The results of this proposed work would help us to determine if 1) there is variability in vertically transmitted seed microbiomes across location and host genotype and 2) what these differences arise from. By fully describing the bacterial community of the leaves and fruits as well, we could determine seed colonists’ origin and address if transmission of leaf epiphytes is possible via the seeds. A similar culture-based approached could be used to determine the protective ability of vertically transmitted microbes, as it is possible that taxonomic identity may differ- but function may be the same. In conclusion, the work described in Chapter 2 makes an important step in filling in the gap in the literature as to the existence and importance of vertically transmitted symbionts in this system. Future work will help us understand the origin of these symbionts and if genotype specific microbes are physically heritable through seed transmission.Breeding for agriculturally beneficial traits, such as disease resistance, is an area of ongoing research in the tomato industry, and agriculture in general. Many resistance genes in modern tomatoes originate from their wild Peruvian relative, S. pimpinellifolium, which has a much larger genetic diversity than modern cultivars. Whether or not there are differences in tomato microbiota due to domestication and the presence/absence of resistance genes is a relatively unexplored topic. My work on adult phyllosphere microbiomes originally began as a way to test the hypothesis that breeding for disease resistance genes in tomatoes would ultimately impact the microbiome of those plants. I was, indeed, able to show that host genotype shaped the phyllosphere community for the first two passages of the experiment. When genotypes are classified “resistant” or “susceptible” , there is no overall effect of resistance in microbiomes that had been fully selected on various tomato varieties.

This lack of an overall effect of resistance leads me to conclude that it may not be the resistance genes themselves, drying and curing bud but rather other genetic differences that existed between the genotypes that drove genotype specificity of the bacterial communities. Although I chose two pairs of near-isogenic lines and one out-group, it is likely that other genetic differences existed between near isogenic lines due to the introgression breeding technique that was used to generate the lines. Furthermore, the pairs of near isogenic lines are different tomato cultivars, further contributing to genetic differences amongst hosts. Future work by others will further test the degree to which differences in host genetics impacts the phyllosphere community, and indeed, some evidence for heritable taxa already has been produced using genome wide association studies in corn and Arabidopsis. Moving forward, if there are taxonomic differences between microbiomes, whole genome metagenomics sequencing will help us determine if the functional capability of the microbiomes differed as well. One of the puzzling results from the microbiome passaging work that we were not able to fully explain is that host genotype had a significant impact on bacterial community at the beginning of the experiment, but this declined over time. Through identifying the specific taxa that were significantly associated with the five genotypes in P1 and P2 it’s clear that it is not only the rare taxa associated with particular genotypes that drove such a genotype effect on the microbiomes, and thus the decline in genotype effect cannot be fully explained simply by an overall decrease in diversity. It seems likely that the microbiome underwent environmental selection driven by three factors: 1) the greenhouse, 2) the tomato phyllosphere, and 3) specific tomato genotypes. It seems reasonable that the relative strength of each selection pressure would change over time, whereby host genotype is important early on, but the community experiences progressively more time in the tomato phyllosphere in the greenhouse, the pressure of those environments overshadows a genotype effect.

Even with well-designed and well-controlled experiments, it is difficult to disentangle the selection pressures at play.Through our microbiome transplant and passaging technique that is biologically relevant to how the phyllosphere is naturally colonized, we were not only able to select upon entire host-associated microbial communities, but we could also experimentally test hypotheses regarding microbiome adaption in subsequent experiments. Again, this is due to the physical accessibility of the phyllosphere community and the ease at which it can be inoculated onto hosts. These findings also shed light on a notable challenge in microbiome research. Our data suggest that when describing the microbiome of an open environment, such as plant surfaces, many of the taxa found there may be transient visitors. In the case of the phyllosphere, there are microbes on leaf surfaces that may have emigrated from air, soil, surrounding plants, or other non-plant habitats and do not necessarily represent an adapted community that is capable of growth and persistence. Passaging of microbiomes on a particular host seems to be a powerful way of differentiating those taxa that are, or can become, well adapted to a plant host environment and those that were present upon sampling, but are not well adapted to the environment. Across all systems, much of the work in microbial ecology is highly descriptive: the community associated with a particular host or ecosystem at a given time is described to be its microbiome, implying strong selection for a particular interactive community- rather than a context-dependent assemblage with many recent immigrants, for example. Our findings raise the question as to if a microbiome should be defined as the community that is merely found there upon sampling, or alternatively, if a true microbiome is only one that is adapted to its host or environment. The latter definition might prove hard to establish in many habitats, but fortunately can be readily addressed in the phyllosphere. Thus we expect that our phyllosphere studies will provide important conceptual contributions to the field as a whole.The final chapters of my dissertation explore the importance of bacteriophages in the phyllosphere community.

There are many challenges facing phage research, and studying environmental phages is an especially difficult field due to lack of a universal marker gene for sequencing, a lack of cultivability, and our nascent understanding of phage genetic diversity. Thus, we are unable to describe the abundance and diversity of phages within our samples without shotgun metagenome sequencing. Some of these challenges I was able to overcome, and others I was not. My work primarily depended on the assumption that there were phages on the leaves used to generate the initial inoculum. I took a “black box approach” in which I isolated the size fraction of the microbiome that should contain most lytic phage particles. I treated this as the “phage fraction”. I then looked for an effect of this fraction on bacterial abundance, composition, and diversity. This approach allowed me to overcome the difficulty of identifying and quantifying phages. My findings show that there is, indeed, an important effect of the phage fraction on the microbiome as a whole. This work also provides empirical support for the theory that phages mediate prokaryotic diversity and contribute to temporal population size dynamics. In order to measure phage abundance in starting samples, I attempted to use both fluorescence microscopy and transmission electron microscopy. Although both approaches yielded images of “phage-like particles”, it was impossible to quantify these particles, primarily due to the amount of background fluorescence that interfered with microscopy and the sheer difficulty of identifying low-abundance phages using electron microscopy. Another alternative for quantifying phage particles is flow cytometry, but this method suffers from the samelimitations due to the presence of intrinsically fluorescent contaminating particles. In Appendix 1, I describe a method that I was able to develop for quantification of known phages. If I were to continue the work described in Chapter 4, I would do so with a defined synthetic community of bacteria and phages, from which I could sensitively measure both bacterial and phage abundance throughout the course of the experiment. There remains much to learn about lytic and temperate phages. My findings in Chapter 5 attempt to disentangle the effect of lytic versus temperate phages on the bacterial community on leaves. This work is an important extension from Chapter 4 in a number of ways. First, I wanted to test if the patterns that I observed of the effect of lytic phages on the bacterial community after only a short time together on leaves were persistent over time. I found that patterns of the effect of phages on bacterial communities differed when examined after 3 weeks compared to 1 week. This has important implications in how we think about the issue of timescale in microbial community interactions. It also begs the question: how many lytic phages were present at the start of the experiments, cannabis drying and how long did they persist on the surface of the plants? If lytic phages do not persist, can selection for lysogenic phages produce some of the same patterns in the bacterial community as well? Interestingly, I did observe that treatment consisting only of bacteria that were passaged on leaves for several weeks, i.e. the treatment in which lysogenic phages would have been selected, had a qualitatively higher alpha bacterial diversity than the other treatments. This may suggest that lysogenic phages are capable of promoting bacterial diversity over longer time scales We were not, however, able to find conclusive evidence for an increased presence of lysogenic phage in the communities of bacteria passaged on plants in the absence of lytic phage.

The questions that I attempted to address in Chapters 4 and 5 are difficult to answer with current, common bacteriophage techniques. Moving forward, the best way to uncover both lytic and temperate phage abundance and diversity in the phyllosphere is likely through shotgun metagenomic sequencing. Sequencing the phage fraction would better describe the lytic phages in the system. Sequencing the bacterial fraction should reveal the prevalence of temperate phages integrated into the bacterial genomes. This approach is not devoid of challenges, however, as it can be difficult to sequence microbiomes associated with plants because of the presence of abundant contaminating plant genetic material. Improvements in high throughput sequencing are allowing us to overcome this limitation, if only by the sheer amount of sequences that can be obtained in environmental samples. I see this approach as the most promising way to comprehensively understand the abundance, diversity, and importance of bacteriophages in the phyllosphere.That the microbiome is an entity that fundamentally influences host health and function has caught the attention of researchers, medical doctors, nutritionists, and every other person interested in the microbial world that exists around and within them. Next generation sequencing and other “omics” approaches have enabled us to address the diversity and complexity of various microbial communities, but there are limitations to these approaches. For example, many labs frequently use 16S rRNA amplicon sequencing to describe bacterial communities. This is the most accessible approach due to reasonable cost, accessibility to protocols, and ease of use of sequence analysis pipelines. However, 16S amplicon sequencing only gives us coarse taxonomic resolution of the community, and it does not give any idea of function . The movement of the field away from the division of bacterial diversity within Operational Taxon Units to Exact Sequence Variants should provide more resolution of bacterial diversity, but still does not provide much insight into the functions of these taxa. It seems likely that as the cost of sequencing continues to fall and more labs have access to both the sequencing and analysis technology required, shotgun metagenomics will surpass 16S amplicon sequencing in popularity. Even so, these advanced sequencing approaches must be coupled with hypothesis-driven experiments and highly controlled experimental design. This, in addition to culture-based approaches and the use of synthetic communities when possible, will enable us to move the field beyond observational and correlational findings.Reconstructing the timing and magnitude of changes in human population size is important for understanding the impact of climatic fluctuation, technological innovation, natural selection, and random processes in the evolution of our species. With census population sizes estimated to be only in the millions during most of the Pleistocene, it is obvious that human population size has increased dramatically towards the present.