| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 University of Paris 5, Faculty of Medicine, INSERM, U571, F-75730 Paris 15, France
2 Université Aix-Marseille III, Faculté St Jérome, IMRN service 342, F-13397 Marseille 20, France
Correspondence
Michel Fons
michel.fons{at}univ.u-3mrs.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Present address: Department of Biological Sciences, Imperial College London, Sir Alexander Fleming Building, South Kensington Campus, London SW7 2AZ, UK.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
; Matic et al., 1997; Oliver et al., 2000; Richardson & Stojiljkovic, 2001; Richardson et al., 2002). When compared to wild-type strains, mutators have an increased mutation rate throughout their genome, due to mutations that disrupt some aspect of DNA replication or repair functions (Friedberg et al., 1995). Most mutators from natural isolates are defective for the methyl-directed mismatch repair system (MMR). Among MMR-defective mutators, the majority are MutS-defective, a key component of the MMR (LeClerc et al., 1996; Matic et al., 1997; Oliver et al., 2002; Richardson et al., 2002). The MMR is the principal post-replicative pathway for DNA replication fidelity. In E. coli, the MMR uses dam DNA methylation (adenine methylation of GATC sequences) to discriminate between parental and neo-synthesized DNA strands. When replication errors occur, MutS recognizes base mismatches as well as insertions/deletions (up to 4 bases) and mediates the fixation of MutL. MutL recruits MutH, which cleaves the newly synthesized, unmethylated strand at the nearest GATC sequence. Consequently, the UvrD (MutU) helicase unwinds the strands and the newly synthesized one is degraded by exonucleases in the direction of the mismatch. Finally, DNA polymerase III fills the resultant gap (Modrich & Lahue, 1996). Many other loci, such as mutY (MutY removes adenine from 8-oxo-G and A-G mismatches) or xthA (XthA is a 5' abasic endonuclease), have been described in the laboratory as conferring a mutator phenotype when affected (Horst et al., 1999).
The frequency of mutators in natural populations is higher than that generated by mutation and counterselection equilibrium. They may be selected for directly or indirectly (Boe et al., 2000; Mao et al., 1997; Ninio, 1991). It has been proposed that mutator loci could be indirectly selected by ‘hitch-hiking’, together with the advantageous mutations generated (Chao & Cox, 1983). The selection for advantageous mutations leads to the selection of conserved gene linkages and hence the selection of the mutator allele. This beneficial effect would counter-balance the negative impact of deleterious and lethal mutations also overproduced in mutator lineages. This scenario is consistent with in silico, in vitro and in vivo observations (Taddei et al., 1997; Chao & Cox, 1983; Giraud et al., 2001). For the in vivo model, the mutS+ genotype was restored in a mutS– strain that had previously evolved over 42 days in a mouse digestive tract (this strain was designated r-mutS+). In competition with the mutS+ ancestor strain, the r-mutS+ strain showed a similar gain to a mutS– strain in competition with the mutS+ ancestor strain (Giraud et al., 2001). These results proved that the selection for mutators is indirect, due to other advantageous mutations and not to the mutator phenotype alone.
Due to their ability to adapt more rapidly than wild-type bacteria, and because their evolutionary clock is accelerated, mutators are frequently used as models by ecologists and evolutionists. Since mutators can be involved in pathologic processes (such as infections and cancers) where multiple adaptive steps seem to be required (Denamur et al., 2002; Loeb, 1998; Oliver et al., 2000; Richardson & Stojiljkovic, 2001), and acquire drug resistance more rapidly than wild-type (Giraud et al., 2002; Oliver et al., 2000), they are considered as risk factors. Thus, an understanding of the conditions that permit mutator selection could be useful in the treatment of infectious diseases and of tumours.
It is generally accepted that mutators are selected when they represent a given proportion of the total population; such a selection is termed frequency-dependent (Chao & Cox, 1983). This conclusion seems paradoxical as it would lead to mutator advantage only when they are already sufficiently frequent. It does not address the initial selective steps for reaching such a situation. A second paradox has been provided by in vivo results: in germ-free mice, competition experiments between the wild-type and mutS– strain led to mutator selection when both populations were inoculated in the same mouse. In contrast, when three mice were kept together – one mono-associated with mutators, one mono-associated with wild-type and the third one initially germ-free – the mutator advantage disappeared in the global population (Giraud et al., 2001). Thus it seems that, for some reason that still remains unclear, allowing migration of strains between mice reduces the mutator advantage in vivo.
In this study, an in vitro model examining competition, under selective pressure, between mutators and wild-type E. coli was developed. mutS- and xthA-defective strains were chosen as mutators (xthA– is a weaker mutator than mutS–). By varying the mutation rates, population sizes and relative frequency, we show that mutator selection depends solely on one parameter: the presence of at least one spontaneous pre-adapted bacterium within the mutator population. We further demonstrate that this presence is linked to the mutator population size, mutation rate and chance. These results suggest that a reduced mutator advantage associated with migration (Giraud et al., 2001) is due to pre-migration fixation of an adapted clone in the wild-type population. This allows for separate expansion in the selective environment, prior to post-migration mutator/wild-type competition. In all cases, the mutator advantage depended directly on the ratio of bacteria carrying beneficial mutations in mutator versus wild-type populations.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. All were derived from the E. coli K12 MG1655, a fully sequenced strain (Blattner et al., 1997). MGS and MGN were spontaneous mutants of MG1655 exhibiting resistance to streptomycin and nalidixic acid, respectively. To construct MGS21 the 
(xth-pncA)90 allele was obtained by P1 transduction (Miller, 1974) in MGS using a phage lysate developed in strain LG101 (Schaaper et al., 1985). (xth-pncA)90 was co-transduced together with the zdh-201 : : Tn10 marker. Transductants were selected on tetracycline-supplemented medium (tetracycline resistance is carried by the Tn10 transposon). The MGS21 mutator phenotype was confirmed by determining the frequency of spontaneously occurring spectinomycin-resistant mutants [f(SpcR)].
|
Determination of the frequency of spontaneously occurring SpcR mutants.
Liquid cultures were performed in 10 ml screw-cap tubes containing 10 ml LB medium with vigorous shaking, incubated at 37 °C overnight. The number of c.f.u. was determined on LB plates and spectinomycin-supplemented LB plates. The f(SpcR) is defined as the ratio c.f.u. on spectinomycin plates/c.f.u. on LB plates. This was done several times for each strain.
Competition experiments.
Liquid cultures supplemented with spectinomycin were inoculated with different dilutions of stationary-phase wild-type and mutator. StrR and NalR markers (see Table 1
) were used to distinguish between wild-type and mutator strains and markers were reversed to control for their effects. Spectinomycin represents a lethal selective pressure in these experiments. Appropriate dilutions of the overnight cultures were plated onto solid medium supplemented with streptomycin, nalidixic acid or spectinomycin to determine population sizes and SpcR mutant initial frequencies in the inoculum. After 24 h of competition, appropriate dilutions were plated onto solid medium supplemented with streptomycin or nalidixic acid to determine wild-type and mutator population size.
Migration experiments.
Wild-type and mutator overnight cultures were diluted separately in LB medium supplemented with spectinomycin. At different times, an equal volume of mutator and wild-type culture was mixed in a new culture tube (in the presence of spectinomycin, the ‘post-migration culture’) and incubated for 24 h at 37 °C. Mutator and wild-type population levels were measured as described above.
Quantitative mutator gain.
This was calculated as the final mutator/wild-type ratio, divided by the initial mutator/wild-type ratio. The log10 of quantitative mutator gain defined a standard estimation of fitness. When this value is >0, mutators have been selected for; when the value is 0 there is no benefit of the mutator clone; and when it is <0, mutators are counterselected.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
a). Hence, the antibiotic resistance markers (NalR and StrR) did not affect competitiveness between the strains. When the initial inoculation ratio of mutator/wild-type was above 10–2, the final mutator/wild-type ratio exceeded the initial ratio, indicating that mutators were selected. When the inoculation ratio was below 10–2, mutators were counterselected. Noticeably, the mutator/wild-type ratio of 10–2 represented a critical threshold where mutators could be selected or counterselected. These results could be interpreted as frequency-dependent selection for the mutator lineage.
|
). This observation can again be seen as evidence for frequency-dependent selection of the mutator lineage.
However, the selection thresholds of both sets of experiments (Fig. 1a and b) are clearly different, 10–2 versus 10–4. Thus, for a given initial wild-type population size, mutator selection appears to be frequency-dependent. When using different initial wild-type population sizes, the same absolute frequency did not lead to the selection of mutators. This paradoxical result led us to look for other explanations.
Mutator selection depends on the presence of a mutator bearing an adaptive mutation
In the two previous sets of experiments, the selective threshold corresponded to the same initial mutator population size (2·4x103 mean for MutStr and MutNal). This population size was similar to the inverse of the frequency of beneficial alleles in the mutators population [1/f(SpcR)=3x103]. Below this population size, mutators were not selected for, whereas above this population size they were. This population size was considered to be a population threshold, showing that the selection of the mutators was population-size-dependent.
Twenty-three competition experiments were performed with a mutator inoculum size comparable to the population threshold. In 16 out of the 23 experiments, mutators were selected, whereas in the other seven they were not. This could be due to the stochastic occurrence of adaptive mutations. To check whether mutator selection could be stochastic, a new set of competitions was initiated. The initial mutator population size was tenfold lower than the population threshold (5x102) and the wild-type population was 5x107 bacteria. As expected, the vast majority (109/117) of the trials led to the decline of mutators. However, in eight trials, mutator selection was observed (Fig. 2). Thus mutator selection depends on chance and the probability of mutators to be selected increases with mutator population size.
|
a) or 107 wild-type bacteria (Fig. 3b). In the majority of experiments, mutators were selected when their initial population size was over the ratio threshold and counterselected when under. These two ratios were different but correspond to a similar initial mutator population size, close to the inverse of the frequency of beneficial alleles in the mutator population [1/f(SpcR)=1x105]. The dependence of mutator selection on population size was thus confirmed. The population threshold of xth– mutators was higher than for mutS– mutators, suggesting that strong mutators (in this case mutS–) would be selected in smaller populations than weak mutators (in this case xth–). This is expected if the presence of a beneficial variant was rate-limiting for mutator lineage selection.
|
, third column), the quantitative mutator gain was higher than in ‘1 : 1 competitions' initiated with 4x107 mutators and 5x107 wild-type bacteria (Fig. 2, comparison between second and third columns). The mean gain of these eight competitions was 14-fold higher than that in the ‘1 : 1 competitions’. However, averaged over the 117 competitions, the quantitative mutator advantage was identical to that of the ‘1 : 1 competitions' experiments (56±11 for ‘1 : 1 competitions' and 54±27 for ‘1 : 105 competitions’; Fig. 2, second and third columns). Independently of the initial mutator/wild-type ratio, the mean mutator quantitative advantages were equal to the ratio of the adaptive mutation frequency in mutators to wild-type (52±6; Fig. 2, first column).
Mutators do not have to pay a migration cost to see their benefit reduced after migration
As we have shown that mutator selection, by second-order selection, is only dependent on the presence of a mutator bacterium bearing an adaptive mutation, it was of interest to test if the results showing a reduction of mutator advantage when migration was allowed between mouse guts (Giraud et al., 2001) could also be explained by the presence or absence of adaptive bacteria in mutator and wild-type populations. To eliminate potential costs associated with migration between mouse guts, an in vitro migration model was developed where bacteria did not pay a migration cost because the initial and final niches were similar and migration was instantaneous. Mutators and wild-type were inoculated in fresh medium supplemented with spectinomycin. At different times, an equal volume of mutator and wild-type cultures was mixed together in a new culture tube (in presence of spectinomycin: the ‘post-migration culture’) and incubated for 24 h at 37 °C. In the initial cultures, mutator and wild-type populations exhibited similar growth patterns and reached the same final population size. However, the mutator population grew faster than the wild-type, due to the presence of more numerous adapted variants at the onset of the experiments (Fig. 4a). In the ‘post-migration cultures’, on average the final mutator/wild-type ratio corresponded to the initial ratio of the adaptive mutation frequency in the mutator to that in the wild-type when migration occurred after up to 15 h of separate culture. With migrations occurring after 18, 21 and 24 h of separate culture, the final mutator/wild-type ratio was reduced to one (Fig. 4b). Without any migration cost, our results showed a reduction of the mutator advantage similar to previous results describing bacterial migration between mice (Giraud et al., 2001); depending on adaptive and migration timing, mutator benefit could be diminished without invoking any migration cost.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Consistent with this, it was observed that mutator selection depended on (i) population size; (ii) frequency of adapted mutants in the mutator population [in this case f(SpcR)]; (iii) chance – in some cases, even when the expected number of adapted mutants in mutator population was <1 (0·08), mutators increased in frequency in 8/117 competition studies. The stochastic and population-size effects on mutator selection had been previously predicted in theoretical studies (Tenaillon et al., 1999) where the presence of one mutant (in our case adapted) in a given population depends on: (i) the size of the population; (ii) the frequency of generation of such mutants; and (iii) stochastic occurrence. From our results, it can be concluded that mutator selection by hitch-hiking solely depends on the presence of at least one adapted clone in the mutator population before replacement by the wild-type population. Note that the above results were obtained with an almost lethal selective pressure. In cases of a non-lethal selective pressure, lack of one adapted mutant in the initial mutator population can lead to mutator selection only if an adapted mutant can appear in the mutator population before its replacement by the wild-type population. This event should again depend on chance, mutation rate and total number of mutator genomes replicated before the fixation of the non-mutator lineage.
In a small subset of populations, chance can allow mutator selection when, on average, the number of adapted mutants is <1. In these cases, mutator selection was associated with a much larger gain (an increase of nearly 800-fold with an adapted mutant inoculum frequency of 0·08). However, averaging all competition, mutator gain was identical regardless of the initial expected number of adapted mutants in the mutator population. As mutator selection depends on the presence of adapted clones, mutator gain corresponds to the ratio of the adaptive mutation frequency in mutators versus wild-type.
Giraud et al. (2001) showed that competition between wild-type and mutS– strains (at a 1 : 1 ratio) in germ-free mice lead to mutator selection when both populations were inoculated in the same mouse. In contrast, when three mice were housed together, one colonized by mutators, one mono-associated with wild-type and the third initially germ-free, mutator advantage in the metapopulation (sum of the three bacterial populations) disappeared (Giraud et al., 2001). Genetic amnesia [i.e. the loss of temporary unused genetic functions in a given environment (Funchain et al., 2000; Giraud et al., 2001)] appeared after a longer period of time in mice and could not explain this reduction (Giraud et al., 2001). The reduction of mutator advantage could be due to an intrinsic migration cost (i.e. a pleiotropic cost of mutS– allele during migration). In the absence of such hypothetical intrinsic migration costs, the reduction of the mutator advantage could also be due to the fact that wild-type and mutator populations were initially in different mice. Actually, separation could allow the adaptation of the wild-type population, as it has enough time to fix beneficial mutations. Our in vitro system allowed us to test the separation hypothesis in a context where mutators do not pay this hypothetical migration cost. In our experiments, mutators grew faster, presumably because the initial inoculum contained more adapted (SpcR) mutants than the initial wild-type inoculum. In the ‘post-migration cultures’, mutators lost their advantage when the wild-type inoculum contained an equal number of pre-adapted mutants to the mutator inoculum. Thus, in the ‘post-migration culture’, the competition would be between pre-adapted mutators and pre-adapted wild-type bacteria. Under such circumstances, mutators were not favoured anymore. The mutators did not pay a migration cost, yet their advantage disappeared when competition happened after the fixation of beneficial alleles. The reduction of mutator gain was due to the fixation of adapted clones in the wild-type population, resulting from separate growth in the selective environment before competition. Under this scenario, and after wild-type adaptation, mutators are not favoured any more, probably because no more beneficial mutations can be obtained or perhaps the fitness gain associated with late-arising mutations is too low to be detected in a short time period.
From an evolutionary point of view, it is important to determine if mutators accelerate adaptive evolution in bacterial populations. In nature, E. coli population size may be extremely variable (Savage, 1977). If the population is fully adapted, at equilibrium, the mutator frequency is stable (Boe et al., 2000; Ninio, 1991). Our results confirm that mutator selection occurs more often in large populations than in small ones. Thus, in large populations the probability that a mutator minority contains pre-adapted clones is higher than in small ones. In addition, the probability of mutator selection by hitch-hiking is also higher. It appears that mutators may be preferentially selected in large bacterial populations, where they moderately accelerate adaptive evolution, because wild-type populations are large enough to contain pre-adapted clones at the outset (Arjan et al., 1999; Tenaillon et al., 1999). In contrast, mutators greatly accelerate adaptive evolution in small bacterial populations where the probability for their selection is weaker (Arjan et al., 1999; Tenaillon et al., 1999). Thus, in this case it could be argued that the role of mutators as accelerators of adaptive evolution may be reduced.
However, we can modify this conclusion by considering sequential adaptation and chance. In infectious diseases, pathogens are submitted to a variety of stresses during the pathogenic process or subsequent sequential antibiotic treatments. Even if the initial population is large, the probability that the wild-type population could rapidly cross several selective barriers resulting in the repeated re-occurrence of narrow bottlenecks, is weaker than for mutators. Mutators can be selected and rapidly fixed during adaptation to the first stress. The large mutator population thus obtained can accelerate adaptation to the following stresses (Denamur et al., 2002; Giraud et al., 2002). In situations comparable to small populations, 8 out of 117 competition experiments (initiated with 102 mutators) showed mutator selection. Thus, by permitting mutator selection in small populations, chance allows mutators to accelerate adaptive evolution. So, sequential adaptation and chance could allow a mutator lineage to accelerate adaptive evolution of clonal populations, including the development of mutator tumours in humans (Loeb, 1998).
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Blattner, F. R., Plunkett, G., 3rd, Bloch, C. A. & 14 other authors (1997). The complete genome sequence of Escherichia coli K-12. Science 277, 1453–1474.
Boe, L., Danielsen, M., Knudsen, S., Petersen, J. B., Maymann, J. & Jensen, P. R. (2000). The frequency of mutators in populations of Escherichia coli. Mutat Res 448, 47–55.[Medline]
Chao, L. & Cox, E. C. (1983). Competition between high and low mutating strains of Escherichia coli. Evolution 37, 125–134.[CrossRef]
Denamur, E., Bonacorsi, S., Giraud, A. & 8 other authors (2002). High frequency of mutator strains among human uropathogenic Escherichia coli isolates. J Bacteriol 184, 605–609.
Friedberg, E. C., Walker, G. C. & Siede, W. (1995). DNA Repair and Mutagenesis. Washington, DC: American Society for Microbiology.
Funchain, P., Yeung, A., Stewart, J. L., Lin, R., Slupska, M. M. & Miller, J. H. (2000). 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.
Giraud, A., Matic, I., Tenaillon, O., Clara, A., Radman, M., Fons, M. & Taddei, F. (2001). Costs and benefits of high mutation rates: adaptive evolution of bacteria in the mouse gut. Science 291, 2606–2608.
Giraud, A., Matic, I., Radman, M., Fons, M. & Taddei, F. (2002). Mutator bacteria as a risk factor in treatment of infectious diseases. Antimicrob Agents Chemother 46, 863–865.
Horst, J. P., Wu, T. & Marinus, M. G. (1999). Escherichia coli mutator genes. Trends Microbiol 7, 29–36.[CrossRef][Medline]
LeClerc, J. E., Baoguang, L., Payne, W. L. & Cebula, T. A. (1996). High mutation frequencies among Escherichia coli and Salmonella pathogens. Science 274, 1208–1211.
Loeb, L. A. (1998). Cancer cells exhibit a mutator phenotype. Adv Cancer Res 72, 25–56.[Medline]
Mao, E. F., Lane, L., Lee, J. & Miller, J. H. (1997). Proliferation of mutators in a cell population. J Bacteriol 179, 417–422.
Matic, I., Radman, M., Taddei, F., Picard, B., Doit, C., Bingen, E., Denamur, E. & Elion, J. (1997). Highly variable mutation rates in commensal and pathogenic Escherichia coli. Science 277, 1833.
Miller, J. H. (1974). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
Modrich, P. & Lahue, R. (1996). Mismatch repair in replication fidelity, genetic recombination, and cancer biology. Annu Rev Biochem 65, 101–133.[CrossRef][Medline]
Ninio, J. (1991). Transient mutators: a semiquantitative analysis of the influence of translation and transcription errors on mutation rates. Genetics 129, 957–962.[Abstract]
Oliver, A., Canton, R., Campo, P., Baquero, F. & Blazquez, J. (2000). High frequency of hypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection. Science 288, 1251–1254.
Oliver, A., Baquero, F. & Blazquez, J. (2002). The mismatch repair system (mutS, mutL and uvrD genes) in Pseudomonas aeruginosa: molecular characterization of naturally occurring mutants. Mol Microbiol 43, 1641–1650.[CrossRef][Medline]
Richardson, A. R. & Stojiljkovic, I. (2001). Mismatch repair and the regulation of phase variation in Neisseria meningitidis. Mol Microbiol 40, 645–655.[CrossRef][Medline]
Richardson, A. R., Yu, Z., Popovic, T. & Stojiljkovic, I. (2002). Mutator clones of Neisseria meningitidis in epidemic serogroup A disease. Proc Natl Acad Sci U S A 99, 6103–6107.
Savage, D. C. (1977). Microbial ecology of the gastrointestinal tract. Annu Rev Microbiol 31, 107–133.[CrossRef][Medline]
Schaaper, R. M., Danforth, B. N. & Glickman, B. W. (1985). Rapid repeated cloning of mutant lac repressor genes. Gene 39, 181–189.[CrossRef][Medline]
Taddei, F., Matic, I., Godelle, B. & Radman, M. (1997). To be a mutator, or how pathogenic and commensal bacteria can evolve rapidly. Trends Microbiol 5, 427–428; discussion 428–429.[CrossRef][Medline]
Tenaillon, O., Toupance, B., Le Nagard, H., Taddei, F. & Godelle, B. (1999). Mutators, population size, adaptive landscape and the adaptation of asexual populations of bacteria. Genetics 152, 485–493.
Received 9 August 2005;
revised 14 October 2005;
accepted 21 October 2005.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | MICROBIOLOGY | J GEN VIROL |
| J MED MICROBIOL | ALL SGM JOURNALS | |
, Competitions between WTStr and MutNal; 
, competitions between WTNal and MutStr.
