Presentation on theme: "Metagenomics and biogeochemistry"— Presentation transcript:
1 Metagenomics and biogeochemistry How do microorganism-driven geochemical cycles affect structure and function of ecosystems?How do we assess structure and function of ecosystems?How about starting by relating microbial assemblage composition to biogeochemical parameters and functions?Can we find predictable relationships? Patterns and scales of variability?Is metagenomics (e.g. shotgun or large-insert libraries) the best way to assess microbial assemblage composition for such studies?Are there faster and cheaper ways that permit analysis of many samples?
2 Amplified Ribosomal Intergenic Spacer Analysis (ARISA) For microbial community fingerprints with high phylogenetic resolutionStart with DNA extracted from a mixed community.PCR spans rRNA operon, 16S to 23S genes. One tagged primer.PCR primersFluorochrome16S rRNA gene23S rRNA geneIntergenic SpacerVariable LengthPCRFragment SizeFluorescenceFragment analysis. Smallest detectable peak ~0.1% of totalShows exact sizes. Each peak = “Operational Taxonomic Unit.”Data based, not gel-based Ref: Fisher and Triplett 199916S-23S clone libraries to identify most peaks:Brown, Hewson, Schwalbach & Fuhrman, Envir. Microbiol 2005
3 16S-ITS Clone Library permits ID from ARISA. Example:USC Microbial Observatory512 clones cover 94% of ARISA peaksBrown et al. 2005, Envir Microbiol.
4 Quantitation from PCR-based Fingerprinting? Real comparison: Prochlorococcus, ARISA vs flow cytometry countsSan Pedro Ocean TimeSeries 4 year datasetNote: we use a highly standardized assay, with eukaryotes removed, and measured amounts of DNAR2=0.86Flow cytometric counts% area from ARISAFingerprint % area is remarkably proportional to counts.Also, SAR11 % clones are close to % cells.Brown, Hewson, Schwalbach & Fuhrman, Envir. Microbiol 2005
5 7 samples from each of 2 North Pacific Gyre Stations Replicate 20L samples have very similar ARISA fingerprints7 samples from each of 2 North Pacific Gyre StationsCompares OTU proportions OTU Presence/absence onlyHewson et al. Aquat Microb Ecol 2006
6 What is an ARISA OTU? Phylogenetic resolution is about 98% 16S rRNA similarity - comparable to “species” levelEasily determined differenceBrown et al, Env Microbiol 2005
7 Near-surface SAR11 subclades as determined by ITS sequences and lengths
8 Temporal Variability in Bacterioplankton Communities How fast do communities change?San Pedro Ocean Time SeriesUSC Microbial ObservatoryMeasured Microbial and Oceanographic properties monthly since 2000, at depths to 880 mAlso, daily measurements near USC Wrigley Marine Science Center on Catalina - open water accessible daily by small boatFollow taxa by ARISA to look for temporal diversity patterns45 km
9 Relative stability over days at one location (open water, Catalina) Abundant taxa vary littleGraphs: all OTU over 6 daysgSAR 11ActinobactdateRarer taxa can vary moreNot just “noise” inmeasurementagProchlorococcusRarest detectable taxa.CFBgSAR 11
10 Monthly observations at SPOTS over 4 years showed some taxa clearly had repeatable seasonal patterns.How about the bacterial community in general?Brown et al. 2005
11 Predictable Annual Bacterial Community Reassembly Fuhrman et al., PNAS with Shahid Naeem171 taxa followed by ARISA over 4.5 yearsDFA scores reflect quantitative distribution of taxa via ARISA
12 DFA showed some subsets of bacterial taxa could predict the month of sampling with 100% accuracy. Multiple Regression with environmental parameters was highly significant (r2 ~0.7)– implies predictability of bacterial communities – even in an open marine system. Different subsets of taxa were predictable from different parameters – implies niches.Highly repeatable and predictable patterns imply little functional redundancy, contrary to common expectation for bacteria. This refers to combinations of functions in a particular taxon.Note- Not all taxa were included in the predictable subsets, but most were.Significant Parameters in MRAtemperature, salinity, nitrite, nitrate, silicate, oxygen, bacterial and viral abundances, bacterial production via leucine and thymidine incorporation, chlorophyll, phaeopigments ARISA richness
13 The taxa that had significant multiple regression coefficients were affected by different parameters – many controlling factors, and different taxa controlled differently (niches).
14 Biogeography on a Global Scale SeaWiFSGlobal survey of bacterioplankton at numerous sites in 3 ocean basins, under Arctic ice cap, and near Antarctica
15 “Things change” Global Diversity Measurements via ARISA Assemblages clearly vary“Things change”Weddell SeaSingaporeCatalina IslandGreat Barrier ReefLong Island NYNorwegian SeaSuva Harbor, FijiVillefranche (Med)BarbadosGerlache Strait, AntarcticaDeception Is, AntarcticaCoral SeaArctic OceanNew CaledoniaPhilippines
16 Bacterioplankton Biogeography LATITUDINAL GRADIENT OF RICHNESS ARISA measured the same way from 78 samples collected in all seasons and both hemispheres over 10 years (opportunistic sampling)Diversity generally highest at low latitudes, lowest in polar environments – like animals and plants (in every general biology textbook)Contrasts sharply with results reported for protistsp<0.005Highly significant(p<0.005) as linear regression, rank correlation, or with potential outliers removed
17 Regional Diversity Patterns Bacterial Community Similarity (via ARISA) vs DistanceNEAR-SURFACE samples“Mixing” curve between Pacific and Indian Basins?Hewson et al 2006 Mar. Ecol. Prog. Ser.
18 Deep-Sea (500-3000 m depth) patterns differ with locations and depth. Cause(s) unknownNorth Atlantic 1000m depth samples were in vicinity of Amazon PlumePacific*Pacific*Hewson et al Limnol. Oceanogr.
19 Example - What does proteorhodopsin do? Go beyond just observing nature - EXPERIMENTATIONExample - What does proteorhodopsin do?Does it provide much energy, and help microbial growth, as many assume? Genomics alone can’t answer.Schwalbach et al. (2005 Aquat. Microb. Ecol. 39: 235 ) did light/dark experiments with oceanic plankton.Water collected from oligotrophic and mesotrophic Pacific Ocean locations, collected and stored in natural light or total darkness for 5-10 days.Bacterial assemblages monitored by the ARISA whole-community fingerprinting approach
20 Database of ARISA OTU Identities EXPERIMENTAL TEST of Significance of Phototrophy.Light Removal Experiments – focus on Bacterial Groups that are supposed to have ProteorhodopsinIncubate bacteria in Light or Dark for 5-10 daysP1 P2110kmLight14:10hrcycleDark24hrDAPICell AbundancesMonitored over timeMesocosms(2x20L)Collect CellsAfter 5-10 daysP3DNA ExtractionBacterial Community CompositionITS Clone Library Construction16sITS23srDNAPCRPCRrDNA16sITS23sClone & SequenceDNAABI 377XLARISADelineate 98% 16s rDNA16S-ITS-23SABI 377XLDatabase of ARISA OTU IdentitiesAutomated Ribosomal Intergenic Spacer Analysis
21 Light Removal Experiments, 5-10 days darkness Schwalbach et al Aquat Microb Ecol 2005Light Removal Experiments, 5-10 days darknessMagnitude of change(n-fold difference)Histogram summarizing magnitude of change in individual taxa, light vs dark treatmentsMost taxa were NOT affected by light removal#ofOTUDark preferenceLight preferenceCyano/PlastidsSar11Sar86CFBRoseobacterSar116Sar406ActinobacterFibrobacterMarinobacterVerrucomicrobiaCyanobacteria & Phytoplankton exhibited consistent preference for light treatmentsMixed Responses, mostly dark preference, in ALL OTHER “phototrophic” groups (e.g. SAR11, SAR86, CFB, Roseobacter)Number of taxa displaying response, ALL experiments
22 Conclusions of Schwalbach et al (2005) : Most taxa (including presumed PR-containing and bacteriochlorophyll a – containing groups) do not decline significantly in extended darkness, unlike cyanobacteria.In fact, most bacterial groups did no differently or much better in extended darkness than in normal light.Suggests no clear direct benefit from light for most organisms.But some organisms do benefit.
23 Pelagibacter, in SAR11 cluster Even the one pure culture that contains proteorhodopsin grows no better in the light than in the darkPelagibacter, in SAR11 clusterOni et al Nature 2005“The Pelagibacter proteorhodopsin functions as a light-dependent proton pump. The gene is expressed by cells grown in either diurnal light or in darkness, and there is no difference between the growth rates or cell yields of cultures grown in light or darkness.”Giovannoni et al. Nature 2005
24 NSF, esp. Microbial Observatories Program AcknowledgementsNSF, esp. Microbial Observatories ProgramUSC Wrigley InstituteDave CaronMark BrownIan HewsonMike SchwalbachJosh SteeleAnand PatelShahid NaeemTony MichaelsDoug CaponeXimena HernandezR/V Kilo MoanaR/V SeawatchAjit SubramaniamBurt Jones
25 Other IssuesQuantitation from Environmental Genomic DataAccurate prediction of biogeochemical (or any other) function from genes. “Genome Rot,” Multifunctional genes, e.g. generic reductases. More important with slow-growing organisms and “streamlined” genomes?
26 Quantitation Issues/Problems PCR Clone Libraries – Copy number bias mentioned yesterday.Primer Choice/Bias, Extension Bias? Yes, but how bad?Example – Marine Archaea compared to Bacteria. DISTANTFuhrman et al. (1992) used universal primers, found 5 of 7 clones from 500 m were Crenarchaeota. DeLong (1992) used archaeal primers with surface waters only, and RNA hybridization to compare to Bacteria. Archaea <2%.Fuhrman and Davis (1997, univ. primers) Archaea were 1/3 of clones from 500 m – 3000 m, Atlantic and PacificFISH results – Fuhrman and Ouverney 1998, Archaea to 40% at 600 m in Pacific, 60% at 200 m in Mediterranean. Karner et al. (2001) – Archaea ~30% below ~ 200m at HOT over > 1 year.Note – If QPCR shows doubling each cycle and if not at the saturation point, anything primed OK should quantify OK
27 DeLong et al. Science, 2006 Metagenomics BIAS? Missing rRNA genes from large-insert libraryAllBLAST hits-%SSU rRNAGenes-Presence/absenceDeLong et al.Science, 2006SAR11