Research

The Koskella lab tests evolutionary and ecological theory using empirical approaches and observational data to better understand:

(i) How the transmission mode of microbiota influences host-microbiome interactions and coevolution over time;

(ii) What role the microbiome plays in shaping a host’s susceptibility to disease; and

(iii) How viruses of bacteria (phages) alter bacterial pathogen phenotype, and shape the assembly, composition, or stability of the microbiome.

We use plants as model systems to address these questions. This includes bacterial and phage communities living naturally within long-lived tree hosts (horse chestnut, oak, and pear trees), and interactions between tomato, phyllosphere (leaf-associated) microbiota, the tomato pathogen Pseudomonas syringae, and phages in the greenhouse/laboratory. 

We combine laboratory-based, experimental evolution techniques with field studies of natural interactions between hosts and symbionts to identify and address fundamental questions about community structure, coevolution, and abiotic environment as driving forces of diversity.

Our overall approach is to use natural systems to identify general patterns in disease ecology and evolution, and then to test for specific processes underlying these patterns in the laboratory using experimental (co)evolutionary techniques.

The current research projects in the lab include:

1) Coevolution between bacteria and phages in both natural and controlled laboratory populations (and the comparison of the two).

We have been using controlled selection experiments and surveys of natural populations to explore the influence of environmental heterogeneity on the trajectory of coevolution for tri-trophic interactions between plants, plant-pathogenic bacteria (including Pseudomonas syringae, and Erwinia amylovora), and lytic bacteriophages (viruses that infect and kill their bacterial hosts). These systems are ideal for investigating the scale and impact of phage adaptation to bacteria because the spread of phages influences both the population size and evolution of bacteria, and thus both the colonization success of bacteria and the interactions of bacteria with their host plants. Because each species is directly influenced by the surrounding environmental conditions, and is itself acting as a changing environment for the other two interacting species, this three-way interaction can be used to clearly examine the implications of biotic landscape heterogeneity for the coevolutionary process.

2) The specificity of bacteria-phage interactions in natural microbial communities and host-associated microbiomes

Parasites have been highlighted as the key missing links in food web research, increasing connectance and altering the symmetry of interactions. However, the impact of parasites on species community networks ultimately depends on how specific these interactions are and how parasites evolve in response to changing community structure. For example, a clear understanding of how parasites are transmitted to new species is key to predicting the probability of host shifts and to effectively managing the health of natural, agricultural and human populations. For parasites in microbial populations (e.g. phages) this question is particularly pertinent because bacterial species from a common environment are both more likely to exchange genetic material and to have experienced the same abiotic and biotic selection pressure than the same bacterial species from different environments. This then leads to the hypothesis that phage host range will be more dependent on shared ecology than would be expected by phylogenetic distance among hosts within and across environments. Understanding the host range of phages in complex microbial communities is of additional importance to mitigating risks associated with “phage therapy” (the use of phages to control pathogenic bacterial populations), as the ecological and evolutionary impact of phages on commensal bacteria will likely significantly affect host fitness.

We are currently generating large datasets on bacteria-phage networks within the phyllosphere to determine how specific natural phages may be, and therefore to predict the effect they may have on the diversity and community structure of microbiota living within eukaryotic hosts. You can see some of our preliminary results here, and a discussion of  the impact this understanding may have on our ability to use phages as a biopesticide here.

A) Species composition of culturable bacterial isolates from on and in leaves of eight horse chestnut trees. B) Mean susceptibility of each species to phages (±SEM) isolated from the leaves. C) Imprint of four representative leaves on KB agar and grown for 24 hours.

Since 2016, we have also been running a survey of 25 Fire Blight-infected Pear trees in Berkeley, CA. We are using these temporal collections to build a new try-trophic disease system in which to examine phage host range in natural microbial communities.

Pictured: Awesome undergraduate researchers Tristan Caro and Callie Cuff sampling one of our diseased pear trees in Berkeley (left) and Britt taking advantage of a day home with a sick child to join in (right).

3) The microbiomes of plants, and the potential link between microbiota and susceptibility to disease.

It is now evident that host-associated microbes are a vital part of the host phenotype and may play a role as a first line of defense against pathogens. However, it is also clear that the benefits of a given bacterial community depend on the host genotype in which they are found. Considering the vast differences in microbial community composition found within hosts across time and across host individuals, the importance of community robustness and stability on host fitness is unclear. Plants face an onslaught of attack from bacterial and fungal pathogens, and trees are particularly vulnerable to pathogens due to their exceptional longevity and slow recruitment. Given the known role of commensal bacteria in plant establishment and growth, insight to the contribution of host-associated microbes to plant defense will move forward our understanding of the complex selection pressures shaping the evolution of hosts and their microbiota in nature. We have been using culture-independent, next generation sequencing to characterize the microbiota of horse chestnut, oak and pear trees, as well as tomato plants in order to determine the ecological factors that may shape these communities (including interactions with phages), as well as the importance of community composition in shaping the plant’s ability to resist pathogens. Much of this work is in collaboration with Drs. Jessica Metcalf from Princeton and Steve Lindow.

For example, we recently showed that application of both naturally occurring tomato leaf-associated microbial communities and a synthetic community can confer protection to tomato plants against colonization by Pseudomonas syringae. However, we also discovered that ‘more’ bacteria (both in terms of diversity and dose of inoculation) doesn’t not necessarily confer more protection. Furthermore, we found that fertilization plants can significantly reduce the protective effects we observed! For further coverage of this recent work, see here.

If you’d like to hear more about our work on the complexities of microbiome-mediated protection, you can watch a recent talk Britt gave at the Evolution Meeting in Montpellier, France here.

4) Using experimental coevolution to test whether bacterial resistance to phages comes at a cost, and whether these costs can be harnessed to design better ‘phage therapy’ treatments.

Fitness costs associated with resisting parasites have been the focus of over 200 papers. The research focus reflects the ubiquity of these costs, the significance of environmental context in determining these costs, and the importance of the shape of costs in predicting the evolutionary outcomes. However, given that natural host populations face selection pressures both from multiple parasite strains/species and from other species with which they interact, there remain many open questions regarding the context of the costs within complex networks. We have thus far demonstrated that the magnitude of these costs depends on the coevolutionary history and alternate selection pressures faced by the bacteria, and have ongoing work examining how these costs might manifest within the plant host environment.  These results have applied significance as they suggests that combination phage therapy, i.e., the use of multiple phages in a single cocktail to treat disease, has great promise for controlling bacterial pathogens, even after resistance evolves.