Regulation of expression of the cyanide-insensitive terminal oxidase in Pseudomonas aeruginosa

Megan Cooper{dagger}, Gholam Reza Tavankar and Huw D. Williams

Department of Biological Sciences, Imperial College London, Sir Alexander Fleming Building, South Kensington Campus, London SW7 2AZ, UK

Correspondence
Huw D. Williams
h.d.williams{at}imperial.ac.uk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The regulation of the cyanide-insensitive oxidase (CIO) in Pseudomonas aeruginosa, a bacterium that can synthesize HCN, is reported. The expression of a cioA–lacZ transcriptional fusion, CioA protein levels and CIO activity were low in exponential phase but induced about fivefold upon entry into stationary phase. Varying the O2 transfer coefficient from 11·5 h-1 to 87·4 h-1 had no effect on CIO expression and no correlation was observed between CIO induction and the dissolved O2 levels in the growth medium. However, a mutant deleted for the O2-sensitive transcriptional regulator ANR derepressed CIO expression in an O2-sensitive manner, with the highest induction occurring under low-O2 conditions. Therefore, CIO expression can respond to a signal generated by low O2 levels, but this response is normally kept in check by ANR repression. ANR may play an important role in preventing overexpression of the CIO in relation to other terminal oxidases. A component present in spent culture medium was able to induce CIO expression. However, experiments with purified N-butanoyl-L-homoserine lactone or N-(3-oxododecanoyl)homoserine lactone ruled out a role for these quorum-sensing molecules in the control of CIO expression. Cyanide was a potent inducer of the CIO at physiologically relevant concentrations and experiments using spent culture medium from a {Delta}hcnB mutant, which is unable to synthesize cyanide, showed that cyanide was the inducing factor present in P. aeruginosa spent culture medium. However, the finding that in a {Delta}hcnB mutant cioA–lacZ expression was induced normally upon entry into stationary phase indicated that cyanide was not the endogenous inducer of the terminal oxidase. The authors suggest that the failure of O2 to have an effect on CIO expression in the wild-type can be explained either by the requirement for an additional, stationary-phase-specific inducing signal or by the loss of an exponential-phase-specific repressing signal.


Abbreviations: CIO, cyanide-insensitive oxidase; kLa, oxygen transfer coefficient

{dagger}Present address: Defence Science and Technology Laboratories, Fort Halstead, Sevenoaks, Kent TN14 7BP, UK.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Pseudomonas aeruginosa is an opportunistic pathogen that causes a variety of nosocomial infections, including pneumonia, urinary tract infections, surgical wound infections, and bloodstream infections (for a review see Deretic, 2000). It causes life-threatening illness in patients with cystic fibrosis. Initially, P. aeruginosa colonizes the airways with other pathogens such as Haemophilus influenzae and Staphylococcus aureus. However, in most of these patients chronic lung disease develops in which the bacterial population consists almost exclusively of P. aeruginosa in the form of biofilms (Govan & Deretic, 1996). P. aeruginosa is a facultative anaerobe that preferentially obtains its energy via aerobic respiration, but it is well adapted to conditions of limited O2 supply (Palleroni, 1984; Davies et al., 1989). It is capable of anaerobic growth with nitrate as a terminal electron acceptor and in the absence of nitrate it is able to ferment arginine, generating ATP by substrate-level phosphorylation (Palleroni, 1984; Davies et al., 1989; Van der Wauven et al., 1984; Zannoni, 1989).

Most facultative anaerobes contain multiple energy-generating pathways, the synthesis of which is often controlled by overlapping regulatory circuits. By controlling the expression of particular electron-transfer components, especially its cytochrome oxidases, the bacterium constructs the most appropriate electron-transfer chain for the prevailing environmental conditions (Poole & Cook, 2000; Richardson, 2000). Analysis of the genome sequence of P. aeruginosa indicates that it has the potential to make a complex, highly branched aerobic respiratory chain comprising up to five terminal oxidases (Stover et al., 2000; M. Cooper & H. D. Williams, unpublished; Fig. 1). Of four putative terminal oxidases belonging to the haem–copper oxidase superfamily, three are predicted to be cytochrome c oxidases and one a quinol oxidase (Fig. 1). The exception is the cyanide-insensitive oxidase (CIO), encoded by the cioAB operon, which is homologous to the cytochrome bd quinol oxidases of Escherichia coli and Azotobacter vinelandii (Cunningham & Williams, 1995; Cunningham et al., 1997; Junemann, 1997), and as such shows no homology to members of the ubiquitous haem–copper oxidase superfamily (Garcia-Horsman et al., 1994). CioA and CioB are homologous to the two subunits of the cytochrome bd quinol oxidase, CydA and CydB, of E. coli. Histidine and methionine residues identified in E. coli cytochrome bd as being ligands to the low-spin haem b558 (H196 and M329 in CioA) and high-spin b595 (H21 in CioA) are conserved, as is a periplasmic loop – the Q-loop – that contains a putative quinol-oxidizing site, although the Q-loop is significantly shorter in CioA than in CydA (Cunningham et al., 1997). However, the distinctive absorption bands of a cytochrome bd quinol oxidase are not present in P. aeruginosa membranes (Zannoni, 1989; Cunningham & Williams, 1995; Cunningham et al., 1997) and it was proposed that a high-spin haem B replaces the usual haem D in this oxidase. It is likely that there is a family of bacterial quinol oxidases related to the cytochrome bd of E. coli and of other bacteria, which differs in a number of important ways, particularly haem composition, from the E. coli paradigm.



View larger version (11K):
[in this window]
[in a new window]
 
Fig. 1. The proposed aerobic electron-transport pathways of P. aeruginosa. This scheme focuses on the terminal oxidases and does not attempt to include all the potential dehydrogenases at the low redox potential end of the pathway. This scheme is based on analysis of the genome sequence (Stover et al., 2000) together with our own analysis of the presence or absence of CuA sites in the oxidases, which is diagnostic of a cytochrome c oxidase. CIO, cyanide-insensitive terminal oxidase; Cyt., cytochrome; DHase, dehydrogenase. There are two complete ccoNOQP gene clusters in the genome sequence, although the ccoQ genes are not annotated on the current sequence, which are predicted to encode cytochrome cbb3-type oxidases. In addition there are two orphan ccoN genes in the genome. The cyoABCDE genes are predicted to encode a quinol oxidase and while the haem composition of this oxidase cannot be deduced with certainty from the gene sequence, it is likely to be a cytochrome bo3. The same is also true for the cytochrome c oxidase encoded by the coxABC genes, which here we have labelled as a cytochrome aa3.

 
Intriguingly, under low-O2 conditions, P. aeruginosa synthesizes HCN as a metabolic product at concentrations of up to 300 µM. At these concentrations cyanide inhibits the function of members of the haem–copper oxidase superfamily of cytochrome oxidases (Cunningham et al., 1997; Blumer & Haas, 2000; Matsushita et al., 1980, 1983; Castric, 1975, 1983, 1994; Castric et al., 1979). P. aeruginosa has evolved a respiratory chain that allows its own aerobic respiration to function in the presence of this potent terminal oxidase inhibitor. Therefore, the CIO has been proposed to have a role in allowing aerobic respiration under cyanogenic growth conditions (Cunningham & Williams, 1995; Cunningham et al., 1997). Cyanide has been detected in tissue samples infected with P. aeruginosa (Goldfarb & Margraf, 1967) and recent experiments using mutants that do not make cyanide showed that cyanide is a virulence factor in a Caenorhabditis elegans model of infection (Gallagher & Manoil, 2001).

Mutation or overexpression of the cioAB genes has profound effects on the biology of P. aeruginosa. Mutation of cioAB leads to temperature sensitivity for growth, difficulty exiting stationary phase, cell division defects and multiple antibiotic sensitivity, probably due to damage to a multidrug efflux pump. This may be partly explained by increases in oxidative stress levels in CIO-defective strains as a result of the inability of these strains to synthesize a specific catalase, leading to oxidative protein damage (Tavankar et al., 2003).

Adaptation to low-O2 or anaerobic environments may be significant in cystic fibrosis patients infected with mucoid strains of P. aeruginosa, where alginate has been shown to restrict diffusion of O2, thus creating a microaerobic or anaerobic environment, where nitrate levels are increased (Hassett, 1996). Furthermore, in cystic fibrosis patients with established lung disease, P. aeruginosa is located within hypoxic mucopurulent masses in airway lumen and steep O2 gradients are present within the mucus on cystic fibrosis epithelial surfaces prior to infection (Worlitzsch et al., 2002). These conditions will inevitably require the bacterium to carefully control the expression of its electron-transfer proteins. Therefore, there is a compelling case for a more detailed study of the respiratory pathways of P. aeruginosa.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Bacterial strains and growth conditions.
The bacterial strains used in this study are listed in Table 1. P. aeruginosa was grown in LB medium at 30 °C. We found it to be important to use cultures that were well adapted to exponential phase in order to get reproducible gene expression data, particularly from the early exponential phase of growth. Therefore, an overnight culture was diluted 1 : 100 into fresh medium and grown to mid-exponential phase, subcultured, and again grown to mid-exponential phase before being used as the inoculum for the experimental flask. To vary the O2 supply to cultures, they were grown at three oxygen-transfer coefficient (kLa) values (Pirt, 1975): high (87·4 h-1), medium (27·8 h-1) and low (11·5 h-1). This was achieved by using identical 250 ml flasks, shaking at 200 r.p.m., but altering the medium volume. The kLa values for medium volumes of 25 ml (high), 75 ml (medium), and 150 ml (low) were estimated by the sulphite oxidation method (Pirt, 1975; Ruchti et al., 1985; Gil et al., 1992; D'Mello et al., 1994).


View this table:
[in this window]
[in a new window]
 
Table 1. P. aeruginosa strains used in this study

 
Anaerobic cultures were grown in 50 ml LB medium with 0·2 % (w/v) nitrate in 50 ml Falcon tubes and incubated at 30 °C, without shaking, in a GasPak anaerobic jar system (Beckton Dickinson). P. aeruginosa was cultured twice anaerobically into exponential phase before inoculation of the experimental culture.

Determination of percentage O2 saturation.
An MI-730 dip-type oxygen microelectrode and OM-4 O2 meter (Microelectrodes Inc.) were used for dissolved O2 tension measurements. So that they could accept the electrode, 250 ml flasks were modified by addition of a glass inlet tube. The inlet tube was continuous with the main flask, and positioned 5 mm from the bottom of the flask to ensure that the probe was submerged at all times by the lowest volume of medium used (25 ml). An O-ring was positioned at the entry point to the flask to hold the tip of the electrode and to prevent culture fluid entering the inlet tube. The electrode was further held in place by a sealed cap at the far end of the inlet tube.

Design of anti-CioA antibody and Western blotting.
A region of the CioA subunit of the CIO enzyme was selected for production of anti-CIO antibodies. The region selected had to be susceptible to antibody binding, and therefore be on the outer surface of the protein. The cytochrome bd oxidase of E. coli has a large periplasmic loop known as the Q-loop, within which a stretch of 11 amino acids forms part of the quinol-binding site (Dueweke & Gennis, 1990). These 11 amino acids were mapped as the epitope for monoclonal antibodies, which bound to the cytochrome bd oxidase and inhibited quinol binding (Dueweke & Gennis, 1990). This region therefore appeared a likely candidate for epitope selection for anti-CioA antibodies. The Q-loop of P. aeruginosa is more than 60 amino acids shorter than that of E. coli, and of the 11 amino acids mapped as part of the quinol-binding site, 5 are completely conserved and a further 3 are conservatively substituted (Cunningham et al., 1997). To check whether this region of the CioA protein was likely to be antigenic, the Protean program of the DNASTAR software suite was used to calculate the Jameson–Wolf antigenic index for regions within the CioA protein. This analysis demonstrated that the 11 amino acid epitope to CydA was not the best site for antibody design in the CioA protein. A different region within the Q-loop was selected with a more suitable antigenic index and the chosen peptide (KIAAMEGHWDN) was then synthesized and used for the production and affinity purification of rabbit anti-CioA antibodies by Research Genetics Inc. The antibody was used in Western and slot blotting using standard procedures.

DNA manipulations.
General DNA manipulations were carried out as described by Sambrook et al. (1989).

Construction of a cioA–lacZ transcriptional gene fusion.
Using the cioAB sequence data (Cunningham et al., 1997), primers were designed flanking the promoter region of the cioAB genes. The forward primer (pcioF, CGGCCAGCGACTTGTATTTC) was located upstream of a convenient PstI site; the reverse primer (pcioR, CTAGGCATGCCCCATGCCGAAGTTGACC) was located downstream of the promoter region, and had a SphI site incorporated onto the 5' end. PCR was performed using pLC2, a plasmid containing the entire cioAB region, including the promoter region (Cunningham & Williams, 1995; Cunningham et al., 1997), as the template. The amplified promoter segment of 995 bp was cloned into the pGEMT-easy vector (Promega) and removed as a PstI–SphI fragment. This fragment was then cloned into a PstI–SphI-restricted pMP220 vector (Spaink et al., 1987), placing it upstream of a promoterless lacZ gene, to construct pMC10. pMC10 was then transformed into E. coli XL-1 Blue, reisolated and sequenced. pMC10 was introduced into P. aeruginosa strains by triparental conjugation (Rothmel et al., 1991). {beta}-Galactosidase assays were carried out according to Miller (1972) and all experiments were repeated at least three times. Representative experiments are shown and all data points are ±SD.

CIO activity measurements.
Succinate-dependent O2 uptake was determined in whole cells in the presence of 1 mM KCN as described previously (Cunningham & Williams, 1995).


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The CIO is induced upon entry into stationary phase
The expression of the CIO was measured in three ways: by using a cioA–lacZ transcriptional fusion (Fig. 2a–c); by Western/slot blotting using a polyclonal antiserum raised to a synthetic peptide sequence from the major periplasmic (Q-) loop of CioA (Fig. 2d); and by determining CIO activity as succinate-dependent O2 uptake in whole cells in the presence of 1 mM KCN (Fig. 2e). Each of these approaches indicated that CIO expression was lowest in exponential phase and increased rapidly to a maximum in stationary phase. cioA–lacZ expression increased immediately as growth stopped and the culture entered stationary phase, reaching a maximum of approximately five times the exponential phase levels about 2·5 h into stationary phase. This change was paralleled by a similar fivefold increase in CIO activity and a clear increase in CioA protein levels.



View larger version (29K):
[in this window]
[in a new window]
 
Fig. 2. Effect of growth phase and varying the kLa on CIO expression. (a, b, c) P. aeruginosa was grown in LB medium at high (a, kLa=87·4 h-1), medium (b, kLa=27·8 h-1) and low (c, kLa=11·5 h-1) kLa values. Growth as OD600 (filled symbols) and {beta}-galactosidase levels (open symbols) are shown for the control vector pMP220 (circles), or pMC10 (cioA–lacZ transcriptional fusion, squares). (d) Protein (10 µg) from exponential-phase (Exp) or stationary-phase (SP) cells grown at high, medium or low kLa as indicated was used in a slot blot assay with anti-CioA antibody. (e) CIO activity measurements [nmol O2 min-1 (mg protein) -1] made on cells from exponential- or stationary-phase cultures grown at the indicated kLa values.

 
The role of oxygen in CIO expression
Oxygen is a major factor in the regulation of terminal oxidases in many bacteria (Richardson, 2000). To investigate its role in the regulation of the CIO, we followed CIO expression throughout the growth cycle in cultures grown with different O2 transfer coefficients (kLa) and at the same time monitored O2 levels directly in the culture flasks using an O2 microelectrode. Firstly we determined the growth properties and the percentage O2 saturation during growth at three different kLa values (Fig. 3a). Decreasing the kLa had no significant effect on the growth rate but it did reduce the final culture optical density (Fig. 3a). At high kLa the O2 levels reached a minimum of approximately 4 % O2 saturation just as the culture was about to enter stationary phase. At medium kLa the O2 levels reached zero after 5 h of growth just as the culture was entering stationary phase, while at low kLa the percentage dissolved O2 saturation reached zero after 2·5 h, when the culture was in mid-exponential phase.



View larger version (18K):
[in this window]
[in a new window]
 
Fig. 3. Effect of kLa on growth and percentage O2 saturation of P. aeruginosa cultures. (a) Wild-type PAO1; (b) PAO6261 {Delta}anr. The percentage O2 saturation (filled symbols) was measured throughout the growth of P. aeruginosa in cultures in LB medium at different kLa values: high (87·4 h-1, circles), medium (27·8 h-1, triangles), low (11·5 h-1, squares). Growth was measured as OD600 (open symbols).

 
Varying the kLa resulted in a modest increase in the stationary-phase {beta}-galactosidase activity from 1400 units at high kLa to 1800 units at low kLa but had no major effect on CIO activity or CioA protein levels in exponential or stationary phase (Fig. 2). Furthermore, the point in the growth curve at which cioA–lacZ expression was induced was the same irrespective of when the dissolved O2 concentrations reached a minimum (Figs 2 and 3). Therefore, based on these data we conclude that under the conditions used here varying O2 has no effect on CIO expression.

P. aeruginosa can grow under anaerobic conditions using nitrate respiration. During anaerobic growth with nitrate as the terminal electron acceptor cioA–lacZ expression levels were constant at about 300 units {beta}-galactosidase throughout the growth curve, which is similar to the basal level seen in exponential phase during aerobic growth. There was a similar, low level of CioA detected by Western blotting throughout growth (data not shown).

The oxygen-sensitive transcriptional regulator ANR is a repressor of the CIO
We previously found that mutation of the O2-sensitive transcriptional regulator ANR led to a marked increase in CIO activity (Ray & Williams, 1997). Therefore, despite the fact that changing O2 availability by varying the kLa did not alter CIO expression (Fig. 2), we looked again at the effect of an anr mutation on CIO expression. The effect of kLa on growth and the percentage dissolved O2 saturation of the anr mutant culture is shown in Fig. 3(b); the growth results are similar to those for the wild-type.

In the {Delta}anr mutant cioA–lacZ expression was derepressed upon entry into stationary phase (Fig. 4). The extent of this derepression was dependent upon the kLa and was most marked at low kLa, a condition under which ANR is most active (Galimand et al., 1991; Sawers, 1991; Zimmerman et al., 1991). In addition, the timing of cioA–lacZ expression was advanced as the kLa was lowered (Fig. 4c). This change in cioA promoter activity was reflected in increased CioA levels detected by Western blotting and CIO activity assays (Fig. 4). P. aeruginosa has a second FNR homologue known as DNR, which has a role in the regulation of anaerobic respiratory pathways involved in denitrification (Arai et al., 1995). However, mutation of dnr had no effect on cioA–lacZ expression or CioA levels, irrespective of the kLa (data not shown), ruling out a role for DNR in CIO regulation.



View larger version (16K):
[in this window]
[in a new window]
 
Fig. 4. Effect of growth phase and varying the kLa on CIO expression in an anr mutant. (a, b, c), P. aeruginosa PAO6261 {Delta}anr was grown in LB medium at high (a, kLa=87·4 h-1), medium (b, kLa=27·8 h-1) and low (c, kLa=11·5 h-1) kLa values. Growth as OD600 (filled symbols) and {beta}-galactosidase levels (open symbols) are shown for the control vector pMP220 (circles), or pMC10 (cioA–lacZ transcriptional fusion, squares). (b) Protein (10 µg) from stationary-phase cells of PAO1 and PAO6261 grown at low kLa was used in a Western blot assay with anti-CioA antibody.

 
What is the stationary-phase inducing factor?
The data above indicate that O2 can generate a signal that leads to induction of the CIO (in the anr mutant), but suggest that it may not be responsible for the stationary-phase induction observed in wild-type cultures. Consequently, we investigated three alternative explanations for the stationary phase induction of the CIO, as follows.

Induction by the stationary-phase sigma factor RpoS.
The sigma factor RpoS is known to have a role in regulating the expression of stationary-phase genes in a wide range of bacteria, including P. aeruginosa (Jorgensen et al., 1999; Latifi et al., 1996; Suh et al., 1999; Tanaka & Takahashi, 1994; You et al., 1998). However, an rpoS mutation had no effect on cioA–lacZ expression or CioA levels during growth at all three kLa values, ruling out a role for this sigma factor in the regulation of CIO (data not shown).

Quorum sensing.
Quorum sensing via the lasRI and rhlRI systems is well established as regulating a range of stationary-phase phenomena in P. aeruginosa (Fuqua & Greenberg, 1998; Withers et al., 2001; Iuchi & Lin, 1993; Swift et al., 2001). We investigated whether quorum sensing was responsible for stationary-phase induction of the CIO, by looking at the effect of adding spent medium from stationary-phase cultures on cioA–lacZ expression (Fig. 5). Addition of spent culture medium increased cioA–lacZ expression in a dose-dependent manner, with a more marked effect on exponential-phase cultures (Fig. 5a). Similar results were obtained at each of the three kLa values tested (data not shown). Fig. 5(b) shows the kinetics of cioA–lacZ induction upon addition of 60 % (v/v) spent culture medium to a culture. A sharp increase in cioA promoter activity is seen in exponential phase, when cioAB expression is usually at its lowest. Expression began to fall after mid-exponential phase and continued to fall after entry into stationary phase, when cioAB expression would usually be rising. The pattern was the same at all three kLa values tested (data not shown). However, the factor inducing the CIO was not one of the two major N-acylhomoserine lactones synthesized by P. aeruginosa, as addition of neither N-butanoyl-L-homoserine lactone (BHL) nor N-(3-oxododecanoyl)homoserine lactone (OdDHL), nor of both added together, at a concentration of 5 µM had any effect on the expression of the CIO (data not shown). A dose–response experiment using concentrations of OdDHL and BHL from 5 to 50 µM also did not elicit activation of cioA–lacZ expression (data not shown).



View larger version (25K):
[in this window]
[in a new window]
 
Fig. 5. Effect of spent culture medium on cioA–lacZ expression. (a) P. aeruginosa PAO1 containing pMC10 (cioA–lacZ) was grown in LB at high kLa to mid-exponential phase (OD600 0·8, Exp) or stationary phase (SP) with various amounts of filter-sterilized spent culture medium and {beta}-galactosidase levels were assayed. (b) cioA–lacZ expression (open symbols) during growth (OD600, filled symbols) of PAO1 at high kLa in LB medium (squares) or in LB medium with 60 % (v/v) spent culture medium (triangles). Note that in this figure the experiment with the control plasmid pMP220 has been omitted for clarity, but growth and {beta}-galactosidase activity of this strain were similar to those in other experiments.

 
Cyanide.
The CIO is fully active in the presence of 1 mM KCN while the alternative respiratory pathways, terminated by haem–copper oxidases, are inhibited at this concentration; indeed they have an IC50 of about 50 µM (Cunningham et al., 1997; Matsushita et al., 1983). It is known that HCN levels peak in stationary phase and under the growth conditions used here reach 200–300 µM (data not shown) and so an important question is whether cioAB and hcnABC expression are coordinated in some way. Therefore, we investigated the possibility that cyanide could be the stationary-phase inducing factor. When cyanide was added to exponential-phase cultures CIO expression increased markedly in a dose-dependent manner (Fig. 6a). Fig. 6(b) shows the effect of cyanide addition on expression during growth at high kLa; similar data were obtained at other kLa values tested (data not shown). Cyanide had a dramatic effect on expression, as expression was elevated up to 10-fold compared to the control culture and was maintained at a high level until late exponential phase, when it started to fall back to levels similar to the wild-type. Addition of cyanide had little effect on expression when added to stationary-phase cultures, in which the CIO is already maximally expressed. Cyanide does not act by disrupting the function of ANR, as cyanide was still able to induce CIO expression in an anr mutant (data not shown).



View larger version (24K):
[in this window]
[in a new window]
 
Fig. 6. Effect of cyanide on cioAB expression. (a) P. aeruginosa PAO1/pMC10 was grown to mid-exponential phase at low kLa, KCN was added to the indicated concentrations and {beta}-galactosidase assayed 30 min later. (b) P. aeruginosa PAO1/pMC10 was grown in LB medium at low kLa and at the time indicated by the arrow KCN was added to half the cultures to a final concentration of 150 µM and growth (filled symbols) and {beta}-galactosidase (open symbols) were measured. Squares, no KCN; triangles, +KCN.

 
To test whether cyanide is the endogenous inducing factor for CIO expression we made use of a {Delta}hcnB mutant, PAO6344. PAO6344 produces no detectable levels of cyanide (unpublished results). Firstly, we showed that cyanide is an extracellular inducing factor, as spent culture medium from PAO6344 did not induce CIO expression to anything like the same level as spent medium from the wild-type (Fig. 7). This indicates that HCN is the major extracellular inducing factor and explains the quorum-sensing-type effect described above. However, it is noticeable that the kinetics of cioA–lacZ induction is advanced with spent medium from the {Delta}hcnB mutant and shows a modest increase in final {beta}-galactosidase activity in stationary phase. So there may be an additional factor in the growth medium that can induce the CIO, but cyanide is the major effector. However, if cyanide is the endogenous stationary-phase inducing factor then CIO induction should not occur in the {Delta}hcnB mutant. However, cioA–lacZ expression and CioA levels were almost identical in wild-type and PAO6344 (data not shown), indicating that while HCN can induce the CIO it is not the endogenous inducing signal during entry into stationary phase.



View larger version (22K):
[in this window]
[in a new window]
 
Fig. 7. Effect of spent culture medium from a hcnB mutant on cioA–lacZ expression. (a) cio–lacZ expression during growth of P. aeruginosa PAO1/pMC10 at high kLa in LB medium (squares) or in LB medium with 60 % (v/v) spent culture medium from either the wild-type PAO1 (circles) or the {Delta}hcnB mutant PAO6344 (triangles). (b) Western blot of protein (10 µg) from PAO1 cultures grown in the presence of 100 % LB (lane 1), 60 % (v/v) spent medium from PAO1 (lane 2) or 60 % (v/v) spent medium from PAO6344 {Delta}hcnB (lane 3).

 

   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Most facultative anaerobes contain multiple energy-generating pathways, the synthesis of which is often controlled by overlapping regulatory circuits. By controlling the expression of particular electron-transfer components the bacterium is able to construct the most appropriate electron-transfer chain for the prevailing environmental conditions (Poole & Cook, 2000; Richardson, 2000; Iuchi & Lin, 1993). While there are some studies of the regulation of anaerobic respiration in P. aeruginosa (Arai et al., 1994, 1995, 1997; Ye et al., 1995), the environmental and genetic factors that regulate its aerobic electron-transport chains have not been well studied. Moreover, as P. aeruginosa has the genetic potential to produce up to five terminal oxidases for use under aerobic conditions, in principle it has enormous flexibility in the choice of electron-transfer routes to O2. However, there is a paucity of information on the function of the electron-transfer chains in this bacterium and so we have initiated a study of the regulation of the CIO of P. aeruginosa.

The CIO was growth-phase regulated as there was a marked increase in cioA–lacZ expression, CIO activity and CioA protein levels in stationary phase and the kinetics of cioA–lacZ expression indicated that this increase in expression started once growth had stopped. One explanation for this is that as the population density increases, oxygen levels are depleted and this generates a signal for CIO induction. We investigated the effect of O2 on growth by varying the rate of O2 supply and measuring the percentage O2 saturation in the culture during growth. This approach has allowed us to distinguish between stationary-phase and O2 effects on gene expression. Changing the rate of O2 supply had a clear physiological effect on the bacteria as seen by the reduced growth yield as the kLa decreased. If O2 regulates the CIO then a correlation between kLa, the point at which O2 levels reached a minimum and the point of CIO induction might be expected; there was none.

The transcriptional regulator ANR, a homologue of FNR in E. coli, controls denitrification, arginine deiminase activity and cyanide production in P. aeruginosa (Galimand et al., 1991; Sawers, 1991; Zimmerman et al., 1991). Another CRP/FNR-related regulator, DNR, has also been shown to be necessary for denitrification, and transcription of the dnr gene is under the control of ANR (Arai et al., 1995, 1997). ANR controls expression of genes in response to intracellular O2 levels, whereas DNR is thought to respond to N-oxides (Hasegawa et al., 1998). The ANR protein is activated under low O2 levels and subsequently activates or represses genes with specific promoter sequences (ANR boxes) (Galimand et al., 1991; Sawers, 1991; Zimmerman et al., 1991). The cioAB promoter sequence contains two putative ANR boxes (Cunningham et al., 1997). Previously, we found that an anr mutant had increased CIO activity (Ray & Williams, 1997). Here we have shown that stationary-phase cioA–lacZ expression and CioA levels increase markedly in the anr mutant in an O2-dependent manner, being greatest at low kLa. This is consistent with ANR repressing the expression of the cioAB operon in response to the low O2 levels encountered in stationary phase. Importantly, it indicates that CIO expression can respond to a signal generated by low O2 levels. So ANR may play an important role in preventing overexpression of the CIO in relation to other terminal oxidases. Indeed, the presence of the cioAB genes on a multicopy plasmid can lead to up to a fourfold increase in CIO activity with all the electron flux going to O2 via the CIO during NADH-dependent O2 uptake, even though significant cytochrome c oxidase activity is present in the cytoplasmic membranes (Cunningham & Williams, 1995; Cunningham et al., 1997).

The alternative sigma factor, RpoS, which has an established role in regulating stationary-phase phenomena (Jorgensen et al., 1999; Latifi et al., 1996; Suh et al., 1999; Tanaka & Takahashi, 1994; You et al., 1998), did not have a role in stationary-phase induction of the CIO. Initial data showing an effect of a spent medium component in activating CIO expression suggested a quorum-sensing-type control mechanism regulating the CIO. However, further experiments indicated that the extracellular inducing factor was cyanide.

We were interested to see whether CIO expression and HCN synthesis were coordinately regulated. Certainly both are expressed maximally in stationary phase and the synthesis of HCN is known to be dependent on the O2 concentration (Blumer & Haas, 2000; Castric, 1994; Castric et al., 1979). Furthermore, ANR regulates them in opposite ways, activating HCN synthesis but repressing CIO expression. However, we found that HCN can induce expression of the CIO at physiologically relevant concentrations. It is interesting that exogenous HCN is much less effective at inducing expression in stationary phase when the CIO is maximally expressed and, as a highly active component of the respiratory chain, is able to maintain electron flux to O2. This is consistent with cyanide inhibition of electron transport generating a signal for CIO induction. Evidence in support of this mechanism comes from work that demonstrated increased CIO activity in a ccmH mutant, which is defective in cytochrome c biogenesis and does not have an active cytochrome c oxidase terminated pathway (Ray & Williams, 1996). Such a mechanism would also be consistent with the effect of low O2 on CIO expression in a {Delta}anr background. Otten et al. (2001) suggested that reduction of flux through the cytochrome c oxidase pathways enhanced activity and expression of the cytochrome ba3 quinol oxidase of Paracoccus denitrificans. They proposed that this effect was due to increased reduction of the quinone pool and that FnrP may have a role in sensing redox state as well as O2, enhancing expression of cytochrome ba3 when the quinone pool is highly reduced (Otten et al., 2001). The E. coli terminal oxidases are regulated in response to oxygen by the ArcAB two-component signal-transduction system, in which ArcB senses the redox state of the quinone pool (Georgellis et al., 2001). However, in Rhodobacter sphaeroides a two-component regulatory system PrrAB is proposed to regulate electron-transfer components by sensing the redox flux through the cytochrome c oxidizing pathway, specifically through a component of the cytochrome cbb3 (Oh & Kaplan, 2000, 2001). Very interesting, in this context, is the recent report of P. aeruginosa RoxR, a response regulator that is related to PrrA, with a role in regulating cioAB (Comolli & Donohue, 2002). The authors showed that mutation of roxR leads to the loss of cyanide and azide resistance in P. aeruginosa PAK, and provided evidence to support the idea that this resulted from CIO deficiency. However, the roxR mutant still showed significant cyanide-dependent induction of CIO activity and cioA–lacZ expression, indicating that there are other regulators still to be identified, with roles in the cyanide induction of CIO expression. In this paper, we have clearly shown that CIO is not induced by endogenously generated cyanide. It is important to note that Comolli & Donohue (2002) performed their experiments on exponential-phase cultures, that is, conditions under which the CIO expression is at a minimum and during which the endogenous stationary-phase inducing signal either is absent or has its effects repressed. So, while an attractive model, it is premature to suggest that RoxR is responsible for the stationary-phase induction of CIO. Clearly it will be important to determine the growth-phase dependence of CIO induction in a roxR mutant.

An intriguing model is that HCN synthesis is switched on by ANR as the O2 concentration drops and the cyanide inhibits electron transport by haem–copper oxidases, generating a signal that stimulates the induction of the CIO, allowing respiration to continue, with RoxR acting as the transcriptional regulator. However, HCN is not the stationary-phase inducing signal in vivo as the CIO was induced normally in a {Delta}hcnB mutant that does not make cyanide. So does cyanide induction have any physiological relevance? Well, in a natural environment if an actively growing P. aeruginosa cell encounters other bacteria making cyanide then it could induce the CIO to protect itself. The identity of the endogenous inducing signal for stationary-phase induction of the CIO is unknown. However, it could be a signal, such as the catabolic reduction charge, that reflects a change in the cell's metabolic status following entry into stationary phase and which would affect both electron flux and the redox state of electron-transfer components. Does O2 have a physiologically relevant role in the regulation of the CIO? How do we explain the apparent ability of low O2 to generate an inducing signal, as evidenced by CIO expression patterns in a {Delta}anr mutant, but the failure of changing the rate of O2 supply to affect CIO regulation in the wild-type? The point in the growth curve at which stationary-phase induction of the CIO takes place is in all cases after O2 levels have reached a minimum. So O2 could be an inducing signal that is working in combination either with another stationary-phase induction signal or with an exponential-phase repressing signal. We favour the latter, as we have recently found that an increase in exponential-phase CIO activity, as a result of the cioAB genes being present on a multicopy plasmid, leads to a number of detrimental effects on the growth and physiology of P. aeruginosa (Tavankar et al., 2003). This suggests a requirement for tight regulation of the CIO in exponential-phase cultures. Therefore, in this model stationary-phase induction would result from a combination of the generation of an inducing signal by low O2 levels and the loss of a repressing signal as the culture ceases exponential growth and enters stationary phase.


   ACKNOWLEDGEMENTS
 
This work is supported by the UK Biotechnology and Biological Sciences Research Council. We are very grateful to H. Arai for strain RM558, to Dieter Haas for strain PAO6344 and to Paul Williams for strain PAOS and samples of N-acylhomoserine lactones.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Arai, H., Igarashi, Y. & Kodama, T. (1994). Structure and Anr-dependent transcription of the nir genes for denitrification from Pseudomonas aeruginosa. Biosci Biotechnol Biochem 58, 1286–1291.[Medline]

Arai, H., Igarashi, Y. & Kodama, T. (1995). Expression of the nir and nor genes for denitrification of Pseudomonas aeruginosa requires a novel Crp/Fnr-related transcriptional regulator, Dnr, in addition to Anr. FEBS Lett 371, 73–76.[CrossRef][Medline]

Arai, H., Kodama, T. & Igarashi, Y. (1997). Cascade regulation of the two Crp/Fnr-related transcriptional regulators (Anr and Dnr) and the denitrification enzymes in Pseudomonas aeruginosa. Mol Microbiol 25, 1141–1148.[Medline]

Blumer, C. & Haas, D. (2000). Mechanism, regulation, and ecological role of bacterial cyanide biosynthesis. Arch Microbiol 173, 170–177.[CrossRef][Medline]

Castric, P. A. (1975). Hydrogen cyanide, a secondary metabolite of Pseudomonas aeruginosa. Can J Microbiol 21, 613–618.[Medline]

Castric, P. A. (1983). Hydrogen cyanide production by Pseudomonas aeruginosa at reduced oxygen levels. Can J Microbiol 29, 1344–1349.[Medline]

Castric, P. (1994). Influence of oxygen on the Pseudomonas aeruginosa hydrogen-cyanide synthase. Curr Microbiol 29, 19–21.

Castric, P., Ebert, R. F. & Castric, K. F. (1979). The relationship between growth phase and cyanogenesis. Curr Microbiol 2, 287–292.

Comolli, J. C. & Donohue, T. J. (2002). Pseudomonas aeruginosa RoxR, a response regulator related to Rhodobacter sphaeroides PrrA, activates expression of the cyanide-insensitive terminal oxidase. Mol Microbiol 45, 755–768.[CrossRef][Medline]

Cunningham, L. & Williams, H. D. (1995). Isolation and characterization of mutants defective in the cyanide-insensitive respiratory pathway of Pseudomonas aeruginosa. J Bacteriol 177, 432–438.[Abstract]

Cunningham, L., Pitt, M. & Williams, H. D. (1997). The cioAB genes from Pseudomonas aeruginosa code for a novel cyanide-insensitive terminal oxidase related to the cytochrome bd quinol oxidases. Mol Microbiol 24, 579–591.[CrossRef][Medline]

Davies, K. J., Lloyd, D. & Boddy, L. (1989). The effect of oxygen on denitrification in Paracoccus denitrificans and Pseudomonas aeruginosa. J Gen Microbiol 135, 2445–2451.[Medline]

Deretic, V. (2000). Pseudomonas aeruginosa. In Persistent Bacterial Infections, pp. 305–326. Edited by J. P. Natarro, M. J. Blaser & S. Cunningham-Rundles. Washington, DC: American Society for Microbiology.

D'Mello, R., Hill, S. & Poole, R. K. (1994). Determination of the oxygen affinities of the terminal oxidases in Azotobacter vinelandii using the deoxygenation of oxyleghaemoglobin and oxymyoglobin – cytochrome bd is a low-affinity oxidase. Microbiology 140, 1395–1402.

Dueweke, T. J. & Gennis, R. B. (1990). Epitopes of monoclonal-antibodies which inhibit ubiquinol oxidase activity of Escherichia coli cytochrome-d complex localize functional domain. J Biol Chem 265, 4273–4277.[Abstract/Free Full Text]

Fuqua, C. & Greenberg, E. P. (1998). Self perception in bacteria: quorum sensing with acylated homoserine lactones. Curr Opin Microbiol 1, 183–189.[CrossRef][Medline]

Galimand, M., Gamper, M., Zimmermann, A. & Haas, D. (1991). Positive FNR-like control of anaerobic arginine degradation and nitrate respiration in Pseudomonas aeruginosa. J Bacteriol 173, 1598–1606.[Medline]

Gallagher, L. A. & Manoil, C. (2001). Pseudomonas aeruginosa PAO1 kills Caenorhabditis elegans by cyanide poisoning. J Bacteriol 183, 6207–6214.[Abstract/Free Full Text]

Garcia-Horsman, J. A., Barquera, B., Rumbley, J., Ma, J. & Gennis, R. B. (1994). The superfamily of heme-copper respiratory oxidases. J Bacteriol 176, 5587–5600.[Medline]

Georgellis, D., Kwon, O. & Lin, E. C. (2001). Quinones as the redox signal for the arc two-component system of bacteria. Science 292, 2314–2316.[Abstract/Free Full Text]

Gil, A., Kroll, R. G. & Poole, R. K. (1992). The cytochrome composition of the meat spoilage bacterium Brochothrix thermosphacta; identification of cytochrome a3- and d-type terminal oxidases under various conditions. Arch Microbiol 158, 226–233.[Medline]

Goldfarb, W. B. & Margraf, H. (1967). Cyanide production by Pseudomonas aeruginosa. Ann Surg 165, 104–110.[Medline]

Govan, J. R. & Deretic, V. (1996). Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiol Rev 60, 539–574.[Medline]

Hasegawa, N., Arai, H. & Igarashi, Y. (1998). Activation of a consensus Fnr-dependent promoter by Dnr of Pseudomonas aeruginosa in response to nitrite. FEMS Microbiol Lett 166, 213–217.[CrossRef][Medline]

Hassett, D. J. (1996). Anaerobic production of alginate by Pseudomonas aeruginosa: alginate restricts diffusion of oxygen. J Bacteriol 178, 7322–7325.[Abstract]

Holloway, B. W., Romling, U. & Tummler, B. (1994). Genomic mapping of Pseudomonas aeruginosa PAO. Microbiology 140, 2907–2929.[Medline]

Iuchi, S. & Lin, E. C. (1993). Adaptation of Escherichia coli to redox environments by gene expression. Mol Microbiol 9, 9–15.[Medline]

Jorgensen, F., Bally, M., Chaponherve, V., Michel, G., Lazdunski, A., Williams, P. & Stewart, G. S. A. B. (1999). RpoS-dependent stress tolerance in Pseudomonas aeruginosa. Microbiology 145, 835–844.[Abstract]

Junemann, S. (1997). Cytochrome bd terminal oxidase. Biochim Biophys Acta 1321, 107–127.[Medline]

Latifi, A., Foglino, M., Tanaka, K., Williams, P. & Lazdunski, A. A. (1996). Hierarchical quorum-sensing cascade in Pseudomonas aeruginosa links the transcriptional activators LasR and RhiR (VsmR) to expression of the stationary-phase sigma factor RpoS. Mol Microbiol 21, 1137–1146.[Medline]

Matsushita, K., Yamada, M., Shinagawa, E., Adachi, O. & Ameyama, M. (1980). Membrane-bound respiratory chain of Pseudomonas aeruginosa grown aerobically. J Bacteriol 141, 389–392.[Medline]

Matsushita, K., Yamada, M., Shinagawa, E., Adachi, O. & Ameyama, M. (1983). Membrane-bound respiratory chain of Pseudomonas aeruginosa grown aerobically. A KCN-insensitive alternate oxidase chain and its energetics. J Biochem 93, 1137–1144.[Abstract]

Miller, J. H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Oh, J. I. & Kaplan, S. (2000). Redox signaling: globalization of gene expression. EMBO J 19, 4237–4247.[Abstract/Free Full Text]

Oh, J. I. & Kaplan, S. (2001). Generalized approach to the regulation and integration of gene expression. Mol Microbiol 39, 1116–1123.[CrossRef][Medline]

Otten, M. F., Stork, D. M., Reijnders, W. N., Westerhoff, H. V. & Van Spanning, R. J. (2001). Regulation of expression of terminal oxidases in Paracoccus denitrificans. Eur J Biochem 268, 2486–2497.[Abstract/Free Full Text]

Palleroni, N. J. (1984). Family I. Pseudomonadaceae. In Bergey's Manual of Systematic Bacteriology, vol. 1, pp. 141–219. Edited by N. R. Krieg & J. G. Holt. Baltimore: Williams & Wilkins.

Pirt, S. J. (1975). Oxygen demand and supply. In Principles of Microbe and Cell Cultivation, pp. 81–116. Oxford: Blackwell.

Poole, R. K. & Cook, G. M. (2000). Redundancy of aerobic respiratory chains in bacteria? Routes, reasons and regulation. Adv Microb Physiol 43, 165–224.[Medline]

Ray, A. & Williams, H. D. (1996). A mutant of Pseudomonas aeruginosa that lacks c-type cytochromes has a functional cyanide-insensitive oxidase. FEMS Microbiol Lett 135, 123–129.[CrossRef][Medline]

Ray, A. & Williams, H. D. (1997). The effects of mutation of the anr gene on the aerobic respiratory chain of Pseudomonas aeruginosa. FEMS Microbiol Lett 156, 227–232.[CrossRef][Medline]

Richardson, D. J. (2000). Bacterial respiration: a flexible process for a changing environment. Microbiology 146, 551–571.[Free Full Text]

Rothmel, R. K., Chakrabarty, A. M., Berry, A. & Darzins, A. (1991). Genetic systems in Pseudomonas. Methods Enzymol 204, 485–514.[Medline]

Ruchti, G., Dunn, I. J., Bourne, J. R. & Vonstockar, U. (1985). Practical guidelines for the determination of oxygen-transfer coefficients (kLa) with the sulfite oxidation method. Chem Eng J Biochem Eng J 30, 29–38.

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Sawers, R. G. (1991). Identification and molecular characterization of a transcriptional regulator from Pseudomonas aeruginosa PAO1 exhibiting structural and functional similarity to the FNR protein of Escherichia coli. Mol Microbiol 5, 1469–1481.[Medline]

Spaink, H. P., Okker, R. J. H., Wijffelman, C. A., Pees, E. & Lugtenberg, B. J. J. (1987). Promoters in the nodulation region of the Rhizobium leguminosarum Sym plasmid pRL1JI. Plant Mol Biol 9, 27–39.

Stover, C. K., Pham, X. Q., Erwin, A. L. & 28 other authors (2000). Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 406, 959–964.[CrossRef][Medline]

Suh, S. J., Silosuh, L., Woods, D. E., Hassett, D. J., West, S. E. H. & Ohman, D. E. (1999). Effect of rpoS mutation on the stress response and expression of virulence factors in Pseudomonas aeruginosa. J Bacteriol 181, 3890–3897.[Abstract/Free Full Text]

Swift, S., Downie, J. A., Whitehead, N. A., Barnard, A. M., Salmond, G. P. & Williams, P. (2001). Quorum sensing as a population-density-dependent determinant of bacterial physiology. Adv Microb Physiol 45, 199–270.[Medline]

Tanaka, K. & Takahashi, H. (1994). Cloning, analysis and expression of an rpoS homolog gene from Pseudomonas aeruginosa PAO1. Gene 150, 81–85.[CrossRef][Medline]

Tavankar, G. R., Mossialos, D. & Williams, H. D. (2003). Mutation or overexpression of a terminal oxidase leads to a cell division defect and multiple antibiotic sensitivity in Pseudomonas aeruginosa. J Biol Chem 278, 4524–4530.[Abstract/Free Full Text]

Van der Wauven, C., Pierard, A., Kleyraymann, M. & Haas, D. (1984). Pseudomonas aeruginosa mutants affected in anaerobic growth on arginine – evidence for a 4-gene cluster encoding the arginine deiminase pathway. J Bacteriol 160, 928–934.[Medline]

Withers, H., Swift, S. & Williams, P. (2001). Quorum sensing as an integral component of gene regulatory networks in Gram-negative bacteria. Curr Opin Microbiol 4, 186–193.[CrossRef][Medline]

Worlitzsch, D., Tarran, R., Ulrich, M. & 12 other authors (2002). Effects of reduced mucus oxygen concentration in airway Pseudomonas infections of cystic fibrosis patients. J Clin Invest 109, 317–325.[Abstract/Free Full Text]

Ye, R. W., Haas, D., Ka, J. O., Krishnapillai, V., Zimmermann, A., Baird, C. & Tiedje, J. M. (1995). Anaerobic activation of the entire denitrification pathway in Pseudomonas aeruginosa requires Anr, an analog of Fnr. J Bacteriol 177, 3606–3609.[Abstract]

You, Z., Fukushima, J., Tanaka, K., Kawamoto, S. & Okuda, K.. (1998). Induction of entry into the stationary growth phase in Pseudomonas aeruginosa by N-acylhomoserine lactone. FEMS Microbiol Lett 164, 99–106.[CrossRef][Medline]

Zannoni, D. (1989). The respiratory chains of pathogenic pseudomonads. Biochim Biophys Acta 975, 299–316.[Medline]

Zimmermann, A., Reimmann, C., Galimand, M. & Haas, D. (1991). Anaerobic growth and cyanide synthesis of Pseudomonas aeruginosa depend on anr, a regulatory gene homologous with fnr of Escherichia coli. Mol Microbiol 5, 1483–1490.[Medline]

Received 24 September 2002; revised 12 February 2003; accepted 13 February 2003.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Cooper, M.
Articles by Williams, H. D.
Articles citing this Article
PubMed
PubMed Citation
Articles by Cooper, M.
Articles by Williams, H. D.
Agricola
Articles by Cooper, M.
Articles by Williams, H. D.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2003 Society for General Microbiology.