[Frontiers in Bioscience 14, 4113-4126, January 1, 2009]
Cooperative production of siderophores by Pseudomonas aeruginosa
Freya Harrison1,2, Angus Buckling1
1
TABLE OF CONTENTS
1. ABSTRACT
The production of iron-scavenging siderophores by the opportunistic animal pathogen Pseudomonas aeruginosa is a textbook example of public goods cooperation. This trait provides an excellent model system with which to study cooperation. Further, the links between siderophore production and P. aeruginosa virulence allow us to investigate how pathogen ecology, social behaviour and pathology might be connected. We present here the results of basic research on the evolution and ecology of siderophore cooperation in this species. In particular, we explore the effects of population and community structure, iron regime and genomic mutation rate on the relative success of siderophore cooperators and cheats. We also present preliminary data on the links between siderophore production and another clinically-relevant social trait, biofilm formation. It is our hope that more realistic laboratory studies of siderophore cooperation in P. aeruginosa will eventually cast light on the roles played by social traits in long-term microbial infections.
2. INTRODUCTION: SIDEROPHORES ARE A 'PUBLIC GOOD.'
Iron is a major limiting factor for bacterial growth (1). Under aerobic conditions, iron exists in the largely insoluble ferric form, and within animal hosts is generally sequestered by high-affinity iron-binding proteins (1, 2). Iron is therefore not easily accessible, especially to pathogenic bacteria. In order to deal with this problem, bacteria have evolved numerous mechanisms to scavenge iron. One common solution is the production of iron-scavenging molecules which are released into the environment in response to iron deficiency (3).
One important and well-characterised group of such molecules is the siderophores (4). These chelate iron from host protein complexes, re-enter the bacterium via specific receptors and release their cargo of iron to cellular metabolism. Their production is facultatively regulated in response to iron availability (1): low iron availability increases siderophore production. As siderophores can be taken up by any cell with the cognate receptor, regardless of its own levels of siderophore production, siderophores represent a form of public goods cooperation: while individually metabolically costly to produce, siderophores confer a group-level benefit (5). This is predicted to make populations of siderophore producers open to invasion by non-producers (see (5-8)). Simple experiments using the opportunistic bacterial pathogen Pseudomonas aeruginosa (9) have confirmed the metabolic cost and growth advantage of siderophore production. First, in iron-limited growth media, monocultures of producers reach higher densities than do monocultures of non-producers. Second, when mixed together, non-producing clones can outcompete producing clones; and non-producing clones reach higher densities in the presence of siderophore producers than they do in monoculture (9). This ability of non-producers to exploit producers allows us to recognise them as social defectors or "cheats" (10, 11).
Production of siderophores by fluorescent Pseudomonads in fact represents a remarkably tractable model system for studying the evolution and ecology of cooperation. The primary siderophore of Pseudomonas spp. is the yellow-green pigment pyoverdine (1, 3, 4, 12). Cells which do not produce pyoverdine form white colonies on iron-limited agar, providing an easy visual score for cooperator versus cheat clones (Figure 1)(9). Further, a simple and cheap colorimetric assay for total siderophore content of culture supernatants exists (13). We have shown that the results of visual scoring and this assay are consistent with one another (8).
The relationship between siderophore production and bacterial growth rates led to the suggestion that siderophore production might enhance bacterial virulence. Consistent with this, mutants deficient in siderophore production have been shown to exhibit reduced virulence (12, 14, 15). Therefore, understanding the evolution and ecology of siderophore production should not only serve as an excellent basis for advancing our understanding of cooperation in general, but also provides us with an opportunity to study direct links between cooperation and virulence. In fact, efficient host exploitation by many microbial pathogens often relies on the cooperative production of siderophores (14-16), other nutrient-scavenging molecules (17), toxins (18), products that interfere with the host immune response (19, 20) or enzymes that degrade antibiotics (21). (See (22) for a review of social behaviours in microbes). That links between pyoverdine production and other virulence-related traits have been demonstrated (23, 24) makes this molecule even more interesting as a model public good.
Understanding how P. aeruginosa virulence is determined is of particular interest as this species is an important opportunistic human pathogen. Acute infections of burn wounds and nosocomial colonisation via medical artefacts (e.g. catheters) are risk factors for hospital patients (25). More importantly, this species is one of the most common and dangerous pathogens for cystic fibrosis (CF) sufferers. Almost all CF patients become chronically colonised by this species and colonisation severely worsens prognosis, promoting more frequent acute exacerbations, a general decline in lung function and significant increases in patient morbidity (26-28). P. aeruginosa can also interfere with the host immune system and chronic colonisation may increase the probability of further pathogenesis (19, 29, 30). Most people with CF die as a result of respiratory failure resulting from chronic microbial infection (31).
We and our co-workers have carried out a number of experiments using P. aeruginosa to elucidate the importance of population and community structure and evolution in determining levels of siderophore cooperation. We have also researched how cooperation mediates the outcome of mixed infections, and have begun to explore the links between siderophore production and another cooperative trait linked with pathogenicity, biofilm formation. In this review, we will present the results of these investigations and make suggestions for future work based on our observations. The path diagram in Figure 2 illustrates the hypothesised connections between the factors investigated.
3. SIDEROPHORE COOPERATION AND VIRULENCE
Many models of parasite virulence and evolution predict that mixed (multi-genotype or low relatedness) infections should be more virulent than single-genotype infections. Historically, this prediction gained wide acceptance, and it was attributed to increased resource competition in low-relatedness infections favouring more voracious host exploitation (32-36). However, empirical support for this hypothesis was, and remains, lacking. One likely explanation for this lack of support is that models often assume that simple resource competition represents the only possible interaction between coinfecting strains, whereas in reality other types of interaction exist: in particular, cooperation via public goods production. If some necessary virulence factors are public goods (14-16, 18-22), then this has serious implications for models of relatedness and virulence: kin selection theory predicts that public goods production, and so virulence, will be decreased at low relatedness.
Relatedness can be defined as the probability of two individuals sharing an allele at a particular locus, relative to the other members of the population (37). Low relatedness infections can thus comprise a mixture of public goods producers and non-producers. The total amount of public good produced, and hence the effect on the host, will be reduced in comparison with high-relatedness infections comprising only producers. Relatedness at loci governing cooperative behaviours is also a major determinant of the evolutionary success of cooperative traits. If the actor and beneficiary of a cooperative act share a cooperative allele, then cooperation can be maintained by natural selection due to its positive effect on inclusive fitness (10). Models of siderophore production and virulence have produced results consistent with the prediction that virulence can indeed be decreased at low relatedness (5).
We wanted to test empirically the hypothesis that virulence is decreased in mixed cooperator + cheat infections when public goods (siderophores) are important for bacterial growth. This investigation is published as (38). We inoculated waxmoth larvae (Galleria mellonella) with cells of a wild type, pyoverdine-producing strain of P. aeruginosa, a pyoverdine-minus mutant, or a 1:1 mixture of the two. Consistent with our prediction, we found that cooperator infections killed caterpillars on average two hours (15%) sooner than did cheat infections, and that mixed infections resulted in an intermediate time to death. These results are illustrated in Figure 3A. The growth rate of single- and mixed-clone infections showed the same pattern: cooperators grew faster than cheats, and mixed infections had an intermediate growth rate. These data demonstrate that the presence of siderophore cheats can reduce the growth rate of a bacterial population, and hence reduce virulence.
We next addressed whether cheats were more likely to be favoured in low versus high relatedness infections. We inoculated larvae with pure strains or a 1:1 mixture as above and incubated them for for eight hours before plating out homogenised tissue. We then calculated the number of cell doublings per gramme of fresh weight for each strain. The results are shown in Figure 3B: as predicted, cheat populations grow more rapidly in mixed, as opposed to single-clone, infections, while the opposite pattern is observed for cooperators. The selective advantage of cheating is therefore predicted to be higher in low relatedness infections. We then investigated the frequency-dependence of cheat fitness by inoculating larvae with mixtures containing 3-90% cheats and calculating cheat relative fitness after a period of growth. In these simple experiments, our 'cheats' never had a detectable selective advantage: at best, their fitness was the same as cooperators. This is in contrast with in vitro work, where the same strain was consistently able to to invade cooperating populations from initially low frequencies (9, 11, 38).
There are three likely explanations for this discrepancy. The first is the greater spatial heterogeneity within a caterpillar compared with a medium-filled tube, which is likely to increase the relatedness of immediate neighbours (37), and reduce the probability of direct cooperator-cheat interactions, bestowing a net benefit on cooperators. However, we have observed the de novo evolution of decreased siderophore production in P. aeruginosa populations growing in waxmoth larvae (D Racey, F Harrison & A Buckling, in preparation). The second possible explanation is the longer periods of time bacteria spend at high densities in tubes compared with larvae: tube assays reached densities of approximately 108 cells / ml (9), while insect assays reached densities of approximately 107 cells / ml. The higher the population density of bacteria, the greater the likelihood cheats will come into contact with cooperators (17, 37). Siderophore cheats have been observed at appreciable frequencies in chronic, clinical P. aeruginosa infections of the CF lung (7), where P. aeruginosa densities can reach 108 - 1010 CFU per ml of respiratory secretions, and pyoverdine production has been known to decrease over the course of chronic infection (39). As most CF patients are initially colonised by a single, environmental clone (40-42), any cheats present will most likely have arisen within the patient and invaded from an initially low frequency. These observations strongly suggest that cheats can enjoy a selective advantage in longer term, high density infections. The extent to which this apparent advantage is frequency dependent is not known. The evolutionary success of de novo cheating mutants in CF lungs and waxmoth larvae lead us to suggest that the apparent discrepancy in our results may reflect a peculiarity of the cheat strain used. This strain may, for instance, have a slightly reduced capacity for siderophore uptake or otherwise differ from naturally-evolved cheats as a result of extra mutations incurred during its manufacture by UV mutagenesis. Further work on the de novo evolution of siderophore cheats in larvae should clarify this issue.
4. TESTING HAMILTON'S RULE USING SIDEROPHORES: THE IMPORTANCE OF POPULATION STRUCTURE
Having established that siderophores are important virulence factors, understanding the ecological factors that influence their production takes on a practical significance. If siderophores are indeed a public good, we might expect the evolutionary dynamics of siderophore production to conform to the predictions of kin selection theory (10).
The P. aeruginosa siderophore model system was, in fact, successfully used to test two key predictions of kin selection theory. Hamilton (10) stated that altruistic cooperation is favoured when rb > c, where r is the genetic relatedness between actor and beneficiary, b is the benefit to the recipient and c is the net direct fitness cost to the actor. Hamilton also suggested that high population viscosity (limited dispersal) should favour cooperation as it will lead to increased r. This sparked a debate in the theoretical literature (see references in (9)), as high population viscosity is also expected to lead to increased competition between relatives: any increase in fitness accruing to the recipient of a cooperative act would therefore be achieved at the expense of relatives. This type of competition, where individuals compete with one another within a habitat patch, can be termed local competition. This contrasts with global competition, where groups inhabiting discrete patches within a metapopulation compete with one another. In this scenario, the probability of a given individual contributing to the next generation depends not only on its growth rate relative to its neighbours, but also on the total growth rate of its local patch relative to neighbouring patches. As the presence of cheats will have a negative impact on population density, subpopulations containing higher frequencies of cooperators are more likely to contribute to the next generation and so global competition is expected to favour cooperation (37, 43).
In the real world, populations are likely to experience a mixture of local and global competition (44, 45), so it is necessary to employ an extension of Hamilton's rule which takes the extent of local competition into account. Frank's (37) extension to Hamilton's rule states that cooperation is favoured when r(b - a(b - c)) - c > 0, where a represents the proportion of competition that is local. Models constructed by other authors confirmed the role of local competition in general terms (43, 46) and with relevance to siderophore production (5).
Griffin, West & Buckling (9) then used experimental populations of P. aeruginosa to examine the roles of r and a in the outcome of in vitro competition between pyoverdine-producing and pyoverdine-negative strains. Replica metapopulations consisting of twelve glass tubes containing iron-limited medium were subjected to high and low relatedness and local or global competition in a factorial design. Populations were maintained for six days (c. 50 bacterial generations), with daily transfer to fresh medium. The results are shown in Figure 4 and were entirely consistent with theory: high relatedness and global competition favoured cooperation. Global, unlike local, competition allowed the more productive cooperative tubes to contribute more to the global pool; and high relatedness prevented cheats from coming into direct contact with cooperators. Under low relatedness conditions on the other hand, cheats were able to exploit the pyoverdin produced by cooperators, and increased in frequency, explicitly demonstrating a cost to pyoverdin production.
5. THE INFLUENCE OF THE ABIOTIC AND BIOTIC ENVIRONMENT ON SIDEROPHORE COOPERATION
We then moved on from investigating population structure to address the role of the benefit:cost ratio (b/c) of siderophore production. b/c is a key determinant of the strength of kin selection (10, 37). The effect of environmental nutrient status on b/c and, by extension, on cheat success, has been demonstrated by computational simulation (43). We wished to explore the ecological factors that might influence b/c, and so the evolution of P. aeruginosa siderophore cheats. Our investigation was specifically designed to allow inferences about CF lung communities to be made.
The most influential aspect of the abiotic environment with regard to b/c is presumably iron availability, as evidenced by the iron-dependent regulation of siderophore gene expression (1, 3). The biotic community in which P. aeruginosa finds itself may also also influence b/c. In the CF airways, P. aeruginosa exists alongside numerous other pathogenic microbes (see (47) for a detailed review of lung microbial ecology). Its most notable coinfections partners are Burkholderia cepacia complex and Staphylococcus aureus (28, 29, 48-51). Interspecific interactions with these and other species may mediate b/c: for instance, P. aeruginosa is capable of lysing cells of S. aureus and utilising the iron released to support its own growth in vitro (52, 53). It is not known how significant this extra iron is in the context of the CF lung, or how this behaviour is related to interspecific competition for environmental iron. We wished to determine whether P. aeruginosa responds to S. aureus as an iron source. Further, we wished to determine whether any effect of S. aureus on P. aeruginosa siderophore production could be dependent on the levels of environmental free iron, as tissue damage due to chronic infection leads to elevated free iron levels in colonised airways (54-56).
This investigation is reported in (57). We founded forty-eight populations of P. aeruginosa (siderophore producers) in casamino acids broth microcosms. Twenty-four microcosms were supplemented with iron and twenty-four made iron limited by adding an iron chelator (human apotransferrin). S. aureus cells were then added to half of the microcosms, yielding a fully factorial balanced design. We had previously ascertained that the P. aeruginosa strain used was capable of lysing the S. aureus strain. The populations were passaged daily into fresh growth medium for 20 days (c. 140 P. aeruginosa generations). Preliminary work had shown that the probability of carrying S. aureus over into the new tubes was minimal and fresh ancestral S. aureus was added at each transfer, removing any potential effect of S. aureus evolution.
We wished to separate the initial, physiological response of bacteria to the environmental benefit:cost ratio from any longer-term response to selection. The physiological response of P. aeruginosa was measured using data on siderophore production collected after the first twenty-four hours' growth in the treatment environments (chromeazurol sulphate assay as described by (13)). The evolutionary response was measured by storing samples of evolving populations at regular intervals and later growing these up in identical, iron-limited microcosms. Measurements of siderophore production in situ reflect a mixture of physiological downregulation and cheat evolution, whereas growth of samples in a common, iron-limited environment forces cells that had downregulated expression to begin producing siderophores once more, thus giving a reliable estimate of the extent of cheat evolution.
Our results showed that both iron regime and the presence of S. aureus affect the physiological regulation of siderophore gene expression and, by inference, b/c (Figure 5A). Consistent with iron-dependent regulation of siderophore expression (1, 3), iron-limited populations produced significantly more siderophores per CFU than did iron-supplemented cultures. Further, siderophore production by iron-limited populations was higher when S. aureus was present. This is not consistent with the hypothesis that P. aeruginosa was using S. aureus as an iron source. In this experiment, S. aureus instead seems to have an analogous effect to iron depletion, further increasing b/c when iron was already scarce.
As shown in Figure 5B, siderophore cheats evolved de novo much more readily under iron-limited conditions - i.e. when high values of b/c created a common pool of highly beneficial siderophore and hence a selective advantage to cheating. In iron-rich environments, siderophore production was physiologically downregulated due to a low b/c, resulting in little selection pressure for cheating. This is consistent with previous work describing short-term competition experiments between siderophore producers and a siderophore deficient mutant across a range of iron-limitation conditions (9). Figure 5B also shows that P. aeruginosa siderophore cheats were more common in iron-limited environments when S. aureus was present. This is consistent with the results shown in Figure 5A, where the presence of of S. aureus led to increased siderophore upregulation in iron-limited populations.
The simplest explanation for the observed effect of S. aureus is that it competes with P. aeruginosa for iron. However, we cannot rule out the possibility that other interactions between were responsible for siderophore upregulation in low-iron environments - cross-species quorum sensing, for example (58). In order to address this issue, we and our collaborators are carrying out experiments designed to determine whether the presence of S. aureus causes alterations in the expression of P. aeruginosa signal molecules (S. Diggle pers. commun.).
It would be interesting to see if S. aureus can ever serve as a significant iron source for P. aeruginosa, and whether this can ever outweigh the effect described above. The role of S. aureus evolution in long-term mixed populations is also a matter for consideration, as is any effect of environmental heterogeneity, and the net effect of the numerous suggested synergistic and antagonistic interactions between these species (31, 52, 53, 59-62). Experiments designed to explore questions such as these could add significantly to our understanding of the complexity of community interactions and the virulence of mixed infections.
6. HYPERMUTABILITY AND COOPERATION
The evolution of social traits should also be considered in the context of a population responding to various selection pressures. Evolutionary changes in non-social traits could influence mean trait values for cooperative behaviours - and vice versa - via pleiotropic effects or genetic linkage. We identified one commonly-selected trait that we thought could affect cooperator and cheat dynamics: mutation rate.
Bacterial populations readily evolve lineages with a 10-1000-fold increase in genomic mutation rate (63, 64). When adaptation is limited by the supply of beneficial mutations, as is the case in novel or changeable environments, such a 'mutator' genotype can hitch-hike with advantageous mutations to reach high frequencies (65-69). Mutators are common in bacterial infections, and Oliver et al (70) report that 37% of 128 CF patients they studied were colonised by a mutator genotype of P. aeruginosa. Hypermutability is believed to aid adaptation to changing host defences and chemotherapy (70-72). However, when a populations is well-adapted to its environment, the increased rate of deleterious mutations in mutator lineages means that hypermutability is selectively disadvantageous (65-67, 73, 74). Mutators may also be disadvantaged when population growth depends upon cooperation, as they are expected to generate cheating genotypes more readily and do decrease the relatedness of evolving populations. Mathematical models predict a negative correlation between mutation rate and cooperation when there is a sufficient element of local competition (34).
We tested this hypothesis in two experiments. The first experiment addressed the relative propensity of mutators to generate cheating mutants. We evolved wild-type and mutator populations of P. aeruginosa in iron-limited growth medium for c. 200 generations. The populations were subjected to entirely local competition to maximise the advantage of defection within a cooperating population. Siderophore cheats arose de novo in all populations, but cheats arose earlier, increased in frequency more rapidly and reached higher maximal frequencies in mutator populations. Furthermore, cheats were able to reach fixation in mutator populations but showed marked fluctuations in frequency in wild-type populations (Figure 6A). Despite the increased levels of cheating in mutator populations, the rate of decline in population density did not differ between wild-type and mutator populations (8). These results suggested that the mutator was able to generate fitter cheating genotypes than was the wild type. We then carried out competition assays to ascertain the ability of evolved cheat clones a) to invade cooperating populations from rare and b) to persist at high frequency in mixed populations.
We plated samples from frozen aliquots of our selection lines and isolated white colonies. These were then assayed for siderophore production in iron-limited medium and ability to invade cooperator populations from rare: clones which produced < 50% of the siderophores produced by their cooperative ancestor and which could invade from rare were taken to be true cheats and were competed with cooperator cells in iron-limited growth medium. Consistent with our suggestion, we found that while both wild-type-derived and mutator-derived cheats could all invade cooperators from a low starting frequency (c.5%), only the mutator produced cheats whose relative fitness exceeded one when mixed with cooperators at a starting frequency of c.50% (6/18 mutator-derived cheats had fitness >1 compared with 0/10 wild-type-derived cheats) (Figure 6B) (75).
Our second experiment (76) tested the hypothesis that mutators should be selected against under conditions that favour cooperation. Mutator and wild-type genotypes were competed in globally-competing metapopulations under conditions of high and low relatedness as described in section 4 and Figure 4. Consistent with predictions, mutators reached lower frequencies in the metapopulation under high relatedness than under low relatedness. Further, when relatedness was low, mutators were more likely to exceed 50% of the metapopulation. For each metapopulation, we then calculated the mean mutator frequency and mean pyoverdine cheat frequency over time. Crucially, the mean frequency of mutators within a metapopulation was positively correlated with the mean frequency of cheats (Figure 7; note the right-shift of the low-relatedness data relative to the high-relatedness data). In both of the mutator studies, data on visually-scored pyoverdine cheats was consistent with data on total siderophore production obtained using the chromeazurol sulphate colorimetric assay (13).
The effect of hypermutability on cooperation may help to explain observed distributions of P. aeruginosa mutators. Mutators are more common in clinical isolates than in conspecific environmental populations (77) and P. aeruginosa mutators are more commonly observed in long-term, chronic infections than in acute infections (70). These observations have previously been attributed to the environmental fluctuations experienced during long-term in vivo growth (host immune responses and medical intervention: (66, 70-72, 78). We would suggest that there may also be weaker selection for cooperation, and hence weaker selection against mutators, in clinical versus environmental, and in chronic versus acute bacterial populations. In CF patients, P. aeruginosa infection generally results from colonisation by a single clone (40) and persists for several years and even decades. Patient-to-patient transmission is relatively rare. Taken together, these observations suggest that lung populations experience local competition. While little is known about the ecology of P. aeruginosa in soil and water, we would tentatively suggest that environmental populations could experience higher relatedness (due to shorter patch longevity) and more global competition (due to inter-patch dispersal) than would clinical populations.
The appearance of hypermutable variants of pathogenic bacteria is associated with increased antibiotic resistance (70, 71, 79) and as such mutators represent a significant risk factor for chronically colonised patients. Understanding the evolutionary ecology of mutators may thus benefit our understanding of chronic infection progression; indeed, understanding the evolution and ecology of pathogen communities may be vital in the development of new and more effective prophylaxis.
7. LINKS BETWEEN SIDEROPHORE PRODUCTION AND BIOFILM FORMATION
The sheer variety of social traits indulged in by any given species of microbe (22, 80) mean that these behaviours cannot be expressed or evolve in isolation from one another. A recent study (23) has shown a genetic link between siderophore production and a second virulence-linked cooperative trait: biofilm formation. Biofilms facilitate the colonisation of new niches, push daughter cells upwards to reach aerobic zones and increase resistance to antibiotics and ciliary clearance (81-83). The production of extracellular polymers necessary for biofilm formation is metabolically costly and so biofilms are open to exploitation by non-producing cheats (84, 85). Iron has been shown to be necessary for biofilm formation by P. aeruginosa (86) and ferrisiderophores act as a signal to switch from planktonic to biofilm growth: as a result of this, siderophore-deficient mutants have reduced biofilm forming ability (23). We wished to expand upon these results and to determine their likely evolutionary consequences, using siderophore cheats that had evolved de novo in our lab. Specifically, we addressed the hypothesis that siderophore cheats could also act as biofilm cheats. The following work is detailed in (75) and will form the basis for a more detailed study of siderophores and biofilm.
Briefly, we grew biofilms of five of the evolved siderophore cheats described in section 6 and their siderophore-producing ancestor, PAO6049 (87) in iron-limited and iron-supplemeted growth medium, using the microtitre plate - peg lid system as described in (88, 89). Biofilm mass was assayed using crystal violet staining and the number of cells in the biofilm and planktonic subpopulations calculated from plate counts. The five cheats had lower total biofilm mass and allocated a smaller proportion of the total population to the biofilm than did their ancestor, both under iron limitation and under iron supplementation (T-tests: all P-values < 0.001 and remained significant after sequential Bonferroni correction). This was not a result of generally poor growth by the pyoverdine cheats, which in fact exhibited better planktonic growth than their ancestor in both treatment environments; this is probably due on the one hand to their previous period of adaptation to iron-limited broth and on the other to the costs of siderophore production when iron is supplied). Finally, all five evolved cheat clones had lower biofilm mass under iron-limited, as compared with iron-supplemented conditions (T-tests: all P-values < 0.006 and remained significant after sequential Bonferroni correction).
We also hypothesised that biofilms could represent an environment where siderophore-deficient cells would represent 'super cheats' (90) - not only do they not contribute to iron scavenging but they also contribute less to the structural integrity of the biofilm, exploiting two group-beneficial behaviours while paying the costs of neither. If this is the case, we would expect siderophore cheats to have a greater relative fitness advantage when competing with siderophore producers in biofilms than in planktonic populations. To test this hypothesis, we grew mixed biofilms consisting of the wild type PAO1 plus each of the evolved cheat clones under iron-limited conditions. The mixtures had significantly lower mean biofilm mass than did the pure wild type culture (T-tests: all P-values < 0.005 and remained significant after sequential Bonferroni correction), but none of the populations had significantly different biofilm versus planktonic cheat frequencies (Wilcoxon signed ranks test on paired data: all P-values ≥ 0.25). Interestingly, pyoverdine cheats evolved de novo in the PAO1 pure culture, but again there was no significant difference in cheat frequency between biofilm and planktonic subpopulations (Wilcoxon signed ranks test: P = 0.30).
These preliminary results confirm the existence of a link between siderophore production and biofilm formation, highlighting the fact that adaptive change in social traits can be constrained by selective pressures on genetically linked traits. This is consistent with previous observations that cooperation can be stabilised by pleiotropic fitness costs of cheating (91). Our results are not, however, consistent with the hypothesis that siderophore cheats act as super cheats in a biofilm environment. We hope to extend this work in the future by investigating a wider range of cheat clones, testing biofilm resistance to antibacterial agents and assessing the ability of cheat clones to invade established biofilm communities. Both siderophore production (7, 39) and biofilm-forming ability (39, 92) have been observed to decrease over time in chronic P. aeruginosa infections of cystic fibrosis patients: are both changes the result of adaptation to the airways? Or might selection for siderophore cheating have the knock-on effect of reducing biofilm, even though biofilms aid persistence? This remains to be seen, but it is certainly an intriguing question.
8. CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK
We have successfully used a bacterial model system to explore the evolution and ecology of a clinically-relevant cooperative trait. From simple in vitro experiments, we have been able to confirm and extend theoretical models of cooperation and to highlight the importance of microbial sociality in mediating pathogen virulence and persistence. In addition to providing the first report of the de novo evolution of siderophore cheats in laboratory populations (8), we have carried out experiments with the following conclusions.
i. Kin selection theory (Hamilton's rule) is an appropriate theoretical framework within which to consider siderophore production by P. aeruginosa.
ii. The extent of local competition is as important a determinant of cooperator success, as are relatedness and b/c.
iii. Low pathogen relatedness leads to reduced virulence in acute infections when pathogenesis is promoted by cooperative behaviour.
iv. Interspecific interactions can affect b/c and thus selection pressure on social cheating.
v. Hypermutability accelerates the breakdown of cooperation under local competition and kin selection for cooperation can mediate mutator dynamics in bacterial metapopulations.
vi. Siderophore production affects investment in biofilm formation.
Figure 8 represents a synthesis of our results as they relate to the evolution, ecology and pathology of siderophore cooperation in P. aeruginosa. The work we have presented in this review suggests several avenues for further investigation, which we hope will not only extend social evolution theory but also help to improve our understanding of multi-genotype microbial infections. The main emphasis of our future work will be on increasing the realism, and by extension practical relevance, of our experiments. For instance, models generally assume that cooperative traits represent a binary choice (cooperate or cheat), but in fact many of these behaviours are continuous traits subject to physiological regulation based on environmental signals. Research on how reaction norms (93) of siderophore production evolve could inform much more realistic models of cooperation. In other words, rather than purely looking at how the in situ level or maximum possible level of siderophore production by cells changes over a period of evolution, it would be advisable to ascertain how siderophore production by specific cells changes in response to iron availability, and to determine whether the nature of this relationship is subject to selection. Investigating the evolution of suites of linked social behaviours, such as siderophore production and biofilm formation, is also an exciting avenue for future research. The ecological realism of laboratory experiments could also be improved: we have already looked at the effect of one inter-specific interaction on siderophore cooperation, but this work is only the initial step in a potentially much larger project to ascertain how community structure and evolution changes selection pressures on cooperation over time periods approximating to long-term, chronic infections of humans. The work we have begun has benefited from numerous collaborations; in the future, we hope that work on both this and comparable systems will be facilitated by the combined efforts of evolutionary biologists, microbiologists, clinicians and theoreticians
In conclusion, our work has begun to illustrate the importance of studying cooperative behaviours within the context of evolving communities inhabiting structured environments. Even among microbes, communities can be dynamic and complex.
9. ACKNOWLEDGEMENTS
FH drafted the manuscript and AB contributed to its preparation. Lucy Browning, Ashleigh Griffin, Jon Paul, Michiel Vos and Stuart West all carried out experimental work included in the present review. Michael Brockhurst, Steve Diggle, Andrew Gardner and Ruth Massey provided useful advice and discussion, as did the reviewers of our original research articles. Roger Levesque, Ruth Massey, Jean-Marie Meyer and Antonio Oliver supplied bacterial strains. We would also like to thank Michael H Perlin for inviting us to contribute to this special issue and two anonymous reviewers for their comments on the manuscript.
10. REFERENCES
1.Colin Ratledge, and Lynn G Dover. Iron metabolism in pathogenic bacteria. Annual Review of Microbiology 54, 881-941 (2000)
doi:10.1146/annurev.micro.54.1.881
2. Shelley M. Payne. Iron acquisition in microbial pathogenesis. Trends in Microbiology 1, 66-69 (1993)
doi:10.1016/0966-842X(93)90036-Q
3. C Wandersman, and P Delepelaire. Bacterial iron sources: From siderophores to hemophores. Annual Review of Microbiology 58, 611-647 (2004)
doi:10.1146/annurev.micro.58.030603.123811
4. P Cornelis, and S Matthijs. Diversity of siderophore-mediated iron uptake systems in fluorescent pseudomonads: Not only pyoverdines. Environ Microbiol 4, 787-798 (2002)
doi:10.1046/j.1462-2920.2002.00369.x
5. S. A. West, and A. Buckling. Cooperation, virulence and siderophore production in bacterial parasites. Proc R Soc Lond B Biol Sci 270, 37-44 (2003)
doi:10.1098/rspb.2002.2209
6. S. P. Brown, M. E. Hochberg, and B. T. Grenfell. Does multiple infection select for raised virulence? Trends Microbiol 10, 401-405 (2002)
doi:10.1016/S0966-842X(02)02413-7
7. D. De Vos, M. De Chial, C. Cochez, S. Jansen, B. Tummler, J. M. Meyer, and P. Cornelis. Study of pyoverdine type and production by Pseudomonas aeruginosa isolated from cystic fibrosis patients: Prevalence of type ii pyoverdine isolates and accumulation of pyoverdine-negative mutations. Arch Microbiol 175, 384-388 (2001)
doi:10.1007/s002030100278
8. F. Harrison, and A. Buckling. Hypermutability impedes cooperation in pathogenic bacteria. Curr Biol 15, 1968-1971 (2005)
doi:10.1016/j.cub.2005.09.048
9. Ashleigh S Griffin, Stuart A West, and Angus Buckling. Cooperation and competition in pathogenic bacteria. Nature 430, 1024-1027 (2004)
10. W. D. Hamilton. The genetical evolution of social behaviour i & ii. J Theor Biol 7, 1-52 (1964)
11. A. Ross-Gillespie, A. Gardner, S. A. West, and A. S. Griffin. Frequency dependence and cooperation: Theory and a test with bacteria. Am Nat 170, 331-342 (2007)
doi:10.1086/519860
12. J. M. Meyer, A. Neely, A. Stintzi, C. Georges, and I. A. Holder. Pyoverdin is essential for virulence of Pseudomonas aeruginosa. Infect Immun 64, 518-523 (1996)
13. B. Schwyn, and J. B. Neilands. Universal chemical assay for the detection and determination of siderophores. Anal Biochem 160, 47-56 (1987)
doi:10.1016/0003-2697(87)90612-9
14. S. W. Bearden, J. D. Fetherston, and R. D. Perry. Genetic organization of the yersiniabactin biosynthetic region and construction of avirulent mutants in yersinia pestis. Infect Immun 65, 1659-1668 (1997)
15. C. M. Litwin, T. W. Rayback, and J. Skinner. Role of catechol siderophore synthesis in vibrio vulnificus virulence. Infect Immun 64, 2834-2838 (1996)
16. I. Nyilasi, T. Papp, M. Tako, E. Nagy, and C. Vagvolgyi. Iron gathering of opportunistic pathogenic fungi. A mini review. Acta Microbiol Immunol Hung 52, 185-197 (2005)
doi:10.1556/AMicr.52.2005.2.4
17. D. Greig, and M. Travisano. The prisoner's dilemma and polymorphism in yeast suc genes. Proc Biol Sci 271 Suppl 3, S25-26 (2004)
doi:10.1098/rsbl.2003.0083
18. E. V. O'Loughlin, and R. M. Robins-Browne. Effect of shiga toxin and shiga-like toxins on eukaryotic cells. Microbes Infect 3, 493-507 (2001)
doi:10.1016/S1286-4579(01)01405-8
19. D. S. Hooi, B. W. Bycroft, S. R. Chhabra, P. Williams, and D. I. Pritchard. Differential immune modulatory activity of Pseudomonas aeruginosa quorum-sensing signal molecules. Infect Immun 72, 6463-6470 (2004)
doi:10.1128/IAI.72.11.6463-6470.2004
20. K. Tateda, Y. Ishii, M. Horikawa, T. Matsumoto, S. Miyairi, J. C. Pechere, T. J. Standiford, M. Ishiguro, and K. Yamaguchi. The Pseudomonas aeruginosa autoinducer n-3-oxododecanoyl homoserine lactone accelerates apoptosis in macrophages and neutrophils. Infect Immun 71, 5785-5793 (2003)
doi:10.1128/IAI.71.10.5785-5793.2003
21. Lee Alan Dugatkin, Michael Perlin, and Ronald Atlas. Antibiotic resistance and the evolution of group-beneficial traits. Ii: A metapopulation model. Journal of Theoretical Biology 236, 392-396 (2005)
doi:10.1016/j.jtbi.2005.03.021
22. Stuart A. West, Ashleigh S. Griffin, Andy Gardner, and Stephen P. Diggle. Social evolution theory for microorganisms. Nature Reviews Microbiology 4, 597-607 (2006)
23. E. Banin, M. L. Vasil, and E. P. Greenberg. Iron and Pseudomonas aeruginosa biofilm formation. Proc Natl Acad Sci U S A 102, 11076-11081 (2005)
doi:10.1073/pnas.0504266102
24. Paolo Visca, Francesco Imperi, and Iain L. Lamont. Pyoverdine siderophores: From biogenesis to biosignificance. Trends in Microbiology 15, 22-30 (2007)
doi:10.1016/j.tim.2006.11.004
25. Robert B. Fick: Pseudomonas aeruginosa the opportunist : Pathogenesis and disease (CRC Press, Boca Raton) (1993)
26. Alfonso E Alvarez, Antonio F Ribeiro, Gabriel Hessel, Carmen S Bertuzzo, and JosŽ D Ribeiro. Cystic fibrosis at a brazilian center of excellence: Clinical and laboratory characteristics of 104 patients and their association with genotype and disease severity. Journal de Pediatria (Rio de Janeiro) 80, 371-379 (2004)
doi:10.2223/1221
doi:10.2223/JPED.1221
27. G. M. Nixon, D. S. Armstrong, R. Carzino, J. B. Carlin, A. Olinsky, C. F. Robertson, and K. Grimwood. Clinical outcome after early Pseudomonas aeruginosa infection in cystic fibrosis. J Pediatr 138, 699-704 (2001)
doi:10.1067/mpd.2001.112897
28. N. T. Petersen, N. Hoiby, C. H. Mordhorst, K. Lind, E. W. Flensborg, and B. Bruun. Respiratory infections in cystic fibrosis patients caused by virus, chlamydia and mycoplasma--possible synergism with Pseudomonas aeruginosa. Acta Paediatr Scand 70, 623-628 (1981)
doi:10.1111/j.1651-2227.1981.tb05757.x
29. K. Riedel, M. Hentzer, O. Geisenberger, B. Huber, A. Steidle, H. Wu, N. Hoiby, M. Givskov, S. Molin, and L. Eberl. N-acylhomoserine-lactone-mediated communication between Pseudomonas aeruginosa and burkholderia cepacia in mixed biofilms. Microbiology 147, 3249-3262 (2001)
30. A. A. Wahab, I. A. Janahi, M. M. Marafia, and S. El-Shafie. Microbiological identification in cystic fibrosis patients with cftr i1234v mutation. J Trop Pediatr 50, 229-233 (2004)
doi:10.1093/tropej/50.4.229
31. J. B. Lyczak, C. L. Cannon, and G. B. Pier. Lung infections associated with cystic fibrosis. Clin Microbiol Rev 15, 194-222 (2002)
doi:10.1128/CMR.15.2.194-222.2002
32. H. J. Bremermann, and J. Pickering. A game-theoretical model of parasite virulence. J Theor Biol 100, 411-426 (1983)
doi:10.1016/0022-5193(83)90438-1
33. S. A. Frank. A kin selection model for the evolution of virulence. Proc Biol Sci 250, 195-197 (1992)
doi:10.1098/rspb.1992.0149
34. S. A. Frank. Kin selection and virulence in the evolution of protocells and parasites. Proc Biol Sci 258, 153-161 (1994)
doi:10.1098/rspb.1994.0156
35. S. A. Frank. Models of parasite virulence. Q Rev Biol 71, 37-78 (1996)
doi:10.1086/419267
36. M. A. Nowak, and R. M. May. Superinfection and the evolution of parasite virulence. Proc Biol Sci 255, 81-89 (1994)
doi:10.1098/rspb.1994.0012
37. Steven A. Frank: Foundations of social evolution (Princeton University Press, Princeton, NJ) (1998)
38. F. Harrison, L. E. Browning, M. Vos, and A. Buckling. Cooperation and virulence in acute Pseudomonas aeruginosa infections. BMC Biol 4, 21 (2006)
doi:10.1186/1741-7007-4-21
39. E. E. Smith, D. G. Buckley, Z. Wu, C. Saenphimmachak, L. R. Hoffman, D. A. D'Argenio, S. I. Miller, B. W. Ramsey, D. P. Speert, S. M. Moskowitz, J. L. Burns, R. Kaul, and M. V. Olson. Genetic adaptation by Pseudomonas aeruginosa to the airways of cystic fibrosis patients. Proc Natl Acad Sci U S A (2006)
40. S. Finnan, J. P. Morrissey, F. O'Gara, and E. F. Boyd. Genome diversity of Pseudomonas aeruginosa isolates from cystic fibrosis patients and the hospital environment. J Clin Microbiol 42, 5783-5792 (2004)
doi:10.1128/JCM.42.12.5783-5792.2004
41. U. Romling, J. Wingender, H. Muller, and B. Tummler. A major Pseudomonas aeruginosa clone common to patients and aquatic habitats. Appl Environ Microbiol 60, 1734-1738 (1994)
42. M. J. Struelens, V. Schwam, A. Deplano, and D. Baran. Genome macrorestriction analysis of diversity and variability of Pseudomonas aeruginosa strains infecting cystic fibrosis patients. J Clin Microbiol 31, 2320-2326 (1993)
43. Steven D. Allison. Cheaters, diffusion and nutrients constrain decomposition by microbial enzymes in spatially structured environments. Ecology Letters 8, 626-635 (2005)
doi:10.1111/j.1461-0248.2005.00756.x
44. Ilik Saccheri, and Ilkka Hanski. Natural selection and population dynamics. Trends in Ecology & Evolution Twenty years of TREE - part I 21, 341-347 (2006)
45. S. A. West, I. Pen, and A. S. Griffin. Cooperation and competition between relatives. Science 296, 72-75 (2002)
doi:10.1126/science.1065507
46. A. Gardner, and S. A. West. Demography, altruism, and the benefits of budding. J Evol Biol 19, 1707-1716 (2006)
doi:10.1111/j.1420-9101.2006.01104.x
47. Freya Harrison. Microbial ecology of the cystic fibrosis lung. Microbiology 153, 917-923 (2007)
doi:10.1099/mic.0.2006/004077-0
48. M M Anzaudo, N P Busquets, S Ronchi, and C Mayoral. Microorganismos pat-genos aislados en muestras respiratorias de ni-os con fibrosis qu'stica. Revista Argentina de Microbiolog'a 37, 129-134 (2005)
49. N. Hoiby. Epidemiological investigations of the respiratory tract bacteriology in patients with cystic fibrosis. Acta Pathol Microbiol Scand (B) Microbiol Immunol 82, 541-550 (1974)
50. J. E. Moore, A. Shaw, B. C. Millar, D. G. Downey, P. G. Murphy, and J. S. Elborn. Microbial ecology of the cystic fibrosis lung: Does microflora type influence microbial loading? Br J Biomed Sci 62, 175-178 (2005)
51. M. A. Santana, E. Matos, M. do Socorro Fontoura, R. Franco, D. Barreto, and A. C. Lemos. Prevalence of pathogens in cystic fibrosis patients in bahia, brazil. Braz J Infect Dis 7, 69-72 (2003)
doi:10.1590/S1413-86702003000100008
52. L. M. Mashburn, A. M. Jett, D. R. Akins, and M. Whiteley. Staphylococcus aureus serves as an iron source for Pseudomonas aeruginosa during in vivo coculture. J Bacteriol 187, 554-566 (2005)
doi:10.1128/JB.187.2.554-566.2005
53. K. L. Palmer, L. M. Mashburn, P. K. Singh, and M. Whiteley. Cystic fibrosis sputum supports growth and cues key aspects of Pseudomonas aeruginosa physiology. J Bacteriol 187, 5267-5277 (2005)
doi:10.1128/JB.187.15.5267-5277.2005
54. B. E. Britigan, M. B. Hayek, B. N. Doebbeling, and R. B. Fick, Jr. Transferrin and lactoferrin undergo proteolytic cleavage in the Pseudomonas aeruginosa-infected lungs of patients with cystic fibrosis. Infect Immun 61, 5049-5055 (1993)
55. S. W. Stites, M. W. Plautz, K. Bailey, A. R. O'Brien-Ladner, and L. J. Wesselius. Increased concentrations of iron and isoferritins in the lower respiratory tract of patients with stable cystic fibrosis. Am J Respir Crit Care Med 160, 796-801 (1999)
56. S. W. Stites, B. Walters, A. R. O'Brien-Ladner, K. Bailey, and L. J. Wesselius. Increased iron and ferritin content of sputum from patients with cystic fibrosis or chronic bronchitis. Chest 114, 814-819 (1998)
doi:10.1378/chest.114.3.814
57. F. Harrison, J. Paul, R. C. Massey, and A. Buckling. Inter-specific competition and siderophore-mediated cooperation in Pseudomonas aeruginosa. ISME Journal 2, 49-55 (2008)
doi:10.1038/ismej.2007.96
58. Laurent Keller, and Michael G. Surette. Communication in bacteria: An ecological and evolutionary perspective. 4, 249-258 (2006)
59. J. L. Burns, J. Emerson, J. R. Stapp, D. L. Yim, J. Krzewinski, L. Louden, B. W. Ramsey, and C. R. Clausen. Microbiology of sputum from patients at cystic fibrosis centers in the united states. Clin Infect Dis 27, 158-163 (1998)
doi:10.1086/514631
60. N. Hoiby, and J. B. Hertz. Quantitative studies on immunologically specific and non-specific absorption of Pseudomonas aeruginosa antibodies in serum from cystic fibrosis patients. Acta Pathol Microbiol Scand (C) 89, 185-192 (1981)
61. S. Qazi, B. Middleton, S. H. Muharram, A. Cockayne, P. Hill, P. O'Shea, S. R. Chhabra, M. Camara, and P. Williams. N-acylhomoserine lactones antagonize virulence gene expression and quorum sensing in Staphylococcus aureus. Infect Immun 74, 910-919 (2006)
doi:10.1128/IAI.74.2.910-919.2006
62. F. Ratjen, G. Comes, K. Paul, H. G. Posselt, T. O. Wagner, and K. Harms. Effect of continuous antistaphylococcal therapy on the rate of p. Aeruginosa acquisition in patients with cystic fibrosis. Pediatr Pulmonol 31, 13-16 (2001)
doi:10.1002/1099-0496(200101)31:1<13::AID-PPUL1001>3.0.CO;2-N
doi:10.1002/1099-0496(200101)31:1<13::AID-PPUL1001>3.3.CO;2-E
63. P. D. Sniegowski, P. J. Gerrish, and R. E. Lenski. Evolution of high mutation rates in experimental populations of e. Coli. Nature 387, 703-705 (1997)
doi:10.1038/42701
64. P. D. Sniegowski, P. J. Gerrish, T. Johnson, and A. Shaver. The evolution of mutation rates: Separating causes from consequences. Bioessays 22, 1057-1066 (2000)
doi:10.1002/1521-1878(200012)22:12<1057::AID-BIES3>3.0.CO;2-W
doi:10.1002/1521-1878(200012)22:12<1057::AID-BIES3>3.3.CO;2-N
65. Mark A. Bedau, and Norman H. Packard. Evolution of evolvability via adaptation of mutation rates. Biosystems 69, 143-162 (2003)
doi:10.1016/S0303-2647(02)00137-5
66. J. A. de Visser. The fate of microbial mutators. Microbiology 148, 1247-1252 (2002)
67. A. Giraud, I. Matic, O. Tenaillon, A. Clara, M. Radman, M. Fons, and F. Taddei. Costs and benefits of high mutation rates: Adaptive evolution of bacteria in the mouse gut. Science 291, 2606-2608 (2001)
doi:10.1126/science.1056421
68. P. B. Rainey. Evolutionary genetics: The economics of mutation. Curr Biol 9, R371-373 (1999)
doi:10.1016/S0960-9822(99)80230-9
69. F. Taddei, M. Radman, J. Maynard-Smith, B. Toupance, P. H. Gouyon, and B. Godelle. Role of mutator alleles in adaptive evolution. Nature 387, 700-702 (1997)
doi:10.1038/42696
70. A. Oliver, R. Canton, P. Campo, F. Baquero, and J. Blazquez. High frequency of hypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection. Science 288, 1251-1254 (2000)
doi:10.1126/science.288.5469.1251
71. A. Oliver, B. R. Levin, C. Juan, F. Baquero, and J. Blazquez. Hypermutation and the preexistence of antibiotic-resistant Pseudomonas aeruginosa mutants: Implications for susceptibility testing and treatment of chronic infections. Antimicrob Agents Chemother 48, 4226-4233 (2004)
doi:10.1128/AAC.48.11.4226-4233.2004
72. Franziska Schaaff, Andrea Reipert, and Gabriele Bierbaum. An elevated mutation frequency favors development of vancomycin resistance in Staphylococcus aureus. Antimicrob Agents Chemother 46, 3540-3548 (2002)
doi:10.1128/AAC.46.11.3540-3548.2002
73. P. Funchain, A. Yeung, J. L. Stewart, R. Lin, M. M. Slupska, and J. H. Miller. The consequences of growth of a mutator strain of Escherichia coli as measured by loss of function among multiple gene targets and loss of fitness. Genetics 154, 959-970 (2000)
74. W Trobner, and R Piechocki. Selection against hypermutability in Escherichia coli during long term evolution. Mol Gen Genet 198, 177-178 (1984)
doi:10.1007/BF00328720
75. E.F. Harrison (2007) Cooperative behaviour in Pseudomonas aeruginosa: Ecology, evolution and pathology. DPhil thesis, University of Oxford.
76. F. Harrison, and A. Buckling. High relatedness selects against hypermutability in bacterial metapopulations. Proc R Soc B 274, 1341-1347 (2007)
doi:10.1098/rspb.2006.0408
77. J. E. LeClerc, B. Li, W. L. Payne, and T. A. Cebula. High mutation frequencies among Escherichia coli and Salmonella pathogens. Science 274, 1208-1211 (1996)
doi:10.1126/science.274.5290.1208
78. Antoine Giraud, Ivan Matic, Miroslav Radman, Michel Fons, and Francois Taddei. Mutator bacteria as a risk factor in treatment of infectious diseases. Antimicrob Agents Chemother 46, 863-865 (2002)
doi:10.1128/AAC.46.3.863-865.2002
79. O. Ciofu, B. Riis, T. Pressler, H. E. Poulsen, and N. Hoiby. Occurrence of hypermutable Pseudomonas aeruginosa in cystic fibrosis patients is associated with the oxidative stress caused by chronic lung inflammation. Antimicrob Agents Chemother 49, 2276-2282 (2005)
doi:10.1128/AAC.49.6.2276-2282.2005
80. G. J. Velicer, and Y. T. Yu. Evolution of novel cooperative swarming in the bacterium myxococcus xanthus. Nature 425, 75-78 (2003)
doi:10.1038/nature01908
81. J W Costerton, Z Lewandowski, D E Caldwell, D R Korber, and H M Lappin-Scott. Microbial biofilms. Annual Review of Microbiology 49, 711-745 (1995)
doi:10.1146/annurev.mi.49.100195.003431
82. William Costerton, Richard Veeh, Mark Shirtliff, Mark Pasmore, Christopher Post, and Garth Ehrlich. The application of biofilm science to the study and control of chronic bacterial infections. J Clin Invest 112, 1466-1477 (2003)
83. J. B. Xavier, and K. R. Foster. Cooperation and conflict in microbial biofilms. Proc Natl Acad Sci U S A 104, 876-881 (2007)
doi:10.1073/pnas.0607651104
84. M. A. Brockhurst, M. E. Hochberg, T. Bell, and A. Buckling. Character displacement promotes cooperation in bacterial biofilms. Curr Biol 16, 2030-2034 (2006)
doi:10.1016/j.cub.2006.08.068
85. P. B. Rainey, and K. Rainey. Evolution of cooperation and conflict in experimental bacterial populations. Nature 425, 72-74 (2003)
doi:10.1038/nature01906
86. P. K. Singh, M. R. Parsek, E. P. Greenberg, and M. J. Welsh. A component of innate immunity prevents bacterial biofilm development. Nature 417, 552-555 (2002)
doi:10.1038/417552a
87. M. Rella, A. Mercenier, and D. Haas. Transposon insertion mutagenesis of Pseudomonas aeruginosa with a tn5 derivative: Application to physical mapping of the arc gene cluster. Gene 33, 293-303 (1985)
doi:10.1016/0378-1119(85)90237-9
88. H. Ceri, M. E. Olson, C. Stremick, R. R. Read, D. Morck, and A. Buret. The calgary biofilm device: New technology for rapid determination of antibiotic susceptibilities of bacterial biofilms. J Clin Microbiol 37, 1771-1776 (1999)
89. S. M. Moskowitz, J. M. Foster, J. Emerson, and J. L. Burns. Clinically feasible biofilm susceptibility assay for isolates of Pseudomonas aeruginosa from patients with cystic fibrosis. J Clin Microbiol 42, 1915-1922 (2004)
doi:10.1128/JCM.42.5.1915-1922.2004
90. Mary Jane West Eberhard. The evolution of social behavior by kin selection. Q Rev Biol 50, 1-33 (1975)
doi:10.1086/408298
91. K. R. Foster, G. Shaulsky, J. E. Strassmann, D. C. Queller, and C. R. Thompson. Pleiotropy as a mechanism to stabilize cooperation. Nature 431, 693-696 (2004)
doi:10.1038/nature02894
92. B. Lee, J. A. Haagensen, O. Ciofu, J. B. Andersen, N. Hoiby, and S. Molin. Heterogeneity of biofilms formed by nonmucoid Pseudomonas aeruginosa isolates from patients with cystic fibrosis. J Clin Microbiol 43, 5247-5255 (2005)
doi:10.1128/JCM.43.10.5247-5255.2005
93. S. C. Stearns: The evolution of life histories (Oxford University Press, Oxford) (1992)
Abbreviations: CF: cystic fibrosis
Key Words: cooperation, Hamilton's rule, iron, microbial ecology, Pseudomonas aeruginosa, Virulence, Review
Send correspondence to: Freya Harrison, Department of Biology and Biochemistry, University of Bath, Bath BA2 7AY UK, Tel: 441225 385437, Fax: 441225 386779, E-mail:freya.harrison@zoo.ox.ac.uk