Research

Bacteria, archaea, and viruses play critical roles in all marine systems, but most of these organisms have not been incorporated into trait-based global biogeochemical models, such as Darwin, for a variety of reasons. As part of the Simons Collaboration on Computational Biogeochemical Modeling of Marine Ecosystems (CBIOMES), the Fuhrman Lab is working to develop information and tools to integrate diverse microorganisms into such models. This includes assessing the functional characteristics, community composition, and global distributions of most such microbes. In particular, we are using time-series analyses, laboratory experiments, and data mining to better describe the fundamental traits and worldwide distributions of microorganisms and the factors controlling them. We also are working with modelers to optimize how these microbes are incorporated into the models.

            Ongoing projects as part of CBIOMES include:

• · Using modern denoising algorithms and basin-scale meta-‘omics datasets to infer the spatial and temporal distribution of microbial taxa as exact amplicon sequence variants. This work will create a stable biogeographic database of global organismal distribution and abundance patterns, which will contribute essential data for modelling work (including “ground truthing”).
• · Using genomics and other information to develop a database of traits of a broad variety of marine microorganisms, including fundamental lifestyle, preferred nutrients and conditions, geographic distributions, interactions with other microbes, etc.
• · Using metatranscriptomics and metagenomics to infer taxon-specific indicators of growth and maximum growth rates. Resulting data will allow use of increasingly-available ‘omics data to assign growth traits of varied marine prokaryotes at different ocean locations and times, and integrate them into global biogeochemical models.
• · Conducting controlled experiments with natural seawater microbial communities to directly observe taxon-specific growth traits and to validate ‘omics-based growth metrics. This work will allow us to better use static meta-‘omics datasets to infer dynamic growth characteristics of natural microbial communities.

Since 2000, in conjunction with the Caron Lab , the Fuhrman Lab has collected and analyzed water at multiple depths from San Pedro Ocean Time Series (SPOT) as part of the USC Microbial Observatory. Using a combination of molecular tools including gene-specific sequencing (16S rRNA, 18S rRNA, myoviral g23), metagenomics, and transcriptomics, we continue to assess the microbial populations (viral, bacterial, archaeal, eukaryotes) in conjunction with environmental parameters and biological metrics (microscopic and flow cytometric bacteria and virus counts, secondary production rates from thymidine and leucine incorporation, etc.).

The monthly time series enables us to assess long-term patterns and networks of seasonality, co-occurrence, and diversity for all players in the microbial food web. These patterns allow us to use real-world variations of natural complex communities to address fundamental questions about the factors controlling microbial biomass, diversity, activities, and biogeochemical roles.

We are also assessing patterns at different temporal (weekly, daily, sub-daily) and spatial scales (regional to global). Daily time series have shown unexpectedly sudden shifts in microbial communities that were previously missed by more routine monthly or weekly sampling. Our recent work showed that the dominant phytoplankton during the spring bloom can change on a near-daily basis, much more rapidly than first thought, and can include multiple harmful algal bloom species in rapid succession. We also found that Euryarchaea can peak to ~40% of the prokaryote community, in other words these poorly-known microbes can briefly bloom to make up the dominant organism in the near-surface ocean.

 We assessed dynamics in relative abundance and potential activities of the entire cellular microbial community via automated samping - by sequencing rRNA genes and rRNA molecules (rRNA) of Bacteria, Archaea, and Eukaryota once to twice daily between March and May 2014 off Catalina. Ostreococcus, Braarudosphaera, Teleaulax, and Synechococcus dominated phytoplankton sequences (including chloroplasts) while SAR11, Sulfitobacter, and Fluviicola dominated non-phytoplankton Bacteria and Archaea. We observed short-lived increases of diatoms, mostly Pseudo-nitzschia and Chaetoceros, with quickly responding Bacteria and Archaea including Flavobacteriaceae (Polaribacter & Formosa), Roseovarius, and Euryarchaeota (MGII), notably the exact amplicon sequence variants we observed responding similarly to another diatom bloom nearby, 3 years prior. We observed correlations representing known interactions among abundant phytoplankton rRNA sequences, demonstrating the biogeochemical and ecological relevance of such interactions: (1) The kleptochloroplastidic ciliate Mesodinium 18S rRNA gene sequences and a single Teleaulax taxon (via 16S rRNA gene sequences) were correlated (Spearman r = 0.83) yet uncorrelated to a Teleaulax 18S rRNA gene OTU, or any other taxon (consistent with a kleptochloroplastidic or karyokleptic relationship) and (2) the photosynthetic coccolithophorid Braarudosphaera bigelowii and two strains of diazotrophic cyanobacterium UCYN-A were correlated and each taxon was also correlated to other taxa, including B. bigelowii to a verrucomicrobium and a dictyochophyte phytoplankter (all r > 0.8). We also reported strong correlations (r > 0.7) between various ciliates, bacteria, and phytoplankton, suggesting interactions via currently unknown mechanisms. These data reiterate the utility of high-frequency time series to show rapid microbial reactions to stimuli, and provide new information about in situ dynamics of previously recognized and hypothesized interactions.

We also analyze our data in the context of regional and global diversity patterns via our own sampling of three regional sites, which include SPOT plus a site in the Port of Los Angeles as well as a matched site near relatively pristine Catalina Island. In addition, we compare our data to those from major global sampling networks, such as the Earth Microbiome Project, and Ocean Sampling Day, both of which Dr. Fuhrman serves as a scientific advisor.

Selected Publications: Chow et al. 2012, Cram et al. 2015, Needham et al. 2013, Needham and Fuhrman 2016,

Needham et al 2018

Although not widely recognized, high quality microbial community composition analysis requires periodic calibration and checking, just like any chemical assay. Small differences in PCR protocols and analytical pipelines can introduce major changes in results. Using a mock community of clones that were generated (by David Needham in our lab) to represent microbes common in mesotrophic and oligotrophic marine environments, we assessed the accuracy and precision of our 16S rRNA gene sequencing pipeline. We have selected a primer pair and analytical pipeline particularly well suited to marine and other microbiome projects. Our 515-926 primer pair not only amplifies the ssu rRNA (16S or 18S) of the vast majority of known organisms in all three domains, but testing with our prokaryotic mock communities shows that the results are remarkably quantitative. When we compared the observed vs expected abundances from our 27-member mock community, a log-log plot shows the r² value is 0.95 (see figure to the left). This compares to an  r² ~0.5 for the popular 515-806 primers used by many labs.

Based on our results, we strongly encourage the use of a mock community for anyone using rRNA “tag sequencing,” and it is important to include the mock communities blindly within sets of samples (not run alone). Not only did we find that popular 515-806 primers poorly quantified ssu rRNA gene abundance, but it also poorly amplified members of the SAR11 cluster, the most abundant bacteria in the global surface ocean; we also found that our downstream analyses, especially clustering, were greatly informed by tracking expected vs. observed classifications and operational taxonomic unit (OTU)-generation of the mock community. Several “standard” aspects of popular analytical pipelines caused incorrect merging or splitting of mock community OTUs, but we found ways to avoid that.

Chloroplasts. Of significant note, the 16S sequences of chloroplasts in eukaryotic phytoplankton (except dinoflagellates, which have aberrant chloroplast genomes) provide a valuable assessment of their community composition, far less affected by copy number variations than more classic 18S rRNA gene sequence analysis.

For a set of recommendations on how to use the Fuhrman Lab’s mock communities and PCR primers , please see our Methods and Publications.

Select Publications: Parada et al. 2015, Needham and Fuhrman 2016, Walters et al. 2016, Yeh et al 2018

Bacterial and archaeal populations are shaped by resource availability as well as viral lysis and protist grazing. Bottom-up and top-down controls are ultimately equally important.  We use dilution experiments, time-series data, network analysis, and statistical tools to simultaneously assess top-down and bottom-up regulation on community-wide abundance and diversity.

Analysis of 5 years of monthly cyanobacterial community (Synechococcus and Prochlorococcus) data from SPOT at extremely high resolution, through a custom ITS high throughput sequencing method, showed that at the ecotype level the cyanobacterial communities are seasonal and appear controlled mostly by environmental conditions. However at the exact sequence (strain) level, communities are not predicted well by seasonal factors but much better relate to the myovirus community composition. We interpret this as meaning that the abiotic environment largely controls what general types of cyanobacteria are present at what abundances, but the strain-level diversity is more controlled by viral infection.

Select Publications: Chow et al. 2014, Cram et al. 2016

Ahlgren, N. A.,  J. Perelman,Yi-chun Yeh, and J.A. Fuhrman. 2019. Multi-year dynamics of fine-scale marine cyanobacterial populations are more strongly explained by phage interactions than abiotic, bottom-up factors. Envir Microbiol. in press

Viral lysis helps to control microbial populations, and viruses impact community composition and microbial evolution.  We are addressing several aspects of the interactions between viruses and cellular microorganisms. In recent years we transitioned from fingerprinting-based analysis of myoviruses via PCR of the g23 gene, to sequencing of that gene in high-throughput manner, to the most recent work on viral metagenomics to assess the entire viral community. We have used time series (daily, monthly) to statistically link viruses with hosts and to assess the fundamental nature of virus-host networks as the contrast with bacteria-grazer networks. We are using a combination of bioinformatics, statistical networks, cultures, time series data, and single amplified genomes (SAGs), to address several questions, including how host ranges differ between viruses and the extents that viruses impact different members of food webs.

We also work with Computational Biology colleagues (labs of Dr. Fengzhu Sun and Dr. Ting Chen) to use bioinformatics and word patterns to identify viral sequences, distinguish viral sequences from those in cellular organisms, and to match viruses with hosts.

Recently we have also been using metatranscriptomics from microbial cells to show active viral infection in field samples. An exciting result is that viral expression of psbA photosynthetic reaction center genes in cyanobacteria is often half or more represented by the viral version of the gene rather than the cells' own version. This indicates massive infection, and also that viruses-encoded genes are responsible for a significant fraction of photosynthesis.

Select Publications: Chow et al. 2012, Chow et al. 2014, Needham et al. 2013

Metatranscriptomics to show viral activity, Sieradzki et al 2019

A major issue in microbial ecology research is how to best cluster similarly-functioning related organisms together and how to split ones that are related but ecologically different. This is essentially trying to define ecological species. While microbiologists frequently use marker genes and apply standard cutoffs of similarity, e.g. 99% similar 16S rRNA (or, 97%, which we consider too coarse), it is not clear what levels are truly most appropriate for ecological research, and how it may differ depending on the environment and scientific questions at hand.
By taking advantage of the most recent sequencing capabilities, and combining time-series data and information from the literature and major databases, we are able to address questions relating to suitable genes for study and the level of resolution needed to relate diversity to ecological processes without excessive lumping or splitting. These are not expected to be the same for all questions and for all organisms. For example, we expect that resource utilization may be studied at a coarser sequence resolution than viral susceptibility.  This work uses both marker genes and metagenomes.

Reference: Needham et al 2017

We maintain several mixed enrichment cultures containing Marine Group I archaeal strains (Thaumarchaeota) that were enriched from SPOT, originally based on cultures and protocols generously shared by Professor Alyson Santoro. Isolates are instrumental in understanding fundamental processes.

We have fully sequenced one such culture, named Candidatus Nitrosomarinus catalina SPOT01, a novel strain that is less warm-temperature tolerant than other cultivated Thaumarchaeota. Using metagenomic recruitment, strain SPOT01 comprises a major portion of Thaumarchaeota (4–54%) in temperate Pacific waters. Its complete 1.36 Mbp genome possesses several distinguishing features: putative phosphorothioation (PT) DNA modification genes; a region containing probable viral genes; and putative urea utilization genes. The PT modification genes and an adjacent putative  restriction enzyme (RE) operon likely form a restriction modification (RM) system for defence from foreign DNA. PacBio sequencing showed >98% methylation at two motifs, and inferred PT guanine modification of 19% of possible TGCA sites. Metagenomic recruitment also reveals the putative virus region and PT modification and RE genes are present in 18–26%, 9–14% and <1.5% of natural populations at 150 m with 85% identity to strain SPOT01. The presence of multiple probable RM systems in a highly streamlined genome suggests a surprising importance for defence from foreign DNA for dilute populations that infrequently encounter viruses or other cells. This new strain provides new insights into the ecology, including viral interactions, of this important group of marine microbes

We are currently investigating archaeal global distributions and virus infectivity, among other aspects.

Reference: Ahlgren et al. 2017

To study how anthropogenic inputs affect the microbial community composition and function, we analyze samples from the Port of Los Angeles, with from the SPOT site  and also a nearshore site near Catalina Island as minimally impacted controls. Using a variety of tools, we are assessing the resistance of the naturally occurring community to pollutants like heavy metals and various aromatic hydrocarbons.

We also apply a cutting edge combination of Stable Isotope Probing (SIP) and high throughput sequencing to reveal which of the community members can metabolize pollutants and incorporate them into their biomass, directly linking diversity and function.

Our time-series samples offer unique opportunities to “go back in time” and re-analyze archived material using the most modern techniques. Using metagenomics and bioinformatics tools such as Anvi’o, we are addressing questions such as “How do genomic variants change over time?” in ways that can resolve these changes at the gene or allele level.

In concert with metagenomics, we are also mapping parallel metatranscriptomes to observe changing patterns of gene expression between adjacent sites and also over seasons. better understand what functions particular organisms are performing under different conditions.

Reference  Sieradzki et al 2019