Forest Rohwer

Coral reefs worldwide are in decline
The dramatic rise in incidences of coral disease over the last two decades has been instrumental in this process. We have hypothesized that most of these diseases are actually opportunistic infections instigated by anthropogenic stressors. Our research is focused around understanding the interactions between the microbial world and coral reefs, and how these systems change following perturbation.

The Coral Holobiont
Corals are host to a wide diversity of organisms, including endosymbiotic algae, protists, fungi, Bacteria, Archaea, and viruses. Together, these organisms make up the coral holobiont. In our lab, we are interested in understanding the physiological roles of these players in their interaction with the coral animal, and how this relates to coral reef health. Incidences of coral death and disease are highly correlated with human impact, and we propose that anthropogenic stresses induce microbes normally associated with the coral to become opportunistic pathogens. Alternatively, opportunistic or specific pathogens from the water column might attack the weakened coral. To differentiate between these possibilities, my lab has had to determine if healthy corals have characteristic microbiotas. To do this, we have employed a variety of techniques ranging from electron microscopy (e.g., Johnston and Rohwer 2007) to metagenomics (e.g., Wegley et al. 2007).

In order to look at the diversity and specificity of coral microbes, our lab used high-throughput sequencing of bacterial 16S rDNAs associated with three coral species. This culture-independent study of coral-associated Bacteria found 430 (mostly novel) bacterial species in 14 samples from 3 coral species. The coral-associated microbial communities were ecologically structured: different coral species had different bacterial communities, even when physically adjacent, while bacterial communities from the same coral species separated by time (~1 year) or space (3000 km) were similar. We also found that some bacterial species were present only in a subset of spatial niches within individual coral colonies (Rohwer, et al., 2002).

In order to look at the function of microbes on corals, we use metagenomic sequencing (454 Life Sciences) to identify the microbes and their functional genes. Our work found that bacteria associated with corals are primarily heterotrophic. Our metagenomic data showed an abundance of sugar and protein utilization and uptake pathways in the microbial community. These microbes are likely utilizing the complex polysaccharides and peptides from the coral mucus. Several types of cyanobacteria were also found associated with the coral, and may be providing fixed carbon and nitrogen to the coral. In addition, an abundance of fungi were associated with corals, including those involved in nitrogen cycling, indicating that fungi may be fixing nitrogen and making it available to members of the coral holobiont (Wegley et al. 2007).

We have also looked at the viruses associated with healthy and bleaching corals, and find viruses with a wide variety of hosts including many of the various members of the coral holobiont. These viruses include plant and algal viruses, herpes-like viruses, and cyano- and vibriophage, to name a few (Wegley et al. 2007, Marhaver et al. 2008). Due to the abundance of viruses and the wide variety of host ranges they possess, we expect that they play an important role in coral health and structuring of the coral holobiont.

In summary, the associations of the coral animal, prokaryotes, zooxanthellae, viruses, fungi, and other undefined components will define the niche that any coral colony occupies on a reef. This system is almost certainly exemplary of many other interactions between microbes and their higher eukaryotic hosts, and our studies will make predictions that can/will be tested in other complex host-microbial flora systems.

Stressors Alter Microbial Dynamics on Corals
An important implication of the coral holobiont model is that disrupting any one of these components may cause the whole community to collapse and lead to coral death. In order to test this hypothesis, we have performed several experiments exposing corals to different stressors and then looked at the changes in microbial dynamics and diversity, as well as coral pathology. In collaboration with Dr. Nancy Knowlton and Davey Kline at the Scripps Institution of Oceanography, we applied stresses to different coral species in the presence and absence of antibiotics. Our data showed that of the many commonly cited stressors of corals, organic carbon (OC) loading is the most problematic. Coral death induced by OC can be delayed with antibiotics. Additionally, OC loading causes the coral-associated microbial communities to grow much faster then normal. This strongly suggests that changes in the bacterial community, and not the stresses themselves, are responsible for coral mortality. (Kline et al. 2006, Kuntz et al. 2005). Additionally, when corals are placed next to algae with a filter impervious to viruses and bacteria, corals mortality is high. This mortality is also inhibited by antibiotics (Smith et al. 2006).

In a separate experiment, corals were exposed to one of four types of stressors currently threatening coral reefs: elevated nutrients, temperature, and organic carbon, and lowered pH. We then isolated the microbial and viral communities and performed whole-genome sequencing (pyrosequencing, 454 Life Sciences) to look at how the diversity and function of these organisms changed following stress. Our data showed that stress led to a shift towards a more pathogenic microbial community in all cases, with pathogen-associated genes also increasing in abundance (e.g. motility, virulence, and secondary metabolite genes) (Vega Thurber et al. Env Micro 2009). The viral assemblages also changed on the coral, with viruses related to the Herpesviridae family greatly increasing in abundance (Vega Thurber et al. PNAS 2008). We found that one herpes-like virus was undetectable by quantitative PCR (qPCR) prior to stress, but then increased dramatically within 1 hr of stress exposure, indicating an increase in production of the virus under stress.

Coral Reef Microbiology in Pristine and Human-impacted Reefs
In 2005, we visited the Northern Line Islands with a group of coral reef experts to look at coral reef health across a gradient of human disturbance. The islands ranged from uninhabited to serving as a home for over 9000 residents. Surveys found that uninhabited islands had high coral cover and fish biomass (Sandin et al. 2008). The microbial community on healthy reefs was evenly split between autotrophs and heterotrophs, while on Kiritimati, the most inhabited island, the microbial community was primarily heterotrophs. Microbes were also ~10 times more abundant on these inhabited reefs, while coral cover decreased and disease was much more prevalent (Dinsdale et al. 2008).

In April 2009, we participated in a cruise to the Southern Line Islands to characterize pristine coral reefs. The group included most of the collaborators from the NLI trip, as well as a group from National Geographic to photo-document the reefs (http://ocean.nationalgeographic.com/). The islands visited on this trip lie just south of the equator, and have been uninhabited for 100 years or more. As such, they are home to some of the last remaining pristine coral reefs on the planet. We are again characterizing microbial and viral communities at these islands.

We are also trying to use metabolic theory to link performance of individual organisms to the whole community ecology of coral reefs, using data collected across the Pacific. We are interested in how fishing-related alterations in trophic structure affect community-level energy use and biomass production. Allometric power laws provide the basis for assessing the relative importance of fish vs. microbial components across coral reefs of varying degrees of health. Metabolic theory predicts that rates of energy and nutrient use should be approximately equal for all size categories within a restricted taxonomic group. However, when both fish and microbes are considered, we want to know if the flow of energy and materials through coral reef ecosystems is dominated by small or large organisms (i.e., microbes or fish).
Collaborators: Dr. Jennifer Smith, Dr. Stuart Sandin