- Email: [email protected]
What is driving the acquisition of mutS and rpoS polymorphisms in Escherichia coli ?
Opinion
TRENDS in Microbiology
457
Vol.11 No.10 October 2003
What is driving the acquisition of mutS and rpoS polymorphisms in Escherichia coli ? Thomas Ferenci School of Molecular and Microbial Biosciences G08, University of Sydney, NSW 2006, Australia
PR
O
O
F
rpoS, and many other strains have variations and mutations in this gene [4]. In every study in which pathogenic or commensal E. coli isolates were analysed, several sequence variants of rpoS were reported and a range of phenotypes attributable to altered rpoS function were exhibited [5– 7]. In the most comprehensive single study, over 20% (13/58) of toxin-producing isolates from diverse sources contained rpoS mutations [5]. The segment of the chromosome between rpoS and mutS is also one of the most polymorphic regions of the genome [8 – 10]. Confirming to the complexity of this region, several studies have shown that mutS mutants also represent more than 1% of independent isolates (e.g. [11]). Because increasing attention is being paid to explaining how patchwork genomes evolve, it is timely to consider general principles that are important for the fixation of these genomic polymorphisms. The environmental conditions that select for strains with mutS and rpoS mutations within populations have been elucidated. Experimental evolution studies [12,13] and a deeper understanding of the role of RpoS (sS) in transcriptional regulation [14 –16] have provided a surprising explanation for the selection of rpoS mutations, and have also indicated the pathway by which these mutations are fixed. To further explain the fixation of polymorphisms, we also need to consider what selective pressures drive the restoration of functions at these loci, which mainly occurs through the lateral transfer of functional genes in both E. coli and Salmonella [10,17]. Both ends of the mutS– rpoS region need to be considered to get a full picture of the selective forces at work; a recent review by Kotewicz et al. [9] discusses the mutS contribution to polymorphisms, the structure of this region and the consequences associated with mutS mutations, including loss of DNA mismatch repair [18]. Here I focus on rpoS and the relative contributions of mutS and rpoS to the hot-spot in polymorphisms.
TE D
Pathogenic and commensal Escherichia coli isolates frequently contain defective alleles of the mutS and rpoS genes, located in a highly polymorphic segment of the chromosome. The environments leading to enrichment of rpoS mutations and the selective advantages of these mutants are becoming apparent. Unexpectedly, rpoS defects occur because of a basic design limitation in cellular regulation. Antagonistic pleiotropy results from the futile competition between different sigma factors associated with the RNA polymerase, and drives the elimination of RpoS (or sS) in environments requiring high levels of transcription that is dependent on RpoD (or sD or s70). Nutrient-limited environments provide an ideal breeding ground for rpoS mutations. By contrast, in other settings, increased stress resistance selects for restoration of rpoS function. Hence extensive polymorphism in the mutS –rpoS region is postulated to result from cycling between environments in which the functional or non-functional genes provide distinct fitness advantages.
U
N
C
O
R
R
EC
Genetic polymorphisms, or sites with frequent sequence difference within a species, need to be explained to allow us to understand how genomes evolve. Recent genomic sequencing projects with Escherichia coli strains and other comparative studies have revealed many such sites [1]. Some polymorphisms are relatively easy to explain; for example, those affecting surface antigens of E. coli reflect extensive antigenic diversity in the species as a consequence of niche adaptation to different hosts [2]. However, polymorphisms also affect genes that are involved in core functions of a cell and therefore essential to bacterial survival. These are more difficult to explain; this discussion attempts to resolve the apparent paradox that a particular gene, which is important for bacterial fitness in many environments, is found to be commonly mutated. The gene in question is rpoS, which encodes the central regulator of general stress resistance [3] (Box 1). Time is ripe to consider why rpoS, so central in preserving bacterial viability, is not highly conserved in both sequence and function. The rpoS gene of E. coli probably sets the record for the number of distinct alleles found in different isolates. Even three ‘wild-type’ K12 strains have sequence differences in Corresponding author: Thomas Ferenci ([email protected]).
Selection of mutations in mutS and rpoS Evidence is growing that the enrichment of mutS mutations within populations is a result of second-order selection in environments where beneficial mutations are occurring [19]. Recent experiments using evolving continuous-culture populations demonstrated that mutators such as mutS and mutY are enriched by hitchhiking with
http://www.trends.com 0966-842X/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.tim.2003.08.003
Opinion
458
TRENDS in Microbiology
Vol.11 No.10 October 2003
Box 1. What are rpoD and rpoS?
What is nutrient limitation, a nutrient-limited chemostat and stress?
F
Bacteria grow at maximum rates only when all nutrients are available at excess concentrations, which is rarely the case in nature. If all nutrients except one are in excess, but the one limiting nutrient (such as glucose) is present at micromolar instead of millimolar levels, the growth rate is reduced to sub-maximal rates. Such conditions can be readily established in chemostat culture. Nutrient-limited bacteria exhibit unique regulatory properties distinct from the feast and famine scenarios [46], including partially elevated expression of RpoS. Because nutrient limitation results in reduced growth rates, it is perceived as stress, even though no external physical or chemical danger surrounds the bacterium.
scavenging [13], but nutrient limitation leads to reduced growth rates and induction of sS [27]. Also reinforcing the competition between sigma factors is an anti-sD protein, known as Rsd, which is expressed from a sS-controlled promoter [28]. When neither extreme of vegetative or stress-induced gene expression is appropriate, a slowgrowth dilemma is imposed (Figure 1). This is a major limitation in environments that are sub-optimal for growth-rate but require high housekeeping gene expression. In this respect, the sD – sS switch is not well designed for dealing with environments that do not exert external stress and do not correspond to the extreme feast or famine scenarios. Therefore escape from the burden imposed by stress-gene expression provides selection pressure for the common occurrence of rpoS mutations in populations of E. coli [13]. The pressure can be extremely strong, and lead to takeover by rpoS mutants within 10 generations of growth under glucose-limited chemostat culture conditions [13]. This trade-off between the high-risk growth strategy in rpoS mutants and the safe but limited growth with sS is also a good example of antagonistic pleiotropy that is relevant to understanding aging in bacteria [29]. Interestingly, the types of rpoS mutations that are selected under nutrient limitation are sensitive to secondary environmental factors. When a mild external stress (such as lowered pH) is combined with glucose limitation, rpoS mutants are still enriched, but are predominantly partial mutants capable of limited stress resistance rather than null mutants, which by contrast are common in the absence of physical stress [13]. Hence the trade-off between stress resistance and vegetative growth is highly sensitive to the environment and can also contribute to the diversity of alleles found in natural populations.
U
N
C
O
R
R
EC
TE D
beneficial mutations in the same cell [20,21]. Given that bacteria readily acquire beneficial mutations that improve fitness in sub-optimal surroundings, it is not surprising that mutator bacteria are common in many settings, including pathogen – host environments [22,23]. The contribution of mutS mutations to nearby polymorphisms is reviewed in [9]. To turn to the other end of the polymorphic region, why have many natural isolates lost or partially lost RpoS function? The answer appears to be that the general stress response controlled by the RpoS s factor protects bacteria under adverse conditions [3], but at the considerable cost of diverting gene expression from housekeeping to stress response [13]. The relative concentrations of sS and sD (RpoD), and that of the signal molecule ppGpp [24], define whether bacteria optimally express housekeeping genes or general stress response genes (Figure 1). Vegetative growth is mainly dependent on sD, but various external stress-producing factors lead to the accumulation of ppGpp and sS [25]. In the absence of chemical or physical stress, a reduction in the growth rate elicited by nutrient limitation also leads to increased ppGpp [26] and sS levels [27]. The concentrations of the core components of RNA polymerase do not change much under various conditions, therefore the increased expression of sS and stress genes leads to a reduction in the expression of housekeeping genes [14 – 16]. In environments that lead to high rpoS expression, but in which housekeeping gene expression is more important, rpoS mutations that escape the limitation imposed on sD-dependent expression are selected [12,13]. The problem is particularly acute under nutrient-limited situations, such as glucose- or nitrogenlimitation [13], or in stationary phase when bacteria are scavenging trace nutrients [12]. It is mostly housekeeping genes expressed with sD that are involved in nutrient
MutS, together with MutH and MutL, constitute the major contributors to the methyl-directed mismatch repair system of E. coli [44]. MutSdependent repair not only corrects mismatches in DNA, but also has a role in maintaining the fidelity of homologous recombination [45]. Mutants lacking MutS function have an approximate 100-fold increase in mutation rates and are therefore known as mutators.
O
A mutation might be beneficial and selected in one setting but have a negative effect on cellular function under other conditions. There is a trade-off between an advantage under the selection condition and a defect under others. Such mutations are pleiotropic because they have multiple effects. For example, an rpoS mutation is pleiotropic because it enhances growth under nutrient limitation and also the expression of many RpoD-dependent genes. The same rpoS mutation exhibits antagonistic pleiotropy because it has a negative effect: reducing
What is mutS?
O
What is antagonistic pleiotropy?
resistance to multiple stresses and preventing expression of many genes.
PR
These genes encode two out of seven different sigma factors in Escherichia coli. RpoD is also called sD or s70, and RpoS is also known as sS or s38. These sigma factors interact with core components of RNA polymerase to initiate RNA synthesis using promoters they have recognised [41]. RpoD is responsible for expression of housekeeping genes, such as those involved in metabolism, macromolecular synthesis and transport. It is also the most commonly used sigma factor in exponentially growing bacteria in rich media. RpoS levels are elevated by a number of stress signals such as low pH, increased osmolarity, and also by signals that indirectly result from reduced growth rate [27,39] (Figure 1). With significant RpoS levels, transcription of about 100 genes is elevated, and the products contribute to enhanced stress resistance such as increased DNA stability (dps in Figure 1), reduced oxygen toxicity (katE) and increased thermo- or osmotolerance through trehalose synthesis (otsAB) [42,43].
http://www.trends.com
Opinion
TRENDS in Microbiology
459
Vol.11 No.10 October 2003
Growth strategy
rpoD
Housekeeping genes e.g. rrn, mgl, his, mal/lamB, ptsG
σD Reduced growth rate e.g. nutrient limitation Core RNA polymerase
ppGpp Stress signals e.g. low pH
σS General stress resistance genes e.g. dps, otsAB, katE
O
Survival strategy
F
rpoS
TRENDS in Microbiology
that the restoration of mutS activity is strongly selected for in either favorable or poor environments. The presence of mutS mutations in a cell and the absence of DNA mismatch repair does ease one aspect of the lateral reacquisition of genes however, namely through the relaxed recombinational fidelity due to mutS [33]. Hence some of the structural polymorphisms in this region might be due to not-quite homologous recombination in the mutS– rpoS region. Perhaps also relevant is the observation that many naturally occurring mutS mutations result from deletions [11], so reversion is often not an option in obtaining functional mutS genes. In contrast to mutS, restoration of rpoS function must be strongly selected for in many environmental settings. The selective conditions for cycling between active and inactive rpoS states are summarized in Figure 2. Studies have shown the increased susceptibility and loss of viability of rpoS mutants resulting from physical or chemical stresses such as hydrostatic pressure, cold stress, oxidative damage, high osmolarity or low pH [7,34 – 37]. RpoS is therefore significant in diverse environments, from survival in seawater [38] to greater acid sensitivity during passage of E. coli O157:H7 through the intestinal tract [6]. The considerable disadvantage of rpoS mutants in stressful environments compared with rpoS þ cells is proposed to provide the selection pressure necessary to regain intact rpoS, and therefore stress resistance. Selection for stress resistance is probably strong and frequent in nature, although rpoS 2 to rpoS þ trend has
U
N
C
O
R
R
EC
TE D
Completing the cycle: restoration of functional mutS and rpoS in populations The driving forces behind the reacquisition of functional mutS and rpoS genes also need to be considered. Given that both mutS (DNA mismatch repair) and rpoS (stress resistance) are involved in important cellular functions, a superficial analysis would suggest that regaining these functional genes would be beneficial. But are the fitness advantages derived from such restoration of function equally important in natural settings? Mutations of the mutS gene cause increased mutation rates in bacteria. The idea that mutations are detrimental and place a heavy burden on the cell is ingrained [30]. However, there is very little experimental evidence that mutators are bad for fitness in laboratory populations. Negative fitness effects only become apparent with extremely high mutation rates when bacteria do sustain growth disadvantages [31]. With weaker mutators, such as mutS or mutY, there is no obvious growth disadvantage [21,32] and these are not rapidly eliminated from growing populations. In the single example where the elimination of mutS mutations was detected in a continuous culture population, mutS became rare owing to a periodic selection event that purged mutators, rather than an inherent fitness disadvantage [20]. In the more complex world outside the laboratory, perhaps mutators do impose a fitness load over extended periods of time, but this will need to be established experimentally. In the narrower context of mutS polymorphisms, it has not been proven
PR
O
Figure 1. Decision making between vegetative growth and survival strategies in Escherichia coli. The content of core RNA polymerase components (a, b and b0 subunits) is relatively constant in a cell, therefore elevation of RpoS (sS) levels leads to competition between sigma factors and reduction in the expression of housekeeping genes [14,39]. Under starvation or extreme stress conditions, RpoS-dependent gene expression leads to a semi-differentiated state with high levels of protection for cellular integrity. During exponential growth, the general stress response is poorly induced and RpoD (sD) does not need to compete for RNA polymerase components. This pattern of decision making is fine under the extremes of feast and famine, but bacteria also need to face intermediate situations in many natural settings. For example, under nutrient limitation with low levels of glucose, ppGpp and rpoS expression is elevated, but housekeeping genes still need to be expressed to high levels so that bacteria can compete for nutrients; mgl, lamB and ptsG are examples of genes essential for nutrient scavenging [40].
http://www.trends.com
Selection environment
Selection for enhanced stress resistance
No external stress; Growth at sub-optimal growth rates, e.g. nutrient limitation; rpoD function more important than rpoS Selection for enhanced growth rate
Defective rpoS TRENDS in Microbiology
Figure 2. Antagonistic pleiotropy in the rpoS polymorphism cycle. The elevated rpoS expression resulting from reduced growth rates under nutrient limitation is a brake on RpoD-dependent gene expression, and is postulated to be the primary selection condition for defective mutations in the rpoS region. Selection is strongest in the absence of ‘external’ stress, defined as a physical or chemical impairment of cell function. In environments where stress survival is more beneficial than nutrient scavenging, RpoS2 bacteria are at a disadvantage. Such an environment selects for either reversion of point mutations or the lateral reacquisition of active rpoS genes when the initial mutation is a deletion.
not been experimentally demonstrated or studied inside the host or other stressful environments.
The author would like to thank Thea King and Andy Holmes for helpful comments on the manuscript and the Australian Research Council for grant support.
References 1 Whittam, T.S. and Bumbaugh, A.C. (2002) Inferences from wholegenome sequences of bacterial pathogens. Curr. Opin. Genet. Dev. 12, 719 – 725 2 Reeves, P.R. (1992) Variation in O-antigens, niche-specific selection and bacterial populations. FEMS Microbiol. Lett. 79, 509 – 516 3 Hengge-Aronis, R. (2000) The general stress response in Escherichia coli. In Bacterial Stress Responses (Hengge-Aronis, R. and Storz, G., eds), pp. 161 – 178, ASM Press 4 Atlung, T. et al. (2002) Characterisation of the allelic variation in the rpoS gene in thirteen K12 and six other non-pathogenic Escherichia coli strains. Mol. Genet. Genomics 266, 873 – 881 5 Waterman, S.R. and Small, P.L. (1996) Characterization of the acid resistance phenotype and rpoS alleles of shiga-like toxin-producing Escherichia coli. Infect. Immun. 64, 2808– 2811 6 Price, S.B. et al. (2000) Role of rpoS in acid resistance and fecal shedding of Escherichia coli O157: H7. Appl. Environ. Microbiol. 66, 632 – 637 7 Benito, A. et al. (1999) Variation in resistance of natural isolates of Escherichia coli O157 to high hydrostatic pressure, mild heat, and other stresses. Appl. Environ. Microbiol. 65, 1564 – 1569 8 Herbelin, C.J. et al. (2000) Gene conservation and loss in the mutSrpoS genomic region of pathogenic Escherichia coli. J. Bacteriol. 182, 5381– 5390 9 Kotewicz, M.L. et al. (2003) Genomic variability among enteric pathogens: the case of the mutS-rpoS intergenic region. Trends Microbiol. 11, 2 – 6 10 Kotewicz, M.L. et al. (2002) Evolution of multi-gene segments in the mutS-rpoS intergenic region of Salmonella enterica serovar Typhimurium LT2. Microbiology 148, 2531– 2540 11 Leclerc, J.E. et al. (1996) High mutation frequencies among Escherichia coli and Salmonella pathogens. Science 274, 1208 – 1211 12 Finkel, S.E. et al. (2000) Long-term survival and evolution in the stationary phase. In Bacterial Stress Responses (Hengge-Aronis, R. and Storz, G., eds) pp. 231 – 238, ASM Press 13 Notley-McRobb, L. et al. (2002) rpoS mutations and loss of general stress resistance in Escherichia coli populations as a consequence of conflict between competing stress responses. J. Bacteriol. 184, 806– 811 14 Farewell, A. et al. (1998) Negative regulation by RpoS – a case of sigma factor competition. Mol. Microbiol. 29, 1039– 1051 15 Kvint, K. et al. (2000) RpoS-dependent promoters require guanosine tetraphosphate for induction even in the presence of high levels of sigma(s). J. Biol. Chem. 275, 14795 – 14798 16 Colland, F. et al. (2002) The interaction between sigma(S), the stationary phase sigma factor, and the core enzyme of Escherichia coli RNA polymerase. Genes Cells 7, 233– 247 17 Brown, E.W. et al. (2001) Phylogenetic evidence for horizontal transfer of mutS alleles among naturally occurring Escherichia coli strains. J. Bacteriol. 183, 1631– 1644 18 Cox, E.C. et al. (1972) Mutator gene studies in Escherichia coli: the mutS gene. Genetics 72, 551 – 567 19 Tenaillon, O. et al. (2001) Second-order selection in bacterial evolution: selection acting on mutation and recombination rates in the course of adaptation. Res. Microbiol. 152, 11 – 16 20 Notley-McRobb, L. and Ferenci, T. (2000) Experimental analysis of molecular events during mutational periodic selections in bacterial evolution. Genetics 156, 1493 – 1501
U
N
C
O
R
R
EC
TE D
Conclusions and prospects The rpoS polymorphisms in natural populations of E. coli can now be explained. E. coli populations shuttle between different environments in which either suboptimal vegetative growth or stress resistance become advantageous, leading to a balance of intact and mutated rpoS genes in natural populations (Figure 2). Indeed, the rpoS results suggest a more general model of how bacteria generate polymorphisms. The antagonistic cycling seen with rpoS is a good model for the generation of polymorphisms in genomes that are subject to more than one environmental pressure (Figure 2), which is the fate of all free-living bacteria. The rpoS situation is only part of the story regarding the mutS– rpoS region. The largely polymorphic nature of the mutS– rpoS segment is the consequence of the close linkage of two loci where mutations occur frequently in bacterial evolution. The individual quantitative contributions of the rpoS and mutS mutations in generating heterogeneity are not easy to estimate, but I suggest that it would be very surprising if the antagonistic pleiotropy of rpoS mutations were not a major player in the chromosomal variation in its neighborhood. Many aspects of the story are yet to be clarified. In natural populations, the driving forces for the restoration of rpoS and mutS function need to be studied. It is also uncertain whether naturally occurring rpoS and mutS mutations are linked. In addition, it is unclear whether the foreign genomic segments that translocate to this region are the result of relaxed recombination in the absence of mismatch repair, or occur through perfectly normal
Acknowledgements
F
External stress in the environment, e.g. low pH; intact rpoS needed to protect viability; rpoD not the major survival determinant
homologous recombination in DNA flanking the mutS – rpoS region. Because there are conserved sequences outside the mutS– rpoS region in enteric bacteria, relaxed recombination is not a necessary condition for lateral transfer to occur. Despite these gaps, the rpoS – mutS region is well on the way to being an example of a wellunderstood polymorphism.
O
Functional rpoS
Vol.11 No.10 October 2003
O
TRENDS in Microbiology
PR
Opinion
460
http://www.trends.com
37
38
39
40 41
42
43 44 45
46
F
36
O
35
of Escherichia coli to cold and is essential for viability at low temperatures. Proc. Natl. Acad. Sci. U.S.A. 99, 9727 – 9732 Eisenstark, A. et al. (1996) Role of Escherichia coli rpoS and associated genes in defense against oxidative damage. Free Radic. Biol. Med. 21, 975– 993 Hengge-Aronis, R. (1993) Survival of hunger and stress: the role of rpoS in early stationary phase gene regulation in E. coli. Cell 72, 165– 168 Robey, M. et al. (2001) Variation in resistance to high hydrostatic pressure and rpoS heterogeneity in natural isolates of Escherichia coli O157: H7. Appl. Environ. Microbiol. 67, 4901– 4907 Munro, P.M. et al. (1995) Influence of the RpoS (KatF) sigma factor on maintenance of viability and culturability of Escherichia coli and Salmonella typhimurium in seawater. Appl. Environ. Microbiol. 61, 1853– 1858 Jishage, M. et al. (1996) Regulation of RNA polymerase sigma subunit synthesis in Escherichia coli – intracellular levels of four species of sigma subunit under various growth conditions. J. Bacteriol. 178, 5447– 5451 Ferenci, T. (1999) Regulation by nutrient limitation. Curr. Opin. Microbiol. 2, 208– 213 Gross, C.A. et al. (1998) The functional and regulatory roles of sigma factors in transcription. Cold Spring Harb. Symp. Quant. Biol. 63, 141– 155 Loewen, P.C. and Hengge-Aronis, R. (1994) The role of the sigma factor sigma(S) (KatF) in bacterial global regulation. Annu. Rev. Microbiol. 48, 53 – 80 Wolf, S.G. et al. (1999) DNA protection by stress-induced biocrystallization. Nature 400, 83 – 85 Horst, J.P. et al. (1999) Escherichia coli mutator genes. Trends Microbiol. 7, 29 – 36 Vulic, M. et al. (1999) Mutation, recombination, and incipient speciation of bacteria in the laboratory. Proc. Natl. Acad. Sci. U.S.A. 96, 7348 – 7351 Ferenci, T. (2001) Hungry bacteria – definition and properties of a nutritional state. Environ. Microbiol. 3, 605– 611
U
N
C
O
R
R
EC
TE D
21 Notley-McRobb, L. et al. (2002) Enrichment and elimination of mutY mutators in Escherichia coli populations. Genetics 162, 1055– 1062 22 Oliver, A. et al. (2000) High frequency of hypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection. Science 288, 1251 – 1253 23 Denamur, E. et al. (2002) High frequency of mutator strains among human uropathogenic Escherichia coli isolates. J. Bacteriol. 184, 605 – 609 24 Jishage, M. et al. (2002) Regulation of Sigma factor competition by the alarmone ppGpp. Genes Dev. 16, 1260 – 1270 25 Gentry, D.R. et al. (1993) Synthesis of stationary-phase sigma factor Sigma-S is positively regulated by ppGpp. J. Bacteriol. 175, 7982 – 7989 26 Teich, A. et al. (1999) Growth rate related concentration changes of the starvation response regulators sigma(S) and ppGpp in glucose-limited fed-batch and continuous cultures of Escherichia coli. Biotechnol. Prog. 15, 123 – 129 27 Notley, L. and Ferenci, T. (1996) Induction of RpoS-dependent functions in glucose-limited continuous culture: what level of nutrient limitation induces the stationary phase of Escherichia coli? J. Bacteriol. 178, 1465– 1468 28 Jishage, M. and Ishihama, A. (1999) Transcriptional organization and in vivo role of the Escherichia coli rsd gene, encoding the regulator of RNA polymerase sigma D. J. Bacteriol. 181, 3768 – 3776 29 Nystrom, T. (2003) Conditional senescence in bacteria: death of the immortals. Mol. Microbiol. 48, 17 – 23 30 Kimura, M. (1967) On the evolutionary adjustment of spontaneous mutation rates. Genet. Res. 9, 23 – 34 31 Trobner, W. and Piechocki, R. (1984) Selection against hypermutability in Escherichia coli during long term evolution. Mol. Gen. Genet. 198, 177 – 178 32 Trobner, W. and Piechocki, R. (1984) Competition between isogenic mutS and mut þ populations of Escherichia coli K12 in continuously growing cultures. Mol. Gen. Genet. 198, 175– 176 33 Denamur, E. et al. (2000) Evolutionary implications of the frequent horizontal transfer of mismatch repair genes. Cell 103, 711 – 721 34 Kandror, O. et al. (2002) Trehalose synthesis is induced upon exposure
461
Vol.11 No.10 October 2003
O
TRENDS in Microbiology
PR
Opinion
Do you want to reproduce material from a Trends journal?
This publication and the individual contributions within it are protected by the copyright of Elsevier. Except as outlined in the terms and conditions (see p. ii), no part of any Trends journal can be reproduced, either in print or electronic form, without written permission from Elsevier. Please address any permission requests to: Rights and Permissions, Elsevier Ltd, PO Box 800, Oxford, UK OX5 1DX. http://www.trends.com
TRENDS in Microbiology
457
Vol.11 No.10 October 2003
What is driving the acquisition of mutS and rpoS polymorphisms in Escherichia coli ? Thomas Ferenci School of Molecular and Microbial Biosciences G08, University of Sydney, NSW 2006, Australia
PR
O
O
F
rpoS, and many other strains have variations and mutations in this gene [4]. In every study in which pathogenic or commensal E. coli isolates were analysed, several sequence variants of rpoS were reported and a range of phenotypes attributable to altered rpoS function were exhibited [5– 7]. In the most comprehensive single study, over 20% (13/58) of toxin-producing isolates from diverse sources contained rpoS mutations [5]. The segment of the chromosome between rpoS and mutS is also one of the most polymorphic regions of the genome [8 – 10]. Confirming to the complexity of this region, several studies have shown that mutS mutants also represent more than 1% of independent isolates (e.g. [11]). Because increasing attention is being paid to explaining how patchwork genomes evolve, it is timely to consider general principles that are important for the fixation of these genomic polymorphisms. The environmental conditions that select for strains with mutS and rpoS mutations within populations have been elucidated. Experimental evolution studies [12,13] and a deeper understanding of the role of RpoS (sS) in transcriptional regulation [14 –16] have provided a surprising explanation for the selection of rpoS mutations, and have also indicated the pathway by which these mutations are fixed. To further explain the fixation of polymorphisms, we also need to consider what selective pressures drive the restoration of functions at these loci, which mainly occurs through the lateral transfer of functional genes in both E. coli and Salmonella [10,17]. Both ends of the mutS– rpoS region need to be considered to get a full picture of the selective forces at work; a recent review by Kotewicz et al. [9] discusses the mutS contribution to polymorphisms, the structure of this region and the consequences associated with mutS mutations, including loss of DNA mismatch repair [18]. Here I focus on rpoS and the relative contributions of mutS and rpoS to the hot-spot in polymorphisms.
TE D
Pathogenic and commensal Escherichia coli isolates frequently contain defective alleles of the mutS and rpoS genes, located in a highly polymorphic segment of the chromosome. The environments leading to enrichment of rpoS mutations and the selective advantages of these mutants are becoming apparent. Unexpectedly, rpoS defects occur because of a basic design limitation in cellular regulation. Antagonistic pleiotropy results from the futile competition between different sigma factors associated with the RNA polymerase, and drives the elimination of RpoS (or sS) in environments requiring high levels of transcription that is dependent on RpoD (or sD or s70). Nutrient-limited environments provide an ideal breeding ground for rpoS mutations. By contrast, in other settings, increased stress resistance selects for restoration of rpoS function. Hence extensive polymorphism in the mutS –rpoS region is postulated to result from cycling between environments in which the functional or non-functional genes provide distinct fitness advantages.
U
N
C
O
R
R
EC
Genetic polymorphisms, or sites with frequent sequence difference within a species, need to be explained to allow us to understand how genomes evolve. Recent genomic sequencing projects with Escherichia coli strains and other comparative studies have revealed many such sites [1]. Some polymorphisms are relatively easy to explain; for example, those affecting surface antigens of E. coli reflect extensive antigenic diversity in the species as a consequence of niche adaptation to different hosts [2]. However, polymorphisms also affect genes that are involved in core functions of a cell and therefore essential to bacterial survival. These are more difficult to explain; this discussion attempts to resolve the apparent paradox that a particular gene, which is important for bacterial fitness in many environments, is found to be commonly mutated. The gene in question is rpoS, which encodes the central regulator of general stress resistance [3] (Box 1). Time is ripe to consider why rpoS, so central in preserving bacterial viability, is not highly conserved in both sequence and function. The rpoS gene of E. coli probably sets the record for the number of distinct alleles found in different isolates. Even three ‘wild-type’ K12 strains have sequence differences in Corresponding author: Thomas Ferenci ([email protected]).
Selection of mutations in mutS and rpoS Evidence is growing that the enrichment of mutS mutations within populations is a result of second-order selection in environments where beneficial mutations are occurring [19]. Recent experiments using evolving continuous-culture populations demonstrated that mutators such as mutS and mutY are enriched by hitchhiking with
http://www.trends.com 0966-842X/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.tim.2003.08.003
Opinion
458
TRENDS in Microbiology
Vol.11 No.10 October 2003
Box 1. What are rpoD and rpoS?
What is nutrient limitation, a nutrient-limited chemostat and stress?
F
Bacteria grow at maximum rates only when all nutrients are available at excess concentrations, which is rarely the case in nature. If all nutrients except one are in excess, but the one limiting nutrient (such as glucose) is present at micromolar instead of millimolar levels, the growth rate is reduced to sub-maximal rates. Such conditions can be readily established in chemostat culture. Nutrient-limited bacteria exhibit unique regulatory properties distinct from the feast and famine scenarios [46], including partially elevated expression of RpoS. Because nutrient limitation results in reduced growth rates, it is perceived as stress, even though no external physical or chemical danger surrounds the bacterium.
scavenging [13], but nutrient limitation leads to reduced growth rates and induction of sS [27]. Also reinforcing the competition between sigma factors is an anti-sD protein, known as Rsd, which is expressed from a sS-controlled promoter [28]. When neither extreme of vegetative or stress-induced gene expression is appropriate, a slowgrowth dilemma is imposed (Figure 1). This is a major limitation in environments that are sub-optimal for growth-rate but require high housekeeping gene expression. In this respect, the sD – sS switch is not well designed for dealing with environments that do not exert external stress and do not correspond to the extreme feast or famine scenarios. Therefore escape from the burden imposed by stress-gene expression provides selection pressure for the common occurrence of rpoS mutations in populations of E. coli [13]. The pressure can be extremely strong, and lead to takeover by rpoS mutants within 10 generations of growth under glucose-limited chemostat culture conditions [13]. This trade-off between the high-risk growth strategy in rpoS mutants and the safe but limited growth with sS is also a good example of antagonistic pleiotropy that is relevant to understanding aging in bacteria [29]. Interestingly, the types of rpoS mutations that are selected under nutrient limitation are sensitive to secondary environmental factors. When a mild external stress (such as lowered pH) is combined with glucose limitation, rpoS mutants are still enriched, but are predominantly partial mutants capable of limited stress resistance rather than null mutants, which by contrast are common in the absence of physical stress [13]. Hence the trade-off between stress resistance and vegetative growth is highly sensitive to the environment and can also contribute to the diversity of alleles found in natural populations.
U
N
C
O
R
R
EC
TE D
beneficial mutations in the same cell [20,21]. Given that bacteria readily acquire beneficial mutations that improve fitness in sub-optimal surroundings, it is not surprising that mutator bacteria are common in many settings, including pathogen – host environments [22,23]. The contribution of mutS mutations to nearby polymorphisms is reviewed in [9]. To turn to the other end of the polymorphic region, why have many natural isolates lost or partially lost RpoS function? The answer appears to be that the general stress response controlled by the RpoS s factor protects bacteria under adverse conditions [3], but at the considerable cost of diverting gene expression from housekeeping to stress response [13]. The relative concentrations of sS and sD (RpoD), and that of the signal molecule ppGpp [24], define whether bacteria optimally express housekeeping genes or general stress response genes (Figure 1). Vegetative growth is mainly dependent on sD, but various external stress-producing factors lead to the accumulation of ppGpp and sS [25]. In the absence of chemical or physical stress, a reduction in the growth rate elicited by nutrient limitation also leads to increased ppGpp [26] and sS levels [27]. The concentrations of the core components of RNA polymerase do not change much under various conditions, therefore the increased expression of sS and stress genes leads to a reduction in the expression of housekeeping genes [14 – 16]. In environments that lead to high rpoS expression, but in which housekeeping gene expression is more important, rpoS mutations that escape the limitation imposed on sD-dependent expression are selected [12,13]. The problem is particularly acute under nutrient-limited situations, such as glucose- or nitrogenlimitation [13], or in stationary phase when bacteria are scavenging trace nutrients [12]. It is mostly housekeeping genes expressed with sD that are involved in nutrient
MutS, together with MutH and MutL, constitute the major contributors to the methyl-directed mismatch repair system of E. coli [44]. MutSdependent repair not only corrects mismatches in DNA, but also has a role in maintaining the fidelity of homologous recombination [45]. Mutants lacking MutS function have an approximate 100-fold increase in mutation rates and are therefore known as mutators.
O
A mutation might be beneficial and selected in one setting but have a negative effect on cellular function under other conditions. There is a trade-off between an advantage under the selection condition and a defect under others. Such mutations are pleiotropic because they have multiple effects. For example, an rpoS mutation is pleiotropic because it enhances growth under nutrient limitation and also the expression of many RpoD-dependent genes. The same rpoS mutation exhibits antagonistic pleiotropy because it has a negative effect: reducing
What is mutS?
O
What is antagonistic pleiotropy?
resistance to multiple stresses and preventing expression of many genes.
PR
These genes encode two out of seven different sigma factors in Escherichia coli. RpoD is also called sD or s70, and RpoS is also known as sS or s38. These sigma factors interact with core components of RNA polymerase to initiate RNA synthesis using promoters they have recognised [41]. RpoD is responsible for expression of housekeeping genes, such as those involved in metabolism, macromolecular synthesis and transport. It is also the most commonly used sigma factor in exponentially growing bacteria in rich media. RpoS levels are elevated by a number of stress signals such as low pH, increased osmolarity, and also by signals that indirectly result from reduced growth rate [27,39] (Figure 1). With significant RpoS levels, transcription of about 100 genes is elevated, and the products contribute to enhanced stress resistance such as increased DNA stability (dps in Figure 1), reduced oxygen toxicity (katE) and increased thermo- or osmotolerance through trehalose synthesis (otsAB) [42,43].
http://www.trends.com
Opinion
TRENDS in Microbiology
459
Vol.11 No.10 October 2003
Growth strategy
rpoD
Housekeeping genes e.g. rrn, mgl, his, mal/lamB, ptsG
σD Reduced growth rate e.g. nutrient limitation Core RNA polymerase
ppGpp Stress signals e.g. low pH
σS General stress resistance genes e.g. dps, otsAB, katE
O
Survival strategy
F
rpoS
TRENDS in Microbiology
that the restoration of mutS activity is strongly selected for in either favorable or poor environments. The presence of mutS mutations in a cell and the absence of DNA mismatch repair does ease one aspect of the lateral reacquisition of genes however, namely through the relaxed recombinational fidelity due to mutS [33]. Hence some of the structural polymorphisms in this region might be due to not-quite homologous recombination in the mutS– rpoS region. Perhaps also relevant is the observation that many naturally occurring mutS mutations result from deletions [11], so reversion is often not an option in obtaining functional mutS genes. In contrast to mutS, restoration of rpoS function must be strongly selected for in many environmental settings. The selective conditions for cycling between active and inactive rpoS states are summarized in Figure 2. Studies have shown the increased susceptibility and loss of viability of rpoS mutants resulting from physical or chemical stresses such as hydrostatic pressure, cold stress, oxidative damage, high osmolarity or low pH [7,34 – 37]. RpoS is therefore significant in diverse environments, from survival in seawater [38] to greater acid sensitivity during passage of E. coli O157:H7 through the intestinal tract [6]. The considerable disadvantage of rpoS mutants in stressful environments compared with rpoS þ cells is proposed to provide the selection pressure necessary to regain intact rpoS, and therefore stress resistance. Selection for stress resistance is probably strong and frequent in nature, although rpoS 2 to rpoS þ trend has
U
N
C
O
R
R
EC
TE D
Completing the cycle: restoration of functional mutS and rpoS in populations The driving forces behind the reacquisition of functional mutS and rpoS genes also need to be considered. Given that both mutS (DNA mismatch repair) and rpoS (stress resistance) are involved in important cellular functions, a superficial analysis would suggest that regaining these functional genes would be beneficial. But are the fitness advantages derived from such restoration of function equally important in natural settings? Mutations of the mutS gene cause increased mutation rates in bacteria. The idea that mutations are detrimental and place a heavy burden on the cell is ingrained [30]. However, there is very little experimental evidence that mutators are bad for fitness in laboratory populations. Negative fitness effects only become apparent with extremely high mutation rates when bacteria do sustain growth disadvantages [31]. With weaker mutators, such as mutS or mutY, there is no obvious growth disadvantage [21,32] and these are not rapidly eliminated from growing populations. In the single example where the elimination of mutS mutations was detected in a continuous culture population, mutS became rare owing to a periodic selection event that purged mutators, rather than an inherent fitness disadvantage [20]. In the more complex world outside the laboratory, perhaps mutators do impose a fitness load over extended periods of time, but this will need to be established experimentally. In the narrower context of mutS polymorphisms, it has not been proven
PR
O
Figure 1. Decision making between vegetative growth and survival strategies in Escherichia coli. The content of core RNA polymerase components (a, b and b0 subunits) is relatively constant in a cell, therefore elevation of RpoS (sS) levels leads to competition between sigma factors and reduction in the expression of housekeeping genes [14,39]. Under starvation or extreme stress conditions, RpoS-dependent gene expression leads to a semi-differentiated state with high levels of protection for cellular integrity. During exponential growth, the general stress response is poorly induced and RpoD (sD) does not need to compete for RNA polymerase components. This pattern of decision making is fine under the extremes of feast and famine, but bacteria also need to face intermediate situations in many natural settings. For example, under nutrient limitation with low levels of glucose, ppGpp and rpoS expression is elevated, but housekeeping genes still need to be expressed to high levels so that bacteria can compete for nutrients; mgl, lamB and ptsG are examples of genes essential for nutrient scavenging [40].
http://www.trends.com
Selection environment
Selection for enhanced stress resistance
No external stress; Growth at sub-optimal growth rates, e.g. nutrient limitation; rpoD function more important than rpoS Selection for enhanced growth rate
Defective rpoS TRENDS in Microbiology
Figure 2. Antagonistic pleiotropy in the rpoS polymorphism cycle. The elevated rpoS expression resulting from reduced growth rates under nutrient limitation is a brake on RpoD-dependent gene expression, and is postulated to be the primary selection condition for defective mutations in the rpoS region. Selection is strongest in the absence of ‘external’ stress, defined as a physical or chemical impairment of cell function. In environments where stress survival is more beneficial than nutrient scavenging, RpoS2 bacteria are at a disadvantage. Such an environment selects for either reversion of point mutations or the lateral reacquisition of active rpoS genes when the initial mutation is a deletion.
not been experimentally demonstrated or studied inside the host or other stressful environments.
The author would like to thank Thea King and Andy Holmes for helpful comments on the manuscript and the Australian Research Council for grant support.
References 1 Whittam, T.S. and Bumbaugh, A.C. (2002) Inferences from wholegenome sequences of bacterial pathogens. Curr. Opin. Genet. Dev. 12, 719 – 725 2 Reeves, P.R. (1992) Variation in O-antigens, niche-specific selection and bacterial populations. FEMS Microbiol. Lett. 79, 509 – 516 3 Hengge-Aronis, R. (2000) The general stress response in Escherichia coli. In Bacterial Stress Responses (Hengge-Aronis, R. and Storz, G., eds), pp. 161 – 178, ASM Press 4 Atlung, T. et al. (2002) Characterisation of the allelic variation in the rpoS gene in thirteen K12 and six other non-pathogenic Escherichia coli strains. Mol. Genet. Genomics 266, 873 – 881 5 Waterman, S.R. and Small, P.L. (1996) Characterization of the acid resistance phenotype and rpoS alleles of shiga-like toxin-producing Escherichia coli. Infect. Immun. 64, 2808– 2811 6 Price, S.B. et al. (2000) Role of rpoS in acid resistance and fecal shedding of Escherichia coli O157: H7. Appl. Environ. Microbiol. 66, 632 – 637 7 Benito, A. et al. (1999) Variation in resistance of natural isolates of Escherichia coli O157 to high hydrostatic pressure, mild heat, and other stresses. Appl. Environ. Microbiol. 65, 1564 – 1569 8 Herbelin, C.J. et al. (2000) Gene conservation and loss in the mutSrpoS genomic region of pathogenic Escherichia coli. J. Bacteriol. 182, 5381– 5390 9 Kotewicz, M.L. et al. (2003) Genomic variability among enteric pathogens: the case of the mutS-rpoS intergenic region. Trends Microbiol. 11, 2 – 6 10 Kotewicz, M.L. et al. (2002) Evolution of multi-gene segments in the mutS-rpoS intergenic region of Salmonella enterica serovar Typhimurium LT2. Microbiology 148, 2531– 2540 11 Leclerc, J.E. et al. (1996) High mutation frequencies among Escherichia coli and Salmonella pathogens. Science 274, 1208 – 1211 12 Finkel, S.E. et al. (2000) Long-term survival and evolution in the stationary phase. In Bacterial Stress Responses (Hengge-Aronis, R. and Storz, G., eds) pp. 231 – 238, ASM Press 13 Notley-McRobb, L. et al. (2002) rpoS mutations and loss of general stress resistance in Escherichia coli populations as a consequence of conflict between competing stress responses. J. Bacteriol. 184, 806– 811 14 Farewell, A. et al. (1998) Negative regulation by RpoS – a case of sigma factor competition. Mol. Microbiol. 29, 1039– 1051 15 Kvint, K. et al. (2000) RpoS-dependent promoters require guanosine tetraphosphate for induction even in the presence of high levels of sigma(s). J. Biol. Chem. 275, 14795 – 14798 16 Colland, F. et al. (2002) The interaction between sigma(S), the stationary phase sigma factor, and the core enzyme of Escherichia coli RNA polymerase. Genes Cells 7, 233– 247 17 Brown, E.W. et al. (2001) Phylogenetic evidence for horizontal transfer of mutS alleles among naturally occurring Escherichia coli strains. J. Bacteriol. 183, 1631– 1644 18 Cox, E.C. et al. (1972) Mutator gene studies in Escherichia coli: the mutS gene. Genetics 72, 551 – 567 19 Tenaillon, O. et al. (2001) Second-order selection in bacterial evolution: selection acting on mutation and recombination rates in the course of adaptation. Res. Microbiol. 152, 11 – 16 20 Notley-McRobb, L. and Ferenci, T. (2000) Experimental analysis of molecular events during mutational periodic selections in bacterial evolution. Genetics 156, 1493 – 1501
U
N
C
O
R
R
EC
TE D
Conclusions and prospects The rpoS polymorphisms in natural populations of E. coli can now be explained. E. coli populations shuttle between different environments in which either suboptimal vegetative growth or stress resistance become advantageous, leading to a balance of intact and mutated rpoS genes in natural populations (Figure 2). Indeed, the rpoS results suggest a more general model of how bacteria generate polymorphisms. The antagonistic cycling seen with rpoS is a good model for the generation of polymorphisms in genomes that are subject to more than one environmental pressure (Figure 2), which is the fate of all free-living bacteria. The rpoS situation is only part of the story regarding the mutS– rpoS region. The largely polymorphic nature of the mutS– rpoS segment is the consequence of the close linkage of two loci where mutations occur frequently in bacterial evolution. The individual quantitative contributions of the rpoS and mutS mutations in generating heterogeneity are not easy to estimate, but I suggest that it would be very surprising if the antagonistic pleiotropy of rpoS mutations were not a major player in the chromosomal variation in its neighborhood. Many aspects of the story are yet to be clarified. In natural populations, the driving forces for the restoration of rpoS and mutS function need to be studied. It is also uncertain whether naturally occurring rpoS and mutS mutations are linked. In addition, it is unclear whether the foreign genomic segments that translocate to this region are the result of relaxed recombination in the absence of mismatch repair, or occur through perfectly normal
Acknowledgements
F
External stress in the environment, e.g. low pH; intact rpoS needed to protect viability; rpoD not the major survival determinant
homologous recombination in DNA flanking the mutS – rpoS region. Because there are conserved sequences outside the mutS– rpoS region in enteric bacteria, relaxed recombination is not a necessary condition for lateral transfer to occur. Despite these gaps, the rpoS – mutS region is well on the way to being an example of a wellunderstood polymorphism.
O
Functional rpoS
Vol.11 No.10 October 2003
O
TRENDS in Microbiology
PR
Opinion
460
http://www.trends.com
37
38
39
40 41
42
43 44 45
46
F
36
O
35
of Escherichia coli to cold and is essential for viability at low temperatures. Proc. Natl. Acad. Sci. U.S.A. 99, 9727 – 9732 Eisenstark, A. et al. (1996) Role of Escherichia coli rpoS and associated genes in defense against oxidative damage. Free Radic. Biol. Med. 21, 975– 993 Hengge-Aronis, R. (1993) Survival of hunger and stress: the role of rpoS in early stationary phase gene regulation in E. coli. Cell 72, 165– 168 Robey, M. et al. (2001) Variation in resistance to high hydrostatic pressure and rpoS heterogeneity in natural isolates of Escherichia coli O157: H7. Appl. Environ. Microbiol. 67, 4901– 4907 Munro, P.M. et al. (1995) Influence of the RpoS (KatF) sigma factor on maintenance of viability and culturability of Escherichia coli and Salmonella typhimurium in seawater. Appl. Environ. Microbiol. 61, 1853– 1858 Jishage, M. et al. (1996) Regulation of RNA polymerase sigma subunit synthesis in Escherichia coli – intracellular levels of four species of sigma subunit under various growth conditions. J. Bacteriol. 178, 5447– 5451 Ferenci, T. (1999) Regulation by nutrient limitation. Curr. Opin. Microbiol. 2, 208– 213 Gross, C.A. et al. (1998) The functional and regulatory roles of sigma factors in transcription. Cold Spring Harb. Symp. Quant. Biol. 63, 141– 155 Loewen, P.C. and Hengge-Aronis, R. (1994) The role of the sigma factor sigma(S) (KatF) in bacterial global regulation. Annu. Rev. Microbiol. 48, 53 – 80 Wolf, S.G. et al. (1999) DNA protection by stress-induced biocrystallization. Nature 400, 83 – 85 Horst, J.P. et al. (1999) Escherichia coli mutator genes. Trends Microbiol. 7, 29 – 36 Vulic, M. et al. (1999) Mutation, recombination, and incipient speciation of bacteria in the laboratory. Proc. Natl. Acad. Sci. U.S.A. 96, 7348 – 7351 Ferenci, T. (2001) Hungry bacteria – definition and properties of a nutritional state. Environ. Microbiol. 3, 605– 611
U
N
C
O
R
R
EC
TE D
21 Notley-McRobb, L. et al. (2002) Enrichment and elimination of mutY mutators in Escherichia coli populations. Genetics 162, 1055– 1062 22 Oliver, A. et al. (2000) High frequency of hypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection. Science 288, 1251 – 1253 23 Denamur, E. et al. (2002) High frequency of mutator strains among human uropathogenic Escherichia coli isolates. J. Bacteriol. 184, 605 – 609 24 Jishage, M. et al. (2002) Regulation of Sigma factor competition by the alarmone ppGpp. Genes Dev. 16, 1260 – 1270 25 Gentry, D.R. et al. (1993) Synthesis of stationary-phase sigma factor Sigma-S is positively regulated by ppGpp. J. Bacteriol. 175, 7982 – 7989 26 Teich, A. et al. (1999) Growth rate related concentration changes of the starvation response regulators sigma(S) and ppGpp in glucose-limited fed-batch and continuous cultures of Escherichia coli. Biotechnol. Prog. 15, 123 – 129 27 Notley, L. and Ferenci, T. (1996) Induction of RpoS-dependent functions in glucose-limited continuous culture: what level of nutrient limitation induces the stationary phase of Escherichia coli? J. Bacteriol. 178, 1465– 1468 28 Jishage, M. and Ishihama, A. (1999) Transcriptional organization and in vivo role of the Escherichia coli rsd gene, encoding the regulator of RNA polymerase sigma D. J. Bacteriol. 181, 3768 – 3776 29 Nystrom, T. (2003) Conditional senescence in bacteria: death of the immortals. Mol. Microbiol. 48, 17 – 23 30 Kimura, M. (1967) On the evolutionary adjustment of spontaneous mutation rates. Genet. Res. 9, 23 – 34 31 Trobner, W. and Piechocki, R. (1984) Selection against hypermutability in Escherichia coli during long term evolution. Mol. Gen. Genet. 198, 177 – 178 32 Trobner, W. and Piechocki, R. (1984) Competition between isogenic mutS and mut þ populations of Escherichia coli K12 in continuously growing cultures. Mol. Gen. Genet. 198, 175– 176 33 Denamur, E. et al. (2000) Evolutionary implications of the frequent horizontal transfer of mismatch repair genes. Cell 103, 711 – 721 34 Kandror, O. et al. (2002) Trehalose synthesis is induced upon exposure
461
Vol.11 No.10 October 2003
O
TRENDS in Microbiology
PR
Opinion
Do you want to reproduce material from a Trends journal?
This publication and the individual contributions within it are protected by the copyright of Elsevier. Except as outlined in the terms and conditions (see p. ii), no part of any Trends journal can be reproduced, either in print or electronic form, without written permission from Elsevier. Please address any permission requests to: Rights and Permissions, Elsevier Ltd, PO Box 800, Oxford, UK OX5 1DX. http://www.trends.com