Multilevel population genetics in antibiotic resistance


Evolutionary approaches to antibiotic resistance have been mostly based on the consideration of the changes in frequency of resistant bacterial cells (organisms) in comparison with their susceptible counterparts. In this context, the term ‘population’ is frequently used to describe the ensemble of bacterial individuals within a species possessing or lacking a resistance trait (‘resistant population’ vs. ‘susceptible population’). Of course this ‘population’ concept is useful in epidemiology, but does not necessarily reflect real populational variation within a bacterial species (Camus & Lima, 2002; Waples & Gaggiotti, 2006; Balloux, 2010). A group of susceptible bacterial individuals might differ from a group of resistant individuals of the same species by only a single nucleotide, a difference that is not necessarily enough to consider them as different biological populations. Of course, the application of molecular techniques based on polymorphic changes in the bacterial genome to define the population structure of bacterial species, such as multilocus sequence typing, produces a much more faithful image of the population diversity of bacterial species. The term ‘susceptible’ or ‘resistant’ populations refers more to intraclonal ‘populations’, for which the term of intraclonal variants would be better. An emerging concept increasingly used in Public Health Microbiology and extensively treated in this Thematic Issue is reflected by the terms ‘high risk clones, or clonal complexes’, referring to highly specialized genetic populations or subpopulations with enhanced ability to colonize, spread and persist in particular niches after having acquired a diversity of adaptative traits that increase their epidemicity and/or pathogenic potential, including antibiotic resistance. These populations, replacing the more heterogeneous antibiotic-susceptible ones, become ‘amplificators’ of antimicrobial resistance in particular environments as hospitals or farms. However, there has been a new turn in this conceptual frame. Is antibiotic resistance a driver for creating populational diversification?

First, antibiotic resistance could be acquired by lateral genetic transfer as part of a package of genes that might profoundly alter the biology of the recipient susceptible cell, creating a ‘quantum leap’ in evolution and the emergence of a well-separated population. Second, the acquisition of antibiotic resistance could substantially increase the absolute number or cells of a particular clone, and its possibilities of transmission or persistence in particular spaces, environments, and such effects might be followed by a cascade of secondary local adaptations leading again to separate, discrete populations.

In short, antibiotic resistance is not only the consequence of genetic variation, but also a cause of genetic variation. Population genetics theory is the approach that Fisher, Haldane, and Wright originated in the mid-last century, and which was brought to maturity by scientists such as Crow and Kimura during the 1970s (Crow & Kimura, 1970; Mayr, 1998) and Hartl and Clark in the late 1980s (Hartl & Clark, 2007). This theory centred on the study of the frequency, causes and mechanisms of genetic variation, and its phylogenetic and quantitative consequences on the population structure of replicating biological entities. Antibiotic resistance is the consequence and cause of genetic variation, and its effects on Public Health depend on the effects of this variation in the population structure of bacteria. Therefore, as Levin and Lipsitch proposed a decade ago, antibiotic resistance should be investigated from a population genetics perspective (Levin et al., 1997).

In the title encompassing all reviews collected in this Thematic Issue, we refer to ‘multilevel population genetics’, and this requires an explanation. For this purpose, we have to return to the central concept of population. Populations are composed of groups of individuals linked by common patterns of genetic (evolutionary cohesion) and demographic connectivity (ecological cohesion), that is, differing by descent and spatial location from other populations (Waples & Gaggiotti, 2006). However, now, what is in this context an individual? Indeed an individual is both a replicator (a reproducing entity) and a variator (let us borrow this term from engineering and computer sciences), so that an individual is always an evolutionary individual. Of course we are ready to accept that a bacterial cell is an evolutionary individual, eventually forming bacterial populations. However, bacterial cells are not the only evolutionary individuals; at subcellular and supracellular levels, we also have evolutionary individuals. For instance, genes or plasmids can be considered as evolutionary individuals; and also clonal complexes or species (Baquero, 2011). Changes in the frequency of each one of these individuals essentially depend on the ‘four Ps’: penetration (ability to reproduce and spread), promiscuity (ability to exchange information), plasticity (variability) and persistence (construction of durable links with its environment). It necessarily follows that these influences should shape gene or plasmid populations, clonal and species populations. This view imposes the application of the population genetics theory at each one of these population levels, and that justifies expression, multilevel population genetics.

Antibiotic resistance evolves and propagates in a complex network of multilevel populations. Key evolutionary processes in population genetics as selection or random drift affects evolutionary individuals (evolutionary units) across all dimensions of the network. Changes in time of the frequency of particular units occurring in one of the levels will produce changes in the frequency of units in other levels, up or down the hierarchy. A consistent amount of theory has been produced in the last decade about trans-hierarchical evolutionary changes, for instance considering the units and levels of selection. Selection acts simultaneously, but independently on different units because of its complex horizontal and vertical networking, influenced by its relative density and cooperation (Cohen et al., 2011). Of course what we can detect as molecular observers of antibiotic resistance are ‘correlated changes’ in the frequency in entities as resistance genes, other genes, plasmids, clones, species or bacterial communities. This certainly requires ‘multilevel epidemiology’ approaches, in which all the evolutionary units involved in antibiotic resistance are converted in ‘units of surveillance’ (Baquero et al., 2010). Then, we should be able to build correlational structures of change among units and levels, in the hope of establishing the real causal structures of the evolution of antibiotic resistance. Appropriate interventions require the knowledge of causes. This will not be an easy task, as we shall expect to have a complex causal structure of antibiotic resistance, involving factors ranging from antibiotic use, to human and animal demography, and many other ecogenetic influences on the Microbiosphere.

This Thematic Issue reflects such an epistemological approach. The review articles cover most of the current knowledge about the population genetics of most important evolutionary individuals (genes, transposons, plasmids, clones) involved in antibiotic resistance, and how the trans-hierarchical effects become possible across the complex networking structure. A number of causes modifying these effects (biological costs, antibiotic or environmental selection) and how bioinformatic methods might clarify these complex issues are treated in this Thematic Issue. Indeed we are obliged to confront the complexity of facts, imagining complex ways of understanding complex processes, and consequently pave the path for future therapeutic interventions based on ecology and evolution (Baquero et al., 2011). We sincerely hope that this Thematic Issue helps to apply late developments of the ‘new evolutionary synthesis’ (Mayr, 1998) to the global problem of antibiotic resistance.


Work in F.B. and T.M.C. laboratories is funded by research grants from the European Commission (LSHM-2006-037410, KBBE-2008-2B-227258, PAR-241476-FPHEALTH 2009), the Instituto de Salud Carlos III, from the Spanish Ministry of Science and Innovation (PS09/02381 and PI10/02588), the CIBERESP research network for Biomedical Research in Epidemiology and Public Health (CB06/02/0053), and the Regional Government of Madrid (deRemicrobiana network (CAM.S-SAL-0246-2006).