I was part of a team project that evaluated the dynamics of SAR11 subclades over the course of a roughly nine year time-series at BATS. This work follows up on that of Carlson et al. 2009, by using a more advanced technique- phylogenetic placement of pyrosequenced V1-V2 regions from the 16S rRNA gene. The paper describes the development of a new pipeline, PhyloAssigner, which utilizes Pplacer from Erik Matsen’s group, combined with a Last Common Ancestor analysis (LCA), developed by Bánk Beszteri. LCA evaluates the statistical significance of a sequence assignment to a given node on a phylogenetic tree and then pushes it back to a more basal position until that significance achieves a given threshold. The primary results of this analysis demonstrate the presence of additional subclades of SAR11 which display unique spatio-temporal patterns, and emphasize the importance of phylogenetic approaches for determining microbial diversity.
The most recent paper from our lab details the discovery of four new phage which infect Pelagibacter (SAR11). What makes this discovery so compelling is that up until now, no phage for SAR11 had been observed, and some even postulated that SAR11 may be immune to viral infection. In fact, not only do SAR11 phage, termed Pelagiphage, exist, one of these, HTVC010P is more abundant than any previously described marine phage. This makes sense intuitively since SAR11 is the most abundant microorganism, but the discovery is a testament to the value of culture-based techniques for discovering novel phage. HTVC010P is extremely divergent from all characterized phage, which is what we believe has prevented its being noticed in metagenomic libraries prior. However, with the genome in hand, we were able to find it everywhere, and sometimes with startlingly high percent identity values, especially for a virus. Dr. Yanlin Zhao isolated three of the phage, including HTVC010P and Dr. Mike Schwalbach isolated a fourth. Dr. Ben Temperton did the bulk of the metagenomic analysis and manuscript preparation, and I contributed phylogenetic analyses and some metagenomics. The wonderful microscopy images were contributed by Thomas Deerinck and Mark Ellisman at the National Center for Microscopy and Imaging Research. Importantly, Dr. Matt Sullivan at the University of Arizona contributed metagenomic datasets for quantitative viral analysis.
Figure 1 from Zhao et al. 2013.
Reference:
Zhao, Yanlin*, Ben Temperton*, J. Cameron Thrash, Michael S. Schwalbach, Kevin L. Vergin, Zachary C. Landry, Mark Ellisman, Tom Deerinck, Matthew B. Sullivan and Stephen J. Giovannoni. (2013) Abundant SAR11 viruses in the ocean. Nature 494(7437): 357-360. doi: 10.1038/nature11921. (*Equal contribution)
See the related Nature News & Views article by Dr. David Kirchman here, and the Economist article here.
In continued collaboration with the Rappé lab, we published a paper describing an in-depth comparative genomics analysis of the seven SAR11 genomes used in our phylogenomics paper, below. Importantly, it included genomes of SAR11 strains that are highly diverged from the subclade Ia organisms that have been previously studied. The most striking feature of SAR11 genomes is how similar they are across wide divergences at the 16S rRNA gene, or in other words, evolutionary time. In spite of spanning up to 18% divergence at the 16S, these organisms still all have small, streamlined genomes, with ~50% of their orthologs in the core genome, which is unprecedented at this time. Gene order (synteny) is also among the most conserved of currently evaluated organisms. 
Depth of branching in the clade and comparisons of 16S rRNA divergence vs. average amino acid identity divergence indicate that there is no unusual evolution happening using these metrics compared to other organisms. We found a paucity of paralogs throughout the genomes and the majority of these were classified as inparalogs, providing support for streamlining as an ancestral feature of the clade. We also discovered a conserved hypervariable region (termed HVR2) in similar positions and size in all genomes, which we hypothesize may serve as a kind of test bed for novel genetic material in otherwise constrained genomes. For example, some unique genes which may confer lineage-specific metabolic capabilities are found in HVR2, and many parallel duplications and other paralogs occur there as well. In spite of the high similarity of these genomes we still extrapolate an open pan-genome, and thus expect a massive reservoir of additional genetic material owing to the immense population sizes of the SAR11 clade. Check out the full, open-access publication at mBio for more!
Reference
Grote, Jana*, J. Cameron Thrash*, Megan J. Huggett, Zachary C. Landry, Paul Carini, Stephen J. Giovannoni, and Michael S. Rappé. (2012) “Streamlining and Core Genome Conservation among Highly Divergent Members of the SAR11 Clade.” mBio. 3(5): e00252-12 (*Equal contribution)
Recently, our ongoing comparative genomics work revealed that several SAR11 isolates share pathways for C1 oxidation. Jing Sun demonstrated that one of our SAR11 isolates, HTCC1062, could conserve energy by oxidizing C1 compounds and methyl groups from methylated compounds (termed methylovory), and other members of the team showed that mineralization of these various compounds can be observed in bulk seawater. My role was to look at the distribution of the genes for these oxidation pathways shown below across the SAR11 genomes and the Alphaproteobacteria, and some of the phylogenetic analyses of individual genes. This is an important study as the exact energy sources for SAR11 in nature are still poorly understood. Further, it illuminates previously undescribed components of the marine DOC pool accessible for microbial oxidation, a finding which will have possibly significant ramifications for carbon cycling models.
C1 oxidation pathways in SAR11
With the continued cultivation of new SAR11 strains by the Giovannoni, Rappé, and Cho labs, there is an ever-growing stock of genome sequences from which to do comparative genomics. My first SAR11 study was to examine the phylogenomic relationship of SAR11 within the Alphaproteobacteria, and to mitochondria. The close relationship of mitochondria to the Alphaproteobacteria has been supported for some time, but the exact rooting of this clade has been (and may continue to be) up for debate based on the antiquity of the divergence. Using whole genomes from mitochondria and new, highly divergent SAR11 strains with the phylogenomic pipeline Hal, we were able to show considerable phylogenomic evidence for the common ancestry of SAR11 and mitochondria. An important feature of this study was the variety of parameters that we explored to test for topological stability. Taxon sampling, G+C biased outgroups, gap-removal strategy, and allowed missing data were all altered. In all, the study looked at 216 complete phylogenomic trees. Some with more than 45,000 characters.
Example of SAR11 as a sister group to mitochondria
Check out the article covering our SAR11/mitochondrial connection by Carl Zimmer in his Discover Magazine blog The Loom.
SAR11 is a clade of bacteria which are the most abundant marine heterotrophs, with an estimated global population size of 10^28 cells. They are ubiquitous in the world’s oceans and can make up 25-50% of the cells in any given sample (Morris et al. 2002). Due to their extraordinary dominance, they are expected to play a major role in biogeochemical cycling, and thus investigation into their physiology, genomics, ecology and evolution is essential for attaining accurate understanding of the marine carbon cycle. Such a combined approach has been a major part of the ongoing work in the Giovannoni Lab for more than two decades. Cultivation of the organism in 2002 (Rappé et al. 2002) with new high throughput culturing techniques (Connon et al. 2002) opened the door to major findings about the organism, such as its extraordinarily streamlined genome (Giovannoni et al. 2005a), the presence of proteorhodopsin and its use during starvation (Giovannoni et al. 2005b, Steindler et al. 2011), the requirement for reduced sulfur (Tripp et al. 2005), the presence of riboswitches as key regulatory elements (Tripp et al. 2009), and decoupling of mRNA and protein expression during iron limitation (Smith et al. 2010). Work by Vergin et al. (2007) and Wilhelm et al. (2007) has highlighted extraordinary features about SAR11 genomics. In particular, Vergin et al. showed extremely high recombination rates among closely related SAR11 strains. Wilhelm et al. described the presence of conserved highly variable regions (HVRs) in SAR11 genomes and showed the mosaic nature of the recombination pattern among the genomes and metagenomes of these organisms. In this work it was postulated that SAR11 genome heterogeneity, especially in the HVRs, may operate as balancing polymorphisms across the population.
*For references, please see the publication list at the Giovannoni Lab.
Theo Dreher and colleagues recently isolated a new freshwater cyanomyophage, S-CRM01 and sequenced its 178,563 bp circular genome. In spite of being a phage of a freshwater Synechococcus strain, gene content and phylogenomics link S-CRM01 to marine cyanomyophages. I was responsible for the phylogenomic characterization of S-CRM01, which was an interesting project. Because phage have such small genomes, and gene content is not necessarily conserved, phylogenomic analysis (with the automated pipeline Hal) allowed us a new means to characterize the relationship of this phage to others, and notice the effects of allowed missing data on the overall topology.
During the course of my graduate work in the Coates Lab, I did some research on oxidation of inorganic electron donors, chemolithotrophy, in the context of anaerobic respiration. I co-authored a study by a post-doc in the lab at the time, Karrie Webber, now a professor at University of Nebraska, Lincoln. She was looking at nitrate-dependent iron and uranium oxidation in a variety of strains, and I did some physiology and phylogeny in collaboration with her. We found that the perchlorate-reducing Magnetospirillum bellicus VDY, which I had isolated from the perchlorate-reducing bioelectrical reactor community, could also oxidize Uranium, as could many other strains (Weber et al. 2011). This research is significant because, while the total amount of Uranium oxidized by most strains was small, it was enough to mobilize this toxic metal to concentrations above that of regulatory limits. Given the ubiquity (determined by microcosm experiments) and phylogenetic diversity of the organisms surveyed that were capable of nitrate-dependent Uranium oxidation, this also has significant implications for strategies to immobilize Uranium using microbial reduction.
I also did research on the ability of strain VDY to oxidize iron, both in the presence of nitrate and perchlorate (Thrash et al. 2011). We knew from my previous characterization of the organism that it could oxidize hydrogen and express RuBisCO specifically under autotrophic conditions, so we postulated that it could oxidize other inorganic electron donors as well. Unfortunately, while the organism showed oxidation of iron relative to controls, it could not grow by this metabolism and at that stage of my dissertation, I was not able to follow-up further on where the activity might be coming from. We postulated that an abiotic reaction with the chlorite-dismuatse or chlorite itself may be the cause, but that was purely speculative.
See a review of this title by Jeffrey Gralnick for Microbe magazine.
My first project in graduate school was to try to apply electrochemistry to microbial perchlorate reduction. Perchlorate is a significant contaminant in water supplies around the country and is problematic because it competitively inhibits iodine uptake by the thyroid glad. Microorganisms can reduce perchlorate completely to harmless chloride, making bioremediation of the compound an attractive solution. Unfortunately, most strategies for stimulating dissimilatory perchlorate-reducing bacteria (DPRB) use chemical electron donor amendment that stimulates growth, as well as perchlorate reduction, leading to biofouling that can cause system failure in both in- and ex-situ operations. Work by another graduate student in the lab, Ian Van Trump, had shown that the electron shuttle, AQDS, could supply electrons to various DPRB for perchlorate reduction without stimulating growth. Thus, my experiments began using an electrochemical cell where the shuttle was reduced in the cathodic chamber which was inoculated with local water samples. Enrichment of DPRB became evident and I was able to eventually isolate two strains from the cathodic chamber. One of these organisms, strain VDY, was isolated from the electrode surface itself and was capable of functioning well as the inoculum source for flow-through reactors. Importantly, while the organism could exist and reduce perchlorate, it did not grow much, and therefore did not cause biofouling. Later characterization of the organism showed that it was most likely autotrophic and therefore could survive at low growth rates without an added carbon source.
