Detection of Shifts in Microbial Functional Gene (nosZ and n/fH) Distribution Due to Long Term Warming of a High Arctic Soil Jennifer KM Walker B.Sc., University of Northern British Columbia, 2003 Thesis Submitted In Partial Fulfillment Of The Requirements For The Degree Of Master of Science in Natural Resources and Environmental Studies (Biology) The University of Northern British Columbia August 2006 © Jennifer KM Walker, 2006 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Library and Archives Canada Bibliotheque et Archives Canada Published Heritage Branch Direction du Patrimoine de I'edition 395 W ellington Street Ottawa ON K1A 0N4 Canada 395, rue W ellington Ottawa ON K1A 0N4 Canada Your file Votre reference ISBN: 978-0-494-28410-0 Our file Notre reference ISBN: 978-0-494-28410-0 NOTICE: The author has granted a non­ exclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell theses worldwide, for commercial or non­ commercial purposes, in microform, paper, electronic and/or any other formats. AVIS: L'auteur a accorde une licence non exclusive permettant a la Bibliotheque et Archives Canada de reproduire, publier, archiver, sauvegarder, conserver, transmettre au public par telecommunication ou par I'lnternet, preter, distribuer et vendre des theses partout dans le monde, a des fins commerciales ou autres, sur support microforme, papier, electronique et/ou autres formats. The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. L'auteur conserve la propriete du droit d'auteur et des droits moraux qui protege cette these. Ni la these ni des extraits substantiels de celle-ci ne doivent etre imprimes ou autrement reproduits sans son autorisation. In compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis. Conformement a la loi canadienne sur la protection de la vie privee, quelques formulaires secondaires ont ete enleves de cette these. While these forms may be included in the document page count, their removal does not represent any loss of content from the thesis. Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant. i*i Canada Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPROVAL Name: Jennifer KM Walker Degree: Master o f Science Thesis Title: Detection o f Shifts in Microbial Functional Gene (nosZ and niJH) Distribution Due to Long Term Warming o f a High Arctic Soil Examining Committee: ---------------------- : —— =------------------------------------ Chair: Dr. Martha MacLeod, Associate Professor Nursing Program University o f Northern British Columbia Su^fvisor: Dr. Keith Egger, Professor Natural Resources and Environmental Studies Program ^University o f Northern British Columbia Committee Member: Dr. Brent Murray^ Associate Professor Natural Resources and Environmental Studies Program University o f Northern British Columbia ---- --,----- ---------------------------------------y f- M . Committee Member: Dr. Michael Rutherford, Associate Professor Natural Resources and Environmental Studies Program University o f Northern British Columbia External Exarniner: Dr. Paul Sanborn, Associate Professor NaturaEResources-ahd Environmental Studies Program University o f Northern British Columbia Date Approved: Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT Climate change is expected to increase annual arctic temperatures by up to 5 °C and winter arctic temperatures as much as 7 °C within the next century. High Arctic coastal lowlands, such as at Alexandra Fiord, Ellesmere Island, have greater productivity and species diversity than the surrounding desert-like plateaus. Warming due to polyethylene open-topped chambers (OTCs) has been shown to increase plant productivity and cause community shifts compared to adjacent control plots. Changes in the soil environment have also been noted and will affect microbial activity and community composition. Long term elevated temperatures may result in changes to the community structure of these organisms. By studying terminal restriction fragment length polymorphisms (TRFLPs) of genes associated with nitrogen cycling, we can examine variations in the nitrogen cycling microbial community. The objective of this study was to detect shifts in denitrifying and nitrogen fixing soil microbial communities by measuring changes in functional gene frequency, abundance and/or genotypic richness of nosZ and nifH respectively. The study area encompassed five high arctic sites that differed by dominant plant community, soil parent material and/or moisture regime and that had been subjected to a thirteen year warming experiment. Four OTCs plus four adjacent control plots were sampled at each of these locations in order to investigate differences in these gene communities due to site, depth, and treatment. Samples were recovered from the top and bottom 5 cm of soil cores; these ranged from 13 cm to 44 cm deep. DNA extractions from soil samples were ii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. tested with two different primer pairs. Functional genes targeted included those that code for nitrous oxide reductase (nos) and nitrogenase (nif). Differences in frequency and relative abundance of terminal restriction fragments (TRFs) were assessed graphically by Non-metric Multidimensional Scaling (NMS) and tested statistically with permutational multivariate ANOVA (PERMANOVA). Genotypic richness was examined by testing the difference in number of TRFs with nested and one-way ANOVA. Functional gene frequency and relative abundance was shown to differ overall by site in NMS ordinations of both denitrifying and nitrogen fixing communities. PERMANOVA tests also suggested a significant difference in the relative abundance of denitrifying TRFs, and both the frequency and relative abundance of nitrogen-fixing TRFs by depth over all sites. Differences in genotype richness were detected between sites, at different depths, and due to treatment for both communities. In general, denitrifying and nitrogen fixing community richness decreased with soil depth and with OTC treatment. iii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS ABSTRACT TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF ACRONYMS ii iv vi vii xi Chapter 1. Literature Review 1.1 Climate Change and arctic systems 1.2 Arctic research sites 1.3 Soil N cycling and plants 1.4 Soil N cycling and microbes 1.5 Microbial communities 1.6 Genetic diversity versus physiological function 1.7 Molecular techniques 1 3 4 7 11 13 15 Chapter 2. Detection of shifts in microbial functional gene (nosZ and nifH) distribution due to long term warming of a high arctic soil 2.1. Introduction 2.2. Methods 2.2.1 Site description 2.2.2 Experimental design Open Topped Chambers (OTCs) Control Plots Randomization Soil sampling 2.2.3 Preservation and transport 2.2.4 Field measurements 2.2.5 DNA extraction and PCR 2.2.6 T-RFLP 2.2.7 Statistical analysis 2.3. Results 2.3.1 G e n e frequency and relative abundance Site effect Depth effect T reatment effect 2.3.2 Genotype richness nosZ nifH iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16 20 23 24 24 24 25 26 27 29 31 32 44 58 61 65 2.4. Discussion 2.4.1 Methodological considerations 2.4.2. Did the denitrifying and nitrogen-fixing communities respond similarly or differently to experimental factors? 2.4.3. What experimental factors most influence microbial community structure? 2.4.4. What measures of community structure are most affected by environmental factors? 69 70 76 90 Chapter 3. Summary 3.1. Conclusions 3.2. Microbial community structure and physiological function 3.3. Future research Literature cited 93 95 96 98 V Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES Table 1: Individual site data collected during other warming experiments at Alexandra Fiord include soil pH, soil water content (SWC) and both air and soil temperatures. 23 Table 2: Nested ANOVA showing overall significant differences in number of nosZ TRFs between sites (p<0.000001) and between depths across all sites (p=0.018). Table 3: Nested ANOVA showing overall significant differences in number of nosZ TRFs between sites (p<0.000001) and between treatment across all sites (p=0.039). Q2 Table 4: Nested ANOVA showing overall significant differences in number of nifH TRFs between sites (p<0.000001) and between depths across all sites (p=0 .0 0 0 0 0 1 ). 65 Table 5: Nested ANOVA showing overall significant differences in number of nifH TRFs between sites (p<0.000001) and between treatment across all sites (p<0 .0 0 0 0 0 1 ). vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES Figure 1: The High Arctic site at Alexandra Fiord (6) is one of a number of arctic, subarctic and alpine ITEX study sites (modified from Arft et al., 1999). 4 Figure 2: Layout of the Alexandra Fiord Study area showing the three lowland sites (Sedge Meadow, Cassiope Heath, and Riverside Willow) and the two upland sites (Upland Granite and Upland Dolomite). Contour intervals are 100m (modified from Sterenberg and Stone, 1994). 22 Figure 3: Arrows indicate significant differences in the frequency nosZ (a) and nifH (b) TRFs between sites. 33 Figure 4: NMS plot of nosZ TRF frequency by Site p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). 34 Figure 5: NMS plot of n/7H TRF frequency by Site p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). 35 Figure 6: Distribution of the top 20 most frequent nosZ TRFs over all Sites. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). 36 Figure 7: Distribution of the top 20 most frequent nifH TRFs over all Sites. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). 37 Figure 8: Arrows indicate significant differences in the relative abundance of nosZ TRFs between sites. 38 Figure 9: NMS plot of nosZ TRF relative abundance by Site showing distribution of the 20 most abundant TRFs p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). Crosses mark TRF location and numbers reflect fragment length in base pairs (bp). 39 Figure 10: Distribution of the top 20 most abundant nosZ TRFs over all Sites. 40 Figure 11: Arrows indicate significant differences in the relative abundance of nifH TRFs between sites. 41 Figure 12: NMS plot of nifH TRF relative abundance by Site showing distribution of the 20 most abundant TRFs p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and 42 vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Upland Dolomite (UD). Crosses mark TRF location and numbers reflect fragment length in base pairs (bp). Figure 13: Distribution of the top 20 most abundant nifH TRFs over all Sites. 43 Figure 14: NMS plot of nosZ TRF frequency by Depth for the Upland Dolomite site showing the top 20 most frequent TRFs (p=0.0196). Depths are upper (U) or lower (L), and all numbered samples are labeled control (C) or OTC (O). Crosses mark TRF location, and bold number reflects fragment length in base pairs (bp). 45 Figure 15: Distribution of the top 20 most frequent nosZ TRFs from Control and OTC samples of the Upland Dolomite Site. 45 Figure 16: NMS plot of nosZ TRF relative abundance by Depth for the Sedge Meadow site showing the top 20 most abundant TRFs (p=0.0196). Depths are upper (U) or lower (L), and all numbered samples are labeled either control (C) or OTC (0) (OTC icons are larger in order to highlight the statistical significance of their separation). Crosses mark TRF location and numbers (bold) reflect fragment length in base pairs (bp). 47 Figure 17: Distribution of the top 20 most abundant nosZ TRFs from OTC samples of the Sedge Meadow Site. 48 Figure 18: NMS plot of nosZ TRF relative abundance by Depth for the Upland Dolomite site showing the top 20 most abundant TRFs (p=0.0196). Depths are upper (U) or lower (L) and all numbered samples are labeled either control (C) or OTC (O) (control icons are larger in order to highlight the statistical significance of their separation). Crosses mark TRF location and numbers (bold) reflect fragment length in base pairs (bp). 49 Figure 19: Distribution of the top 20 most abundant nosZ TRFs from Control samples of the Upland Dolomite Site. 50 Figure 20: NMS plot of nifH TRF frequency by Depth for the Sedge Meadow site showing the distribution of the top 20 most frequent TRFs (p=0.0196). Depths are upper (U) or lower (L) and numbered samples are labeled either control (C) or OTC (O). Crosses mark TRF location and numbers (bold) reflect fragment length in base pairs (bp). 51 Figure 21: Distribution of the top 20 m ost frequent nifH T R F s from Control and OTC samples of the Sedge Meadow Site. 51 Figure 22: NMS plot of nifH TRF frequency by Depth for the Upland Granite site showing the top 20 most frequent TRFs (p=0.0196). Depths are upper (U) or lower (L) and numbered samples are labeled control (C) or OTC (O) (control icons are larger in order to highlight the statistical significance of their 53 viii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. separation). Crosses mark TRF location and numbers (bold) reflect fragment length in base pairs (bp). Figure 23: Distribution of the top 20 most frequent nifH TRFs from Control samples of the Upland Granite Site. 54 Figure 24: NMS plot of nifH TRF relative abundance by Depth for the Sedge Meadow site showing the distribution of the top 20 most abundant TRFs (p=0.0196). Depths are upper (U) or lower (L) and all numbered samples are labeled control (C) or OTC (O). Crosses mark TRF location and numbers (bold) reflect fragment length in base pairs (bp). 55 Figure 25: Distribution of the top 20 most abundant nifH TRFs from Control and OTC samples of the Sedge Meadow Site. 56 Figure 26: NMS plot of nifH TRF relative abundance by Depth for the Upland Granite site showing the top 20 most abundant TRFs (p=0.0392). Depths are upper (U) or lower (L) and all numbered samples are labeled Control (C) or OTC (O) (control icons are larger in order to highlight the statistical significance of their separation). Crosses mark TRF location and numbers (bold) reflect fragment length in base pairs (bp). 57 Figure 27: Distribution of the top 20 most abundant nifH TRFs from Control samples of the Upland Granite Site. 58 Figure 28: NMS plot of nosZ TRF relative abundance for all sites showing that samples do not separate by treatment (p=0.0196). Treatments are control (C) or OTC (O), and crosses mark the top 20 TRFs over all sites. Numbers reflect fragment length in base pairs (bp). 59 Figure 29: NMS plot of nifH TRF relative abundance for all sites showing that samples do not separate by treatment (p=0.0196). Treatments are control (C) or OTC (O), and crosses mark the top 20 TRFs over all sites. Numbers reflect fragment length in base pairs (bp). 60 Fig 30: Mean number of nosZ genotypes (TRFs) at each site (p<0.001, confidence interval = 0.95). 62 Fig 31: The mean number of nosZ genotypes (TRFs) differed over all sites by depth (p = 0 .0 1 8 , C l= 0 .9 5 ). Significant differences at individual sites are marked with an asterisk (p<0.01). 63 Fig 32: The mean number of nosZ genotypes (TRFs) differed by treatment over all sites (p=0.039, Cl=0.95). Significant differences at individual sites are marked with an asterisk (p<0.05). 64 IX Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Fig 33: Mean number of nifH genotypes (TRFs) at all sites (p<0.001, confidence interval = 0.95). 66 Fig 34: The mean number of nifH genotypes differed by depth over all sites (p<0.000001, Cl=0.95). Significant differences at individual sites are marked with an asterisk (p<0.05). 67 Fig 35: The mean number of nifH genotypes differed by treatment over all sites (p<0.000001,CI=0.95). Significant differences at individual sites are marked with an asterisk (p<0.005). 68 x Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF ACRONYMS SM CH RW UG UD Sedge Meadow Cassiope Heath Riverside Willow Upland Granite Upland Dolomite nir nos nif amo nitrite reductase nitrous oxide reductase dinitrogenase reductase ammonium monooxygenase OTCs T-RFLPs TRFs PCR RT-PCR DGGE DEA ARA GC CEC OM SWC open-topped chambers terminal restriction fragment length polymorphisms terminal restriction fragments polymerase chain reaction real time PCR denaturing gradient gel electrophoresis denitrification enzyme assay acetylene reduction assay gas chromatography cation exchange capacity organic matter soil water content NMS PERMANOVA Non-metric Multidimensional Scaling permutational multivariate ANOVA ITEX IPCC ACIA International Tundra Experiment Intergovernmental Panel on Climate Change Arctic Climate Impact Assessment xi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 1. Literature Review 1.1 Climate change and arctic systems Climate change is expected to disproportionately affect arctic and antarctic latitudes (Hassol 2004, Maxwell etal. 1992). Positive feedbacks due to changes to polar weather patterns and ecosystems will further affect global climate. Current climate change models assume a doubling of atmospheric carbon dioxide (CO2 ) concentration in less than fifty years. This prediction is based upon rates of CO2 increase measured recently; the concentration rose steadily after the last ice age, but has risen dramatically from the 19th century to present day. It is the accumulation of radiative gases such as CO2 in the atmosphere that reinforces the greenhouse effect. Radiative gases have the capacity to absorb outgoing thermal energy and re-radiate it back toward the earth’s surface. This blanket of gases is what makes our planet habitable, but it is at the root of excess global warming. The effects are dramatic: within 100 years, annual arctic temperature may be 3-5 °C higher, with winter temperatures as much as 4-7 °C higher over land than they are currently (Hassol 2004, Maxwell et at. 1992). There are other gases implicated in climate change, and although they are present in much smaller concentrations, their thermal capacity on a per molecule basis, and their persistence in the atmosphere is far greater (Maxwell et al. 1992). Nitrous oxide (N2 O), a product of nitrogen transformation cycles, such as denitrification and nitrification, is one of these. - 1 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Though many climate change models may agree that certain variables are crucial inputs, the outcomes predicted are not always the same and in fact are often contradictory (Maxwell et at. 1992). The exact changes, their feedbacks, and their implications are highly debatable. Most researchers agree that overall annual global precipitation will increase, and, like temperature, will be greater at the poles. Higher winter temperatures and rainfall in the arctic will decrease residual spring snow cover and promote permafrost thawing (Hassol 2004). Contradictions arise regarding net soil moisture; some suggest an increase due to melting of the permafrost layer and earlier snowmelt, while others hint at a decrease due to better drainage and increased evaporative losses (Hassol 2004, Kane et al. 1992, Maxwell et al. 1992). A combination of these two predictions may be possible with wet, coastal areas getting wetter, and dry, inland areas getting drier (Maxwell etal. 1992). Increased surface temperature resulting in heat transfer to the soil via conduction has also been assumed (Kane et al. 1992). The greatest warming is predicted to occur in the winter, but the soil at -15 cm should still remain approximately 3°C warmer than usual throughout the summer (Kane et al. 1992). Recent studies have shown that air temperature can be raised to predicted levels by experimental warming with greenhouses without a proportional increase in soil temperature (Jonasson et al. 1993, Robinson 2002). This challenges the assumption of a deeper active layer and more nutrient cycling with warming. - 2 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.2 Arctic research sites Some coastal lowlands in the high arctic have greater productivity and species diversity due to different temperature and moisture regimes than the surrounding desert-like plateaus (Freedman etal. 1994). These unique ecosystems provide excellent study sites for the predicted changes due to global warming because of seasonal extremes in solar energy and precipitation, and because these areas are expected to experience the greatest transitions (Freedman et al. 1994, Maxwell et al. 1992). Alexandra Fiord is a High Arctic research site established by the International Tundra Experiment (ITEX) (Figure 1). See Arft et al. (1999), Rustad et al. (2001) and Walker et al. (2006) for comprehensive summaries of climate change experiments at this and other subarctic and alpine sites. Polyethylene open-topped chambers (OTCs) have been installed across various gradients at the Alexandra Fiord site, and many studies have taken place over the last decade. Warming due to these OTCs, which generate a small greenhouse effect, has been shown to increase plant productivity and create community shifts compared to adjacent control plots on wet, moist, and dry lowland, and on granitic and dolomitic upland sites (Freedman et al. 1994, Rolph 2003). -3 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 1: The High Arctic site at Alexandra Fiord (6) is one of a number of arctic, subarctic and alpine ITEX study sites (modified from Arft et al. 1999). 1.3 Soil N cycling and plants The soils of a polar oasis can differ dramatically from those of the surrounding polar desert (Muc etal. 1994). Even slight increases in both temperature and precipitation can lead to deeper active layers, higher rates of chemical transformations and ultimately, more nutrient availability and plant productivity. The primarily granitic-derived soils of the Alexandra Fiord lowland, although less developed than those of temperate regions, have more organic matter, higher soil moisture levels and increased cation exchange capacity (CEC) than upland -4 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. samples. Despite this, all lowland sites and both the granitic and dolomitic upland sites are still considered nutrient (especially nitrogen) limited (Muc et al. 1994). Nitrogen gas (N2) comprises 78% of our atmosphere, and its fixation by microbes is the major source of new nitrogen (N) in soils (Paul and Clark 1996). Other inputs of N such as fertilization and atmospheric deposition are not relevant in remote arctic regions (Paul and Clark 1996). It is possible that future arctic environments will be less nitrogen limited due to a predicted increase in nitrogen fixation (Chapin and Bledsoe 1992). The potential increase is attributed to favourable enzymatic activity at higher temperatures, and to elevated levels of carbon dioxide and moisture that will benefit the aquatic and photosynthetic nitrogen fixers (diazotrophs) (Chapin and Bledsoe 1992). Nutrient cycling in this arctic environment could be enhanced with the additional warming predicted by climate change models (Berendse and Jonasson 1992). This increase in soil fertility and nitrogen availability will affect plant-soil dynamics and predominant soil nutrient cycles such as net mineralization versus net immobilization (Berendse and Jonasson 1992). Mineralization is simply the degradation of organic forms of elements such as nitrogen to their mineral forms; ammonium (NH4+) is then immobilized for microbial growth as required, and the excess may accumulate in the soil (Paul and Clark 1996). Arctic plants are already adapted to low temperatures, so the predicted increased productivity will be limited by nutrient availability which is in turn mediated by soil response to -5 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. warming (Chapin III et al. 1992). It is likely that changes in soil characteristics, especially moisture, will drive the changes in other biogeochemical cycles (Chapin III etal. 1992). In general, net soil mineralization is low or even negative during the arctic growing season with any excess nutrients sequestered in microbial biomass (Jonasson et al. 1999, Nadelhoffer et al. 1992). Microbes compete strongly for nutrients throughout the growing season, but as their populations decline in the winter, it is possible that nutrients are released and available to plants (Jonasson et al. 1999). These plants are adapted to low nutrient levels and therefore any significant changes in availability can lead to changes in plant productivity and species composition. Although increased net N mineralization with warming has been predicted (Nadelhoffer et al. 1992), recent studies have not confirmed that it will be affected by short term increases in soil temperature of a few degrees Celsius (Jonasson et al. 1999, Schmidt et al. 1999). The variability in net mineralization between soils from different plant communities is often more than any differences observed due to temperature treatments (Schmidt et al. 1999). Although in past studies increases in soil temperatures of 1-1.5°C at 5-7 cm deep were accomplished with OTCs, greater net N mineralization was only observed at a one high altitude site in the winter; no changes in microbial immobilization were noted for any treatments. The techniques used in this study suggest that microbes are very strong competitors for soil nutrients in all seasons, and that an - 6 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. increase in mineralization with their population decline at the end of a growing season cannot be considered a general rule (Schmidt et al. 1999). Longer term warming experiments have revealed increases in microbial immobilization due to changes in soil properties that are attributed to sustained elevated temperatures (Rolph 2003). Shorter term changes in N cycling and availability have modified the plant community composition over time and the feedbacks are now notable as changes in litter quality and plant productivity (Rolph 2003). 1.4 Soil N cycling and microbes Changes in soil moisture status (associated with global warming) will impact microbial processes such as decomposition in all arctic ecosystems; the resulting combination of moisture and temperature will dictate organic matter (OM) turnover rates, and ultimately N H / availability via mineralization (Paul and Clark 1996, Nadelhoffer et al. 1992). The predicted and measured changes in the soil environment with increased temperature are not limited to nutrient mineralization and subsequent availability to plants; microbial activity and community composition are affected (Nadelhoffer et al. 1992). Depending upon moisture, nutrient availability, and timing, the process of nitrogen fixation is shared across a range of soil organisms (Paul and Clark 1996). There exist both symbiotic and free-living diazotrophs that can reduce N2 to ammonia (NH3) in almost all habitats. The reaction is inhibited by oxygen and -7 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. by the presence of nitrate (N 0 3 ). Nitrogen fixation is mediated by a number of genes including those that code for the nitrogenase enzyme, which exists in a complex with two Fe-S proteins, nifH is the structural gene for dinitrogenase reductase, while additional nif genes function as transcriptional regulators, code for cofactors, or are involved in electron transport. Production of the nitrogenase enzyme is inhibited by the presence of its product, NH3. Overall, nitrogen fixation is not favoured in the presence of available N. An increase in mineralization with warming could stall this process in the arctic. Once the NH3 is transformed to NH4+, it is available for microbial growth and plant uptake, becomes adsorbed to minerals and organic matter in the soil, or is lost due to leaching or as a volatile gas. Additionally, NH4+ is used as an energy source by nitrifying bacteria (Paul and Clark 1996). Nitrification is detectable in arctic soils, and there are different degrees of nitrate use by each plant community (Paul and Clark 1996, Nadelhoffer et al. 1992). It is a temperature sensitive process that likely occurs only in the growing season and that could be greatly affected by potential soil warming (Paul and Clark, 1996). Direct increases in nitrification due to higher temperatures may be in addition to the increases attributed to a higher availability of substrate if warming also promotes mineralization (Nadelhoffer et al. 1992). Nadelhoffer et al. (1992) predicted that arctic ecosystems with an intermediate moisture regime will see the greatest increases in nitrification with warming; the drier systems lack the - 8 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. deep OM layer required for large changes in the soil environment, and biological processes at the wetter sites are limited by anaerobic conditions. The nitrification pathway from ammonium (NH4+) to nitrate (N 03‘) is mediated by specific microbes; NO3' is preferred by plants, but is easily lost due to high solubility in water (Paul and Clark 1996, Prescott et al. 1999). This process is accomplished in two stages primarily by chemoautotrophic aerobes that oxidize NH4+ to hydroxylamine or that oxidize nitrite (NO2 ) to NO3'. These low energyyielding reactions are mediated by the enzymes ammonium monooxygenase, hydroxylamine oxidoreductase, and nitrite dehydrogenase (Paul and Clark 1996). The ammonium oxidizers (Nitroso- spp.) are members of the beta or gamma proteobacteria while the nitrite oxidizers (Nitro- spp.) occupy the alpha proteobacterial group (Prescott et al. 1999). Nitrification is sensitive to pH and does not occur below pH 4.5 in agricultural soils; forest soils that harbor heterotrophic nitrifiers can support nitrification even in slightly acidic conditions. This process requires an aerobic, mesic environment and is slow below 5°C (Paul and Clark 1996). Nitrate can be reduced by other members of the Proteobacteria (Rhizobium, Alcaligenes, Pseudomonas) and Archaea (Haloarcula) (Prescott etal. 1999), and lost as environmentally detrimental gaseous byproducts of denitrification (Paul and Clark 1996). Denitrification is anaerobic respiration that uses oxidized inorganic forms of nitrogen as electron acceptors (Paul and Clark 1996, Prescott -9 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. et al. 1999). Steps in this pathway are mediated by the enzymes nitrate reductase, nitrite reductase, nitric oxide reductase, and nitrous oxide reductase; the final product is nitrogen gas (N2) (Paul and Clark 1996). This process is inhibited in an oxygenated and/or acidic environment and has been observed in some diazotrophs but is temporally separated from nitrogen fixation because the enzyme nitrogenase is repressed by N 0 3\ Denitrification has also been observed in some heterotrophic nitrifiers; the process is spatially separated but can occur at the same time as nitrification. If carbon supply is limited, incomplete reduction of NO3' can occur with a resulting accumulation of the nitrogen oxide gases. Denitrification potential decreases linearly with soil temperatures below 15°C (to a minimum at 5°C), and is optimal in oxygen limited soils with pH 6 - 8 (not detected below pH 4) (Paul and Clark 1996). Nitrification and denitrification may be closely associated, especially at aerobic/anaerobic interfaces where the nitrification product, NO3' is readily available for reduction (Nicolaisen et al. 2004). Gaseous losses of nitrogen have been attributed to these coupled cycles, but Nicolaisen etal. (2004) suggest that plant roots are strong competitors for available nitrogen, so that in their presence loss to the atmosphere is minimized. - 10 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.5 Microbial communities Nitrogen fixers are represented across most prokaryotic groups, and include organotrophic and phototrophic, aerobic and anaerobic microorganisms that possess the enzyme nitrogenase (Paul and Clark 1996, Zehr et al. 2003). The free-living aerobes such as Azotobacter and Beijerinckia can be found on the surface of roots and in the adjacent soil. Their high rates of respiration are one strategy to protect the nitrogenase enzyme from oxygen. The microaerophiles {Azospirillum, Bacillus) and anaerobes (Clostridium, Desulfovibrio) already occupy the anoxic environment required by this enzyme. There are also a significant proportion of diazotrophic cyanobacteria (Nostoc, Anabaena) above and below the soil surface that contribute greatly to nitrogen fixation in the arctic (Chapin and Bledsoe 1992, Liengen 1999, Paul and Clark 1996). Other aquatic species include the nitrogen fixing green and purple sulfur bacteria. Symbiotic actinomycetes such as Frankia fix nitrogen in association with the arctic plant Dryas (Paul and Clark 1996). Based upon modes of nutrition, aerobic nitrifiers should be found at and just below the soil surface, where litter and humus are mineralized providing net available N H / (Paul and Clark 1996). It follows that the heterotrophic denitrifiers would be found at a depth where anaerobic conditions prevail, and where there is a carbon source, for example, below the rhizosphere (Paul and Clark 1996). Nicolaisen et al. (2004) confirm that abundance and activity of nitrifiers follows a -11 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. distribution based upon the presence of 0 2 and NH4+ in the soil. By separating their samples into surface, bulk, or rhizosphere zones, they noted highest numbers and nitrification potential in the surface soil (2-3 cm deep) with lowest numbers and nitrification potential in the bulk soil (5-6 cm deep). The authors note that by overlooking the differences between these soil depths, the surface may not be appropriately recognized as an important nitrification site. In addition, molecular techniques can detect nitrifiers in deeper samples although they may simply be in a dormant state due to unfavourable anoxic conditions. Despite these findings, differences in nitrifier community structure were not apparent across the three zones; in fact, diversity was low overall when compared to other soil environments (Nicolaisen et al. 2004). Avrahami and colleagues (2002), also show that increased NH4+ availability promotes an increase in nitrous oxide production, especially via nitrification, without a corresponding community shift in the nitrifier community. However, when the denitrifier population was examined with Terminal Restriction Fragment Length Polymorphisms (T-RFLP) of the nirK gene, changes in diversity were detected that appeared to mirror changes in nitrous oxide (N20 ) contribution by the different microbial communities (Avrahami et al. 2 0 0 2 ). Further studies by the same group showed a shift in the ammonia oxidizer community with temperature manipulations in the lab (Avrahami and Conrad 2003). Their previous measurements of increased nitrifier activity levels with fertilizer additions could not be attributed to a community shift; there may instead - 12- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. have been a physiological adaptation of the existing community. Avrahami and Conrad (2003) once again measured nitrification activity in fertilized treatments with different soil temperatures, moisture levels, and pH; the lowest activity was noted in the coolest soils. Using the gene target amok, the researchers showed that acidic soils had a higher diversity of ammonium oxidizers, and that both acidic and alkaline soils experienced a change in community structure with temperature manipulations. Most importantly, the researchers concluded that temperature was selecting for different ammonium oxidizing communities, and that long term temperature changes will result in changes to the community structure of these organisms in the natural soil environment (Avrahami and Conrad 2003). 1.6 Genetic diversity versus physiological function The link between changes in microbial community structure, measured by changes in gene diversity, and corresponding changes in the processes mediated by these communities has not been satisfactorily established. Community shifts that do not affect microbial activity have been documented (Avrahami and Conrad 2003, Deslippe 2004) and changes in physiological function have been noted without affecting diversity at the genetic level (Gomez et al. 2004, Nicolaisen et al. 2004, Rich and Myrold 2004). It has been suggested that biotic and abiotic environmental factors can structure the genetic composition of a system, but that gene selection is not the only driver of microbial -1 3 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. diversity (Zehr et al. 2003). Zehr et al. (2003) explain that a habitat not limited by nitrogen availability should not select for the functional genes required for nitrogen fixation. In such a habitat, the diversity of these genes should diminish, but the opposite has been documented, suggesting that function and diversity are not related in this system. Alternately, Rich et al. (2003) found a significant relationship between denitrifier community composition and denitrification activity on the basis of vegetation type. The proportional abundance of at least two dominant terminal restriction fragments clearly differed in a forested area of low denitrification versus a meadow with high activity (Rich etal. 2003). Further studies by the same group confirmed unique nosZ TRF communities at each of three distinct sites, but denitrifying activity measurements did not vary with microbial community structure (Rich and Myrold 2004). PCR plus denaturing gradient gel electrophoresis (DGGE) also failed to detect denitrifier community shifts corresponding to increasing N2 O, but the same genetic marker revealed changes in diversity associated with the rise in N2O when gene expression was measured by real time PCR (RT-PCR) (Sharma et al. 2006). Although changes in the distribution of functional gene markers may not yet reliably predict changes in physiological function, techniques that measure the presence and/or relative abundance of functional genes are ideal for assessing microbial community structure (Tiedje et al. 1999, Zehr et al. 2003). -14- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.7 Molecular techniques Terminal Restriction Fragment Length Polymorphism (T-RFLP) is a molecular technique that allows researchers to examine the diversity of the microbial genes associated with important soil processes (Tiedje et al. 1999). The procedure uses polymerase chain reaction (PCR) amplification with fluorescently labeled primers and subsequent restriction digests of the PCR product to generate terminal restriction fragments (TRFs) that are representative of unique taxonomic units. The presence and relative abundance of these fragments are interpreted by an automated sequencer, and their patterns can be used to describe differences in gene distribution and richness (Dunbar et al. 2000, Tiedje et al. 1999). This assessment of microbial community structure can be used to compare the differences in potential gene function based upon larger scale environmental change especially where species diversity is low or moderate (Engebretson and Moyer 2003). The technique is rapid and repeatable, but may overestimate similarity and thus underestimate diversity when compared to cloning and sequencing methods (Dunbar et al. 2000). It is highly dependent on the choice of restriction enzyme; accurate measures of diversity require that all unique restriction sites are identified (Dunbar et al. 2000, Engebretson and Moyer 2003). - 15 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 2. Detection of shifts in microbial functional gene (nosZ and nifH) distribution due to long term warming of a high arctic soil. 2.1 Introduction Climate change is expected to disproportionately affect arctic and antarctic latitudes (Hassol 2004, Maxwell etal. 1992). Positive feedbacks due to changes to polar weather patterns and ecosystems will further affect global climate. Current climate change models assume a doubling of atmospheric carbon dioxide (CO2) concentration in less than fifty years. The effects are dramatic: within 100 years, annual arctic temperature may be 3-5 °C higher, with winter temperatures as much as 4-7 °C higher over land than they are currently (Hassol 2004, Maxwell et al. 1992). Most researchers agree that overall annual global precipitation will increase, and, like temperature, will be greater at the poles. Contradictions arise regarding net soil moisture; some suggest an increase due to melting of the permafrost layer and earlier snowmelt, while others hint at a decrease due to better drainage and increased evaporative losses (Hassol 2004, Kane etal. 1992, Maxwell et al. 1992). Increased surface temperature resulting in heat transfer to the soil via conduction has also been assumed (Kane et al. 1992). T h e re are coastal lowlands in the high arctic with g re ater productivity and species diversity due to different temperature and moisture regimes than the surrounding desert-like plateaus (Freedman etal. 1994). These unique ecosystems provide excellent study sites for the predicted changes due to global warming because of - 16- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. seasonal extremes in solar energy and precipitation, and because these areas are expected to experience the greatest transitions (Freedman et al. 1994, Maxwell et al. 1992). The research site at Alexandra Fiord is on the northern side of Johan Peninsula, on the eastern coast of Ellesmere Island, Nunavut (Freedman et al. 1994). The study area encompasses a lowland outwash plain, bordered by ocean to the north and glacier to the south, and by the cliffs of an upland plateau, up to 750 m higher on the east and west. Polyethylene opentopped chambers (OTCs) have been established across various gradients at this site, and many studies have taken place over the last decade. Warming due to these small greenhouses has been shown to increase plant productivity and create community shifts compared to adjacent control plots on wet, moist, and dry lowland, and on granitic and dolomitic upland sites (Freedman etal. 1994, Rolph 2003, Walker et al. 2006). All lowland sites and both the granitic and dolomitic upland sites are considered nutrient (especially nitrogen) limited, but even slight increases in both temperature and precipitation can lead to deeper active layers, higher rates of chemical transformations and ultimately, more nutrient availability (Berendse and Jonasson 1992, Muc et al. 1994). Changes in soil moisture status will impact microbial processes such as decomposition in dry, moist, and wet arctic ecosystems; the resulting combination of moisture and temperature will dictate organic matter turnover rates, and ultimately ammonium (NH4+) availability via mineralization (Paul and Clark 1996, Nadelhoffer et al. 1992). It is possible that - 17- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. future arctic environments will be less nitrogen limited due to a predicted rise in nitrogen fixation with increased enzyme activity and levels of carbon dioxide (Chapin and Bledsoe 1992) but in general, nitrogen fixation is not favoured in the presence of available nitrogen (N) therefore an increase in mineralization with warming could stall this process in the arctic (Paul and Clark 1996). Direct increases in nitrification due to higher temperatures may be in addition to the increases attributed to a higher availability of substrate if warming does indeed promote mineralization (Nadelhoffer et al. 1992). Nitrification and denitrification may be closely associated, especially at aerobic/anaerobic interfaces where the nitrification product, nitrate (NO3') is readily available for reduction (Nicolaisen et al. 2004). If soil warming is favourable for nitrification, it may also stimulate denitrification and subsequent nitrous oxide (N20 ) production (Paul and Clark 1996). We know that experimental nutrient amendments increase the rates of N20 production, but it still has not been confirmed if this is correlated with the genetic structure of denitrifier communities (Rich and Myrold 2004). Additionally, a habitat with excess nitrogen should not select for the functional genes required for nitrogen fixation (Zehr et al. 2003) but the diversity of the nitrogen fixing community has not been shown to diminish in such cases (Piceno and Lovell 2000a). - 18- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The objective of this study was to investigate changes in the frequency, relative abundance and richness of the functional genes nosZ and n/7H as markers for denitrification and nitrogen fixation respectively. We were looking for shifts in the denitrifying and nitrogen fixing soil microbial communities from sites with different dominant plant communities, on different parent material and across different moisture regimes after a thirteen year warming experiment in the High Arctic. Our null hypothesis was that long term warming treatments would not alter microbial community structure. - 19- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.2. Methods 2.2.1. Site description The research site at Alexandra Fiord is on the northern side of Johan Peninsula, on the East coast of Ellesmere Island, Nunavut (78°53’ N, 75°55’ W) (Freedman et al. 1994). The study area encompasses a lowland outwash plain, bordered by ocean to the north and glacier to the south, and by the cliffs of an upland plateau, up to 750 m higher on the east and west. The topography of the area leads to warmer temperatures and greater accumulations of water in the lowland plain than the surrounding landscape; this qualifies the lowland as a true polar oasis (Freedman et al. 1994). There is a relatively large amount of organic matter in the young soils, and diverse vegetation including cushion plants and dwarf shrubs. In summer, the dark soil surface results in a low albedo which contributes to higher air and soil temperatures at this site than those measured at the next closest weather stations in Eureka and Resolute (Labine 1994). The regoliths of the upland area are drier overall than the lower site, but support some of the same plant species on both granitic and dolomitic parent rock (Freedman et al. 1994). During the growing season this area experiences below zero air temperatures and less absorbed solar radiation than the coastal lowland (Labine 1994). - 20- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Five distinct sites are investigated in this study (Figure 2). They differ primarily by dominant plant community, and each corresponds to a particular soil moisture gradient and/or soil parent material (Muc et al. 1989). All soils at the three lowland sites are derived from granitic material. The Sedge Meadow site (SM) is characterized as the wettest (hydric) with a thick organic layer over mineral soil and a pH range of 6.6 in surface soils to 5.9 in deeper layers. It is dominated by Sedge, Cushion Plant, and Dwarf Shrub species that include Carex stans, Polygonum viviparum and Vaccinium uliginosum plus hummocks of Salix arctica and Dryas integrifolia. The Cassiope Heath site (CH) has an average of 3-5 cm of organic soil over coarse mineral soil with a pH range of 4.9 - 5.4; the site is described as hydric-mesic. Cushion Plant and Dwarf Shrub species at this location include Cassiope tetragona, which dominates, plus S. arctica, D. integrifolia, and Saxifraga oppositifolia. The driest (mesic-xeric) lowland site is Riverside Willow (RW). Sandy mineral soils here range in pH from 5.2 to 4.6, and support the greatest plant diversity of all sites. Deciduous dwarf shrubs and graminoid species are present, primarily Salix arctica and including Festuca brachyphylla. The Upland Granite (UG) and Upland Dolomite (UD) study sites are distinguished by the origin of their mineral soils and this is reflected in their soil pH. UG has an acidic pH range of 4.9 - 5.5, while the alkaline UD has an average pH of 7.9. Both sites are xeric, though UD is somewhat drier, and are dominated by the - 21 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. deciduous dwarf and semi-evergreen shrubs S. arctica and Dryas integrifolia (Klady 2006, personal correspondence, Muc et al. 1989). Site data collected in previous experiments at this study area are summarized in Table 1. 75' 45' 75° 40' 2000 m NOT ALEXANDRA FIORD MAPPED 2 00 R.C.M.P. OUTPOST Sedge Meadow \W ve rs id e . Willow ' Cassiope Heath 60 0 , rE x te n s iv e + Felsen m eer*' y■Ll l— \ <+ + _+ &£££ u| -uU oJ 78* 51* TWIN Figure 2: Layout of the Alexandra Fiord Study area showing the three lowland sites (Sedge Meadow, Cassiope Heath, and Riverside Willow) and the two upland sites (Upland Granite and Upland Dolomite). Contour intervals are 100 m (modified from Sterenberg and Stone 1994). - 22 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Site SM CH RW UG UD T reatment Control OTC Control OTC Control OTC Control OTC Control OTC pHa 6.3 5.2 4.9 5.2 7.9 sw cb 86.3 91.3 37.6 40.0 39.6 39.0 0.120 0.125 0.083 0.070 Air temp (°C)C 8.35 9.81 8.61 8.90 8.48 9.55 n/a Soil temp (°C)c 6.40 7.05 9.16 9.36 9.19 9.66 n/a n/a n/a a pH average o ver upper and low er soil sam ples (Klady, persona correspondence) b approxim ate vo lu m e tric SW C (%) (low land) and g ra vim e tric SW C (g H20/g soil) (upland) averaged o ver grow ing season (data com piled from Rolph 2003) c air tem perature at +10 cm and soil tem perature at -10 cm (SM ) or -2 cm (CH and RW ) averaged o ve r grow ing season (Rolph 2003) Table 1: Individual site data collected during other warming experiments at Alexandra Fiord include soil pH, soil water content (SWC) and both air and soil temperatures. 2.2.2. Experimental design Open Topped Chambers (OTCs) All lowland treatments plots had 1.8m2 hexagonal transparent fiberglass chambers with 0.5m high inclined sides in place where limited non-destructive sampling had occurred since their installation in 1992 (Rolph 2003). OTCs were installed at the upland site the following year, but these sites were faced with an increased exposure to wind and some of the treatment plots were missing a warming chamber, while others had been replaced with a smaller version of the original OTC in 2002 (Klady, personal correspondence June 2006). These structures remained in place throughout the year and did not cause a significant -23 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. difference in SWC when this factor was measured in 2001 (Rolph 2003). Based upon data averaged over the 2001 growing season, OTCs at the lowland sites increased air temperatures by 0.3 to 1.5°C and soil temperatures up to 0.7°C at 10 cm (Table 1). Control plots For all lowland sites, adjacent control plots were 25 paces from their associated OTC in a direction perpendicular to the overall site layout in order to preserve the original random location of each treatment plot. For both upland sites, the same method was used, but was limited to 10 paces due to size constraints of the area. Randomization Random sampling was accomplished with a random numbers table plus a 50 cm X 50 cm quadrat. The quadrat was always placed as close to the center of each OTC as was practical, or was dropped at 25 or 10 paces from the OTC for every control plot. The quadrat was divided into 5 cm X 5 cm squares that easily accommodated the soil corer, and the numbers table was used to choose rows and columns for sampling. Soil sampling Four OTC pairs (one OTC plus one adjacent control plot) were sampled at each of the five sites: three on the lowland (Sedge Meadow, Cassiope Heath and Riverside Willow) and two on the upland (Granitic and Dolomitic). Sampling - 24- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. occurred once at each site over a period of peak plant growth (Rolph 2003) from July 18 to 27, 2004. Six 2 cm diameter soil cores were taken from each OTC and from each of the corresponding control plots. A 45 cm long soil corer was used, and samples were recovered from the top 5 cm (upper) and the bottom 5 cm (lower) layers based upon the depth of the core. This corer adequately represented all sites as the active layer never extended beyond its reach. Upper samples were always taken from the top 0-5 cm of the soil core, while the average depths of lower samples were as follows: Sedge Meadow 39-44 cm, Cassiope Heath 38-43 cm, Riverside Willow 34-39 cm, Upland Granite30-35 cm, and Upland Dolomite 8-13 cm. Cores that were not at least 10cm deep were rejected so that a separation could always be made between the top and bottom 5 cms. Approximately 1 g of soil was taken from each of the six replicates at the two different depths. This yielded 12 x 1 g soil samples from each OTC plus 12 x 1 g soil samples from each control plot. The twenty-four samples from each OTC pair, for four pairs at each of five treatment sites, resulted in 480 x 1 g samples frozen for transport back to UNBC for DNA extraction. 2.2.3. Preservation and transport The 1 g samples remained frozen on site in an underground ‘permafrost refrigerator’ in 2 mL microcentrifuge tubes and in sealed airtight bags until transport to UNBC. Every attempt was made to keep the samples frozen at -20°C during transport and upon return to UNBC until ready for DNA extraction. -25 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.2.4. Field measurements Air and soil temperature, soil water content (SWC), and soil chemical analysis were to be provided by Dr. GHR Henry (Alexandra Fiord HEX Research Site Supervisor, UBC Department of Geography) upon return to UBC in the fall of 2004. Additional soil cores were collected for this purpose from representative plots across all sites during the 2004 growing season. This information was to include (at least) air and soil temperature at +10, -2 and -10 cm from data collected by thermocouples and data loggers at each site, along with SWC, pH, and C:N. Data from the 2004 growing season is currently being compiled and is not yet available as of June 2006 (Klady, personal correspondence June 2006). Data from throughout the 2001 growing season was collected as follows (Rolph 2003): Soil and air temperature at the Sedge Meadow site were measured at -10cm and +10cm, respectively. Hobo® Pro Temperature loggers (H8, Onset Computer Corp. MA, USA) with thermistors connected to Pocket Data Loggers (XR220, Pace Scientific, NC, USA) were employed for this purpose. Thermocouples connected to data loggers (CR-10, Campbell Scientific Inc., UT, USA) were used to collect soil and air temperature data at -10 cm, -2 cm, and +10 cm at both the Cassiope Heath and Riverside Willow sites. Volumetric SWC was measured with Hydrosense™ probes (Campbell Scientific Inc., UT, USA) at lowland sites while gravimetric SWC was assessed by soil cores from upland sites. Measurements were made four times throughout the season at both control and OTCs with the - 26 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. exception of the upland sites; in this case OTCs were sampled only once at the end of the season. 2.2.5. DNA extraction and PCR DNA extractions were performed with the MoBio UltraClean™ Soil DNA Isolation Kit (MoBio Laboratories, Inc., CA, USA) according to the “Alternative Protocol for maximum yields”. Spectrophotometric analysis of extractions revealed final DNA concentrations of 50-150 ng/pl. Each extraction was amplified with two different primer pairs. Functional genes targeted included those that code for the denitrification enzyme nitrous oxide reductase (nosZ), and for the nitrogen fixing enzyme nitrogenase (n/ftH). Degenerate primer pairs were designed, tested and optimized for the genes nosZ (Throback et al. 2004), and nifH (Deslippe 2004). The half-nested nosZ amplification utilized the primers nosZ-F (5’-CG(C/T) TGT TC(A/C) TCG ACA GCC AG-3’) (Kloos et al. 2001 [In Throback et al. 2004]) and nosZ1622R (5-CGC (G/A)A(C/G) GGC AA(G/C) AAG GT(G/C) CG-3’) (Throback et al. 2004) for the primary amplification. The secondary reverse primer Nos1773R (5’-AAC GA(A/C/G) CAG (T/C)TG ATC GA(T/C) AT-3’) (Throback et al. 2004) was labeled with Light Sabre Blue (D4) dye (Synthegen, LLC). Each 30pl PCR reaction contained 3 pL 1:10 dilutions of genomic DNA, 1X PCR Buffer, 0.2 mM dNTPs, 2.0 mM MgCI2, 0.04 pM of each primer, and 0.75 U -27- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Platinum Taq DNA Polymerase (Invitrogen). The secondary PCR mix differed only in MgCh concentration (2.125 mM). Thermocycler conditions were the same for both reactions: a 2 minute denaturation step at 94°C was followed by 35 cycles of denaturing, annealing and extension at 94°C for 30 seconds, 55°C for 1 minute, and 72°C for 1 minute respectively. The final extension required 10 minutes at 72°C. The half-nested n/flH protocol used Nh21 F (5’-GCIWTITAYGGNAARGGNGG-3’) and WidNhR (5’-GCRTAIABNGCCATCATYTC-3’) for the primary PCR reaction (Widmer et al. 1999) and Nh428R (5’-CCRCCRCANACMACGTC-3’) for the second amplification (Deslippe 2004)(sequences follow standard IUPAC notation for mixed bases). The reverse primer Nh428R was labeled with Synthegen Light Sabre Green (D3) dye. Each 31.2 pi PCR reaction contained 4.5 pL 1:10 dilutions of genomic DNA, 1X PCR Buffer, 0.2 mM dNTPs, 2.0 mM MgCb, 0.04 pM of each primer, and 0.75 U Platinum Taq DNA Polymerase (Invitrogen). Thermocycler conditions were the same for both reactions: a 1 minute denaturation step at 94°C was followed by 35 cycles of denaturing, annealing and extension at 94°C for 45 seconds, 53°C for 45 seconds, and 72°C for 1 minute 30 seconds respectively. The final extension required 10 minutes at 72°C. PCR product success and quality was assessed by 1% agarose gel electrophoresis and visualized by staining with ethidium bromide. Bands of expected size (approximately 250bp for nosZ and approximately 400bp for n/7H) - 28- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. were cleaned via ethanol precipitation and then resuspended in pH 8.0 TrisEDTA Buffer. 2.2.6. T-RFLP Restriction enzymes used included Hha\ for nosZ product and Mbo\ for nifH plus corresponding REACT 2 buffer for both endonucleases (Invitrogen). Enzymes were selected based upon number of restriction sites targeted (and therefore fragments generated); a series of endonucleases were tested on replicate samples and those that identified the greatest amount of variation were chosen. For each reaction, 6 pL of PCR product was digested with 2.5 U enzyme and 1X buffer. Digests were incubated at 37°C for at least 5 hours and the reactions were terminated at 65°C for 10 minutes. Digested fragments were desalted by ethanol precipitation and resuspended in formamide. Fragments were prepared for analysis as suggested by the manufacturer for the Beckman-Coulter CEQ™ 8000 Fragment Analysis System (Beckman-Coulter Inc.) for 40pl non-multiplexed samples, although resuspended fragments were not diluted 1:10 prior to loading. For each reaction, 2.5 pL of dye-labeled, digested, desalted product was combined with 37.0 pL of Sample Loading Solution (SLS) and 0.5 pL of 400 bp size standard. Fragments were binned and analyzed in the AFLP program of the CEQ™ 8000 Sequencer (Beckman-Coulter Inc.). Analysis parameters were as per the 400 size standard cubic model with minimum relative peak height set at 1% and a bin -29- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. width of 3 bp. At least four of six field replicate samples from each of two soil depths were analyzed individually for all four treatment and all four control plots from all five sites. Gene community profiles were constructed using only peak heights generated by the sequencer that passed analysis. These peaks represented distinct terminal restriction fragments (TRFs) which in turn corresponded to unique genotypes. Samples that failed due to contamination during fragment analysis or undetected size standard were deleted so that they were not counted with zero-activity samples. TRF frequency was determined for each sample by averaging the binary data for all successful field replicates. For example, if a TRF was present in 3 of 6 sample replicates, its frequency was designated 0.50 for that sample. This made it possible to compare samples with only 4 successful field replicates after PCR to those with 5 or 6. In order to establish the relative abundance of TRFs, the fluorescent signal strength of each peak was relativized to total peak area for each successful field replicate. Once the relative abundance of TRFs was determined for each sample rep, the average was calculated to determine the relative abundance of TRFs for that sample. 2.2.7. Statistical analysis Community composition was investigated graphically with Nonmetric Multidimensional Scaling (NMS) calculated on the basis of a Sorensen distance -30- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. measure. All ordinations were run with PC-ORD 4.0 software (McCune and Mefford 1999) using random starting coordinates. A stepwise reduction in dimensionality was used to minimize stress and 40 runs with real data were accomplished in autopilot mode. Stress is defined as excellent below 5, good below 10, and fair above ten; caution is advised when interpreting ordinations with values approaching 20 (McCune and Grace 2002). Any differences observed in frequency and/or abundance of TRFs were tested statistically with nested permutational multivariate ANOVA (PERMANOVA) in order to examine the effects of Site, Treatment, and Depth (Anderson 2005). A large number of unique permutable units allowed us to establish significance using permutation P-values versus Monte Carlo P-values. Genotypic richness was assessed by comparing differences in the number of TRFs at each site, depth and treatment (Dunbar et al. 2000). Each TRF represents a unique genotype so that an increase or decrease in these numbers reflects an increase or decrease in genotype richness (Tiedje et al. 1999). Nested ANOVA (Statistica 6.0) was used to detect overall significant differences between sites and due to depth or treatment at all sites. Once this was determined, one-way ANOVA was then used to test statistical significance between sites pairwise, and to investigate differences due to depth or treatment unique to each site. For all statistical tests, a significance level of p < 0.05 was accepted. -31 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.3. Results 2.3.1 Gene frequency and relative abundance Site effect TRF frequency and relative abundance was shown to differ overall by site in NMS ordinations of both nosZ and nifH functional genes. Additionally, PERMANOVA suggested significant overall support for an effect of the factor “Site” upon examining the frequency data of both functional genes (p=0.0001). PERMANOVA analysis also confirmed a site effect for the relative abundance data of both nosZ (p=0.0001) and n/YH (p=0.0012). More specifically, PERMANOVA showed significant differences in nosZ TRF frequency between the Sedge Meadow and all other sites: SM and CH (p=0.0269), SM and RW (p=0.0276), SM and UG (p=0.0274), and SM and UD (p=0.0307). Additionally, dissimilarities were detected between CH and both upland sites (UG p=0.0287 and UD p=0.0281) plus RW and both upland sites (UG p=0.0305 and UD p=0.0301). Interestingly, the upland sites also differed from each other (p=0.0285). These differences are illustrated in Figure 3a. An NMS graph (Figure 4) shows nosZ frequency data grouping clearly by Site. It illustrates the unique nature of SM, and the similarities between CH and RW. The recommended three dimensional solution required 400 iterations and resulted in a good final stress of 9.07458 and a final instability of 0.00411. Together Axes 1 and 3 explain 72.4% of the variance, while Axis 2 adds 23.1% (cumulative r2=0.954). - 32- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Specific differences in n/'fH TRF frequency were detected by PERMANOVA between SM and CH (p=0.0265), SM and UD (p=0.0281), CH and RW (p=0.0281), CH and UD (p=0.0285), RW and UG (p=0.0279) and finally RW and UD (p=0.0289). These dissimilarities are illustrated more clearly in Figure 3b. Figure 5 shows n/flH frequency data grouping generally by Site for RW and UD, with weak differentiation of the other three sites. This illustrates the unique nature of UD, but that the other sites are not enormously different from each other. The recommended two dimensional NMS solution required 400 iterations and resulted in a fair final stress of 13.543 and a final instability of 0.00401. Together these two axes explain 92.2% of the variance but the image should be interpreted cautiously due to high stress and instability values. Figures 6 and 7 show only the top 20 most frequent TRFs over all sites, and how they are distributed. The former shows differences in nosZ TRF distribution between all fives sites, while the latter reflects the same for n/flH. Figure 3: Arrows indicate significant differences in the frequency of nosZ (a) and n/YH (b) TRFs between sites. -33 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Site A sm CH RW y UG O UD & £r* Axis 3 (r2=0.275) Axis 1 (1^=0.449) Figure 4: NMS plot of nosZ TRF frequency by Site p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). - 34- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Site A SM CH RW UG O UD A O Figure 5: NMS plot of nifH TRF frequency by Site p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (U G ), and U pland Dolom ite (U D ). - 35- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Terminal Restriction Fragment Figure 6: Distribution of the top 20 most frequent nosZ TRFs over all Sites. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). Sedge Meadow was defined by the presence of a few dominant TRFs; only 6 nosZ TRFs showed a frequency greater than 0.50 (Figure 6). These are present in two main clusters: 196/198/200 and 249/251/252. This site is structured by TRF 198 (0.943), TRF 196 (0.749), TRF 251 (0.724), TRF 252 (0.611), TRF 249 (0.568), and TRF 200 (0.565). A greater number of distinct TRFs were present at Cassiope Heath and Riverside Willow and many were shared between both sites. The most frequent CH TRFs are 251 (0.889) 183 (0.866), 200 (0.802), 203 (0.801), and 198 (0.795). These are similar to those at RW: 251 (0.948), 193 (0.780), 183 and 198 (both 0.734), and 249 (0.715). TRF 251 dominated both upland sites as well (UG, 0.669 and UD, 0.889) but all other UG TRFs were far less frequent: 249 (0.481), 123 (0.455), 198 (0.310), and 223 (0.306). UD also contained TRFs 123 (0.700), 196 (0.651), 198 and 223 (both 0.643). -36- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Terminal Restriction Fragment Figure 7: Distribution of the top 20 most frequent nifH TRFs over all Sites. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). Subtle changes in the distribution of the most frequently present individual nifH TRFs explained some of the differences between sites (Figure 7). nifH TRFs with the greatest freq at the SM were 83 (0.874), 192 (0.846), 141 (0.840), 92 (0.838), and 122/222 (both 0.806) (Figure 7). At the CH, TRF83 (0.888) was also the most frequent, but then TRF distribution differs: 102 (0.698), 141 (0.670), 122 (0.658), and 162 (0.627). RW TRFs present with the highest frequency are 222 (0.895), 83 (0.893), 182 (0.881), 162 (0.874), 141 (0.858). Most frequent at UG is TRF 83 (0.884), followed by 92 (0.838), 192 (0.767), 141 (0.721), and 162 (0.701). The UD site is different from most with a very uniform frequency of common TRFs: 282 (0.935), 122, 182 and 322 (all 0.931), and 222 (0.928). -37- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 9 shows nosZ relative abundance data separating loosely by site and includes an overlay of the most abundant nosZ TRFs. This view of a three dimensional NMS solution explains 76.7% of the variance; axis 3 explains an additional 16.7% (cumulative r2=0.934). Final stress was fair (10.46472) and final instability was 0.00001 after 108 iterations. Pairwise PERMANOVA comparisons confirm dissimilarity between the Sedge Meadow site and all others: SM and CH (p=0.0274), SM and RW (p=0.0304), SM and UG (p=0.0291), and SM and UD (p=0.0267). There were significant differences when the two additional lowland sites were compared individually to the upland sites and when the upland sites were compared to each other: CH and UG (p=0.0309) or UD (p=0.0255), RW and UG (p=0.0278) or UD (p=0.0301), and finally UG and UD (p=0.0310). These relationships are compared in Figure 8. Differences in the distribution of the top 20 most abundant nosZ TRFs over all five sites are shown in Figure 10. Figure 8: Arrows indicate significant differences in the relative abundance of nosZ TRFs between sites. -38- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A Site A 1 a sm CH RW o A A ° A A ▼UG O * 198 A , A A ° Oo OUD o ,205 0 A A A _£00 ^46 _|165 O 0 o |100 'y 193 O 251 ▼ _j103 jjl 42 Axis2 (r^O .290) ▼ 0 oo ▼▼ ▼ 23 ▼ _£43 ▼ ▼ ▼ ▼ ▼ _£23 ^4 ▼ ▼ Axisl (1^=0.477) Figure 9: NMS plot of nosZ TRF relative abundance by Site showing distribution of the 20 most abundant TRFs p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). Crosses mark TRF location and numbers reflect fragment length in base pairs (bp). - 39- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0.5 0.45 0.4 ■ SM 1 0.3 ■S 0.25 j ■ CH | < l > 02 1 □ RW re ■ UD | □ UG j §. 0.05 Terminal Restriction Fragment Figure 10: Distribution of the top 20 most abundant nosZ TRFs over all Sites The Sedge Meadow is clearly dominated by TRF 198 (45.4%) while TRF 251 only makes up 14.4% of the gene community followed by TRF 252 (9.0%), TRF 249 (8.1%), and TRF 196 (7.8%) (Figure 10). TRF 251 is the most abundant at CH (35.1%), RW (37.7%) and UD (28.0%). The next most abundant at CH is TRF 249 (15.8%) followed by TRF198 at 10.5%, TRF 123 (7.4%), and TRF 200 (7.2%). TRFs 249 (17.8%), 198 (17.3%), 123 (7.7%), and 200 (7.0%) complete the RW community. The remainder of the UD site is completed by TRFs 198 (11.5%), 196 (11.1%), 252 (9.4%), and 249 (8.0%). At the UG site, TRF 123 is slightly more abundant than TRF 251 (28.6% vs. 26.8%). Other TRFs with substantial abundance are 249 (12.2%), 223 (6.1%), and 103 (6.0%). -40- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 12 is an NMS graph showing nifH relative abundance data separating clearly by site; it includes an overlay of the most abundant nifH TRFs across all sites. This view of a three dimensional solution explains 51.4% of the variance; axis 3 explains an additional 26.1% (cumulative ^=0.775). Final stress was fair (14.91999) and final instability was 0.00001 after 86 iterations. Pairwise PERMANOVA comparisons confirmed dissimilarity between the Sedge Meadow site and all others: SM and CH (p=0.0314), SM and RW (p=0.0273), SM and UG (p=0.0292), and SM and UD (p=0.0283). In addition, there were significant differences between CH and RW (p=0.0298), CH and UD (p=0.0293), RW and UG (p=0.0313), and finally RW and UD (p=0.0280). These dissimilarities are illustrated more clearly in Figure 11. Figure 13 shows the top 20 most abundant nifH TRFs over all five sites, and how they were distributed. SM CH UD RW UG Figure 11: Arrows indicate significant differences in the relative abundance of nifH TRFs between sites. -41 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. o Site A SM CH RW o V ug <>UD 384 + A A 382 + 257 + A A Axis 2 (1^=0.336) A A A A ▼ 296 A 222 + A 292 + 192A 92 + A ° A 282 + + , 260 141T - M l 82 16^. A 299 242 + 83 78 + +2^0 801+ + 278 Axis 1 (1^=0.178) Figure 12: NMS plot of n/7H TRF relative abundance by Site showing distribution of the 20 most abundant TRFs p=0.0196. Sites are: Sedge Meadow (SM), Cassiope Heath (CH), Riverside Willow (RW), Upland Granite (UG), and Upland Dolomite (UD). Crosses mark TRF location and numbers reflect fragment length in base pairs (bp). - 42- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 0.5 0.45 --- cu o O' 0.4 0.35 ■ SM 0.3 ■ CH 0.25 □ RW 0.2 □ UG i 0.15 ■ UD j 0.1 0.05 0 I gO'' 9^ 'S ' - V id , i . 9 h i Etn fffli <■' < y ^ ^ ^ cp*' ^ ciy t?'' ^ JBfU. M G^ ; r ' ...... 0.8 < D 0.7 O C (0 -a 0.6 c 0.5 ■ Upper n3 < 0.4 a > > ■ Lower 0.3 & 0.2 0.1 0 A%A r A< 12 Fig 30: Mean number of nosZ genotypes (TRFs) at each site (p<0.001, confidence interval = 0.95). - 62- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Significant differences were detected overall in the number of nosZ genotypes between upper and lower samples (Figure 31). With the exception of the UG site, there was greater genotype richness in surface soils. One-way ANOVA confirmed differences in genotype number at SM (7.3 vs. 5.2, p=0.0078) and UD (16.8 vs. 12.4, p=0.0019). 3 10 UD UPPer 151 Lower Fig 31: The mean number of nosZ genotypes (TRFs) differed over all sites by depth (p=0.018, Cl=0.95). Significant differences at individual sites are marked with an asterisk (p<0.01). -63- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Significant differences were detected in the number of nosZ genotypes between control and OTC plots (Figure 32). In general, there were fewer genotypes in OTC treatment samples. One-way ANOVA confirmed differences in genotype number at SM (7.1 vs. 5.2, p=0.018), UG (6.7 vs. 4.7, p=0.016) and UD (16.1 vs. 12.9, p=0.023). c/> CD Q. O c CD 05 CD n E D C C CD CD UD ~<5~ Control 3 1 OTC Fig 32: The mean number of nosZ genotypes (TRFs) differed by treatment over all sites (p=0.039, Cl=0.95). Significant differences at individual sites are marked with an asterisk (p<0.05). nifH Significant differences in the number of nifH genotypes were detected overall between sites with ANOVA tests (Figure 33). Pairwise investigation showed that - 64- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. richness was much lower at the CH site (15.2) versus SM (21.6, p=0.0001), RW (23.1, p<0.0001), and UD (21.4, p=0.0001), while the UG site displayed intermediate richness relative to the others (18.0). Mean number of genotypes at both upland sites were significantly lower than those at the RW (vs. UG, p=0.0001 and vs. UD, p=0.0068). Nested ANOVA, in addition to showing overall site differences, described significant overall depth (Table 4) and treatment (Table 5) effects. Table 4: Nested ANOVA showing overall significant differences in number of nifH TRFs between sites (p<0.000001) and between depths across all sites (p=0.000001) DF MS P F | Intercept 73795.51 1 73795.51 1152.939 0.000000 2928.94 4 732.23 11.440 0.000000 f Site Depth(Site) [2444.07 5 488.81 7.637 0.000001 Error 18881.89 295 ss o CD Table 5: Nested ANOVA showing overall significant differences in number of n/7H TRFs between sites (p<0.000001) and between treatment across all sites (p<0.000001) f SS [ DF MS F P Intercept 111240.5 1 111240.5 [1755.050 0.000000 Site 2640.1 ]4 [660.0 (10.413 0.000000 Treat(Site) (2627.9 |5 525.6 [8.292 0.000000 [ I Error 18698.0 295 63.5 -65- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sfl 22 Fig 33: Mean number of nifH genotypes (TRFs) at all sites (p<0.001, confidence interval = 0.95). Significant differences were detected in the number of nifH genotypes between upper and lower samples (Figure 34). Unexpectedly, there were more genotypes in lower soil samples at the CH and at both Upland sites. There was significantly greater genotype richness in surface soils at SM (27.3 vs. 16.3, p<0.001) and RW (25.1 vs.21.5, p=0.012). - 66 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 32 30 28 w 26 CD Q. >. o 24 c 0 £22 1 18 ^ 16 00 0 14 12 10 j £ I Upper ~5~ Lower Site Fig 34: The mean number of nifH genotypes differed by depth over all sites (p<0.000001, Cl=0.95). Significant differences at individual sites are marked with an asterisk (p<0.05). Significant differences were detected in the number of nifH genotypes between control and OTC plots (Figure 35). With the exception of the UG site, there were fewer genotypes in OTC treatment samples. One-way ANOVA confirmed differences in genotype number at SM (25.4 vs. 17.4, p=0.0029), RW (25.3 vs. 20.7, p=0.0012) and UG (14 vs. 23.6, p=0.0006). -67- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. $ Control Z5Z OTC Fig 35: The mean number of nifH genotypes differed by treatment over all sites (p<0.000001,0=0.95). Significant differences at individual sites are marked with an asterisk (p<0.005). - 68- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.4. Discussion 2.4.1. Methodological considerations We are assuming that each terminal restriction fragment (TRF) reflects a different bacterial genotype so that an increase or decrease in the frequency, abundance or overall number of these TRFs is a valid measure of genotypic frequency, abundance and/or richness (Tiedje et al. 1999). We also assume that uncut gene fragments TRF252 (nosZ) and TRF384 (nifH) mean that no restriction site exists in these genotypes. It is possible that more than one genotype corresponds to one restriction fragment, including the uncut fragment, so that genotypic diversity based upon the number of TRFs could be underestimated (Dunbar et al. 2000; Wolsing and Prieme 2004). However, there is no reason to believe that any underestimate would be biased according to site, depth or treatment, so even if genotypic diversity were underestimated, comparisons should remain valid. Soil cores were retrieved from plots that had undergone experimentation for over a decade (Freedman et al. 1994). The destructive nature of soil sampling could result in disturbance that affects microbial communities. However, Deslippe et al. (2005) conducted disturbance experiments at Alexandra Fiord sites and confirmed that previous soil excavation had no effect on nitH community structure, so we would not expect this to be a factor at our study sites. -69- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The Sedge Meadow soils had high organic matter content and DNA extractions required 1:100 dilution in order to overcome the interference of humics in PCR reactions of both functional genes. A humic acid purification step may have allowed amplification without dilution, but the process can result in decreased product yield (Brodie et al. 2002).The Upland soils were very dry with little organic matter and although consistent PCR results were obtained for nosZ, few or no nifH amplification products were retrieved from lower samples, even with multiple attempts. We assume that this was not a failure in our methods, but that the number nifH genes were often below detectable limits or simply not consistently present in samples from this depth at UD. 2.4.2. Did the denitrifying and nitrogen-fixing communities respond similarly or differently to experimental factors? The overall differences between sites were similar for both nosZ and nifH, but were not as clearly defined for nitrogen fixers, especially when assessed by gene frequency. Where nosZ TRF frequency was distinct between sites due to one or a few dominant genotypes, nifH TRFs were shared across all sites with small changes in the frequency distribution. In contrast, differences in the relative abundance of shared nifH TRFs were more apparent between sites. Subtle changes in the frequency and abundance of nifH TRFs between sites were consistent with findings of uniform g e n e distribution and identical sequences despite large geographic distance found by other researchers (Poly et al. 2001, Rosch et al. 2002). Alternately, enormous variation in nosZ TRF distribution at distances of centimeters, meters and especially kilometers has been found in -70- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. marine studies suggesting that denitrifier genotypes are unique at each location (Scala and Kerkhof 2000). For both genes, the SM site stood apart from the other four, but nosZ genotype richness was very low where nifH richness was high. There were no significant differences between CH and RW (nosZ) or UG and UD (nifH). nosZ genotype richness was high at both CH and RW, but nifH richness was low at CH. Alternately, nifH genotype richness was intermediate at both UG and UD, but nosZ richness was very low at UG. nosZ genotypic richness at each site seemed to affect the number of genotypes at other levels. For example, sites with low or moderate nosZ richness (SM, UG, and UD) showed significant or unexpected changes in the number of TRFs at two depths and between treatments. These homogeneous sites have unique ecological characteristics that only a few, dominant nosZ genotypes can exploit (anaerobic at SM, acidic at UG, and dry at UD). As a result of this, the community changed enormously over vertical microsites and with warming. Sites with high nosZ richness (CH, RW) showed little or no significant change at other levels. This suggests that sites with high nosZ richness and uniform vertical gene distribution are less likely to experience a change in community structure due to disturbance. This could be due to the fact that the denitrifier communities at CH and RW, due to the heterogeneous nature of these sites, already contain members that are able to exploit all reasonable niches on the lowland, including those that have changed due to OTC treatments. -71 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. nifH genotype richness at each site seemed to have some influence on number of genotypes at two depths or due to treatment, although the effects were not consistent with those described for nosZ. For example, sites with low or moderate nifH richness (CH and UG) showed either no significant changes or showed differences in richness that were not predicted. Sites with a high number of nifH genotypes (SM and RW) showed significant changes due to depth or treatment. This suggests that the majority of these genotypes were rare and able to exploit only a narrow environmental niche. Both nosZ and nifH genotypes were affected overall by depth, particularly at the SM site where each community experienced a significant decline in genotype richness in lower samples. At SM, the frequency of nifH genotypes clearly shifted among common TRFs; changes in the most abundant nifH TRFs show that one distinct genotype structured upper or lower samples. The frequency of nosZ TRFs was not significantly affected at SM, but the relative abundance of TRFs also shifted dramatically to one dominant genotype in lower samples. There was a significant decrease in nosZ genotype richness in lower samples at the UD site, but this could not be confirmed for nifH as limited PCR amplification of lower samples caused enormous variation in the nifH data. Where there was a shift in the distribution of both frequent and abundant nosZ genotypes between upper and lower samples at the UD site, there was, in addition to a shift, a loss of nifH genotypes from lower samples at the UG site. Interestingly, there was a corresponding gain in the mean number of nifH genotypes in lower samples at -72- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. this site, plus a loss of genotype richness in lower samples from the RW site that was not detected in NMS ordinations. The presence or absence of an alternative denitrification marker (nirS) has been used to investigate changes in community structure in ocean sediment cores based upon the transition to an anaerobic environment as depth increased (Braker et al. 2001). Only a slight decrease in denitrifier diversity was detected in deeper samples even though a strong redox gradient was present suggesting that the microbial community was not vertically structured. Enzyme analysis and gene probing of forest soils does suggest that denitrifier abundance decreases with depth even though the deepest samples (-25 cm) had very low levels of oxygen and should have selected for denitrifiers (Mergel et al. 2001, Rosch et al. 2002). The SM soil cores were uniformly anaerobic throughout, so any differences detected could have been due to the change from organic to mineral soils and the subsequent changes in soil chemistry and nutrient availability. The UD soils were uniformly mineral, but a shallow rhizosphere suggests a lack of nutrient availability beyond 8 cm from the surface. Studies of N-poor acid forest soils have indicated a distinct difference in nifH RFLP patterns between litter layers and soil samples (Widmer et al. 1999), and a decrease in pattern complexity from shallow to deep soil samples (Shaffer et al. 2000). In some cases genes appeared to be more abundant in litter versus soil, although the gene marker was shared by both samples, in others the genotype -73 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. was simply not present in the soil sample. Gene probing of forest soils did suggest that nifH was more abundant in the top 5cm than in deeper samples (Mergel etal. 2001, Rosch etal. 2002). There was a significant overall difference in both nosZ and nifH genotype richness between control plots and OTCs. In general, nosZ genotype richness decreased with warming treatments at all sites, but the change was only significant at SM, UG, and UD. The mean number of nifH TRFs also decreased significantly with treatment at SM and RW, and was lower, though not significantly, at CH and UD. We did not expect to see the significant increase in genotype richness with treatment at UG. If the soils of Alexandra Fiord are N limited (Muc etal., 1994), and OTCs effectively raise soil temperatures enough to promote organic matter decomposition and net N mineralization (Berendse and Jonasson 1992, Paul and Clark 1996, Nadelhoffer etal. 1992), the increase in nutrient availability would affect denitrifier and diazotroph communities differently based upon their physiological requirements. Direct increases in both denitrification and nitrogen fixation are possible due to higher enzymatic activity at increased temperatures (Chapin and Bledsoe 1992, Paul and Clark 1996). Additionally, anaerobic denitrifiers and aquatic nitrogen fixers should flourish in the wetter environment that is predicted but not accomplished with OTCs. Photosynthetic diazotrophs, such as the cyanobacteria that dominate arctic systems (Chapin and Bledsoe - 74- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1992, Liengen 1999, Paul and Clark 1996), should also benefit from increases in C 0 2 production associated with OM turnover (Chapin and Bledsoe 1992). It is possible that the same conditions that promote microbial growth in these changed systems may also inhibit the processes mediated by them. For example, net N mineralization is predicted, but N-fixation is inhibited by the presence of available N (Paul and Clark 1996). It is possible that the N-fixation process and diazotroph community structure are unrelated (Piceno and Lovell 2000a), but in theory an N rich habitat should result in low nifH diversity (Zehr et al. 2003). Alternately, N H / is required by nitrifying bacteria that oxidize it to NO3 , the substrate for denitrification (Paul and Clark 1996, Nadelhoffer et al. 1992). Any potential increases in this process will be moderated by the success of aerobic nitrifier communities in an environment predicted to be wetter (Nadelhoffer et al. 1992). If excess NO3' is the result, then anaerobic soil conditions greater than 5 °C, with a pH range of 6 - 8 should be ideal for denitrification to occur (Paul and Clark 1996). Additionally, changes in the proportional abundance of dominant TRFs have been detected between different levels of denitrification activity in two different vegetation types (Rich et al. 2003), although this finding was not repeatable (Rich and Myrold 2004). Recent studies have not confirmed that net N mineralization will be affected by short term increases in soil temperature of a few degrees Celsius (Jonasson et al. 1999, Schmidt etal. 1999). Rolph (2003) could not show that an increase in -75 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. mineralization had occurred due to long term OTC treatment at the Alexandra Fiord site although data from the 2001 growing season indicate that soil temperatures were raised by less than one degree Celsius. 2.4.3. What experimental factors most influenced community structure? Site and depth were more important than treatment in structuring both denitrifying and nitrogen-fixing microbial communities; of the first two factors, site was more important than depth. Site The sites vary in both biotic and abiotic properties, which results in homogeneous or heterogeneous environments with a range of substrate availability. These dissimilarities combine to create the unique nature of SM, the commonalities between CH and RW, and the divergent nature of the upland sites from those on the lowland and from each other. As discussed, the community structure of each functional gene responded differently to these changes reflecting the requirements of either denitrifier or nitrogen fixer groups. The SM site is unique within the study area due to the hydric and often flooded conditions. The dominant sedge vegetation created a distinctive nosZ community in terms of gene frequency, abundance, and richness. Low nosZ genotype numbers at this site may be a function of homogeneous conditions that do not necessitate a diverse denitrifier community to exploit them (Callaghan et al., - 76- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2004). Alternately, high C:N plus anaerobic conditions may not favour nitrification and thus limit the substrates required for denitrification (Nadelhoffer et al. 1992). nifH genotypic richness was high at the SM site where diverse, aquatic cyanobacteria may be dominant (Callaghan et al. 2004, Chapin and Bledsoe 1992, Zehr et al. 1998) and where there was sufficient C plus N-limitations due to a large amount of poorly decomposed plant material (Klady, personal correspondence June 2006). It is not uncommon to find diverse nifH communities in anaerobic environments (Ueda et al. 1995, Zehr etal. 1998). CH and RW shared many nosZ genotypes in common, and grouped together when gene frequency and abundance were examined. Though the RW has the highest plant diversity of all sites in the study area, nosZ richness was similar between CH and RW; comparable soil moisture regimes, common plant species, and similarly low soil pH could contribute to these parallels. It is possible that greater variation in the soil environment at these sites has required a greater variety of nosZ in order to exploit all opportunities for denitrification. The observations of Brodie et al. (2002) based upon bacterial communities in disturbed grassland, and Stres et al. (2004) based upon denitrifiers in forest soils were not mirrored at CH or RW as denitrifier richness was relatively high despite a mixed plant community and acidic soils. Low nifH genotype richness at CH could not be explained by soil moisture, organic matter content, or pH as the site has intermediate values of these factors between all three on the lowland. The hydric SM soils have an average pH of 6.25 while the mesic-xeric soils of RW - 77- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. have an average pH of 4.9; CH is hydric-mesic with an average pH of 5.15. The RW site had the greatest number of nifH TRFs due possibly to the influence of high plant diversity. Variable soil conditions due to plant feedbacks at this site may also have promoted the development of a diverse nifH community. The Upland sites are remarkably different from those on the lowland due to the dry nature of their soils and relatively sparse vegetation. They share a few plant species in common, but differ greatly in soil pH; the granitic site averages pH 5.2 and the dolomitic site is much more alkaline at pH 7.9. The acidic nature of the UG soils could have contributed to the lack of nosZ genotype richness when compared to the adjacent UD site, although low pH did not structure nosZ communities on the lowland. These observations suggest abiotic properties of the dry soil environment may have had more influence on the structure of upland nosZ communities. Soil pH may also explain lower nifH richness at UG when compared to adjacent UD as the plant communities, mineral soils, and moisture regimes are similar at these sites. Past experiments have used frequency and abundance of the TRFs from other denitrification gene markers to explore diversity across spatial scales (Braker et al. 2000, 2001). They confirmed that denitrifier gene communities were unique to the environment from which they were sampled, with only a few genes shared between locations. Similarly, nifH TRFs that were found to be dominant at one - 78- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. site were often less frequent, or less abundant, or not present at all at adjacent sites with different plant and soil characteristics (Shaffer etal. 2000). Meadow soils with lower C:N and higher pH than adjacent forest plots have demonstrated distinct nosZ communities (Rich et al. 2003). Few TRFs were shared between the meadow and forest sites which differed in predominant vegetation as well as N availability. An increase in nosZ diversity in disturbed, amended soils was also related to higher pH than adjacent native plots (Stres et al. 2004). Brodie et al. (2002) found that the number of bacterial TRFs increased as the grassland community became more uniform and as both pH and N, P, and K increased. Other groups have noted that bacterial community diversity decreases when the environment is more homogeneous (Callaghan et al., 2004). Higher pH and available N seem to be more consistent factors associated with high microbe diversity than site vegetation. Moreover, it appears that denitrifier communities are structured by not only the amount of N available, but also whether it is in organic or mineral form (Wolsing and Prieme 2004). Acidic forest soils showed less abundance and diversity of denitrification genes than adjacent marsh soils with high organic matter content (Prieme et al. 2002). This contradicts the finding of low genotypic richness of nosZ in the thick organic and alkaline soils of the Sedge Meadow site, where the anaerobic conditions should be an advantage for denitrifiers. That this site stood out from all others in - 79- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. terms of frequency and abundance of TRFs is consistent with literature that suggests nosZ is habitat specific (Rich et al. 2003, Rich and Myrold 2004, Rosch et al. 2002, Stres et al. 2004), but the adjacent lowland sites are drier and more acidic, yet show greater nosZ richness. On the lowland, it appears that pH and organic matter content are not strong drivers of high nosZ richness, but it is important to note that large seasonal shifts have been detected in denitrifying gene communities (Wolsing and Prieme 2004) and that this study examined samples taken only once during the summer growing season. nifH gene communities are also known to vary over large and small spatial scales (Poly et al. 2001, Rosch et al. 2002) but the factors that control this variation are complex. Community structure is related to soil management and soil texture that influences inorganic N availability (Poly et al. 2001). nifH communities are habitat specific (Shaffer et al. 2000, Zehr et al. 1998) and most similar when their site characteristics are the same but neither plant cover nor soil chemistry could completely explain this (Poly et al. 2001). Depth Depth of the sample (i.e. upper versus lower section of the soil core) was also an im portant factor differentiating nosZ and nifH communities. Although both genes were affected, this separation was not detected at all sites, nor was it consistently found at treatment and OTC plots simultaneously. - 80- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. nosZ relative abundance data showed dissimilarity between upper and lower samples at Sedge Meadow OTCs only. One SM nosZ TRF was most abundant at both depths but its relative contribution was much greater in lower samples. nosZ data for the Upland Dolomite site showed that the distribution of most frequent TRFs changed between depths at both control plots and OTCs but the most abundant nosZ TRFs were different only at control plots. Overall, upper samples appeared to have a more uniform distribution of nosZ TRFs, while lower samples were dominated by one (TRF 251). nifH frequency and relative abundance also varied between depths at the SM site; differences were detected at both control plots and OTCs. Each depth was dominated by one particularly abundant genotype and had a unique distribution of less abundant TRFs. Upland Granite nifH frequency and relative abundance differed at control plots only; overall, the most frequent or abundant TRFs in upper samples were not represented at all in the lower soils. In general, there was greater genotype richness in surface soils. Significant differences were detected in the number of nosZ genotypes between upper and lower samples at SM and UD and in the number of nifH genotypes between upper and lower samples at SM and RW. The SM site had a large range in terms of depth between upper and lower soil samples (0-5 cm deep and 39-44 cm deep, respectively) but this did not differ -81 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. greatly from other lowland sites. More importantly the soil core itself was not as uniform as those from CH and RW; the upper section was wet organic peat, while the lower part of the core had a greater proportion of mineral soil. Differences in the relative abundance and genotypic richness of nosZ TRFs between upper and lower samples reflected the distinct habitats at either end of the SM soil core. UD cores, in contrast, were very shallow and uniform; the lower samples were only 813 cm from the surface. It is interesting that differences were found even though soil samples were in relatively close proximity. This site was the driest in the study area and it is possible that there was little interaction between surface and subsurface soil layers. Differences in nifH gene frequency and abundance between upper and lower samples at the SM, plus differences in genotype richness at SM and RW may be partly explained by the vertical distance between sample locations as explained for nosZ. Additionally, a greater proportion of cyanobacteria would be expected in surface samples due to their phototrophic mode of nutrition (Paul and Clark 1996). The UG site had slightly more moisture than the adjacent dolomitic site, and the depth of soil cores averaged 30-35 cm. The distance between upper and lower samples in a relatively dry soil with a shallow root zone may have contributed to the differences in gene frequency and relative abundance detected. A unique community of organotrophic N fixers would be found adjacent to plant roots versus the soil beyond the influence of plant inputs (Paul and Clark 1996). - 82- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Treatment The main objective of this study was to detect shifts in denitrifying and nitrogen fixing microbial communities after long term warming of arctic soils. Soil warming is a factor that can structure microbial communities. Increases in soil moisture and organic matter availability due to warmer temperatures in arctic systems may lead to higher rates of decomposition and nutrient cycling (Berendse and Jonasson 1992, Paul and Clark 1996, Nadelhoffer et al. 1992). It is important to note that increases in arctic air temperatures predicted by climate models and accomplished with warming experiments do not necessarily translate into proportional increases in soil temperatures (Jonasson 1993, Robinson 2002). At times, only the surface soils are affected by direct heating unless the site is very wet, and shading due to increased plant growth can prevent warming of the soil (Rolph 2003). This means that although enzymatic processes in soils may sometimes increase due directly to higher temperatures, and biogeochemical cycling may be stimulated by increases in substrate availability, there are indirect changes caused by warming that also contribute to changes in the soil environment. Warming can cause shifts in plant, fungal and protozoan communities that further impact microbial communities. Favourable conditions that increase the abundance and activity of soil protozoa can result in increased grazing of soil microbes, directly affecting their number and distribution (Ruess et al. 1999). The -83 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. biomass of both saprotrophic and mycorrhizal fungi is positively affected by moderate increases in temperature and moisture, and changes in the community composition and function of fungal decomposers has been noted (Robinson 2002). Experiments have indicated that increases in fertilization and/or temperature can lead to greater plant productivity (Arft et al. 1999, Muc et al. 1994, Shaver et al. 2001) but loss of diversity (Chapin III et al. 1995). Passive warming experiments that raise air temperatures by 1 to 3°C resulted in decreased plant diversity and species evenness at a number of HEX sites (Walker et al. 2006). Complex feedbacks to the soil environment due to shifts in plant and fungal community structure depend on changes in carbon allocation (Piceno and Lovell 2000b), species composition and litter quality (Shaver et al. 2001), and nutrient use efficiency (Berendse and Jonasson 1992), driven by long term changes in N mineralization and availability (Chapin III et al. 1995, Rolph 2003, Schmidt et al. 2002). Changes in N mineralization and availability were investigated after ten years of warming at the Alexandra site (Rolph 2003). Net N mineralization was not affected by warming treatments, but inorganic N availability was higher in OTCs throughout the growing season except at the upland dolomitic site and NH4+ increased with warming at the heath and willow sites (Rolph 2003). Contrary to expectations, warming had the least effect of all factors on denitrifier and nitrogen fixer community structure. There was no separation of samples due to treatment detected at any site for either nosZ or nifH frequency or relative - 84- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. abundance data. It is possible that NMS ordinations and PERMANOVA tests were not sensitive to changes in the frequency and/or abundance of rare genotypes. When genotypic richness was explored by comparing the number of TRFs between control plots and OTCs, a treatment effect was evident. Overall, it appears that rare species were lost and dominant species prevailed when disturbed by OTC treatments. In general, there were fewer genotypes in treatment samples except for n/YH TRFs at the UG site. The large, significant increase in the number of genotypes at UG OTCs is surprising since Rolph (2003) detected higher inorganic N at this upland site with warming suggesting that a diverse diazotroph community is not required. Alternately, the loss of genotype richness was significant for both genes at SM where Rolph (2003) found the greatest response to warming when nitrogen transformations were compared between control plots and OTCs. Specifically, both i s la n d N immobilization increased at SM OTCs supplying more substrate for denitrification and suggesting an increase in microbial growth (Paul and Clark, 2006). It is possible that although warming at the SM created ideal denitrification conditions, only a few genotypes (from an already limited number) could exploit the altered environment. The nitrogen fixing community, although initially genotype rich at this site, experienced a large decrease in the number of genotypes suggesting an abundance of rare genotypes that are sensitive to disturbance. Significant TRF losses due to OTC disturbance were - 85- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. also detected by changes in the number of nosZ genotypes at UG, and UD, and in the number of nifH genotypes at RW. Long term disturbance of forest soils due to clearcutting leads to distinctly different nifH communities with new dominant TRFs and losses of previously dominant genotypes; additionally, these communities exhibit atypical seasonal variation (Shaffer et al. 2000). Based upon previous work, the authors suggested that a loss in diversity of nifH genotypes reduces the capacity for nitrogen fixation in disturbed systems. Deslippe (2004) also noted that nifH communities disturbed by warming experiments are more seasonally variable; we presume then, that although these communities are very adaptable, they are also less stable. Disturbance of marsh soils by long term fertilization leads to changes in nifH community composition and absence of genotypes due possibly to the loss of competitive advantage over other N-limited microbes (Piceno and Lovell 2000a). This artificial change in nutrient availability mirrors the changes predicted by long term soil warming and may explain why we observed a loss of genotype richness with OTC treatments. Deslippe et al. (2005) suggested that ni1H communities were structured by warming late in the growing season, and NMS ordinations showed differences between control plots and OTCs at lowland sites. Although samples for this investigation were collected at approximately the same time of year (July 18-27) at adjacent plots in the same study area, we did not detect a similar shift in nifH - 86- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. gene frequency using the same statistical methods. This suggests that there may be seasonal shifts in these communities that are detectable with multiple sampling times over the growing period, but that the changes do not persist from year to year. In this case, short term changes due to dynamic soil processes such as N transformations may be more of a factor than long term community shifts due to persistent changes in soil chemistry from altered plant inputs (Shaver, 2000). This has not been found experimentally for nifH] the seasonal community structure appears to be very stable despite repeated sampling over weeks to months (Poly et al. 2001, Shaffer et al. 2000, Widmer et al. 1999), even with increased nitrogen (Piceno and Lovell 2000a) or reduced carbon availability (Piceno and Lovell, 2000b). In contrast, enormous seasonal variation has been detected in nosZ gene markers from ocean sediment samples measured over the course of one year (Scala and Kerkhof 2000) and studies of soil thawing effects on denitrifying communities have demonstrated that short term community shifts can also be detected by measuring gene expression with microbial RNA (Sharma et al. 2006). Sharma and colleagues (2006) found that the diversity of an alternate denitrification gene marker (nirS) appeared to increase as thawing progressed and microbial activity was stimulated; although nosZ DNA was also present, no nosZ transcripts were detected as N2 O accumulated. It has been suggested that seasonal and/or geographic temperature variation can structure nitrifier communities either directly or by affecting N availability (Avrahami et al. 2003, - 87- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Avrahami and Conrad 2003). It would seem that the successful detection of seasonal variation in nitrogen cycling gene community structure (due directly or indirectly to temperature manipulations) depends on the timing and frequency of sampling, plus the molecular technique and gene marker employed. Net N mineralization is greater in wet sedge-dominated tundras than in other arctic ecosystems and has been shown to increase with warming, subsequently doubling the inorganic N pool and resulting in an excess of mineral N but this increase appeared to benefit the plant community more than it did the microbes (Schmidt et al. 2002). This explains why even though the microbial community experienced a strong disturbance effect at SM, gene richness was not affected positively by the increased nutrient availability associated with OTC treatments. Other subarctic studies have shown no significant increases in net N mineralization with experimental soil warming of 1-2°C (Jonasson et al. 1993) and only moderate average increases with soil temperature changes of 0.3-5.1°C (Rustad et al. 2001). Arctic mineralization rates could more than double if soil temperatures reach or exceed 10°C especially where tundra soils have high organic matter quality (Nadelhoffer et al. 1991). Interestingly, no significant changes in gene richness were detected at CH where average soil temperatures measured in 2001 at Alexandra Fiord (Rolph 2003) were increased by only 0.2°C with OTC treatment. TRF loss was noted only for n/7H at RW where the same data shows an average increase of 0.5°C and genotype richness decreased for both genes at SM where the greatest temperature increase was recorded - 88- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (0.7°C). Hollister et al. (2006) also found that OTC effect was variable at different tundra sites, and differed greatly from day to day and from year to year. Average soil temperatures at -10 cm in July ranged slightly from 0.3°C to 0.6°C at dry heath sites, and enormously from -0.8°C to 0.7°C at wet sedge sites. Increases in net mineralization and the subsequent availability of NH4 are highly dependent upon rates of microbial immobilization (Jonasson etal. 1993, Rustad et al. 2001), and will affect soil nitrogen cycles such as nitrification, denitrification, and nitrogen fixation. Nitrogen fixation should be inhibited by the presence of its products NH4+ and NO3' (Paul and Clark 1996), but rates were not affected by the presence of nitrate in field samples of high arctic cyanobacteria (Liengen 1999). It is known that nitrification and denitrification are stimulated by the availability of these substrates (Nadelhoffer etal. 1992, Nicolaisen etal. 2004), and that the amount of NH4-N can be far greater than NO3-N during the growing season in arctic soils (Schmidt et al. 2002). Although enzymatic processes are accelerated in general by increases in temperature, it is possible that a warmer soil environment will become more favourable for denitrification but select against nitrogen fixers (Paul and Clark 1996). -89- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2.4.4. What measures of community structure are most affected by environmental factors? We measured diversity in three ways: genotype richness, genotype frequency and genotype relative abundance. Although genotype richness was the coarsest measure, it was the most sensitive to rare genotypes, and the only one that was able to distinguish a treatment effect. Relative abundance measures are confounded by the presence of rare species because they create noise in the data, and frequency measures are biased towards evenly dispersed sample units (McCune and Mefford 1999, McCune and Grace 2002). We propose that NMS ordinations and PERMANOVA tests did not detect changes in rare genotypes or clustered individuals because, in these analyses, the data are equally weighted and more strongly influenced by dominant TRFs (McCune and Grace 2002, Anderson 2005). By measuring the differences in absolute number of TRFs (richness) we saw how the contribution of all genotypes was affected by site, depth, and treatment (Dunbar et al. 2000). Genotype richness is strongly affected by the environmental factors that make each site unique. The mean number of nosZ TRFs ranged from only 5.8 at the Upland Granite site to 18.8 at the Cassiope Heath site while the mean number of nifH TRFs also varied between sites, from 15.2 at Cassiope Heath to 23.1 at Riverside Willow. Depth also affected both nosZ and nifH richness. Significantly - 90- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. fewer genotypes were found in lower samples at SM (both genes), UD (nosZ), and RW (n/'flH). A significant warming effect on gene richness was detected at three of five sites for both microbial communities. The number of nosZ genotypes decreased with treatment at SM, UG, and UD while n/'fH richness declined at SM and RW, but increased at UG. Genotype frequency was not as sensitive as relative abundance to differences between sites, but it did give us more information than genotype richness about the presence or absence of individual TRFs. For example, although the number of different genotypes may have been low in each sample from one site (resulting in low mean number of genotypes for that site), the frequency data showed us that when pooled, many more TRFs were actually represented. Because the most frequent TRFs (ie. the top 20) were often at least present across all sites, the differences between sites is due to small changes in the frequency distribution of common genotypes. The relative abundance data shows more clearly how each site is structured by one or a few dominant TRFs. The dominant TRF(s) plus a unique distribution of other genotypes in very low abundance create(s) a distinct community at each site. Genotype richness, frequency and relative abundance were all sensitive to sample depth, but not at all sites, and not to the same degree. There was a significant decrease in the number of genotypes with increasing depth at two of -91 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. five sites for each gene. Interestingly, a loss of approximately two nosZ genotypes at SM was not detected as a change in frequency, while a loss of approximately four nosZ genotypes at UD was documented in frequency ordinations. All of the top 20 most frequent UD nosZ TRFs were represented in upper and lower samples, but their distribution was altered. A significant loss of n/'flH genotype richness was detected at SM and RW, but the change was fewer than four TRFs at the latter site and was not reflected in the frequency data. Instead, the frequency distribution of the top 20 SM and UG n/'flH TRFs was shown to have shifted between upper and lower samples. Overall, n/'flH TRF frequency was uniform across upper samples and variable in lower samples; at the UG site, some dominant TRFs were no longer present with increased depth. Depth affected nosZ relative abundance data at both the SM and UD site. Unlike the frequency data, the distribution of TRFs did not simply shift among the top 20 Upper samples were initially more uniform, but there remained only one or a few dominant TRFs in the deeper samples while the rest were lost. The relative abundance of n/'flH TRFs was structured by depth at both the SM and UG sites. At SM, none of the top 20 TRFs were lost, but two very different genotypes were dominant in upper versus lower samples. Two different pairs of TRFs dominated either upper or lower samples from UD. - 92- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter 3. Summary 3.1. Conclusions The objective of this study was to detect shifts in denitrifying and nitrogen fixing soil microbial communities by measuring changes in functional gene frequency, abundance and/or genotypic richness using the genetic markers nosZ and n/'flH respectively. The study area encompassed five high arctic sites that differed by dominant plant community, parent material and/or soil moisture and that had been subjected to a long term warming experiment. We investigated differences in these gene communities due to site, depth, and treatment (warming). The denitrifying and nitrogen-fixing communities respond both similarly and differently to experimental factors The overall differences between sites were similar for both nosZ and n/'flH, but were not as clearly defined for nitrogen fixers. nosZ communities were more distinct between sites due to one or a few dominant genotypes while n/'flH TRFs were shared across all sites with subtle changes in their distribution. For both genes the SM site was unique, but nosZ genotype richness was very low where n/'flH richness was high. Both communities experienced a significant decline in genotype richness in lower samples and due to warming treatments at this site. Sites with low or moderate nosZ richness showed significant or unexpected changes in the number of TRFs at two depths and between treatments; alternately, sites with a high number of n/'flH genotypes showed significant changes due to depth or treatment. -93- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Site most influenced community structure The community structure of nosZ and nifH functional genes responded differently to the biotic and abiotic properties of each site reflecting the requirements of either denitrifier or nitrogen fixer groups respectively. This was most apparent at the SM site where homogeneous anaerobic conditions created distinct TRF communities. Denitrifier diversity may have been limited by the absence of a favourable nitrifier environment and subsequent lack of substrate for denitrification. Diverse diazotrophic cyanobacteria appear to have benefited from the aquatic, N-poor habitat. The unique nature of the nosZ communities at each site confirmed that the distribution of this gene is habitat specific (Rich et al., 2003; Rich and Myrold, 2004; Rosch etal., 2002; Scala and Kerkhof, 2000; Stres etal., 2004). Overall, nosZ richness was lowest where ecological conditions were most extreme and relatively high where conditions were moderate. That each site shared common genotypes in varying proportions confirms that dominant nifH TRFs at one site may be absent, less frequent or less abundant at an adjacent site when predominant vegetation and soil properties differ (Poly et al., 2001; Rosch et al., 2002; Shaffer et al., 2000). Alexandra sites with the greatest nifH richness were those with highly selective conditions. - 94- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Genotype richness was sensitive to all environmental factors Genotype richness was affected by the environmental factors that made each site and depth unique. For both functional genes, the number of TRFs differed between sites and decreased overall in lower samples. When genotypic richness was explored by comparing the number of TRFs between control plots and OTCs, it appeared that rare species were lost and dominant species prevailed when disturbed by warming treatments. 3.2. Microbial community structure vs. physiological function Callaghan et al. (2004) suggest that arctic environmental conditions restrict the metabolic potential of very diverse arctic microbial communities; enzymatic activity would easily approach that of boreal forests with forecasted increases in temperature. Unfortunately, the link between rates of microbially mediated processes and changes in microbial community structure (measured in terms of genetic diversity) are not well established. Changes in physiological function have been noted without affecting diversity at the genetic level (Gomez et al. 2004, Nicolaisen et al. 2004), and community shifts that do not affect microbial activity have been documented (Avrahami and Conrad 2003, Deslippe 2004). It has been suggested that biotic and abiotic environmental factors can structure the genetic composition of a system, but that gene selection is not the only driver of microbial diversity (Zehr et al. 2003). -95 - Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3.3. Future research This study would have benefited from a number of additional and parallel experiments that should be incorporated into future research. Plate counts to determine number of culturable bacteria, substrate-induced respiration for total microbial biomass and Biolog assessment of community profiles would be useful additions if only to help characterize the arctic microbial community. Physiological activity could be measured with enzyme analyses or other assays for total microbial activity (ie. denitrifying enzyme assay (DEA) for denitrification potential, gas chromatography (GC) for N2 O release, and acetylene reduction assay (ARA) for rate of N fixation). Phylogenetic information via culturing, cloning and sequencing would also be a fundamental addition to the study, and real-time PCR for gene expression would be an ideal tool if the objective is to truly link genetic diversity with physiological function. Soil chemical analyses from the 2004 sampling season (pH, C:N, OM, NH4, NO3, P, K) will be invaluable additions as overlays to NMS ordinations in order to more thoroughly compare sites and elucidate factors that drive the differences that we detected. We have learned that both nosZ and nifH communities are strongly structured by the environmental conditions of the site from which they are sampled. Although it is not clear exactly what biotic or abiotic factors drive these patterns, it is worthwhile examining study sites that are very similar to one another if the - 96- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. objective is to measure only the effects of warming treatments. The magnitude of OTC disturbance is unique to each habitat and significant temperature changes below the soil surface must be confirmed. Additionally the method of measurement in important to determine relevant changed to microbial communities. For example, a change in the relative abundance of a particular genotype may indicate a community shift but not necessarily a change in the physiological process it is associated with. Alternately, mean loss of genotypes, although a coarse measure, may give more information about the resilience of a community to disturbance and the potential for microbial activity. - 97- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Literature Cited Arft AM, Walker MD, Gurevitch J, Alatalo JM, Bret-Harte MS, Dale M, Diemer M, Gugerli F, Henry GHR, Jones MH, Hollister RD, Jonsdottir IS, Laine K, Levesque E, Marion GM, Molau U, Molgaard P, Nordenhall U, Raszhivin V, Robinson CH, Starr G, Stenstrom A, Stenstrom M, Totland 0 , Turner PL, Walker LJ, Webber PJ, Welker JM, Wookey PA. 1999. Responses of tundra plants to experimental warming: Meta-Analysis of the International Tundra Experiment. Ecological Monographs. 69: 491-511. Anderson MJ. 2005. PERMANOVA: Permutational multivariate analysis of variance. University of Auckland, Department of Statistics. Avrahami S, Conrad R, and Braker G. 2002. Effect of soil ammonium concentration on N2O release and on the community structure of ammonia oxidizers and denitrifiers. Applied and Environmental Microbiology. 68: 56855692. Avrahami S, Liesack W, Conrad R. 2003. Effects of temperature and fertilizer on activity and community structure of soil ammonia oxidizers. Environmental Microbiology. 5: 691-705. Avrahami S, Conrad R. 2003. Patterns of community change among ammonium oxidizers in meadow soils upon long-term incubation at different temperatures. Applied and Environmental Microbiology. 69: 6152-6164. Berendse F, Jonasson S. 1992. Nutrient use and nutrient cycling in northern ecosystems. In Arctic Ecosystems in a Changing Climate. Chapin III, Jefferies, Reynolds, Shaver and Svoboda, Eds. Academic Press, Toronto. 337-356. Braker G, Zhou J, Wu L, Devol AH, Tiedje JM. 2000. Nitrite reductase genes (n/'rK and nirS) as functional markers to investigate diversity of denitrifying bacteria in Pacific Northwest marine sediment communities. Applied and Environmental Microbiology. 66: 2096-2104. Braker G, Ayala-Del-Rfo HL, Devol AH, Fesefeldt A, Tiedje JM. 2001. Community structure of denitrifiers, Bacteria, and Archaea along redox gradients in Pacific Northwest marine sediments by Terminal Restriction Fragment Length Polym orphism analysis of am plified nitrite reductase (nirS) and 16S rR N A genes. Applied and Environmental Microbiology. 67: 1893-1901. Brodie E, Edwards S, Clipson N. 2002. Bacterial community dynamics across a floristic gradient in a temperate upland grassland ecosystem. Microbial Ecology. 44:260-270. 98 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Callaghan TV, Bjom LO, Chernov Y, Chapin T, Christensen TR, Huntley B, Ims RA, Johansson M, Jolly D, Jonasson S, Matveyeva N, Panikov N, Oechel W, Shaver G, Elster J, Henttonen H, Laine K, Taulavuori K, Taulavuori, E, Zockler C. 2004. Climate change and UV-B impacts on arctic tundra and polar desert ecosystems: biodiversity, distributions and adaptations of arctic species in the context of environmental change. Ambio 33: 404-417. Chapin DM, Bledsoe CS. 1992. Nitrogen fixation in arctic plant communities. In Arctic Ecosystems in a Changing Climate. Chapin III, Jefferies, Reynolds, Shaver and Svoboda, Eds. Academic Press, Toronto. 301-319. Chapin III FS, Jefferies RL, Reynolds JF, Shaver GR, Svoboda J. 1992. Arctic plant physiological ecology in an ecosystem context. In Arctic Ecosystems in a Changing Climate. Chapin III, Jefferies, Reynolds, Shaver and Svoboda, Eds. Academic Press, Toronto. 441-451. Chapin III FS, Shaver GR, Giblin AE, Nadelhoffer KJ, Laundre JA. 1995. Responses of arctic tundra to experimental and observed changes in climate. Ecology. 76: 694-711. Deslippe J. 2004. Will climate change alter arctic nitrogen budgets? Impacts of warming and fertilization on nitrogen fixing microbial communities at Alexandra Fiord, Ellesmere Island, Nunavut. MSc Thesis. UNBC. Deslippe JR, Egger KN, Henry GHR. 2005. Impacts of warming and fertilization on nitrogen-fixing microbial communities in the Canadian High Arctic. FEMS Microbiology Ecology. 53: 41-50. Dunbar J, Ticknor LO, Kuske CR. 2000. Assessment of microbial diversity in four southwestern United States soils by 16S rRNA gene terminal restriction fragment analysis. Applied and Environ. Microbiol. 66: 2943-50. Engebretson JJ, Moyer CL. 2003. Fidelity of select restriction endonucleases in determining microbial diversity by Terminal-Restriction Fragment Length Polymorphism. Applied and Environmental Microbiology. 69: 4823-4829. Freedman B, Svoboda J, Henry GHR. 1994. Alexandra Fiord - An ecological oasis in the polar desert. In Ecology of a Polar Oasis. Svoboda and Freedman, Eds. Captus University Publications, London. 1-9. Gomez EV, Garland JL, Roberts MS. 2004. Microbial structural diversity estimated by dilution-extinction of phenotypic traits and T-RFLP analysis along a land-use intensification gradient. FEMS Microbiol. Ecol. 49: 253-259. Hassol SJ. 2004. ACIA, Impacts of a warming arctic: Arctic Climate Impact Assessment. Cambridge University Press. 99 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Hollister RD, Webber PJ, Nelson FE, Tweedie CE. 2006. Soil thaw and temperature response to air warming varies by plant community: results from an open-top chamber experiment in northern Alaska. Arctic, Antarctic and Alpine Research. 38: 206-215. Jonasson S, Havstrom M, Jensen M, Callaghan TV. 1993. In site mineralization of nitrogen and phosphorus of arctic soils after perturbations simulating climate change. Oecologia 95: 179-186. Jonasson S, Michelsen A, Schmidt IK. 1999. Coupling of nutrient cycling and carbon dynamics in the Arctic, integration of soil microbial and plant processes. Applied Soil Ecology. 11: 135-146. Kane DL, Hinzman LD, Woo M, Everett KR. 1992. Arctic hydrology and climate change. In Arctic Ecosystems in a Changing Climate. Chapin III, Jefferies, Reynolds, Shaver, and Svoboda, Eds. Academic Press, Toronto. 35-57. Klady R. 2006. MSc Candidate. Geography Dept. UBC: Supervisor GHR Henry. Labine C. 1994. Meteorology and climatology of the Alexandra Fiord lowland. In Ecology of a Polar Oasis. Svoboda, Freedman, Eds. Captus University Publications, London. 23-40. Liengen T. 1999. Environmental factors influencing the nitrogen fixation activity of free-living terrestrial cyanobacteria from a high arctic area, Spitsbergen. Canadian Journal of Microbiology. 45: 573-581. Maxwell B. 1992. Arctic climate: potential for change under global warming. In Arctic Ecosystems in a Changing Climate. Chapin III, Jefferies, Reynolds, Shaver, and Svoboda, Eds. Academic Press, Toronto. 11-34 McCune B, Grace JB. 2002. Analysis of ecological communities. MjM Software Design, Gleneden Beach, Oregon, USA. McCune B, Mefford MJ. 1999. PC-ORD. Mulitivariate analysis of ecological data, Version 4. MjM Software Design, Gleneden Beach, Oregon, USA. Mergel A, Schmitz O, Mailman T, Bothe H. 2001. Relative abundance of denitrifying and di-nitrogen fixing bacteria in layers of a forest soil. F E M S Microbiology Ecology. 36: 33-42. Muc M, Svoboda J, Freedman B. 1989. Vascular plant communities of a polar oasis at Alexandra Fiord, Ellesmere Island. Can. J. Botany. 67: 1126-1136. 100 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Muc M, Svoboda J, Freedman B. 1994. Soils of an extensively vegetated polar desert oasis, Alexandra Fiord, Ellesmere Island. In Ecology of a Polar Oasis. Svoboda, Freedman, Eds. Captus University Publications, London. 41-50. Nadelhoffer KJ, Giblin AE, Shaver GR, Laundre JA. 1991. Effects of temperature and substrate quality on element mineralization in six arctic soils. Ecology. 72: 242-253. Nadelhoffer KJ, Giblin AE, Shaver GR, Linkins AE. 1992. Microbial processes and plant nutrient availability in arctic soils. In Arctic Ecosystems in a Changing Climate. Chapin III, Jefferies, Reynolds, Shaver and Svoboda, Eds. Academic Press, Toronto. 281-300. Nicolaisen MH, Risgaard-Petersen N, Revsbech NP, Reichardt W, Ramsing NB. 2004. Nitrification-denitrification dynamics and community structure of ammonia oxidizing bacteria in a high yield irrigated Philippine rice field. FEMS Microbiology Ecology. 49: 359-369. Paul EA, Clark FE. 1996. Soil Microbiology and Biochemistry. 2nd Ed. Academic Press. Toronto. Piceno YM and Lovell CR. 2000a. Stability in natural bacterial communities. I. Nutrient addition effects on rhizosphere diazotroph assemblage composition. Microbial Ecology. 39: 32-40. Piceno YM and Lovell CR. 2000b. Stability in natural bacterial communities. II. Plant resource allocation effects on rhizosphere diazotroph assemblage composition. Microbial Ecology. 39: 41-48. Poly F, Ranjard L, Nazaret S, Gourbiere F, Jocteur Monrozier L. 2001. Comparison of nifH gene pools in soils and soil microenvironments with contrasting properties. Applied and Environmental Microbiol. 67: 2255-2262. Prescott LM, Harley JP, Klein DA. 1999. Microbiology. 4th Ed. McGraw-Hill. Toronto. Ch 22. Prieme A, Braker G, Tiedje JM. 2002. Diversity of nitrite reductase (n/'rK and nirS) gene fragments in forested upland and wetland soils. Applied and Environmental Microbiology. 68: 1893-1900. Rich JJ, Heichen RS, Bottomly PJ, Cromack Jr. K, Myrold DD. 2003. Community composition and functioning of denitrifying bacteria from adjacent meadow and forest soils. Applied and Environmental Microbiology. 69: 5974-5982. Rich JJ, Myrold DD. 2004. Community composition and activities of denitrifying bacteria from adjacent agricultural soil, riparian soil, and creek sediment in 101 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Oregon, USA. Soil Biology and Biochemistry. 36: 1431-1441.Robinson CH. 2002. Controls on decomposition and soil nitrogen availability at high latitudes. Plant and Soil. 242: 65-81. Rolph SG. 2003. Effects of a ten-year climate warming experiment on nitrogen cycling in the high arctic tundra. MSc Thesis, University of British Columbia. Rosch C, Mergel A, Bothe H. 2002. Biodiversity of denitrifying and dinitrogenfixing bacteria in an acid forest soil. Applied and Environmental Microbiology. 68: 3818-3829. Ruess L, Michelsen A, Schmidt IK, Jonasson S. 1999. Simulated climate change affecting microorganisms, nematode density and biodiversity in subarctic soils. Plant and Soil. 212: 63-73. Rustad LE, Campbell JL, Marion GM, Norby RJ, Mitchell MJ, Hartler AE, Cornelissen JHC, Gurevitch J. 2001. A meta-analysis of the response of soil respiration, net nitrogen mineralization and aboveground plant growth to experimental ecosystem warming. Oecologia 126: 543-562. Scala DJ, Kerkhof LJ. 2000. Horizontal heterogeneity of denitrifying bacterial communities in marine sediments by Terminal Restriction Fragment Length Polymorphism analysis. Applied and Environmental Microbiology 66: 19801986. Schmidt IK, Jonasson S, Michelsen A. 1999. Mineralization and microbial immobilization of N and P in arctic soils in relation to season, temperature and nutrient amendment. Applied Soil Ecology. 11: 147-160. Schmidt IK, Jonasson S, Shaver GR, Michelsen A, Nordin A. 2002. Mineralization and distribution of nutrients in plants and microbes in four arctic ecosystems: responses to warming. Plant and Soil. 242:93-106. Shaffer BT, Widmer F, Porteous LA, Seidler RJ. 2000. Temporal and spatial distribution of the nifH gene of N2-fixing bacteria in forests and clearcuts in western Oregon. Microbial Ecology 39:12-21. Sharma S, Szele Z, Schilling R, Munch JC, Schloter M. 2006. Influence of freezethaw stress on the structure and function of microbial communities and denitrifying populations in soil. Applied and Environm ental Microbiology 72: 2148-2154. Shaver GR, Canadell J, Chapin III FS, Gurevitch J, Harte J, Henry G, Ineson P, Jonasson S, Melillo J, Pitelka L, Rustad L. 2000. Global warming and terrestrial ecosystems: a conceptual framework for anlysis. Bioscience 50: 871-882. 102 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Shaver GR, Bret-Harte MS, Jones MH, Johnstone J, Gough L, Laundre J, Chapin III FS, 2001. Species composition Interacts with fertilizer to control long-term change in tundra productivity. Ecology. 82: 3163-3181. Sterenberg VZ, Stone WE. 1994. Geology of the Alexandra Fiord lowland area Ellesmere Island, N.W.T. In Ecology of a Polar Oasis. Svoboda and Freedman, Eds. Captus University Publications, London. 18, 19. Stres B, Mahne I, Avgustin G, Tiedje JM. 2004. Nitrous oxide reductase (nosZ) gene fragments differ between native and cultivated michigan soils. Applied and Environmental Microbiology. 70: 301-309. Tiedje JM, Asuming-Brempong S, Nusslein K, Marsh TL, Flynn SJ. 1999. Opening the black box of soil microbial diversity. Appl. Soil Ecol. 13: 109-122. Throback IN, Enwall K, Jarvis A, Hallin S. 2004. Reassessing PCR primers targeting nirS, nirK AND nosZ genes for community surveys of denitrifying bacteria with DGGE. FEMS Microbiology Ecology. 49: 401-417. Ueda T, Suga Y, Yahiro N, Matsuguchi T. 1995. Remarkable N2-fixing bacterial diversity detected in rice roots by molecular evolutionary analysis of nifH gene sequences. Journal of Bacteriology. 177: 1414-1417. Walker MD, Wahren CH, Hollister RD, Henry GHR, Ahlquist LE, Alatalo JM, BretHarte MS, Clef MP, Callaghan TV, Carroll AB, Epstein HE, Jonsdottir IS, Klein JA, Magnusson B, Molau U, Oberbauer SF, Rewa SP, Robinson CH, Shaver GR, Suding KN, Thompson CC, Tolvanen A, Totland 0 , Turner PI, Tweedie CE, Webber PJ, Wookey PA. 2006. Plant community responses to experimental warming across the tundra biome. PNAS. 103:5 1342-1336. Widmer F, Shaffer BT, Porteous LA, Seidler RJ. 1999. Analysis of nifH gene pool complexity in soil and litter at a Douglas Fir forest site in the oregon cascade mountain range. Applied and Environmental Microbiology 65: 374-380. Wolsing M, Prieme A. 2004. Observation of high seasonal variation in community structure of denitrifying bacteria in arable soil receiving artificial fertilizer and cattle manure by determining T-RFLP of nir gene fragments. FEMS Microbiology Ecology. 48: 261-271. Z e h r JP , M ellon M T, Zani S. 1998. N e w nitrogen-fixing m icroorganism s detected in oligotrophic oceans by amplification of nitrogenase {nifH) genes. Applied and Environmental Microbiology. 64:3444-3450. Zehr JP, Jenkins BD, Short SM, Steward GF. 2003. Nitrogenase gene diversity and microbial community structure: a cross-system comparison. Environmental Microbiology 5: 539-554. 103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.