Genomes to Life Contractor-Grantee Workshop III
February 6-9, 2005, Washington, D.C.
Genomics:GTL Program Projects
Lawrence Berkeley National Laboratory
6
VIMSS Applied Environmental Microbiology Core Research on Stress Response Pathways in Metal-Reducers
Terry C. Hazen*1 (tchazen@lbl.gov), Carl Abulencia3, Gary Andersen1, Sharon Borglin1, Eoin Brodie1, Steve van Dien5, Matthew Fields6, Jil Geller1, Hoi-Ying Holman1, Rick Huang1, Janet Jacobsen1, Dominique Joyner1, Martin Keller3, Aindrila Mukhopadhyay1, David Stahl5, Sergey Stolyar5, Jun Sun3, Dorothea Thompson2, Judy Wall4, Denise Wyborski3, Huei-che Yen4, Grant Zane4, Jizhong Zhou2, and Beto Zuniga5
1Lawrence Berkeley National Laboratory, Berkeley, CA; 2Oak Ridge National Laboratory, Oak Ridge, TN; 3Diversa, Inc., San Diego, CA; 4University of Missouri, Columbia, MO; 5University of Washington, Seattle, WA; and 6Miami University, Oxford, OH
Field Studies
Identification of Different Relationships Between Contaminated Groundwater Samples Based Upon Geochemical Data or Multiple Gene Sequences from Microbial Communities. Factor analysis was used to identify a subset of variables that may explain a majority of the observed variance between the contaminated groundwater sites, and principal components analyses were then used to compare the sites based upon geochemistry, phylogenetic markers (n=353), and functional markers (n=432). The clonal libraries of the multiple genes (SSU rRNA gene, nirK, nirS, amoA, pmoA, and dsrAB) were constructed from groundwater samples (n=6) that varied in degrees of contamination. When geochemical characteristics were analyzed, the data suggested that the samples could be differentiated based upon pH, nitrate, sulfate, nickel, aluminum, and uranium. Similar relationships between the sites were observed when 107 analytes were used, but more resolution was achieved between the more contaminated sites. In addition, a majority of the variance between the acidic samples could be accounted for by tetrachloroethene, 99Tc, SO4, Al, Th and 1,1,2-trichloro-1,2,2-trifluoroethane. The analysis based on a phylogenetic marker resulted in different groupings for background and the two circumneutral sites compared to the geochemical analysis, and analyses of the OTU distributions for the functional genes each predicted different relationships between the sites. A tripartite PCA explained 76% of the variance and grouped the background sample with the three, heavily contaminated sites. When all gene OTUs were used in the analyses, the sites were more similar than in any other comparison, 94% of the observed variance cold be explained, the background site was grouped with the contaminated sites, and possible key populations were identified by factor analysis. The data suggested that even though the background site was phylogenetically and geochemically distinct from the acidic sites, the extreme conditions of the acidic samples might be more analogous to the limited-nutrient conditions of the background site.
Biopanning/Clone libraries. Diversa extracted high molecular weight DNA from organisms present in contaminated soil sediment samples using a method that preserves the integrity of the DNA. Because the number of organisms in these samples was low, the genomic DNA was amplified using a phage polymerase amplification system. 16S rRNA analysis was then used to examine the microbial diversity of the samples. The amplified DNA was also used in the construction of large and small insert DNA libraries. These libraries were then screened for the presence of histidine kinase genes with homology to a subfamily of Desulfovibrio vulgaris histidine kinases. Genomic DNA has been extracted and amplified from nine different sites at the NABIR field research center. 16S rRNA analysis revealed the presence of distinct bacterial phyla, including proteobacteria, acidobacteria, and planctomycetes. Small and large insert libraries were constructed for all samples and examined for clonal diversity. Plaque hybridization of these libraries to histidine kinase homologous probes resulted in multiple positive clones. These clones will be compared and used to develop a better understanding of cellular responses to different environmental factors. These experiments have furthered the understanding of how the biological organisms in a contaminated system are organized, regulated and linked.
Enrichments. Nine D. vulgaris-like bacteria (DP1-9) were isolated from a metal impacted field site (Lake DePue, Illinois) as an additional reference set for comparative stress analyses. All had identical 16S rRNA and dsrAB genes that were virtually identical to the orthologous genes of D. vulgaris Hildenborough (DvH). However, pulse field gel electrophoretic analysis of I-CeuI digests identified a large deletion in the genomes of all isolates. Complementary whole-genome microarray hybridization revealed that approximately 300 deleted genes were distributed in six regions of the chromosome, annotated as conserved/ hypothetical or phage related genes in DvH. These deletions were also confirmed by PCR analysis, using primers complementary to regions flanking the deletions. Continuing collaboration with Judy Wall (U Missouri) has shown that one of the “phage-deficient” D. vulgaris strains (DP4) serves as host for latent viruses of D. vulgaris Hildenborough, identifying two phage morphotypes by EM. MPN enrichments from FRC area 2 sediments were developed using a PIPES buffered B2 medium supplemented with: 1) lactate, 2) lactate plus ethanol, 3) acetate, 4) propionate 5) pyruvate or 6) hydrogen plus carbon dioxide. All showed sulfate reducing activity within a range of 10-1 to 10-4 dilutions. Thirty isolates from the lactate medium were shown by 16S rRNA sequence to be affiliated with the “Firmicutes”. A Gram-negative sulfate reducer (curved-rod morphology) maintained on an H2/CO2 plus acetate medium was also isolated.
Dual culture systems. The kinetics and stoichiometry of syntrophic growth were determined in batch culture by quantifying each population, substrate consumption (lactate), evolution of metabolic intermediates (H2 and acetate), and end-product accumulation (CO2 and methane). D. vulgaris monocultures were grown at generation times comparable to syntrophic batch cultures (24 and 36 hours) in sulfate-limited chemostats for comparative transcription analyses. Fermentative growth D. vulgaris on a lactate medium (sulfate minus) with continuous headspace purging was also developed for comparison. Transcription analyses of co-cultures identified a preliminary set of D. vulgaris genes either up or down regulated with syntrophic association, including periplasmic and cytoplasmic hydrogenases. These analyses are now being replicated at ORNL. A metabolic stoichiometric model was constructed using flux balance analysis (FBA) to complement and direct experimental studies on the physiology of D. vulgaris growing either alone or in co-culture. The network for each organism was based primarily on the annotated genome sequences, supplemented by available biochemical knowledge. The Desulfovibrio model consists of 86 reactions and 73 internal metabolites, while that of the methanogen contains 84 reactions and 72 metabolites.
Stress Experiments
High Throughput Biomass Production. Producing large quantities of high quality and defensibly reproducible cells that have been exposed to specific environmental stressors is critical to high throughput and concomitant analyses using transcriptomics, proteomics, metabolomics, and lipidomics. Culture of D. vulgaris is made even more difficult because it is an obligate anaerobe and sulfate reducer. For the past two years, our Genomics:GTL VIMSS project has developed defined media, stock culture handling, scale-up protocols, bioreactors, and cell harvesting protocols to maximize throughput for simultaneous sampling for lipidomics, transcriptomics, proteomics, and metabolomics. All cells for every experiment, for every analysis are within two subcultures of the original ATCC culture of D. vulgaris. In the past two years we have produced biomass for 38 integrated experiments (oxygen, NaCl, NO3, NO2, heat shock, cold shock, pH) each with as much as 30 liters of mid-log phase cells (3 x 108 cells/ml). In addition, more than 40 adhoc experiments for supportive studies have been done each with 1-6 liters of culture. All cultures, all media components, all protocols, all analyses, all instruments, and all shipping records are completely documented using QA/QC level 1 for every experiment and made available to all investigators on the VIMSS Biofiles database (http://vimss.lbl.gov/perl/biofiles). To determine the optimal growth conditions and determine the minimum inhibitory concentration (MIC) of different stressors we adapted plate reader technology using Biolog and Omnilog readers using anaerobic bags and sealed plates. Since each well of the 96-well plate produces an automated growth curve, over more than 200 h, this has enabled us to do more than 4,000 growth curves over the last two years. Since the Omnilog can monitor 50 plates at a time, this allows us to do more than 5,000 growth curves in a year. We have also developed chemostat techniques using a specially made extremophile fermentor (FairManTech) that has no internal metal parts. With this system we can get D. vulgaris to steady state from the freezer in less than 80 h in turbidostat mode, with a dilution rate of 0.25 l/h. Each reactor has a useable volume of 3 liters, with our current two reactors this enables production of 6 liters of steady state culture twice a week. We have also developed new harvesting techniques to minimize the stress caused by sample preparation for shipping. Since the volumes being centrifuged are large, the cells were not cooling fast enough to ensure high quality samples, so we devised a sampling apparatus that draws the cells from the culture vessel through capillary tubing in a MgCl ice bath that lowers the sample to 4°C in less than 20 sec. These procedures have maximized our reproducibility and throughput for the 8 labs involved.
Phenotypic Responses. Phenotypic Microarray™ analysis is a recently developed analytical tool to determine the phenotype of an organism. The plates, which are commercially available from Biolog™ (Hayward, CA), consist of 20 96-well plates. The first eight plates test a variety of metabolic agents, including electron donors, acceptors, and amino acids. Plates 9 and 10 cover a pH and osmotic stressors, while plates 11-20 contain a variety of inhibitors, including toxic agents and antibiotics. We have developed the ability use these plates under anaerobic conditions by inoculating plates in an anaerobic chamber, and heat-sealing them in polyethylene bags containing an anaerobic sachet. Using this technique, anaerobic conditions were maintained for up to a week. It was found that preconditioning of the cells in specialized media was required for the different types of plates in order to get a valid phenotype. The plates have been successfully used to characterize the phenotype of the D. vulgaris ATCC strain and are currently being applied to mutant strains to provide rapid screening of mutant phenotypic changes, for rapid pathway analyses and modeling. See (https://vimss.lbl.gov/~jsjacobsen/cgi-bin/Test/HazenLab/Omnilog/home.cgi) for sample data sets and analyses.
Synchrotron FTIR Spectromicroscopy for Real-Time Stress Analysis. LBNL’s newly developed synchrotron radiation-based (SR) Fourier-transform infrared (FTIR) spectromicroscopy beamline allows the study of many biochemical and biophysical phenomena non-invasively as the processes are happening. It has been enabling for our high throughput determinations of optimal sampling times stress experiments. We can observe real-time changes of major biomolecule pools within cells as they are exposed to different stressors. This allows us to pick optimal sampling points during the stress response for transcriptome, protemome, meatabolome, and lipidome analyses. It also allows us to verify the purity and state of the culture for QA/QC.
* Presenting author
