nai

Project Progress

1. Research group summaries

Banfield group

Iron concretions and biosignatures. This project is the Ph.D. research of Ms. Claudia Jones, and has been conducted in collaboration with Dr. Jochen Brocks, Australian National University. We have been studying iron concretions actively forming at Lake Tyrrell as analogs for those which formed at Meridiani Planum, Mars at a time when ascending groundwaters flowed across that planet’s surface and deposited ferric iron minerals. It is possible that oxide deposition on an early Mars may have been microbially-mediated; if so, preserved lipid biomarkers from those minerals would provide valuable clues about the nature of these organisms. However, it is far from certain that reduced carbon compounds could survive the oxidizing conditions of ferric oxide formation. Results from a modern, analogous environment could be used to test whether lipid biomarkers can survive oxic lithification, and focus the search for signs of past life on Mars.

Acidic, hypersaline seeps at Lake Tyrrell carry anoxic, iron-rich ground waters to the surface, where they are oxidized and result in ferricrete deposition. The ferricretes are composed of quartz-rich lake sediments cemented into meter-scale rocks by iron oxides (magnetite) and oxyhydroxides (goethite) (Fig. 1)

These ferricretes are underlain by typical acid-saline depositional facies, very similar to that encountered at Meridiani Planum on Mars by the rover Opportunity. To determine whether molecular markers for life can coexist with oxidized mineral deposits, we extracted samples of ferricrete and underlying sediments to establish the presence and provenance of free and bound lipid biomarkers.

In order to assess the effects of oxic lithification on redox sensitive lipids, we monitored the concentration of phytanol and its oxidation product, phytanic acid, in ferricretes and surrounding sediment samples. Results indicate that the jarosite-rich (KFe³⁺₃[(OH)₃SO₄]₂) sediment directly underlying the concretion is a poor matrix for lipid preservation: only small concentrations of phytanol were evident, and phytanic acid was below detection limits (Fig. 2).

By contrast, both the goethite-rich layer of the concretion and the reduced sulfide-rich sediments surrounding it showed greater concentrations of each compound (20x and 250x, respectively). Interestingly, the ratio of phytanol to phytanic acid is approximately equal within the oxic concretion and the reduced sediment, indicating that abiotic oxidation is not likely to be a relevant diagenetic pathway for phytanol in this system.

Differences in compound concentrations between samples demonstrate the differential preservation of lipids within the ferricrete and the underlying sediment. While the concentrations of lipids are ~10 times lower in ferricrete than in sediment, their presence indicates that biomarker molecules may survive the oxidizing conditions of ferricrete formation broadly analogous to those that existed on the Martian surface. It is possible that the acidity of jarosite destroys lipids more quickly than do iron oxides/oxyhydroxides; however, further samples are needed to test this hypothesis. We estimate Lake Tyrrell concretions to be tens to hundreds of years in age; therefore, to determine whether lipid biomarkers can survive in such formations over geologic time, we plan to examine progressively older, unmetamorphosed samples from the geologic record. If biomarkers can be extracted from formations millions to billions of years in age, it would seem reasonable to pursue the search for Martian biomarkers within hematite nodules, or within the iron-oxide rich sediments of Meridiani Planum.

Community genomics and 3-dimensional characterization of novel, uncultivated, ultra-small Archaea in pyrite-based microbial communities. The use of random shotgun genomic sequencing has been very successful in an extremely acidic mine at Iron Mountain, in northern California, due to the limited numbers of species in the communities. The first sample sequenced led to the discovery of a cluster of Archaea, named “ARMAN”, that eluded detection by PCR due to their unusual 16S rRNA genes (Baker et al., 2006, Science, 314,1933-1935). The ARMAN was shown to be at low abundance, but ubiquitous throughout the mine, and likely an important member of these pyrite-oxidizing biofilms.
Three dimensional cryo-electron tomographic characterization of the uncultivated ARMANs (Fig. 3)

revealed that they are right at the cell size predicted to be the lower limit for life, 0.009 μm³ to 0.04 μm³. Intriguingly, we were able to image several ARMAN cells under viral predation. Perhaps the most surprising finding from this study is that these cells only contain ~92 ribosomes per cell (compared to ca. 10,000 for E. coli in culture). This, and the highly organized internal structure (ribosomes positioned around the inside of the inner membrane), provide clues to the strategies of life at its lower size limits. Since the ARMAN are smaller than other inhabitants of this subsurface chemoautotrophic ecosystem, filtration of biofilms (using a <450 nm filters) resulted in enrichments of the ARMAN groups. DNA was extracted from the filtrate and subjected to multiple displacement amplification (MDA) in order to obtain sufficient quantities for a library construction for genomic sequencing. To obtain genomic information from other ARMAN groups, we removed sequencing reads belonging to the abundant bacteria (Leptospirillum groups II and III) from another small-insert library, UBA (ca. 117 Mb). Then we combined this dataset with sequences from an Archaea-rich sample (UBA BS, ca. 100 Mb) and assembled this. From these two assemblies we utilized emergent self organized mapping (ESOM) of tetranucleotide frequencies (Dick et al., in prep.) to reconstruct the genomes of three different ARMAN groups (2, 4 and 5). All three of the genomes are ca. 1Mb in size, smaller than any previously sequenced genome of a free-living microbe (Baker et al., in prep.). Nearly all genes for ribosomal structure and function were identified, yet many genes for key metabolic pathways could not be identified. This is likely due to the fact that the ARMAN groups are highly divergent from Archaea that have been previously sequenced. Of the 1033 annotated proteins in ARMAN-2, 687 (67%) belong to archaeal COGs (arCOGs). This is less than all other sequenced Archaea including Nanoarchaeum equitans (72%), with the exception of the symbiont Crenarchaeum symbiosum (58%). Thirteen proteins in ARMAN-2 had readily identifiable bacterial, but no archaeal orthologs. This number is greater than the 7 found in Candidatus Koryarchaeum cryptofilum which is considered one of the most deeply branched Archaea. These observations point to the high level of evolutionary divergence of this unusual extremophile.

Emerson group

Our group is investigating both natural populations and pure cultures of iron-oxidizing bacteria in an attempt to better understand how their physiology and ecology influences the mineralogy and geochemistry that are hallmarks of these organisms in the environment. One focus of our work has been to gain a better understanding the contribution that neutrophilic, oxygen-dependent iron-oxidizing bacteria (FeOB) make to iron oxidation kinetics both in-situ and in the laboratory. We have done this work in collaboration with George Luther’s group, and also have on-going collaborative work with the Roden group at Wisconsin.

This has been a transitional year for co-I Emerson, since he moved his lab from the ATCC in Virginia to the Bigelow Laboratory in Maine in late August, 2007. Significant time has been devoted to setting up a new laboratory and hiring new staff, since none of the ATCC staff were able to make the move. In spring 2008, the new lab was fully operational and a postdoc, Emily Fleming was hired to work on this project in May, 2008, in addition to a technician who is funded through another project.

Focus has continued to be on field related studies in both freshwater and the marine system at Loihi Seamount. A new freshwater field site has been established close to the laboratory in Maine that is dominated by mats of Fe-oxidizing bacteria, principally sheath-forming Leptothrix. We are currently in the process of developing cultivation-independent methods for determining the taxonomic relatedness of these organisms to other Fe-oxidizing bacteria, as well as assessing the abundance and dynamics in the field. We are also attempting to cultivate them. This site will also be used for comparative purposes to the Dragon Ridge site that was studied quite extensively in Virginia. We are in the process of doing high density phylogenetic analysis of the Dragon Ridge site using high throughput sequencing (454 technology) of these communities to compare with a small clone library comparing summer (August 2006) with fall (November 2006) at Dragon Ridge. The results of the latter clone library (Fig. 4)

show an interesting result: in summer, the mat is dominated by the sheaths of Leptothrix (similar to our Maine site), and is phylogenetically complex, with close to half the clones belonging to unclassified groups of bacteria, and relatively few beta-Proteobacteria, which is where the known freshwater Fe-oxidizing bacteria cluster. In November, the mat is dominated by Gallionella-type stalks, and the community is dominated (50%) by Betaproteobacteria, a number of which are close relatives to known Fe-oxidizers. Although temperature may be a factor, it is not clear what drives these shifts in community composition, or if a much more diverse group of Fe-oxidizing bacteria is present in the summer that we do not recognize through association with known phylotypes. We intend that more analysis of Dragon Ridge data, coupled with new data from the new site will allow us to develop a comprehensive understanding of the ecological dynamics of neutrophilic Fe-oxidizing bacterial communities.

Work by Clara Chan has focused on the ultrastructure and behavior of a unique Fe-oxidizing bacterium, Mariprofundus ferrooxydans, isolated from a deep-sea hydrothermal vent that had extensive mats of Fe-oxidizing bacteria. This organism produced an iron oxide-rich stalk, microscopy and spectroscopy work showed that as the cells grow, they excrete iron and organic-rich fibrils that make up the stalk, at a rate of ~2 microns/hr. Stalk growth appears to be parallel to the direction of Fe and oxygen gradients. We are following up on this work to determine if a unique type of mechano-taxis may be involved.

Other recently published work done in collaboration with George Luther at U. Delaware, establishes the kinetic effects that Fe-oxidizing bacteria have Fe-oxidation at circumneutral pH, both in the lab and in the field. This work establishes how well these bacteria are adapted to their specialized niche of growing at aerobic-anaerobic interfaces. It is important to point out that the paper by Druschel, et al, contains a supplementary figure, which is a movie made by Peter Suchecki, of Fe-oxidizing bacteria responding to gradients of Fe and oxygen. This time-lapse film is part of a larger E/PO project directed by Kevin Cuff and Herb Their on using information about these bacteria in middle and high school science experiments. This illustrates very nicely important points made in the manuscript, and is an example of how an E/PO project can help move scientific discoveries forward, as well as explain important scientific concepts to young students.

Luther group

Our group has been examining the in situ Fe and S geochemistry of several environments where microbially-driven oxidation of Fe(II) oxidation takes place, including sites where free sulfide [H2S and HS-] is not present (creeks in VA and Chocolate Pots, Yellowstone National Park) as well as sites where it is [local Delaware Inland Bays and the hydrothermal vents at Kilo-Moana, located on the East Lau Spreading Centre, in the Lau Basin, SW Pacific Ocean].

During the past year we have published four papers on in situ the rates of Fe(II) oxidation. Work by Trouwborst et al. (2007, Geochim. Cosmochim Acta, 71:4629-4643) described novel in situ techniques (Fig. 5)

for determination of the kinetics of Fe(II) oxidation in cyanobacterial mats of Yellowstone National Park. The results provide constraints on possible biological mechanisms for the generation of Banded Iron Formations on the ancient Earth. A second paper by Renz et al. (2007) (Environ. Sci. Technol., 41:6084-6089) describes collaborative work with the Emerson group on biotic vs. abiotic Fe(II) oxidation in microbial mats containing active populations of neutrophilic Fe(II)-oxidizing organisms. Another collaboration with Emerson’s group (Druschel et al., 2008, Geochim. Cosmochim Acta, 72:3358-3370) presented field measurements of Fe(II) and oxygen distributions in a neutral-pH creek, which together with experimental studies of the kinetics of Fe(II) oxidation in pure cultures demonstrated that neutrophilic Fe(II)-oxidizing bacteria occupy an ecological niche where oxygen concentrations are less than 50 μM. This is the first work to accurately describe the in situ environmental conditions where Fe(II)-oxidizing bacteria reside in nature. A fourth paper (Mullaugh et al., 2008, Electroanalysis 20:280-290.) presents additional in situ kinetic details on the oxidation of Fe(II) by cyanobacteria, and on potential for biogenic Fe(III) phases to oxidize hydrogen sulfide to thiosulfate at diffuse flow hydrothermal vents. Based on these collective studies, we can now distinguish between three Fe(II) oxidation processes using real time measurements: (1) strictly abiotic; (2) biotic mediated by cyanobacteria in the light; and (3) biotic mediated by lithotrophic Fe(II)-oxidizing bacteria. All processes yield Fe(III) (oxy)hydroxide solid phases and thus have implications for banded iron formations (BIF) and the types of soils found on Mars.

In addition to the above papers, we have continued to work on metal sulfide clusters as a source of reduced metals including Fe(II) that may be used by microorganisms instead of the aqueous Fe(II) cation (Druschel et al, 2008). Recent work on inert CdS clusters was presented at the American Society of Mass Spectrometry conference. Here both positive and negative ion mass spectra were obtained after introduction of gaseous hydrogen sulfide with aqueous Cd(II) solutions into a mixing cell or chamber at the inlet of the mass spectrometer (FT-MS with ICR cell). The positive ion data show that water molecules are still attached to the metal in the gas phase so that the ICR cell provides solution like reaction coordinate data. The negative ion spectra show that sulfide displaces water molecules and that clusters with up to four cadmium and sulfide atoms form. These data clearly show that clusters form before nanoparticles.

Roden group

Groundwater Fe seep. Microbial redox cycling of iron (Fe) was studied in a circumneutral pH groundwater seep in north central Alabama (Fig. 6A).

Incubation of freshly collected seep material under anoxic conditions revealed the potential for rapid Fe(III) oxide reduction. Culture-based enumerations revealed significant numbers of organic carbon and hydrogen-oxidizing dissimilatory iron-reducing microorganisms. Three isolates with the ability to reduce Fe(III) oxides by dissimilatory or fermentative metabolism were obtained. MPN analysis also revealed the presence of abundant microaerophillic Fe(II)-oxidizing microorganisms. A 16S rRNA gene clone library was dominated by representatives of the Betaproteobacteria including Gallionella, Lepthotrix and Comamonas species. More detailed molecular analysis of the seep microbial community (pyrosequencing of thousands of 16S rRNA genes) is underway through the DDF project (see Cross-team collaborations section below). Collectively our results suggest that active microbial Fe redox cycling takes place within this habitat, and support previous conceptual models for how coupled microbial Fe oxidation and reduction can accelerate Fe turnover in surface and subsurface sedimentary environments. Analogous studies are planned (Fall 2008) for a groundwater Fe seep in Bloomington, IL which is currently under investigation through a NASA Exobiology project awarded to Juergen Schieber and Flynn Picardal at Indiana University. These DDF-supported studies will include in situ geochemical, mineralogical, culture-based microbiological, and molecular biological analysis of Fe redox cycling phenomena.

Volcanic tuff. The Pliocene-aged, highly-weathered volcanic tuff unit in Box Canyon, ID (Fig. 7A,B)

is a potential analog to ancient weathering environments on Mars that have been identified by the recent rover and orbiter missions. The unit contains substantial quantities of magnetic mineral phases, which are comprised primarily of nanometer-sized maghemite crystallites (Fig. 7C) in close association with smectitic clay particles. These phases are analogous to the pedogenic magnetic components of soils. Although some of the larger maghemite crystals contain substantial titanium (Ti), indicative of a high-temperature lithogenic (abiotic) origin, a sub-population of the crystals is free of Ti and therefore may be of biogenic origin. The mineralogical analyses reported here were part of the M.S. thesis work of Emily Freeman at the University of Wisconsin. Culture-based studies indicate that dissimilatory Fe(III)-reducing microorganisms are present in the weathered basalt, and that these organisms are capable of converting the reddish-brown maghemite crystals into black magnetite nanoparticles (Fig. 7D,E). Previous studies have attributed the presence of maghemite (and magnetite) in soils to the reaction of microbially-produced Fe(II) with amorphous Fe(III) oxide phases such as ferrihydrite. Our findings suggest that this process was responsible for the production of the very small, Ti-free maghemite crystallites present in the volcanic tuff. Thus it appears that these weathered basalts contain extant mineralogical biosignatures of microbial Fe redox processes. Analyses of organic (lipid) markers associated with the magnetite mineral phases, along with a detailed pyrosequencing-based census of the extant Bacterial communities, are underway through the DDF collaboration.

Lake Tyrell. We have conducted a culture- and molecular-based analysis of microbial communities across biogeochemical gradients in this hypersaline environment, with a focus on organisms associated with iron and sulfur redox metabolism. Sediment geochemical survey results from 2006 indicated an microbial iron oxidation-based system in acidic (pH ~ 3) sediments near the southern end of the lake (site A), and a sulfur-cycling based system in circumneutral (pH ca. 7) sediments near central part of the lake (Site F). General aerobic most probable number (MPN) counts revealed pH- and salinity-adapted microbiota at both sites, with highest numbers found at pH 5.5 for site A and pH 7 for site F. Highest cells counts were obtained at salt (NaCl) concentrations between 10% and 15%. Additional MPN and molecular data suggest a complete microbial driven sulfur cycle at the neutral pH site. Enrichment cultures of hydrogen utilizing sulfate reducers were obtained, and sulfur-disproportionating, acetate/lactate utilizing, and aerobic sulfide oxidizing enrichments are in progress. Other studies examined the potential for biomass produced by microbial mats in the lake to serve as the main carbon and energy source for Fe(III) and sulfate reduction. Stable enrichments of acidic pH Fe(III) reducers and neutral pH sulfate reducers were established. Acidic Fe(II) oxidizers were found only in acidic sediments, and from the highest positive MPN tubes a stable co-culture of Thiobacillus prosperus and Marinobacter was obtained. Detailed pyrosequencing-based 16S rRNA gene analyses of bacterial (but not archaeal) communities at the acidic and neutral sediments are underway through the DDF collaboration. Initial results for the acidic sediments at site A (Fig. 8)

indicate a highly simplified bacterial community dominated by Gammaproteobacteria, including high numbers (ca. 19,000 out of ca. 48,000 total reads) of sequences related to known lithotrophic S- and Fe-oxidizing organisms from the family Ectothiorhodospiraceae.

Keswick Reservoir. Iron (Fe) inventories and isotopic compositions were documented for sediments in the Spring Creek Arm of the Keswick Reservoir (SKACR) downstream of the Iron Mountain acid mine drainage site in northern California, USA (Fig. 9A,B).

The high concentration of Fe(III) oxides compared to sulfate in SCAKR sediments (Fig. 9C) allows microorganisms that respire by Fe(III) reduction to outcompete those which respire by sulfate reduction, as indicated by the absence of acid volatile sulfides and the very low abundance of Cr(II)-extractable reduced inorganic sulfur (pyrite and/or elemental S) compared to dilute HCl-extractable Fe(II) (Fig. 9D). The SCAKR sediments are therefore a better analog for ancient (Archean and early Proterozoic, ca. 2-3 billion year old) Banded Iron Formations (BIFs) compared to modern marine sediments. Our results demonstrate production of large (millimolar) quantities of isotopically light (compared to bulk HCl- and HF-extractable Fe pools) aqueous Fe(II) by microbial Fe(III) oxide reduction (Fig. 9E), and suggest pathways whereby microbial Fe(III) reduction could have led to the formation of isotopically light Fe minerals in BIFs.

Nitrate-dependent Fe(II)-oxidizing enrichment culture. The composition and function of the lithoautotrophic Fe(II)-oxidizing, nitrate-reducing enrichment culture described by Straub et al. (1996) (Appl. Environ. Microbiol., 62:1458-1460) was studied by a combination of molecular and cultivation techniques. This culture was shown to be capable of sustained autotrophic growth coupled to oxidation of ferrous iron in the insoluble phyllosilicate mineral biotite (Fig. 10).

A 16S rRNA gene clone library (60 clones) obtained with universal bacteria primers revealed sequences related to Comamonas badi, Rhodanobacter thiooxidans, Parvibaculum lavamentivorans and Gallionella. The library was dominated (45 clones) by one phylotype with 94% similarity to Gallionella ferruginea. Pure cultures of Comamonas, Rhodanobacter, and Parvibaculum were obtained from with acetate and nitrate and showed 100% sequence similarity to the corresponding sequences from the clone libraries.. Neither of these strains alone or in combination were able to oxidize Fe(II) with nitrate under autotrophic conditions. Only in the presence of an additional carbon source (0.25mM acetate) Rhodanobacter and Parvibaculum oxidized up to 20% of the Fe(II) (10 mM) and a green rust is formed. A microorganism capable of oxidizing Fe(II) with nitrate (presumably Gallionella) was isolated using roll tube methods. This isolate showed low Fe(II)-oxidizing activity in the first transfer to liquid culture (only 10-20% of the added 10 mM Fe(II) was oxidized after 8 weeks), and a stable Fe(II)-oxidizing isolate was not obtained. Experiments examining the patterns of growth of the three organisms during nitrate-dependent Fe(II) oxidation by the enrichment culture are under way.

Director’s Discretionary Fund (DDF) Project

This interdisciplinary project (see Cross Team Collaborations) is examining mineralogical and organic signatures of life in three different Fe redox-based microbial ecosystems on Earth in order to provide priority targets and guide analytical tactics for life detection in future missions to Mars. We are evaluating in detail the biogeochemistry of field sites that represent a range of scenarios in which Fe-based microbial ecosystems may produce identifiable biosignatures analogous to those to be sought on Mars: acidic Fe-rich sediments and concretions in Lake Tyrell, Victoria, Australia; partially-saturated, weathered Pliocene-age basalts in Box Canyon, Idaho; and neutral pH groundwater Fe springs in Alabama, Virginia, and Indiana. The fundamental goal is to understand how mineral phases formed in these environments may preserve (either directly as unique mineralogical alterations and/or fossils, or indirectly in the sequestering of organic compounds) indicators that life is or was present in a given habitat.

This project started in the fall of 2007, and to date we have conducted field trips to make in situ measurements and collect samples from Box Canyon (September 2007) and Lake Tyrell (August 2008). We have also extended (by way of pyrosequencing of Bacterial 16S rRNA genes) previous studies of neutral pH groundwater Fe seeps in Alabama and Virginia, and are making plans for a comprehensive field campaign at a groundwater Fe seep in Indiana in September 2008. The latter system is the focus of a current NASA Exobiology project (“Investigating Morphological and Isotopic Biosignatures of Terrestrial Iron Bacteria – A Potential Mars Analog”) led by Juergen Schieber at Indiana University. Our most recent DDF-supported studies of materials from Box Canyon (see above) has confirmed previous findings that dissimilatory Fe(III)-reducing bacteria (both indigenous populations and other pure-culture strains) are capable of converting putative biogenic nanophase maghemite back to magnetite crystallites from which they were presumably derived. Detailed studies of the microbiological and mineralogical composition of materials from Lake Tyrell and the Indiana Fe seep will be conducted in the near future. These studies will include a combination of (1) targeted culture-based work as illustrated by the findings reported above for the Alabama groundwater Fe seep, Box Canyon and Lake Tyrell field sites; (2) organic biomarker analyses similar to those are already underway (see above) with the Box Canyon (collaboration with Jennifer Eigenbrode from NASA Goddard and Marilyn Fogel of the Carnegie Institution of Washington) and Lake Tyrell (collaboration with UCB Ph.D. student Claudia Jones and Dr. Jochen Brooks of Australian National University) materials; and (3) pyrosequencing of large numbers (ca. 50,000 reads per sample) of Bacterial 16S rRNA genes (collaboration with Mitchell Sogin at Marine Biological Laboratory) using procedures that have recently (summer 2008) been applied to DNA extracts from nine previously-collected samples from Fe-based microbial ecosystems. An example of the data generated by pyrosequencing is given in Fig. 8 for a sample of surface sediment from Lake Tyrell. Other samples that were recently analyzed include: four samples from Box Canyon (two collected in 2005 and two collected in 2007); two samples from the Dragon Ridge groundwater Fe seep in Virginia (collected in 2006); one sample from the groundwater Fe seep in Alabama (collected in 2006); and one additional sample from Lake Tyrell (collected in 2007). Co-Is Roden and Emerson, together with postdoctoral researchers Marco Blothe and Evgenya Shelobolina, are currently developing strategies for data reduction and presentation of the pyrosequencing results.