, K. J. and J. BECKWITH, 1993  The Cs sec mutants of Escherichia coli reflect the cold sensitivity of protein export itself. Genetics 133:763-773.[Abstract]

PRINZ, W. A., D. H. BOYD, M. EHRMANN, and J. BECKWITH, 1998  The protein translocation apparatus contributes to determining the topology of an integral membrane protein in Escherichia coli.. J. Biol. Chem. 273:8419-8424.[Abstract/Free Full Text]

QI, H. Y. and H. D. BERNSTEIN, 1999  SecA is required for the insertion of inner membrane proteins targeted by the Escherichia coli signal recognition particle. J. Biol. Chem. 274:8993-8997.[Abstract/Free Full Text]

RAMAMURTHY, V., V. DAPÍC, and D. OLIVER, 1998  SecG and temperature modulate expression of azide-resistant and signal sequence suppressor phenotypes of Escherichia coli secA mutants. J. Bacteriol. 180:6419-6423.[Abstract/Free Full Text]

RIGGS, P. D., A. I. DERMAN, and J. BECKWITH, 1988  A mutation affecting the regulation of a secA-lacZ fusion defines a new sec gene. Genetics 118:571-579.[Abstract/Free Full Text]

ROLLO, E. E. and D. B. OLIVER, 1988  Regulation of the Escherichia coli secA gene by protein secretion defects: analysis of secA, secB, secD, and secY mutants. J. Bacteriol. 170:3281-3282.[Abstract/Free Full Text]

RUSSO, F. D., J. M. SLAUCH, and T. J. SILHAVY, 1993  Mutations that affect separate functions of OmpR, the phosphorylated regulator of porin transcription in Escherichia coli.. J. Mol. Biol. 231:261-273.[Medline]

SAKO, T. and T. IINO, 1988  Distinct mutation sites in prlA suppressor mutant strains of Escherichia coli respond either to suppression of signal peptide mutations or to blockage of staphylokinase processing. J. Bacteriol. 170:5389-5391.[Abstract/Free Full Text]

SCHATZ, P. J. and J. BECKWITH, 1990  Genetic analysis of protein export in Escherichia coli.. Annu. Rev. Genet. 24:215-248.[Medline]

SCHIEBEL, E., A. J. DRIESSEN, F. U. HARTL, and W. WICKNER, 1991  µH⁺ and ATP function at different steps of the catalytic cycle of preprotein translocase. Cell 64:927-939.[Medline]

SCHMIDT, M. G., E. E. ROLLO, J. GRODBERG, and D. B. OLIVER, 1988  Nucleotide sequence of the secA gene and secA(Ts) mutations preventing protein export in Escherichia coli.. J. Bacteriol. 170:3404-3414.[Abstract/Free Full Text]

SCHMIDT, M., H. DING, V. RAMAMURTHY, I. MUKERJI, and D. OLIVER, 2000  Nucleotide binding activity of SecA homodimer is conformationally regulated by temperature and altered by prlD and azi mutations. J. Biol. Chem. 275:15440-15448.[Abstract/Free Full Text]

SCOTTI, P. A., Q. A. VALENT, E. H. MANTING, M. L. URBANUS, and A. J. DRIESSEN et al., 1999  SecA is not required for signal recognition particle-mediated targeting and initial membrane insertion of a nascent inner membrane protein. J. Biol. Chem. 274:29883-29888.[Abstract/Free Full Text]

STADER, J., L. J. GANSHEROFF, and T. J. SILHAVY, 1989  New suppressors of signal-sequence mutations, prlG, are linked tightly to the secE gene of Escherichia coli.. Genes Dev. 3:1045-1052.[Abstract/Free Full Text]

THOMPSON, R. C., D. B. DIX, and A. M. KARIM, 1986  The reaction of ribosomes with elongation factor Tu.GTP complexes. Aminoacyl-tRNA-independent reactions in the elongation cycle determine the accuracy of protein synthesis. J. Biol. Chem. 261:4868-4874.[Abstract/Free Full Text]

TRIPLETT, T. L., A. R. SGRIGNOLI, F. B. GAO, Y. B. YANG, and P. C. TAI et al., 2001  Functional signal peptides bind a soluble N-terminal fragment of SecA and inhibit its ATPase activity. J. Biol. Chem. 276:19648-19655.[Abstract/Free Full Text]

VAN DER WOLK, J. P., P. FEKKES, A. BOORSMA, J. L. HUIE, and T. J. SILHAVY et al., 1998  PrlA4 prevents the rejection of signal sequence defective preproteins by stabilizing the SecA-SecY interaction during the initiation of translocation. EMBO J. 17:3631-3639.[Medline]

WANG, L., A. MILLER, and D. A. KENDALL, 2000  Signal peptide determinants of SecA binding and stimulation of ATPase activity. J. Biol. Chem. 275:10154-10159.[Abstract/Free Full Text]

WILD, J., P. ROSSMEISSL, W. A. WALTER, and C. A. GROSS, 1996  Involvement of the DnaK-DnaJ-GrpE chaperone team in protein secretion in Escherichia coli.. J. Bacteriol. 178:3608-3613.[Abstract/Free Full Text]

YAHR, T. L. and W. T. WICKNER, 2000  Evaluating the oligomeric state of SecYEG in preprotein translocase. EMBO J. 19:4393-4401.[Medline]

	NOTE ADDED IN PROOF

The structure of SecA was recently published (J. F. HUNT, S. WEINKAUF, L. HENRY, J. J. FAK, P. MCNICHOLAS et al., 2002, Nucleotide control of interdomain interactions in the conformational reaction cycle of SecA. Science 297: 2018–2026).

This article has been cited by other articles:

				S. Bieler, F. Silva, C. Soto, and D. Belin Bactericidal Activity of both Secreted and Nonsecreted Microcin E492 Requires the Mannose Permease. J. Bacteriol., October 1, 2006; 188(20): 7049 - 7061. [Abstract] [Full Text] [PDF]

				E. Papanikou, S. Karamanou, C. Baud, M. Frank, G. Sianidis, D. Keramisanou, C. G. Kalodimos, A. Kuhn, and A. Economou Identification of the Preprotein Binding Domain of SecA J. Biol. Chem., December 30, 2005; 280(52): 43209 - 43217. [Abstract] [Full Text] [PDF]