Previous Article | Next Article ![]()
Applied and Environmental Microbiology, March 2003, p. 1739-1747, Vol. 69, No. 3
0099-2240/03/$08.00+0 DOI: 10.1128/AEM.69.3.1739-1747.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Bacteriology Group, International Centre for Genetic Engineering and Biotechnology, 34012 Trieste, Italy
Received 16 October 2002/ Accepted 16 December 2002
|
|
|---|
|
|
|---|
Quorum sensing is a mechanism for regulating gene expression in response to changes in cell density of a bacterial population (15). In gram-negative bacteria, N-acyl homoserine lactone (HSL) autoinducers appear to be the most commonly used signaling molecules; in most cases they are produced by an autoinducer synthase protein belonging to the LuxI family. A transcriptional activator belonging to the LuxR family forms a complex with the cognate autoinducer at high threshold levels to induce transcriptional activation of target genes. The autoinducers produced by different bacterial species differ in the length and structure of the acyl chain, and they are believed to be readily diffusible across the cell envelope into the growth medium where they accumulate. Accumulation continues until the cell density, and consequently also the HSL concentration, reaches a quorum, thereby activating the LuxR type protein, which then elicits the desired cellular response. Quorum sensing has been implicated in the regulation of biofilm formation, plasmid transfer, and motility and in several virulence factors (53).
Another important regulatory component involved in regulation in response to high cell density is the stationary-phase sigma factor RpoS (25, 30). RpoS is an alternative sigma factor which directs transcription of a large number of genes involved in adaptation to nutrient-limiting conditions and several environmental stresses as well as inducing the production of virulence factors (30, 47). In Escherichia coli and Pseudomonas spp., RpoS levels are induced as bacterial cultures enter the stationary phase (14, 26, 27), and it has also been reported that quorum sensing and RpoS cross-regulate their gene expression in Pseudomonas aeruginosa (28, 51).
The quorum-sensing system of a B. cepacia genomovar III cystic fibrosis respiratory isolate has been identified and characterized and consists of cepI and cepR genes. This bacterial isolate, designated K56-2, synthesizes an N-octanoyl HSL (C8-HSL) and an N-hexanoyl HSL (C6-HSL) (29, 32). The CepI/R quorum-sensing system of B. cepacia K56-2 has been implicated in the negative regulation of the siderophore ornibactin and in the positive regulation of a secreted protease. Similarly, the CepI/R quorum-sensing system from another B. cepacia genomovar III cystic fibrosis respiratory isolate, designated H111, has been recently identified and characterized and shown to be involved in regulating biofilm formation and swarming motility (23). In addition, quorum-sensing systems are present and conserved among the heterogeneous B. cepacia complex (19). The stationary-phase sigma factor RpoS, on the other hand, has to our knowledge not been reported in any bacterial strains belonging to the B. cepacia complex.
In this study we report the identification and characterization of the quorum-sensing system of the onion pathogen B. cepacia type strain ATCC 25416 which belongs to genomovar I. This strain has been isolated from a rotten onion, and it has a genome of 8.1 Mb composed of four circular replicons of 3.65 Mb, 3.17 Mb, 1.07 Mb and 200 Kbp (39). We have constructed a cepR-PcepI-lacZ reporter plasmid and shown (i) that a cepR knockout mutant results in a 70% decrease in HSL production, CepR responded to C6-HSL with only 15% the molar efficiency of C8-HSL, and CepR responds best to long chain HSL autoinducers and (ii) that cepI and cepR knockout mutants are attenuated in onion pathogenicity. The rpoS gene of B. cepacia ATCC 25416 has also been identified and characterized. B. cepacia RpoS did not display very high identity in its primary structure to RpoS belonging to the
-Proteobacteria. RpoS was not involved in HSL production, and quorum sensing had a negative effect on rpoS expression.
|
|
|---|
(21), and HB101::Tn5 (33) and were grown in LB medium (36) at 37°C. Antibiotics were added as required at the following final concentrations: tetracycline, 10 (for E. coli), 20 (for C. violaceum), or 300 (for B. cepacia) µg/ml; gentamicin, 10 (for E. coli) or 300 (for B. cepacia) µg/ml; ampicillin, 100 µg/ml (for E. coli); kanamycin, 50 (for E. coli) or 300 (for B. cepacia) µg/ml; and streptomycin, 100 µg/ml (for C. violaceum). |
View this table: [in a new window] |
TABLE 1. Plasmids used in this study
|
Cloning of quorum-sensing genes of B. cepacia ATCC 25416.
About 4 x 109 cells (each) of E. coli HB101 harboring the B. cepacia ATCC 25416 cosmid library and E. coli(pRK2103) and 2 x 108 cells of C. violaceum CVO26 were mixed. The suspension was applied to a 0.45-µm-pore-size membrane filter (Millipore Corp.) on an LB plate. After overnight incubation at 30°C, the cells were resuspended and spread on LB plates containing ampicillin (100 µg/ml), kanamycin (100 µg/ml), streptomycin (100 µg/ml), and tetracycline (20 µg/ml). Strain CVO26 is naturally resistant to ampicillin and streptomycin and resistant to kanamycin due to the mini-Tn5 present on the chromosome. Tetracycline will select transconjugants that have received the pLAFR3-based cosmid clone. These plates were incubated for 48 h at 30°C, and transconjugants that turned purple were further assayed. Three cosmids (pCQS1 to pCQS3) from the cosmid library could restore purple pigmentation, and they only shared a 9-kb EcoRI common insert. Further subcloning experiments confirmed that the 9-kb EcoRI fragment cloned in pLAFR3 (creating pLIR5) could restore pigmentation on strain CVO26. Thin-layer chromatographic analysis with extracts from spent supernatants of strain CVO26(pLIR5) showed that C6- and C8-HSL were now synthesized.
Transposon Tn5 mutagenesis.
Transposon Tn5 insertions within recombinant plasmid pLIR5 were obtained as described previously (33), with E. coli HB101::Tn5 as the source of the transposon. E. coli HB101 cells containing Tn5 insertions within plasmid pLIR5 were identified by purifying plasmid DNA from HB101::Tn5(pLIR5), using it to transform E. coli DH5
, and selecting for plasmids having tetracycline and kanamycin resistance. These recombinant plasmids in DH5
were delivered by triparental conjugation to C. violaceum CVO26 as described above. Transconjugants of strain CVO26(pLIR5::Tn5) lacking the expression of violacein (i.e., those which remained white) were purified, and the position of Tn5 was mapped within the 9-kb EcoRI insert. Three insertions were located in the cepR gene (data not shown).
Construction of a B. cepacia ATCC 25416 cepI and cepR knockout mutants.
Plasmid pLCIKm was constructed as follows. A 3.5-kb HindIII DNA fragment from pSUP2021containing the Kmr gene of transposon Tn5 was cloned into the corresponding site in cepI in plasmid pLIR5. Plasmid pLIR::Tn53 (which contain a Tn5 insertion in the cepR gene harbored in pLAFR3) and plasmid pLCIKm (which contains a kanamycin resistance gene cloned in the cepI gene) were homogenized with the corresponding target regions of the genome of B. cepacia ATCC 25416 by a marker-exchange procedure (9). Plasmid pPH1JI was used as the incoming IncP1-incompatible plasmid, and selections were made on LB plates containing kanamycin and gentamicin. This generated two genomic mutants, designated B. cepacia 25416-I and B. cepacia 25416-R, which harbored a Kmr cassette in the cepI gene and a Tn5 insertion in the cepR gene, respectively. The fidelity of each marker-exchange event was confirmed by Southern analysis (data not shown).
Construction of a B. cepacia CepR-based HSL-detecting plasmid.
The cep genes of B. cepacia ATCC 25416 were used to construct a plasmid detecting HSL molecules. A BamHI-HindIII DNA fragment of 3.5 kb from pBIR containing the whole cepR gene and the cepI promoter together with the first 120 bp of cepI were inserted into the corresponding sites of the vector pQF50 (11), yielding pSCR1. Plasmid pQF50 contains a promoterless lacZ gene, and in plasmid pSCR1, the transcription of the lacZ gene is under the control of the cepI promoter. The ß-galactosidase activity was determined as described previously (34, 43).
Purification, detection, and visualization of autoinducer molecules (HSLs).
The purification, detection, and visualization of HSL inducer molecules from culture supernatants were performed essentially as described previously (25, 34). Synthetic HSLs (C4-HSL, C6-HSL, C7-HSL, C8-HSL, C10-HSL, and C12-HSL) were purchased from Fluka Chemie AG (Buchs, Switzerland). For quantification of CepR activity, overnight E. coli DH5
(pSCR1) cultures were normalized to an optical density at 600 nm (OD600) of 0.1 in a volume of 20 ml of LB containing the desired HSL at the desired concentration. Cultures were then grown with agitation at 37°C for 6 h, and ß-galactosidase activities were determined. The presence of HSL was also detected on solid media by growing E. coli(pSCR1) on 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside (X-Gal) and ampicillin media in close proximity to the tester strain. The presence of HSL was observed when E. coli(pSCR1) turned blue (Fig. 1a).
![]() View larger version (14K): [in a new window] |
FIG. 1. (a) Activation of bacterial sensor strain in cross-streak experiments. The sensor strain E. coli DH5 (pSCR1) was cross streaked on LB agar plates (containing ampicillin and X-Gal) with B. cepacia ATCC 25416 used as a tester strain. (b) HSL bioassay with E. coli DH5 (pSCR1). DH5 (pSCR1) was grown in the presence of variousconcentrations of either C6-HSL ( ) or C8-HSL ( ), and ß-galactosidase activities were determined after 6 h. The values were determined with LB medium, the means of triplicate experiments are given, and the standard deviations are shown. (c) CepR-HSL response. E. coli DH5 (pSCR1) was grown for 6 h in the presence of a 100 nM concentration of either C4-, C6-, C8-, C10-, or C12-HSL, and ß-galactosidase activities were determined. The values were determined with LB medium, the means of triplicate experiments are given, and the standard deviations are shown. The values are expressed as percentages of the activity determined with C8-HSL.
|
![]() View larger version (24K): [in a new window] |
FIG. 4. (a) Production of HSLs at different growth stages from B. cepacia ATCC 25416 (filled bars) and from B. cepacia ATCC 25416-RPOS (open bars). Values were determined as described in Materials and Methods. (b) rpoS promoter activity at different growth stages measured in B. cepacia ATCC 25416(pRPR2) (filled bars), B. cepacia 25416-I(pRPR2) (open bars), B. cepacia 25416-I(pRPR2) plus 100 nM C8-HSL (shaded bars), and B. cepacia ATCC 25416(pMP190) (striped bars). Values were determined as described in Materials and Methods.
|
(pCOSRPOS-2C), the resulting DH5
(pCOSRPOS-2C)(pRBS-3Km) was grown overnight, and the culture was used in a triparental conjugation into Pseudomonas putida WCS358 (16) with E. coli(pRK2013) as a helper. After appropriate selection, pCOSRPOS-2CKm was selected. Transfer of the Kmr cassette by double-crossover homologous recombination from pRBS-3Km to pCOSRPOS-2 was verified by restriction and Southern analysis. The plasmid pCOSRPOS-2CKm was then used in a marker-exchange technique, as described above, in order to introduce site-specific insertion mutations into the rpoS gene of B. cepacia ATCC 25416. The fidelity of the marker-exchange event in the B. cepacia rpoS::Kmr mutant was confirmed by Southern analysis (data not shown). This mutant was designated B. cepacia 25416-RPOS.
Exoenzyme assays, siderophore production, in vitro maceration of onion tissue, and stress response assays.
Proteolytic, lipolytic, and chitinolytic activity were determined on the appropriate indicator plates (23) and also determined spectrophotometrically (29). Polygalacturonase activity was determined as previously described (18). In order to determine the ability to macerate onion tissue, overnight LB cultures were normalized at an OD600 of 1, and 100 µl of culture was inoculated onto the surface of onion tissue prepared as described previously (52). Clean, disease-free onions (Allium cepa) were wiped with 90% (vol/vol) alcohol before being cut aseptically into slices. Onion slices were placed into sterile petri dishes, and nicks of approximately 2 mm were made in the tissue surface. Plates were incubated at 30°C, and readings were taken after 48 h of incubation. The measurement of cell viability and ability to survive heat stress and osmotic stress and sensitivity to hydrogen peroxide were determined as described previously (25).
DNA sequence determination and analysis.
The DNA sequence of the cepI/R locus was determined by using pBIR as template. The DNA sequence of the rpoS gene was determined by using plasmids pRBS-2 and pRBS-3 as templates. Nucleotide sequences were determined by the dideoxy chain-termination method (41) by using [35S]dATP
S for labeling and 7-deaza-dGTP (Pharmacia) instead of dGTP.
Nucleotide accession number.
The sequence of cepI-cepR locus and the rpoS gene have been deposited in the GenBank/EMBL/DDBJ database under accession numbers AJ422183 and AJ457984, respectively.
|
|
|---|
Since B. cepacia ATCC 25416, like many other strains of the B. cepacia complex (19), produced C6- and C8-HSL molecules, we tested the activity of these two molecules by using plasmid pSCR1. CepR responded to C6-HSL with only 15% of the molar efficiency of C8-HSL; the plasmid detected the latter molecule at a concentration of less than 10 nM and reached its highest ß-galactosidase activity when a 100 nM concentration of the molecule was present. Higher concentrations did not result in higher reporter enzyme activity (Fig. 1b). The C6-HSL, on the other hand, did not result in a good response with CepR (Fig. 1b), even at very high concentrations as high as 1,000 nM (data not shown).
It was of interest to determine the response of the CepR-based sensor plasmid with respect to different HSL molecules. This was determined by measuring the ß-galactosidase activities of E. coli(pSCR1) exposed to 100 nM concentrations of various HSL molecules. It was decided to use 100 nM since it was the minimum concentration of C8-HSL (the major molecule produced by B. cepacia) necessary to achieve a good response (see Fig. 3a). As expected, the constructed biosensor responded best to C8-HSL when compared to C4-, C6-, C7-, C10-, and C12-HSL. In Fig. 1c, the percentage of the CepR response to C8-HSL is presented, and as shown, CepR displayed a good response to long-chain signal molecules and very poor one to short-chain signal molecules.
![]() View larger version (9K): [in a new window] |
FIG.3. (a) Map of the 5.5-kbp SmaI DNA fragment from B. cepacia ATCC 25416 isolated in this study. Shown are several enzyme restriction sites and the location of the rpoS gene within this fragment. Also shown is the position where a Kmr-containing BamHI fragment derived from pUC4K was cloned in the corresponding site of the rpoS gene to create pLCIKm. (b, c, and d) Effect of rpoS on stress responses of B. cepacia ATCC 25416. (b) Response to heat shock (50°C). Viability is expressed as a percentage of the number of CFU at time zero. (c) Response to osmotic shock (2 M NaCl). Viability is expressed as a percentage of the number of CFU at time zero. (d) Effect of rpoS mutation on oxidative stress. The sensitivity to H2O2 was measured on cells grown for 16 h in LB at 30°C. The zones of inhibition were measured in millimeters. Filled symbols, WT strain; open symbols, B. cepacia 25416-RPOS strain.
|
CepIR regulates onion pathogenicity in B. cepacia ATCC 25416.
It was observed that 25416-I and 25416-R displayed similar lipolyitc and chitinolytic activities and siderophore production when compared to the wild-type parent strain, but the proteolytic activity, on the other hand, was significantly lower in the two mutants. B. cepacia ATCC 25416 was originally isolated as a pathogen responsible for the rot of onions (5), and it was therefore of interest to check if quorum sensing was related to rot. Experiments with in vitro maceration of the onion A. cepa resulted in both 25416-I and 25416-R mutants having attenuated maceration compared to the wild type; introduction in trans of pMPIR carrying the cepI/R locus resulted in more-severe onion pathogenicity than in the wild type (Fig. 2). The attenuated maceration in the onion could be attributed to the decrease in production of extracellular enzymes involved in the onion maceration. One such example could be polygalacturonase, which has been identified and characterized in this strain and implicated in onion disease development (18). We determined polygalacturonase activity in spent culture supernatants of ATCC 25416 and mutant 25416-I. The enzyme activity in inducing conditions was reduced 40% in 25416-I, whereas in the complemented mutant, 25416-I(pMPIR), the activity was 140% of the wild-type levels (Table 2). Quorum sensing was therefore implicated in onion pathogenicity in this strain, and this is at least in part due to the hydrolytic secreted enzyme polygalacturonase. Similar attenuation and complementation levels of enzyme activity were obtained with 25416-R (data not shown).
![]() View larger version (65K): [in a new window] |
FIG. 2. Role of quorum sensing in onion rot. One hundred microliters of a culture of the indicated strain with an OD600 of 1 was inoculated on the right half of a sterile onion as described in the text. This picture was taken after 48 h of incubation at 30°C.
|
|
View this table: [in a new window] |
TABLE 2. Polygalacturonase activity from spent culture supernatants of B. cepacia ATCC 2516 and 25416-I
|
-Proteobacteria (e.g., P. aeruginosa, E. coli, and Vibrio cholerae). In order to investigate the physiological role that rpoS played in B. cepacia, we insertionally inactivated rpoS in strain ATCC 25416, yielding an rpoS::Kmr knockout mutant called B. cepacia 25416-RPOS. We tested the response of 25416-RPOS to various environmental conditions, as the RpoS sigma factor is known to confer cross-protection against several stresses in other gram-negative bacteria (17, 38, 47). We tested the resistance of the parent strain and 25416-RPOS against heat shock, increased osmolarity, and hydrogen peroxide. Strain 25416-RPOS was more sensitive to heat shock and oxidative stress than the wild-type strain; however, there was no significant difference when cells were exposed to an increase in osmotic pressure (Fig. 3b, c, and d)
Do rpoS and quorum sensing cross-regulate each other in B. cepacia?
It was of interest to investigate whether RpoS and quorum sensing cross-regulate each other since the genes regulated by these two systems are maximally expressed at the stationary phase.
In order to determine whether rpoS influenced the accumulation of HSL signal molecules, we quantified HSL production in 25416-RPOS and compared it to the values obtained with the wild type strain. Signal molecules were extracted from spent supernatants of the wild-type and mutant strains at different growth stages, and the extracts were assayed with E. coli(pSCR1). Figure 4a depicts the results showing that the B. cepacia mutant lacking RpoS synthesizes at different growth stages approximately the same amount of signal molecules produced by the wild-type parent strain, demonstrating that the absence of rpoS does not influence the production of HSL molecules.
In order to determine whether the quorum-sensing system is regulating rpoS expression, the rpoS promoter was cloned as a 2-kb BamHI fragment from pCOSRPOS-2C in the BglII site of ß-galactosidase promoter probe vector pMP190, creating pRPR2. rpoS promoter activity was then assayed at different growth stages in the 25416-I mutant and compared to the promoter activity obtained with the wild-type strain. As depicted in Fig. 4b, rpoS expression in the wild-type strain was relatively constant during the early, exponential, and stationary phases. It was then determined that in 25416-I, rpoS expression increased two- to threefold in all growth stages; this increase in activity was retained even when a 100 or 1,000 nM concentration of C8-HSL was added to the growth media (Fig. 4b). A similar increase in rpoS promoter activity was also observed in the cepR mutant 25416-R (data not shown). In view of the fact that rpoS expression increased in the quorum-sensing mutants, the resistance against heat shock and hydrogen peroxide was determined in mutant 25416-I. It was observed, however, that the resistance to the two stresses was comparable to what was observed with the wild-type strain (data not shown).
|
|
|---|
-Proteobacteria (22). These two loci, cepI/R and rpoS, involved in stationary-phase gene expression were inactivated and studied. B. cepacia ATCC 25416, like other members of the B. cepacia complex, synthesizes C6- and C8-HSL molecules (19). The synthesis of both appears to require the CepI protein reported here, since a cepI knockout mutant did not synthesize either (data not shown). It is possible that the cepI synthase directs the synthesis of C8-HSL and lower traces of C6-HSL by recognizing two different acyl-acyl carrier proteins (37). It was observed that our CepR-based biosensor in E. coli has a stronger response to long-chain HSLs than to short-chain HSLs, and it displays the best response to the C8-HSL, the major molecule synthesized by B. cepacia. It must be stressed, however, that these results will not necessarily reflect what the in vivo situation is in B. cepacia, since it has been reported that for the LuxR homologue CarR of Erwinia carotovora, the protein concentration as well as the bacterial host can influence its specificity towards different HSL molecules (50). However, the fact that the C8-HSL has the strongest response in E. coli is an indication that the results obtained here are likely to also occur in B. cepacia. It is believed that HSL-protein interaction reduces susceptibility to proteolysis as well as allowing multimerization important for activating transcription (31, 50). It is likely that CepR has a binding pocket to fit molecules with a longer acyl chain. Recently, similar experiments were performed by using a cep-gfp-based biosensor showing that CepR from B. cepacia H111 responds best to C8 and other long-chain HSL autoinducers (46). CepR displayed a very low response towards the C6-HSL, questioning the function of this molecule in B. cepacia. However, it cannot be excluded that in the genome of B. cepacia there is another luxR family gene coding for another LuxR family protein, which displays a better response to the C6-HSL autoinducer; in this case, this molecule might have a biological function. The presence of this molecule could slightly affect our results regarding our estimation in E. coli(pSCR1) that a 100 nM concentration of C8-HSL is regarded as the minimum necessary for the E. coli(pSCR1) response (Fig. 1b), since C6-HSL molecules could interfere with CepR activity.
Various extracellular phenotypes were tested in relation to quorum sensing, and most did not display any alteration in the quorum-sensing-deficient mutant, with the exception of protease activity. Proteolytic activity was significantly lower in the 25416-I mutants; this observation was similar to what was observed in two different B. cepacia genomovar III quorum-sensing mutants (23, 29). Interestingly, the two B. cepacia genomovar III quorum-sensing mutants displayed either an increase or decrease in siderophore production, whereas in the genomovar I B. cepacia strain used in this study there was no alteration of siderophore production. The discrepancies can be attributed to the fact that the strains belong to different genomovars, highlighting the fact that there is no common quorum-sensing phenotype found within the B. cepacia complex, with the sole exception of protease activity. Interestingly, polygalacturonase activity was significantly reduced in the quorum-sensing mutants (Table 2), and this was also indirectly observed by the attenuated onion maceration activity (Fig. 2). Polygalacturonase activity in B. cepacia ATCC 25416 has been associated with onion maceration, and the plasmid-borne pehA gene is a virulence factor in this strain (18). The complemented mutants harboring the cep locus in trans resulted in higher polygalacturonase activity and higher onion maceration, thus further indicating the involvement of quorum sensing in the onion pathogenicity of this strain. It was not possible, however, to restore wild-type levels of onion maceration activity of the cepI and cepR mutants by dispersing synthetic C8-HSL on the onion (data not shown).
The stationary-phase sigma factor RpoS has also been implicated in several gram-negative bacteria in the change in gene expression necessary for the adaptation to stationary phase (17, 38, 47), and it has also been shown to be involved in the regulation of virulence factors (8, 24, 42). We have observed that RpoS in B. cepacia is necessary for adaptation to heat and oxidative stress, whereas it was not important for osmotic shock adaptation. It was of interest to study whether RpoS and quorum sensing cross-regulate each other since the gene(s) regulated by these two systems are maximally expressed at the stationary phase. It was determined that a genetic background without RpoS has no influence in the accumulation of the HSL molecules produced by B. cepacia (Fig. 4a); on the other hand, rpoS promoter activity is positively influenced by the absence of the quorum-sensing system (Fig. 4b). This doubling in promoter activity did not reflect an increased resistance to RpoS-regulated stresses; this might be because increasing the promoter activity does not automatically result in an increasing amount of RpoS protein as posttranscriptional regulation of RpoS might occur (22). This increase in rpoS promoter activity in the cepI mutant 25416-I could not be restored to wild-type levels by the addition to the growth media of synthetic C8-HSL (Fig. 4b). The reason for this is not known; in addition, it was observed that rpoS promoter activity was higher in 25416-I already at low cell densities and activity increased as the cells entered the stationary phase (Fig. 4b). In the wild-type strain, however, the expression of rpoS remained relatively constant during bacterial growth; this is in contrast to rpoS regulation in Pseudomonas where transcription dramatically increases at the onset of the stationary phase (23). It is similar, however, to what happens with E. coli since RpoS levels are mainly controlled at the posttranscriptional and posttranslational levels (22). In R. solanacearum it has been reported that RpoS regulates the HSL production by the decrease in solR and solI expression (13). Similarly, in P. aeruginosa, quorum sensing and rpoS cross-regulate each other (28, 51). Future work will determine how quorum sensing regulates rpoS and whether the two systems are involved in the regulation of similar phenotypes of B. cepacia.
|
|
|---|
S (RpoS) subunit of RNA polymerase. Microbiol. Mol. Biol. Rev. 66:373-395.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»