Update

TRENDS in Microbiology

Vol.14 No.1 January 2006

11

oomycete species, these pathogens are set to have a major role in the future of plant disease research.
Acknowledgements
We are grateful for the funding of the Scottish Executive Environment and Rural Affairs Department, the Biotechnology and Biological Sciences Research Council and the National Science Foundation. 12

13 14

References
1 Kamoun, S. (2003) Molecular genetics of pathogenic oomycetes. Eukaryot. Cell 2, 191199 2 Dangl, J.L. and Jones, J.D.G. (2001) Plant pathogens and integrated defence responses to infection. Nature 411, 826833 3 Martin, G.B. et al. (2003) Understanding the functions of plant disease resistance proteins. Annu. Rev. Plant Biol. 54, 2361 4 Espinosa, A. and Alfano, J. (2004) Disabling surveillance: bacterial type III secretion system effectors that suppress innate immunity. Cell. Microbiol. 6, 10271040 5 Jha, G. et al. (2005) Bacterial type two secretion system secreted proteins: double-edged swords for plant pathogens. Mol. Plant Microbe Interact. 18, 891898 6 Rooney, H.C. et al. (2005) Cladosporium Avr2 inhibits tomato Rcr3 protease required for Cf-2-dependent disease resistance. Science 308, 17831786 7 Jia, Y. et al. (2000) Direct interaction of resistance gene and avirulence gene products confers rice blast resistance. EMBO J. 19, 40044014 8 Dodds, P. et al. (2004) The Melampsora lini AvrL567 avirulence genes are expressed in haustoria and their products are recognised inside plant cells. Plant Cell 16, 755768 9 Kemen, E. et al. (2005) Identication of a protein from rust fungi transferred from haustoria into infected plant cells. Mol. Plant Microbe Interact. 18, 11301139 10 Tian, M. et al. (2004) A Kazal-like extracellular serine protease inhibitor from Phytophthora infestans targets the tomato pathogenesis-related protease P69B. J. Biol. Chem. 279, 2637026377 11 Tian, M. et al. (2005) A second Kazal-like protease inhibitor from

15

16 17 18 19

20

21 22

23

Phytophthora infestans inhibits and interacts with the apoplastic pathogenesis-related P69B of tomato. Plant Physiol. 138, 17851793 Shan, W. et al. (2004) The Avr1b locus of Phytophthora sojae encodes an elicitor and a regulator required for avirulence on soybean plants carrying resistance gene Rps1b. Mol. Plant Microbe Interact. 17, 394403 Allen, R. et al. (2004) Hostparasite co-evolutionary conict between Arabidopsis and downy mildew. Science 306, 19571960 Rehmany, A. et al. (2005) Differential recognition of highly divergent downy mildew avirulence gene alleles by RPP1 resistance genes from two Arabidopsis lines. Plant Cell 17, 18391850 Armstrong, M. et al. (2005) An ancestral oomycete locus contains late blight avirulence gene Avr3a, encoding a protein that is recognized in the host cytoplasm. Proc. Natl. Acad. Sci. U. S. A. 102, 77667771 Hiller, N.L. et al. (2004) A host-targeting signal in virulence proteins reveals a secretome in malarial infection. Science 306, 19341937 Marti, M. et al. (2004) Targeting malaria virulence and remodeling proteins to the host erythrocyte. Science 306, 19301933 Przyborski, J. and Lanzer, M. (2004) The malarial secretome. Science 306, 18971898 Meyers, B. et al. (1998) Receptor-like genes in the major resistance locus of lettuce are subject to divergent selection. Plant Cell 10, 18331846 van der Biezen, E. and Jones, J.D.G. (1998) Plant disease resistance proteins and the gene-for-gene concept. Trends Biochem. Sci. 23, 454456 Innes, R. (2004) Guarding the goods. New insights into the central alarm system of plants. Plant Physiol. 135, 695701 Rose, L.E. et al. (2004) The maintenance of extreme amino acid diversity at the disease resistance gene, RPP13, in Arabidopsis thaliana. Genetics 166, 15171527 Huang, S. et al. (2005) Comparative genomics enabled the isolation of the R3a late blight resistance gene in potato. Plant J. 42, 251261

0966-842X/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.tim.2005.11.007

Adaptive evolution by optimizing expression levels in different environments

Organisms adapt to environmental changes through the xation of mutations that enhance reproductive success. A recent study by Dekel and Alon demonstrated that Escherichia coli adapts to different growth conditions by ne-tuning protein levels, as predicted by a simple cost benet model. A study by Fong et al. showed that independent evolutionary trajectories lead to similar adaptive endpoints. Initial mutations on the path to adaptation altered the mRNA levels of numerous genes. Subsequent optimization through compensatory mutations restored the expression of most genes to baseline
Corresponding author: Babu, M.M. ([email protected]). Available online 13 December 2005
www.sciencedirect.com

levels, except for a small set that retained differential levels of expression. These studies clarify how adaptation could occur by the alteration of gene expression.

Environmental change and adaptation All organisms actively maintain homeostasis despite living in changing environments. However, this equilibrium can break down if organisms encounter challenges that exceed their innate adaptations to cope with atypical conditions [1]. Nevertheless, some individuals in a population might have particular genetic mutations that enable them to survive in unfamiliar environmental conditions. Accordingly, these

12

Update

TRENDS in Microbiology

Vol.14 No.1 January 2006

mutants would have a higher reproductive success (i.e. a greater tness) than would individuals without these mutations, resulting in a higher frequency of the mutations in the surviving gene pool. Thus, the process of natural selection drives organisms to adapt better to new and changing environmental conditions. Adaptive mutations can alter a wide range of physiological processes. For example, adaptation to a new toxic compound in the environment could arise because of mutations that cause: (i) overproduction of a protein that neutralizes the effect of the compound; (ii) an increase in the efciency of a transporter that removes the compound from the cell; or (iii) an increase in the catalytic efciency of the enzyme that processes the toxin. Hence, to understand how organisms adapt, one must identify the underlying mutations and determine how they might confer reproductive success.

The bacterial model of adaptive evolution Bacterial models have been crucial for the establishment of the neo-darwinian paradigm of evolution. The advantages of easy cultivation in well-dened conditions, short generation times and viability offrozen stocks mean that bacteria can be used to track and compare changes in tness between ancestor and descendent strains [2]. Importantly, bacterial models can be used to investigate directly the evolution of organisms that are exposed to environmental changes. This can be done by subjecting a stock culture of bacteria to a range of new test environments to isolate survivors and identify the molecular changes that have enabled them to adapt to the new environment. Bacteria can adapt by optimizing protein expression levels Previous work has shown that adaptive mutations typically alter the physicochemical properties of proteins

(a) No lactose, no IPTG No production of LacZ LacI

lacZYA operon

No cost in producing LacZ No benefit from consuming lactose

(b) Lactose LacI LacZ

Lactose Glucose + galactose

No IPTG, lactose present Production of LacZ

lacZYA operon

Cost in producing LacZ Benefit from consuming lactose

(c) IPTG LacI LacZ IPTG present, no lactose Production of LacZ

lacZYA operon

Cost in producing LacZ No benefit from consuming lactose

(d) Lactose IPTG LacI LacZ Glucose + galactose

IPTG present, lactose present Production of LacZ

lacZYA operon

Cost in producing LacZ Benefit from consuming lactose

Figure 1. Regulation of the lacZYA operon. (a) In the absence of lactose or IPTG, the LacI repressor protein (blue) binds to the promoter and represses expression of the lacZYA operon (red arrow). (b) When lactose is present in the medium, small amounts of the carbohydrate bind to the LacI repressor protein and release it from the promoter. This enables expression of the LacZ enzyme (red circle), which converts lactose to glucose and galactose to derive energy, reected in the increase in growth rate of cells. (c) In the presence of IPTG, a constitutive inducer that cannot be metabolized by LacZ, the LacI repressor binds to IPTG and is released from the promoter. Energy is spent producing LacZ, with no benet because lactose is absent from the environment. This is reected in the decrease in cell growth rate. (d) In the presence of IPTG and lactose, energy is spent producing LacZ and, simultaneously, energy is gained from using lactose in proportion to its availability in the environment. In such instances, producing a sufcient amount of enzyme derives maximum energy and drives growth. No change, no product and no benet denoted by black text; change, presence of product and benet incurred denoted by green text; cost incurred denoted by red text.
www.sciencedirect.com

Update

TRENDS in Microbiology

Vol.14 No.1 January 2006

13

[3]. Several studies have also shown that mutations affecting the amount of protein expressed are an alternative adaptive mechanism [2,4,5]. However, the biological consequences of this mechanism were not well understood until Dekel and Alon [6] recently provided a quantitative model of the mechanism. The authors demonstrated the role of alterations in protein production by measuring the costs and benets to bacteria of expressing proteins at varying levels in different environments. This work focused on the use of lactose as an energy source by Escherichia coli by monitoring the expression levels of the b-galactosidase enzyme (LacZ), which hydrolyzes lactose into glucose and galactose (Figure 1). The study proceeded in three stages (Figure 2). Stage 1: measuring the cost and benet of expressing proteins Dekel and Alon estimated the cost to E. coli of producing LacZ when no lactose was present in the medium. This cost, arising from the production of a superuous protein, draws cellular resources away from essential functions,
(a) (i) Benefit (B) Cost (C) (ii)

thereby reducing growth rates. Cost was measured as the change in growth rate of cells concomitant with the induction of varying levels of LacZ, caused by systematically changing the concentration of isopropyl-b-Dthiogalactopyranoside (IPTG) a constitutive LacZ inducer. By contrast, the production of sufcient amounts of protein to assimilate a substrate, when available, drives growth and provides a benet to the organism. Thus, the benet to E. coli of expressing LacZ was measured as the increase in growth rate in response to systematic changes in lactose concentration in the medium. High levels of LacZ expression were ensured by using saturating levels of the LacZ inducer IPTG (Figure 2a). Stage 2: optimizing the costbenet function to predict optimum expression levels Having calculated the cost and benet to E. coli of expressing LacZ in varying concentrations of lactose, the authors calculated the optimum expression level of LacZ that would maximize the growth function for environments containing low, medium or high lactose concentrations (Figure 2b). The growth function is the difference between the cost and the benet of expressing LacZ in a given lactose environment. Stage 3: bacteria evolve and adapt by changing protein expression levels Finally, Dekel and Alon carried out adaptive evolution experiments in which E. coli cells were cultured in different conditions (low, medium or high concentrations of lactose) over many generations. The authors found that, after w500 generations, cells evolved to reach a state in which they expressed LacZ at levels that were predicted by the costbenet optimization model. Cells grown in low lactose concentrations ceased lacZ expression by introducing a deletion near the promoter. However, cells grown in medium lactose concentrations expressed moderate amounts of the enzyme, and cells grown in high lactose concentrations evolved to express high levels of LacZ (Figure 1c). These mutations occurred within the coding region and resulted in silent or conservative changes to LacZ, implying that the mutations altered mRNA or protein stability or modied the rate of translation. For understanding the diversication of protein expression levels, this study established the use of simple econometric principles, which have proved useful elsewhere in evolutionary biology. Thus, bacterial cells adapt by rapidly (within w500 generations in 3035 min) optimizing protein expression levels to maximize growth. Many paths lead to effectively similar adaptive endpoints A recent study by Fong et al. [7] helped to clarify the sequence of changes that leads to environmental adaptation. Upon transferring fresh E. coli cells to multiple parallel replicates of lactate or glycerol media, the authors used microarrays to monitor mRNA expression proles to track the changes in E. coli cells during the course of adaptation to their new environment. The organism was considered to have adapted to the new medium if its growth

C(Z) = c1Z/ (1Z /M)

B(Z) = b1 [ZLin]

LacZ expression level (Z) (b) Growth (G) (c) Relative LacZ expression level

Lactose concentration (Lin)

G(Z) = B(Z)C(Z)
Optimal Z

LacZ expression level (Z)

Generations (in adaptive evolution experiment)

Figure 2. Use of the costbenet function to predict optimal LacZ expression level. (a) Cost, C(Z), and benet, B(Z), functions of LacZ expression (Z) in Escherichia coli cells. (i) Cost of LacZ expression is measured as the reduction in relative growth rate with respect to wild-type cells when varying amounts of LacZ (variable Z) are expressed in the absence of lactose and with variable amounts of IPTG. (ii) Benet of LacZ expression is measured as the increase in the relative growth rate of E. coli cells that express maximal levels of LacZ (constant Z) when grown in varying amounts of lactose and with saturating amounts of IPTG. Forms of the best-t equations of the observed experimental data for the cost and benet functions are shown. M, c1 and b1 are constants estimated from experimental observations. (b) Prediction of optimal LacZ level with the predicted relative growth rate difference, G(Z), as a function of LacZ expression. Optimization of the costbenet function predicts that, for each environment with varying amounts of lactose, there is an optimum level of LacZ expression that maximizes growth. Low lactose concentration shown in red, medium lactose concentration shown in black and high lactose concentration shown in green. (c) Observed levels of LacZ expression in adaptive evolution experiments over 500 generations. E. coli cells were cultured to enable adaptation to different environments of lactose concentration. Cells accumulated mutations that abolished or decreased LacZ expression in an environment with a low lactose concentration (red), showed similar levels of LacZ expression to wild-type cells in an environment with a medium lactose concentration (black), or increased LacZ expression in an environment with a high lactose concentration (green), as predicted by the costbenet optimization model of Dekel and Alon [6].
www.sciencedirect.com

14

Update

TRENDS in Microbiology

Vol.14 No.1 January 2006

rate was stable in successive generations. It was observed that: (i) initially, numerous genes were differentially expressed, suggesting that the rst xed mutations affected global regulatory processes; (ii) several subsequent compensatory mutations restored the expression of most genes to baseline level, leaving only a few genes with altered expression that were essential for adaptation; and (iii) endpoints that were similarly adapted could result from the alteration of expression of different gene sets. This study illustrates that mutations with generic effects on gene expression are more likely to occur than are specic mutations in a single pathway that lead to a tailor-made adaptation. Whereas pleiotropic mutations cause widespread alterations of gene expression, it is possible that the inherent robustness of regulatory networks provides a buffer for survival, thereby enabling a rapid coarse adaptation followed by ne-tuning through compensatory mutations. Practical implications and future directions In another study, Fong et al. [8] showed that strains with different genetic backgrounds can evolve to optimize growth and maximize lactic acid production. The authors used the metabolic network of E. coli to predict and create knockout strains that resulted in high levels of lactate production. These mutants were cultured in a constant environment that enabled them to adapt to optimize growth, coupled with maximum levels of lactate production. In this context, the study by Dekel and Alon [6] might enable an objective calculation of growth functions (e.g. for engineered mutants) to optimize growth for maximal production of a given biomolecule. Alternatively, the application of such studies to understand changes in protein expression during the adaptation of bacterial pathogens to host niches could aid the design of therapeutics. The monitoring of adaptive changes in gene expression using microarrays [79] could also be extended to more-

complex organisms such as humans to investigate geographically distinct populations that have adapted to local environmental and pathogenic challenges over varying time periods. For example, such studies could elucidate the role of subtle evolutionary optimizations of gene expression that might underlie conditions such as diabetes and obesity. Thus, genome-scale analyses of gene expression help the understanding of evolution in real time.
Acknowledgements
We gratefully acknowledge the intramural research program of the National Institutes of Health for funding our research. We thank S. Balaji for useful comments about the manuscript.

References
1 Smith, J.M. (2000) The Theory of Evolution, 3rd edn, Cambridge University Press 2 Elena, S.F. and Lenski, R.E. (2003) Evolution experiments with microorganisms: the dynamics and genetic bases of adaptation. Nat. Rev. Genet. 4, 457469 3 Hall, B.G. (1983) Evolution of new metabolic functions in laboratory organisms. In Evolution of Genes and Proteins (Nei, M. and Koehn, R.K., eds), pp. 234257, Sinauer Associates 4 Ibarra, R.U. et al. (2002) Escherichia coli K-12 undergoes adaptive evolution to achieve in silico predicted optimal growth. Nature 420, 186189 5 Orr, H.A. (2005) The genetic theory of adaptation: a brief history. Nat. Rev. Genet. 6, 119127 6 Dekel, E. and Alon, U. (2005) Optimality and evolutionary tuning of the expression level of a protein. Nature 436, 588592 7 Fong, S.S. et al. (2005) Parallel adaptive evolution cultures of Escherichia coli lead to convergent growth phenotypes with different gene expression states. Genome Res. 15, 13651372 8 Fong, S.S. et al. (2005) In silico design and adaptive evolution of Escherichia coli for production of lactic acid. Biotechnol. Bioeng. 91, 643648 9 Ferea, T.L. et al. (1999) Systematic changes in gene expression patterns following adaptive evolution in yeast. Proc. Natl. Acad. Sci. U. S. A. 96, 97219726
0966-842X/$ - see front matter Published by Elsevier Ltd. doi:10.1016/j.tim.2005.11.005

Reproduction of material from Elsevier articles

www.sciencedirect.com