|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 AQUATIC NITROGEN FIXATION: PATTERNS, RATES AND CONTROLS IN A SHALLOW, SUBTROPICAL LAKE By MOSHE DORON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR T HE DEGREE OF MASTER OF SECIENCE UNIVERSITY OF FLORIDA 2010
2 2010 Moshe Doron
3 To my parents with all my love
4 ACKNOWLEDGMENTS My deepest gratitude is to my advisor Dr. Patrick W. Inglett., who introduced me to biogeochemistry and believe d in me even during stressful times. His expertise and ability to find my independent thinker provided priceless guidance and instruction throughout my graduate education that were crucial to this thesis. I would like to thank the members of my committee D r. A ndrew V. Ogram who was like a mentor for me during my undergraduate studies and Dr. E dward J. Phlips for their continued help encouragement and scientific assistance. Their input provided insightful comments and contributions to my work. Special than ks to John Hendrickson from the St. Johns River Water Management District for sampling, transporting and analyzing Lake George water and for supplying helpful data and financial support for this research. Also, thanks to Mary F. Cichra for her m icroscopic analysis I would like to express my gratitude to Dr. Kanika S. Inglett for her scientific insights, technical assistance and help in the writing of the thesis and manuscript. Dr. James Cole of IFAS Statistics and Dr. Rongzhong Ye are acknowledged for the ir assistance in statistical analysis. My thanks to Dr. Abid al Agely and Ms. Yu Wang for their laboratory assistance and to Gavin Wilson for his assistance troubleshooting equipment malfunctions. I also like to send my love and appreciation to my friends in the lab, Dr. Hiral Gohil, Dr. Haryun Kim Dr. Bae HeeSung, Lisa Stanley Cassandra Medvedeff, Benjamin Hogue, Xiaolin Liao and of course Patricia Gonzalez. Most importantly, none of this would have been possible without the love and support of my imme diate family. My family, to whom this thesis is dedicated to, has been a constant source of love, support and strength during the last 12 years and helped me keep my faith and remember who I am during these hectic times I would li ke them to know they are always in my heart
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 7 LIST OF FIGUR ES .............................................................................................................................. 8 ABSTRACT ........................................................................................................................................ 10 CHAPTER 1 INTRODUCTION ....................................................................................................................... 12 nifH and nifD ............................................................................................................................... 18 Need for Research ....................................................................................................................... 19 Lake George as a Model System ................................................................................................ 20 Research Objec tives .................................................................................................................... 21 2 BIOGEOCHEMICAL ANALYSIS ........................................................................................... 25 Nutrient Limitation by N and P .................................................................................................. 25 Algal N2 Fixation ........................................................................................................................ 26 Eutrophication of Downstream Systems .................................................................................... 26 Research Objectives .................................................................................................................... 27 Materials and Methods ................................................................................................................ 29 Site Description .................................................................................................................... 29 Sample Collection ................................................................................................................ 29 Nitrogenase Activity ............................................................................................................ 30 Physicochemical and Algal Taxonomic Analysis ............................................................. 31 Statistical Analysis ............................................................................................................... 32 Results and Discussion ............................................................................................................... 33 Limnological Parameters ..................................................................................................... 36 Evaluation of Nutrient Limitation ...................................................................................... 38 Microscopic Analysis .......................................................................................................... 41 Conclusions ................................................................................................................................. 44 3 MOLECULAR ANALYSIS ...................................................................................................... 61 Nitrogenase Gene ........................................................................................................................ 62 Cyanobacteria .............................................................................................................................. 64 Heterocystous Cyanobacteria .............................................................................................. 66 Nostocaceae .......................................................................................................................... 67 Proteobacteria ............................................................................................................................. 67 Lake George ................................................................................................................................ 68 Materials and Methods ................................................................................................................ 69
6 Site Description and Sample Collection ............................................................................. 69 DNA Extraction ................................................................................................................... 7 0 Amplification of nifD by Polymerase Chain Reaction (PCR) .......................................... 70 Sequencing and Phylogenetic Analysis .............................................................................. 71 Results .......................................................................................................................................... 73 Discussion .................................................................................................................................... 76 4 SUMMARY AND CONCLUSION ......................................................................................... 107 LIST OF REFERENCES ................................................................................................................. 115 BIOGRAPHICAL SKETCH ........................................................................................................... 124
7 LIST OF TABLES Table page 2 1 Characteristic values of selected chemical parameters in Lake George water .................. 49 2 2 Type 3 test of fixed effects based on nitrogenase activity rate (ARA) ............................... 49 2 3 Nitrogenase activity (measured using ARA) at various types of aquatic ecosystems. ...... 50 2 4 Model parameters for stepwise multiple regression analysis .............................................. 51 3 1 Values of nifD diversity and richness in Lake George water .............................................. 80 3 2 Population simila rity P values for comparison of nifD clone libraries .............................. 81 3 3 Correlation between diazotrophic composition and enviro nmental parameters ............... 82
8 LIST OF FIGURES Fi gure page 1 1 Map showing the location of Lake George in Florida. ........................................................ 23 1 2 Autochthonous N loading and export in Lake Ge orge. ....................................................... 24 2 1 Map showing the location of Lake George in Florida ......................................................... 52 2 2 Temporal distri bution of nitrogenase activity under three light level ................................. 53 2 3 Spatial and temporal distribution of nitrogenase activity under high light incubation ...... 53 2 4 The effect of l ight intensity on acetylene reduction assay rates .......................................... 54 2 5 Temporal patterns of environmental parameters ................................................................. 55 2 6 Temporal patterns of N:P ra tio in Lake George water -column ........................................... 56 2 7 Temporal patterns of labile N: ammonium (NH4 +) and nitrite/nitrate (NOx). .................... 57 2 8 Spatial and temporal patterns of cyanobacteria biovolume : ................................................ 58 2 9 Spatial and temporal patterns of heterocystous cyanobacteria biovolume : ........................ 59 2 10 Separation of lake transect locations by Canonical analysis ............................................... 60 3 1 Selected dates where samples were obtained for molecular analysis ................................. 82 3 2 Rarefraction analysis for nifD c ollected from Lake George water ..................................... 83 3 3 Phylogenetic tree of genomic DNA nifD from 4/22/08 east transect. ................................ 84 3 4 Phylogenetic tree of genomic DNA nifD from 4/22/08 center transect. ............................ 86 3 5 Phylogenetic tree of genomic DNA nifD from 4/22/08 west transect. ............................... 88 3 6 Phylogenetic tree of genomic DNA nifD from 7/29/08 east transect. ................................ 90 3 7 Phylogenetic tree of genomic DNA nifD from 7/29/08 c enter transect ............................. 92 3 8 Phylogenetic tree of genomic DNA nifD from 7/29/08 west transect. ............................... 94 3 9 Phylogenetic tree of genomic DNA nifD from 8/19/08 east transect. ................................ 96 3 10 Phylogenetic tree of genomic DNA nifD from 8/19/08 center transect.. ........................... 98 3 11 Phylogenet ic tree of genomic DNA nifD from 8/19/08 west transect. ............................. 100
9 3 12 Phylogenetic tree of genomic DNA nifD from 9/02/08 east transect. .............................. 102 3 13 Phylogenetic tree of genomic DNA nifD from 9/02/08 west transect. ............................. 104 3 14 Separation of samples by principle component analysis. .................................................. 106
10 Abstract of Thesis Presented to the Graduate School of the University of Florida i n Partial Fulfillment of the Requirements for the Degree of Master of Science AQUATIC NITROGEN FIXATION: PATTERNS, RATES AND CONTROLS IN A SHALLOW, SUBTROPICAL LAKE By Moshe Doron August 2010 Chair: Partick W. Inglett Major : Soil and Water Science Shallow lakes can act as sinks or sources of nutrients such as nitrogen (N) which is often a key nutrient limiting downstream coastal systems. Biological N2 fixation is one of the key processes affecting lake N, but more research is needed to evaluate this process in relation to algal blooms, particularly in subtropical systems. To evaluate these changes, a project was undertaken to investigate the spatial and temporal patterns of nitrogenase activity and diazotrophic community composition in a large, shallow, subtropical lake (Lake George, Florida, USA). This lake may be limited by N rather than phosphorus (P), making it a good candidate for the study of diazotrophs in algal bloom formation. The objectives of this study were threefold: (1) to c haracterize and q uantify N2 fixation in Lake George water (2) to i dentify environmental parameters related to N2 fixation, (3) to c haracterize the diazotrophic commu nity composition in Lake George It was hypothesized that environmental conditions affecting carbon, N and P would impact the diazotrophic community (abundance, composition, and activity rate) and, thus, directly affect lake N budgets. We measured, diazotr ophic community size (microscopy), composition (using nifD ), and function (nitrogenase activity via acetylene reduction) from April to September, 2008. Nitrogenase activity ranged from 3 to 95 nmol l1 h1 during the study and
11 had a significant importance to the lakes N budget. Nitrogenase activity correlated with N limitation (estimated by dissolved inorganic N: total P ratio ( DIN:TP ratio )), heterocystous cyanobacterial blooms, and was positively influenced by light. The chemical variation between sites was significant also showing a seasonal change. Analysis of ni fD clone libraries revealed a diazotrophic community with low diversity that was dominated by filamentous heterocystous cyanobacteria (order Nostocales ). Microscopy and nutrient patterns showed these blooms were episodic, possibly due to availability of P relative to N, and that each bloom was influenced by different parameters, with the later blooms depending on the outcome of those prior. The results of this study documented the potential for N2 fixation to transform the lake into a N source to downstream systems, but further study is needed to determine the role of diazotrophs in the formation of algal bloom events.
12 CHAPTER 1 INTRODUCTION As a vital component of amino and nucleic acids, the building blocks of proteins and DNA/RNA, nitrogen (N) is a key element required for all living organisms. Although the majority of the planets N is found in the lithosphere in the form of nitrate, it is not accessible or bioavailable T he second largest N pool, dinitrogen gas (N2) in the atmosphere, is also unavailable for direct use or uptake by organisms T his inaccessibility of N is the primary reason why many ecosystems are found to be limited by N, and only a limited number of prokaryotes and archea ca n convert atmospheric N2 into a bioavailable form (ammonium, NH4 +) through the process of biological N2 fixation. In N -limited ecosystem s, N2 fixation has the potential to supply N and increase primary production ( Flett, 1980; Smith 1983) As a result, m icroorganisms with the ability to fix N2 (diazotrophs) are somewhat relieved of N limitation, and are expected to demonstrate increased growth under N limiting conditions due to this competitive advantage (Karl, 1997). I n turn, a community shift towards di azotrophs is expected and also explained by their ability to out compe te other organisms (Schindler, 1977; Tilman 1982). This concept holds for aquatic systems, where d iaz otrophs proliferate under N limited conditions. For example, in many N limited Lake s, the diazotrophic density (total cells counts per ml) can typically represent about 2030% of the total phytoplankton community during the growth season with a maximum recorded up to 60% ( Gao 2005). Higher abundance of diazotrophs usually results in highe r rate s of N2 fixation ( Vitousek and Howarth 1991). In eutrophic lakes f or example, N2 fixation constituted for up to 82% of total N loading (Howarth et al. 1988) while in shallow wetlands like the Everglades water column N2 fixation was 3 times higher in nutrient impacted sites (Inglett et al., 2009 ).
13 Presumably, the relative proportions of nutrients in the water column are a key factor shaping the composition, growth and activity rates (per unit of the limiting nutrient) of the phytoplankton community N one theless, the identification of nutrient limitation in natural environments by coupling nutrient concentrations to communit y growth is difficult due to the complexity and dynamics of natural systems (Hecky and Kilham, 1988) and internal storage of bot h N and P by phytoplankton (Lohman et al., 1991). In aquatic systems, more emphasis is put on N or phosphorus (P) because under normal conditions, carbon supply is generally found to be adequate. In m ost marine ecosystems, primary production is most freque ntly limited by N (Tilman et al., 1982), while most fresh water systems are generally limited by P (Schindler 1977). For this reason, P additions can also increase the demand for N in freshwater systems, and thus lead to increasing N limitation ( Paerl 1987; Mills 2004). Availability of P can come through a variety of factors. Although anthropogenic loading is a major source of P that shifts fresh water systems into N limitation, some ground and surface water can deliver natural high concentration s of P ( Sims et al., 1998) leading some lakes and rivers to be n aturally rich in P (Smith 1983; Havens et al., 2007). Temporal changes in precipitation, temperature, wind, salinity, pH, residence time, and light intensity can also have an impact on nutrient abun dance patterns and distribution (Thayer, 1971) D ue to the overall abundance of P contained in sediments, a summer shift in environmental conditions ( e.g. salinity) that converts P into labile forms may increase the demand for N and its fixation. Such pat terns may be related to phytoplankton proliferation and N2 fixation rate (Paerl et al., 1996). In either case, the importance of P for N2 fixation regulation should be reflected by an N:P ratio that evaluate s which of the two elements limits primary produc tion and affects diazotrophic activity
14 Shallow lakes are more abundant than deep lakes and are receiving increasing attention because of their eutrophication around the globe (Dokulil and Teubner 2 003). In general, CO2 limitation is rare in shallow lake s because of its high rate of introduction from the atmosphere and sediment (Jensen et al. 1994) T he depth of a lake is an important parameter that has many implications for lake processes, including the amount of stratification and wind induced sediment resuspension (Bachmann et al. 2000) T hus, it must be considered when examining the nutrient cycle in such systems ( e.g. resuspension of P rich sediment may shift system to N limitation ). Unlike deep lakes where P concentration in the water column represents an internal pool, in shallow lakes the mobile sedimentary P may be directly available for primary producers growth (Istvnovics et al., 2000) Frequently, aquatic systems are hydrologically linked in the landscape, and therefore, processes occurrin g in systems higher in the watershed are likely to influence other downstream systems. For this reason, upstream eutrophic systems that contain many nutrients (including N ) that move through the landscape (mostly in dissolved forms) can act as pollut ion so urces to downstream systems. In this manner, N that was fixed in such systems may be transported to downstream systems; this is particularly true for shallow lakes (especially with P rich sediments) that are ideal environment for diazotrophic cyanobacteria Several factors affect the degree to which eutrophic systems act as pollution sources to downstream systems. In general, lakes and rivers show greater N exports in periods of high discharge, probably due to stored N in the landscape during dry periods an d flushing during wet periods (Howarth et al., 2006). This export can range from 10 to 15% of the N inputs in drier watersheds to over 35% in the wetter w atersheds (Howarth et al., 2006). As this exported N reaches downstream systems, it changes nutrient equilibrium and can result in eutrophication.
15 This is especially true when downstream systems are coastal/marine because N inputs drive eutrophication in N limited systems (Howarth, 1988). In fact, eutrophication of coastal systems is considered to be the biggest pollution problem in the coastal waters of the U.S. (Howarth et al., 2000; NRC 2000). For example, about 4000 estuaries in U.S. are severely degraded from eutrophication driven mostly by N and 67% is degraded to some extent (Bricker et al., 1999; EPA 2001). By shifting nutrient limitation from P to N, P inputs promote N2 fixation that leads to increased levels of N. Thus, P inputs to inland waters may lead to increased levels of N in downstream coastal systems indicating that control on both nutri ents is needed to control eutrophication in these systems (Paerl, 2009; Schindler et al., 2008). In order to prevent such advert environmental consequences, it is imperative to study the dynamics of N2 fixation in these upstream lakes. From an ecological p oint of view, N2 fixation has been measured in wide range of environments and conditions (Henson, 2003; Zehr 1998). Thus, in order to simulate the processes that take place in the environment, an understanding of the environment and the community structur e and function are essential. The first milestone in the path to understanding N2 fixation was placed in 1895, when Winogradsky isolated the first free living N2 fixing bacterium using a pure culture. Most of the diazotrophs in the phylum P roteobacteria a re facultative or obligate anaerobic, chemoautotrophs, or heterotrophic and the few phototrophs that exist do not produce O2. An example is the species Rhodospirillum centenum (also known as Rhodocista centenaria), a heterotrophic bacterium that ca n fix N2 under aerobic growth conditions and may be common and broadly distributed in many environments including lakes Additional diazotrophs are members of other phylum, including c yanobacteria and Actinobacteria ; for example, Frankia are
16 filamentous bacteria which form root nodules live in symbiosis with plants and were found in aquatic systems. The phylum c yanobacteria which its domination in fresh water systems has become synonymous with eutrophication, contains many diazotrophic members. Among diazotrophs, cyanobacteria are exceptional oxygenic photo autotrophs that inhabit almost every illuminated environment on Earth. L ike other primary producers, they are essential for the foundation of the food web, but the ir ability to use atmospheric N2 and their developed mechanisms for P management (by vertical movement to the benthic layer and internal storage of P ) make them an important subgroup ( Scheffer et al. 1997 ). Moreover, they are the only known organisms that can form heterocysts, which are specialized cells for N2 fixation with several mechanisms of protect ion from O2 that may enter from the environment or from nearby vegetative cells that produce O2. Hence, it is not surprising that diazotrophic communities of subtropical inland water systems are usually dominated by cyanobacteria. As mentioned abov e, diazotrophs may be very distinctive from each other by many aspects including their activity rates, regulation mechanisms and preferred niches; thus, it is important to identity the diazotrophic community components and understand their activities and d ynamics. A common feature to all diazotrophs is the production of the nitrogenase enzyme which is the only known enzyme that fixes (reduces) N2 (Henson 2003; Zehr 1998) Due to the extremely high energy demands of N2 fixation ( Witz et al., 1966), diazotr ophs use a cascade of regulatory mechanisms for nitrogenase gene expression. The nitrogenase enzyme is composed of two separate protein components: 1) dinitrogenase reductase which donates two high potential electrons to dinitrogenase, and 2) dinitrogenase which actually catalyzes N2 reduction. The nitrogenase complex is irreversibly inhibited by O2 and its manufacturing usually does not take
17 place in the presence of O2 and reduced N (Wang et al., 1985). The nitrogenase gene encodes the nitrogenase enzyme a nd is divided into several subunits called nif (e.g., nifH and nifD ). In general, there are challenges in cultivating m any micro b es because some cannot be g rown in vitro (some would die, become non -viable, or form spores) while others require specific medi a or co -culture with other species. Even the microbial organisms that can be cultured might not be identified or adequately reflect the ir corresponding environment al conditions T he vast majority of the in vitro analysis is done on pure cultures of organis ms and can not supply an authentic picture of the processes as they occur in the environment. Moreover, even direct technique s for measurement of N2 fixation is accomplished by measuring process rate and enzyme activities ; thus even though it is not depende nt on culturing, i t gives only the fixation rate and ignores the composition of the diazotrophic community. Knowledge about the diazotrophic community composition is valuable for its characterization and can be used to understand their dynamics and potenti al activities, as well as the environmental conditions leading to observed patterns. Before molecular biology advanced and PCR (polymerase chain reaction) was invented, it was difficult to identify the N2-fixing organisms Most methods that were used to identify diazotrophs in phytoplankton populations were culture based and included microscopy (morphology) and biomass estimates. Other relied on biochemical analysis or on the composition of cell membranes (fatty acids). Other studies used biomarker pigments as means for monitoring large shifts in species composition of phytoplankton, but although freshwater cyanobacteria can be discriminated by the presence of specific biomarker photosynthetic pigment, it cannot separate diazotrophs (Rowan, 1989) I n addition, monitoring the distribution of biomarker pigments in spatially complex and temporally transient waters remains a significant obstacle. In
18 many cases, these parameters were correlated to chlorophyll a (chl -a ) that by itself does not offer information abo ut community composition or the presence of diazotrophs. In fact, even the initial attempts to use PCR with 16S ribosome did not yield enough information to identify specific diazotrophs ( Zehr et al., 1989). This was due to the fact that many of the diaz otrophs do not share similar qualities apart from the ability to fix N2 (S tr a cke et al ., 2002). Only later, when specific primers were designed to target the subunits of the nitrogenase gene did quality information about the specific organisms, their roles and interactions start to accumulate ( Zehr et al., 1989). In other studies only a single gene (usually nifH which encodes for the dinitrogenase reductase subunit) was investiga ted, and some of its regulators were identified, although not all members of the diazotrophic community were identified ( Moisander et al., 2006). Some drawback s of the molecular approach may a rise when working with diazotrophs possess ing more than one set of nitrogenase gene s (for example, alternative or additional nitrogenase gene s (Ludden et al., 1989)). In addition results from phylogenetic analysis must be treated with caution because of the possibility that nitrogenase was transferred by lateral gene transfer ( in contrasts to the Darwinian model of vertical descent, where gene s are inherited from the preceding generation) (Young, 1992). E .g. nifH suggested that nitrogenase gene had been horizontally transferred from a Proteobacterium to cyanobacteria nifH and nifD At the present, the nifH database contains one of the largest non -ribosomal gene datasets, yet it includes a relatively high number of uncultivated organisms (Zehr 2003). R ecent studies used nifD gene ( encodes for the dinitrogenase subunit which is the actual site of N2 reduction) sequences due to their ability to supply additional information that can be used to construct the phylotypes of the N2-fixing community. Due to the fact that nifD was used less extensively, its
19 gene database contains a relatively low number of gene sequences (especially when compared to th e popular nifH database). Although the small database of sequences hampers the use of nifD for phylogenetic analysis, when used as a phylogenetic marker, it usually grants a higher resolution between closely related diazotrophs (especially cyanobacteria). In fact, Roselers et al., (2007) demonstrated that nifD better distinguishes between members of the nif gene family even in cases that organisms use one of the alternative nitrogenase enzyme (like nitrogenase -vanadium). When Roselers et al., (2007) were tr ying to develop a molecular technique that would detect N2fixing cyanobacteria in environmental sample s they realized that the gene sequences of nifD had conserved regions that permit the design of PCR primers specific for cyanobacterial nifD The use of nifD primers and the construction of a genetic library are essential before real time quantative PCR (Q PCR) techniques can be used to specifically qu antify and identify spatial and temporal variation in nitrogenase transcription in complex microbial comm unities. Need for Research The latest assessment of many eutrophic lakes around the globe concluded that P only reduction strategies would not be effective to control phytoplankton bloom (Conley et al., 2009). In these lakes, P fluxes between sediment and water, could potentially supply the phytoplankton community, which wa s frequently dominated by cyanobacteria. Using their ability to vertically migrate and consume excess P at the sediment -water interface, and then rise to the water surface enable d them t o form blooms in period s of presumed P limitation Constant and simultaneous measurements and control of both P and N are necessary in these lakes in order to successfully control algal blooms and diazotrophic activity (Schindler et al., 2008). Fresh water lake/reservoir systems are ideal sites for phytoplankton proliferation dominated by diazotrophs that introduce new N into their and downstream systems. In order to
20 evaluate the importance of N2 fixation to the N budget of upstream systems and its influenc e on downstream systems, it is impetrative to identify its major regulators in the system. By measuring enzyme activities under various environmental conditions that the system is exposed to, it may be possible to identify pattern of N2 fixation and quanti fy the amount of N t hat is fixed in the entire lake Lake George as a Model System Lake George is l ocated in subtropical region and is the second largest lake in Florida. Its estimated size is about 19.4 km long and 9.7 km wide covering about 18,60 0 ha (Fi gure 1 1). It is a part of the SJR ecosystem in which it may serve as a source and a sink for various nutrients A s the river move s slowly (mean residence time of the lake is 84 days) northward, it flows through Lake George and continues 200 km on a low gr adient (compared to its upstream 300 km) path to the Atlantic Ocean Lake George is a shallow (mean depth 2.5 m) lake that lies in four counties with different environmental characteristics such as P deposits, natural springs and a nearby forest. Lake Geor ge, like many systems in the SJR and Florida has high natural concentrations of dissolved P due to large deposits of phosphate -rich sediments (Odu 1952). In addition, Lake George has received large amounts of P and N pollution from point source pollution (among them septic tanks, fertilizers runoff and effluent from wastewater treatment plant) and continues to receive nutrients from nonpoint source s (agricultural runoff, particularly from dairy and livestock production facilities) ( Stewart et al., 2006) Like other shallow lakes, Lake George does not stratify completely and its water is relatively homogeneous vertically. Due to different inflow sources, its water chemistry can var y spatially, and like other eutrophic lakes, P levels in Lake George are rela tively high and shift s nutrient lim itation from to N I n such systems, N2 fixation is a key pathway that can supply the
21 demand for labile N. As a result the lake has bec o me dominated by cyanobacteria (both diazotrophic and non -diazotrophic). The growth of these diazotrophs could act as a source of N to downst ream river systems and result in an accelerated rate of eutrophication in particular to marine systems. Recently, this area received attention due to its apparent connection with adverse phenomenon se en in the Jacksonville coastline and in the Atlantic Ocean (nuisance algal blooms and formation of hypoxic zones) ( SJRWMD report 2008) Similar processes are occurring elsewhere and have similar connections with inland eutrophication and sources of nutrie nts. Lake George is an ideal site for this analysis for several reasons. I t is the largest water body on the S t. Johns River which is th e major input of freshwater to the nutrient -sensitive estuary near the city of Jacksonville Florida (NRC, 2000). Becaus e its low residence time is ideal for cyanobacterial proliferation and its relative high P promotes diazotrophic dominance and higher rates of N2 fixation, Lake George is also suitable for the study of cyanobacterial bloom dynamics and N2 fixation,. In add ition, the historical data on the lake and its regular monitoring by the SJR Water Management District (SJRWMD), including their experiments and reports, supply important information about the system For example, data obtained by the SJRWMD has estimated the amount of N being added to Lake George water (calculated by difference between the inflow and outflow) at approximately 600 metric tons (MT) per year ( Figure 1 2). This N may originate from various source s, including springs, leaching and biological N2 fixation Research Objectives Under typi cal summer conditions, downstream N limitation is a fairly consistent feature of many aquatic system s and may be satisfied by N derived from N2 fixation in upstream water bodies The morphology of such upstream sys tems (mostly size and depth) and their trophic status play an important role on the degree of nutrient export; in general, big eutrophic water
22 bodies have greater potential of exporting while shallow lakes are better environments to diazotrophic cyanobacte ria T hus, a eutrophic lake that is both large and shallow may export higher amounts of N to downstream systems. The economic and environmental impacts of exported N has prioritize d the need to determine nutrient thresholds and establish nutrient criteria such as Total Maximum Daily Loads (TMDLs). B efore it is possible to control eutrophication and pollution, however, the complexity of the system and its variables must be underst ood. Th e need to evaluate the portion of N being added to the system through biological processes is essential to understanding the N budget in lakes and its significance to exported N. The examination of different parameters that affect N2 fixation can be used to understand the process under natural conditions and in different time periods. Therefore, the specific objectives for this study were to: 1 Characterize and q uantify N2 fixation in Lake George water by evaluating the influences of location, date and light and by estimating its annual contribution to the N budget of the lake. It is hypothesized that due to this lakes shallow water eutrophic characteristics, N2 fixation is a significant source of N 2 Identify environmental parameters related to N2 fixation. It is hypothesized that rate s are affected by light and seasonality (th rough changes in N, P and their ratio). 3 Characterize the diazotrophic community composition in Lake George as it relates to lake location, date and N2 fixation rate. It is hypothesized that lake conditions will affect the overall composition of the diazot rophic community
23 Figure 1 1. Map showing the location of Lake George in Florida including its main water source and output (SJR).
24 5,124 MT N y14,519 MT N y1 Figure 1 2. Autochthonous N loading and export in Lake George. A comparison between N inputs from SJR and Artesian Springs (west) to exported N through the SJR (adapted from Hendrickson, SJRWMD, 2008).
25 CHAPTER 2 BIOGEOCHEMICAL ANALYSIS Due to its importanc e as a building block of biological systems and its role in countless biochemical pathways N availability is often coupled to biogeochemical processes like photosynthesis and mineralization, thus shaping aquatic ecosystem ecology. As discussed in Chapter 1, as a system becomes more limited by N, higher rate s of N2 fixation are generall y observed due to a competitive advantage diazotrophs have a over other organisms that do no t possess this ability ( Vitousek and Howarth, 1991; Karl et al. 1997). Nitrogen fixation by diazotrophs is the only biological activity (non anthropogenic) that h as the potential to introduce new N into the system and alleviate limitation of aquatic primary production ( Flett et al. 1980; Smith 1983) Nutrient Limitation by N and P Conventional views of nutrient limitation state that P limits primary production in most freshwater ecosystems, while N is limiting in coastal and marine waters ( e.g., Hecky and Kilham, 1988). L imitation by P is common in many inland fresh water systems, but once P enters the system, the limiting nutrient often shifts to N resulting in a competitive advantage to diazotrophs (Scheffer et al., 1997; Havens, 2003). When observing eutrophic systems, a direct relationship between nutrients and N2 fixation would demonstrate its regulators and importance to N export budgets and proliferation of diazotrophs A lthough high in nutrients, eutrophic aquatic systems can experience both N and P limitation from shifts in nutrient loading ratios and biologic al controls on nutrient cycling. Thus in eutrophic systems, N2 fixation may be an important compo nent of the N budget especially in periods of N limitation (Keirn and Brezonik, 1971; Ashton and Hoare, 1981).
26 Algal N2 F ixation Algal blooms have caused much economic and ecological damage globally during the last de c ades (Riegman 1998; Cloern 2001). N itrogen fixation by a l gae has been documented in natural waters for some time now (Hutchinson 1941) and this primarily by cyanobacteria which are the dom i nant diazotrophs in most aquatic environments ( Scheffer et al., 1997; Havens, 2003). Cyanobacteria a re superior competitors under eutrophic conditions and frequently thrive in nutrient -enriched waters (Vincent 1987). The se bacteria can achieve optimal growth under low light conditions, are buoyant (Reynolds 1987; Reynolds and Walsby 1975) have a rela tively slow growth rate (Paerl 1998) and some members can differentiate their cells into specialized structure s called heterocyst s where N2 fixation is protected from O2. For the above reasons, c yanobacteria are considered to be the most i m portant aquat ic diazotrophs (Keirn and Brezonik, 1971). Eutrophication of Downstream Systems Evidence that u pstream eutrophic systems may accelerate eutrophication i n downstream N limited and marine systems (Howarth, 1988) indicates that a control on both N and P may be the best approach for long term management of eutrophication in these systems (Paerl, 2009). In general, watersheds show greater N exports in periods of high discharge, probably by releasing N that was sequestered during dry periods. This export can ran ge from 10 to 15% of the N inputs in dr y watersheds to over 35% in wet watersheds. As exported N reaches downstream systems, it changes nutrient equilibrium and can accelerate eutrophication. In fact, eutrophication of coastal systems is considered to be t he biggest pollution problem in the coastal waters of the U.S. (Howarth et al. 2000; NRC 2000). For example, about 4000 estuaries in U.S. are severely degraded from eutrophication, and 67% are degraded to some extent (Bricker et al. 1999; EPA 2001). In ternal sources may also supply the required P for the phytoplankton community, and
27 thus, sequestered P in these systems may be released seasonal l y (e.g., salinity and light levels) thereby increas ing the demand for N and N2 fixation. In m any shallow subtro pical lakes P does not limit primary producers and as a result the se lake s became inoculum s for pest cyanobacteria (both diazotrophic and no n -diazotrophic). Cyanobacterial growth and activity can act as a nutrient source (especially N) to downst ream syste ms and result in an accelerated rate of eutrophication Recently, such systems have received attention due to the ir apparent connection to adverse phenomena in many coast a l and marine systems around the globe (hypoxic zones and toxic algal blooms). In fact it is believed that restoration of such systems is crucial for the reduction of these global events. In Florida, the S t. J ohns River Water Management District (SJRWMD) attempts to reduce the frequency and magnitude of these events in the Jacksonville coa stline (a nutrient -sensitive estuar y) by evaluating N cycle processes and loads within the watershed of the St. Johns River (SJR; Hendrickson, 2008). One key focus of this study is waterbodies capable of N2 fixation through blooms of cyanobacteria. Several shallow lakes exist in this watershed and meet this criterion for study, among them is Lake George. Lake George is a good model for the investigation of N2 fixation for several reasons including: 1) i t is the largest watershed on the SJR (NRC, 2000), 2) t his lake has a relative high P level promoting diazotrophic dominance, 3) it has a relatively low residence time, which is ideal for cyanobacterial proliferation, and 4) there is a large amount of historical data on Lake George including numerous experimen ts, reports and regular monitoring by the SJRWMD with over 70 years of data on water quality and over a decade of data on phytoplankton composition Research Objectives Under typi cal summer conditions, downstream N limitation is a fairly consistent featu re of the lower SJR system and may be alleviated by N derived from N2 fixation in Lake George
28 (SJRWMD report 2002). For many years, the area downstream from Lake George, including the Jacksonville region, has been experiencing algal blooms and zones of hypoxia. The economic and environmental impacts of exported N prioritize the need to determine nutrient thresholds and establish nutrient criteria such as Total Maximum Daily Load (TMDLs) in this region. The fact that Lake George is the biggest water body al ong the SJR, and its relative proximity to Jacksonville makes it one of the most important systems in this regard. The SJRWMD estimated the amount of annual N being added to Lake George water between 19962006 (calculated by difference in the inflow and outflow total N ) as about 600 metric tons (MT) ( Figure 1 2). This additional N may originate from various sources, including surface discharge, groundwater springs, biological N2 fixation, etc. One of the most important variables needed for understanding t he N budget of Lake George is the portion of N being added to the system through biological N2 fixation. In turn, the examination of different parameters that affect N2 fixation can be used to better understand the process under natural conditions and in d ifferent time periods. Rough estimation of the amount of fixed N in Lake George done by the SJRWMD indicates that it may be a significant source of N to downstream systems ( Figure 1 2), but because it was calculated indirectly, it may not represent actual values (Hedrickson ; SJRWMD, 2008); thus, the overall goal of this study was to evaluate temporal patterns and quantify the amount of N that enters the system through N2 fixation. Lake George is an ideal site for this analysis for the reasons mentioned above and its N2 fixation rates were within the range s observed in other shallow subtropical eutrophic lakes measured between the years 20002002 (Gao 2005). The specific objectives for this section were to: 1 Characterize and q uantify N2 fixation in Lake George water by evaluating the influences of location, date and light and by estimating its annual contribution to the N budget of the
29 lake. It is hypothesized that due to this lakes shallow water eutrophic characteristics, N2 fixation is a significant sour ce of N 2 Identify environmental parameters related to N2 fixation. It is hypothesized that rates are affected by light and seasonality (through changes in N, P and their ratio). Materials and Methods Site Description Lake George is a large (186,000 ha), sh allow (average depth 2.5 m), subtropical lake l ocated in Florida, USA ( Figure 2 1). Its estimated size is about 19.4 km long, 9.7 km wi de, and it is a part of the SJR ecosystem in which it is the biggest water body. Its e utrophication h as been accelerated by its geological characteristics, nearby human activity, and inflow from the SJR (SJRWMD report, 2008) L ake George does not stratify completely, and subsequently it is categorized as Class III by the state (this class supports healthy fish and wildlife populations). Sample Collection Integrated water samples were collected weekly during the summer of 2008 (22 weeks from April 8 to September 2) from nine locations in Lake George (three sites in each of the east, center, and west portions of lake, Figure 2 1). The sites were chosen to represent a range of water chemistry and environmental conditions in Lake George during the study period. The nine water samples were transported to the laboratory in 1 l carboys where they were composited into three transect samples based on lake location corresponding to East, West and Center transects ( Figure 2 1). The three tr ansect samples were mixed in 4 l carboys by constant stirring, and processed within two hours from sampling except on two dates (5/20 and 7/29). Phytoplankton biomass was concentrated from between 150 to 450 ml (depending on water turbidity) of site water by vacuum filtration through 0.7 m glass fiber prefilters (Millipore CAT No. APFF04700). Nitrogenase activity was assayed on triplicate filters immed iately following filtration.
30 Nitrogenase A ctivity Measurements of nitrogenase activity (NA) were performed using a modified version of the Acetylene (C2H2) Reduction Assay (ARA) by Stal (1988). The principle behind this method is based on the fact that the nitrogenase enzyme also reduces C2H2 gas into C2H4 gas. Because both C2H2 and C2H4 are trace elements, it is more accurate measuring low concentration of them compared to the abundant N2. A molar ratio of 3:1 between C2H4 to N2 can be used to estimate the amount of N2 that was fixed (Howarth et al., 1988b). Filters containing phytoplankton biomass were placed in 50 -ml Erlenmeyer flask (60 ml total volume), covered with filtered site water (10 -ml), and capped with a rubber turnover stopper. Two water controls without filters were used; similar to the samples, they contained a volume of 10 ml water: blank controls flasks were filled with 10 -ml D.I. water, while site water controls flasks were filled with 10 ml filtered site water. Using a syringe, a volume of 5 ml of headspace was removed from each flask and followed by injection of 6 ml acetylene gas. Triplicates of ea ch sample were incubated at 23 26oC inside a temperature-controlled shaker (100 rpm) and under three different light treatments. Dark incubation (0 M photons m2 s1 PAR) flasks were placed inside an aluminum foil covered box, low light (66 M photons m2 s1 PAR) flasks were covered by a screen mesh material (1 mm square mesh) and high light flasks were exposed to direct light (115 M photons m2 s1 PAR) from a fluorescent light source. Following incubation (2 3 hours), flasks were shaken to equilibrate the gas phases and sampled for headspace gas samples (4 ml) which were stored in evacuated, 3.5 -ml serum bottles with gray, butyl rubber stoppe rs and aluminum crimp seals. All gas samples were analyzed for ethylene within one week of incubation using a Shimadzu GC 8A gas chromatograph equipped with flame ionization detector (110oC). G as separation was done at 80oC using a six foot,
31 Poropak N colu mn (Supelco, Bellefonte, PA). Ethylene concentrations were verified using a premixed standard gas (10 ppm) (Scott Specialty Gases, Inc., Plumsteadville, PA). For the calculation of total AR per flask, ethylene production was derived from both headspace and aqueous phase volumes. The volume of ethylene in the aqueous phase was calculated by multiplying the headspace ethylene concentration by a tabulated solubility constant (Henrys Law under standard temp erture and pressure) which was multiplied by the water volume of the sample. Using the Universal Gas law (under standard temperture and pressure of 1.1 atmosphere) the volume of ethylene was converted to moles of ethylene produce which was then based per volume of water filtered and length of incubation. P= a bsolute pressure of ethylene (1) V= volume of ethylene generated n = number of moles ethylene generated R= 0.08205784 L atm K1 mol1 T= absolute temperature in Kelvin (K) The result yielded the number of moles of ethylene produced in one liter of lake wat er per hour. Finally, the theoretical conversion was used to estimate the actual rate of N2 fixation (Howarth et al., 1988b). The result was thus expressed as nmol N fixed per liter per hour of incubation. Physicochemical and Algal Taxonomic Analysis Sampl es of integrated lake water were taken and analyzed for physicochemical and nutrient p a rameters by SJRWMD according to their protocols (SJRWMD 2002). Total organic C (TOC) was calc u lated indirectly as the difference between total C (TC ) and total inorganic C (TIC); particulate organic carbon (POC) was calc u lated as the difference between TOC and dissolved organic carbon (DOC) ; nitrate/nitrite ( NOx) was calc u lated as the difference between total N ( TN ) and total Kjeldahl N (TKN) ; dissolved organic N was calc u lated as the difference
32 between dissolved Kjeldahl N (TKN D) and ammonium ( NH4 +); particulate organic N (PON) was calc u lated as the difference between TKN and TKN D; dissolved inorganic N (DIN) was calculated as the sum of NH4 + and NOx; dissolved organic P was calc u lated as the difference between total dissolved P (TP D) and dissolved reactive P (DRP) M icroscopic analysis of densities and biovolume of cyanobacteria, heterocystous cyanobacteria and heterocysts were performed by M. Cichra at the School of F orest Resources and Conservation (Fisheries and Aquatic Science), UF, Gaines ville Statistical Analysis For s tatistical analysis of seasonal patterns the study period was divided into two sub periods that each included two AR peaks (4/08 6/17 and 6/24 9/02). Statistical analysis was done using SAS and JMP, Version 6 (SAS Institute Inc., Cary, NC). T ype III tests of fixed effects ANOVA model based on AR rate with season, transect location and light treatment as main effects. Multiple comparison of mea ns were conducted for significant light treatment ANOVA result (P value <0.05) using Duncans Multiple Range procedure. Canonical analysis was used to separate the transect sites on the basis of their nutrient composition and environmental parameters. Log transformations were used to improve normality based on standard skewness and kurtosis values. Variables included in the data matrices were DOC, log NH4 +, log DRP, DON, log TN:TP, and conductivity. To evaluate the order of influence of water chemistry on A R rate, both forward and backward, stepwise multiple regression were done after screening auto -correlated limnological variables (among them chl -a and biological oxygen demand ( BOD )). Several analyses including 3 -way ANOVA, stepwise analysis, and canonica l correlation were done after interchanging different variables such as water temperature, pH, biovolume of heterocystous cyanobacteria, and concentration ratios of particulate and dissolved C, N and P (POC:PON, PON:PP, PON:PP, DOC:DON, DIN:DRP, and DOC:DO P )
33 Results and Discussion The mean ARA at all sites demonstrated the high variability of nitrogenase activity (NA) during the study period, with a range of 0 165 nmol C2H4 l1 h1 (depending on conditions). The estimated rate of N2 fixation was expresse d on the basis of the lakes water volume ( daily average of 550 g N m3 h1, or 20 nmol N2 l1 h1). The conversion of AR rates to the lakes surface area was done by dividing the AR rate per m3 by the lakes average depth (2.5) resulting in a daily avera ge of 220 g N m2 h1. These measured N2 fixation rates were within the range s observed in other subtropical eutrophic lakes measured ( Table 2 3). Nitrogenase activity showed a pattern consisting of two sub -periods of high AR rates separated by low activi ty ( Figure 2 2) Rates during t he first sub -period (5/20 6/17) w ere higher (138 and 165 nmol C2H4 l1 h1, respectively ) than those of the second sub -period (8/6 9/2 ). Like any met hod, ARA has some disadva n tages. It measures the potential rate of N2 f ixation but it might introduce some biases such as exposure to O2 during filtration, variation in bioenergetics and reducing cofactor levels, presence of other non N2 fixing organisms that can reduce acetylene (e.g. methanogens (Flett et al., 1975)) or oxi dize ethylene (e.g., methane oxidizing bacteria) and grazing of diazotrophs. After considering these disadvantages, ARA still remains a reliable method and its results in this study were consistent with literature. Thus, it was possible to use AR rate s r ecorded during this study (2008) to roughly quantify the total amount of N that was fixed in Lake George water during the study period and examine its contribution to yearly N output A cetylene reduction rate measured in each transect w as scaled up using a model of Lake George prepared by SJRWMD which separates the lake into 320 sections with associated depths and volumes T his model was divided into three lake portions to correspond to the three transects used for sampling. The volume of each transect was multiplied by its corresponding AR rate at
34 each of the 22 weeks of study and multiplied by the number of hours per week (the number of hours used is explained next). Among the assumptions used in this calculated estimate were: a) AR rate under high light a pproximate s fixation rate during the day; b) AR rate under dark incubation represent s fixation rate during the night; c) light penetrated through the entire water column; d) both light and dark periods were 12 hours per day. Results from c alculation of the yearly estimate were expressed as metric tones (MT, or 106 g) of N. Results suggested that during the study from late Spring through early Fall, N2 fixation rate in the dark amounted to 135 MT of N, while daytime N2 fixation accounted for 462 MT N to the lake F or a rough estimate of the amount of N fixed during the rest of the year (outside the study period) an additional 131 MT N was estimated (based on an assumed constant rate of 28 ng N l1h 1). Thus, the total N fixed during the study period is estim ated to be between 600 730 MT during the entire year of 2008. This quantity was of the same magnitude, but slightly larger than the estimated yearly amount of exported N (600 MT SJRWMD ). On an annual basis t his equates to approximately 3 4 g N m2 adde d to Lake George Although a similar pattern of AR was observed at all the sites, both the west and center sites appeared to be more closely related (except on 5/20) compared to the east site ( Figure 2 3). The highest AR rate was measured on 5/20 at the w e st site During the period of 6/07 to 7/15, AR remained consistently be low 30 nmol C2H4 l1 h1. When looking at effectors on AR rate results of three way ANOVA were almost identical in both sub -periods ( Table 2 2 ) and showed significant (P< 0.005) effec ts by date, site and their interaction (except to interaction between site s and light that was not significant during the first sub period ). These results support ed our hypothes is that light level, spatial distribution and dat e /season are potential regula tors of N2 fixation
35 Depending on light level AR also varied greatly, and trends of AR in both sub -periods were similar ( F ig ures 2 2 and 2 4 ). Using selected light treatments enabled us to estimate both the amount of N2 fixation in a 24 h cycle and the le vel of PAR required to achieve light saturation Before comparing the effect of light on AR, rates were normalized by dividing each AR rate by the highest rate measured on that day. Normalized AR rate s under the different light incubations showed similar t emporal pattern yet varied by their activity rate ( Figure 2 4 ). These results under different light incubations showed a clear difference between N2 fixation in the dark (with a mean of 20 nmol C2H4 l1 h1 or normalized mean of 0.14 nmol C2H4 l1 h1) v er sus light (with a mean of 54 nmol C2H4 l1 h1 or normalized mean of 0.48 nmol C2H4 l1 h1) ( Figure 2 4 ). This result reinforces the assumption that the dominant diazotrophs in Lake George use energy from photosynthesis to fix N2 and are able to protect t heir nitrogenase from photosynthetically generated O2. The lack of significant difference between AR under low (with a mean of 54 nmol C2H4 l1 h1 or normalized mean of 0.48 nmol C2H4 l1 h1) and high (with a mean of 58 nmol C2H4 l1 h1 or normalized me an of 0.6 nmol C2H4 l1 h1) light incubations ( Figure 2 4) indicates that light saturation of this diazotrophic community was likely achieved within the light levels of this incubation (up to 115 mol m2 s1 PAR) S aturation at such low light levels is g enerally indicative of cyanobacterial dominance in the phytoplankton (Grimshaw et al., 1993; Havens 2003). Moreover, low light levels might be ideal for organisms that dwell in the deeper areas or in the sediment and are more sensitive to radiation (Grims haw et al., 1997) and indirect damage from high temperatures reduces AR by inhibition of the O2 protective mechanisms surrounding nitrogenase (Gallon et al., 1993).
36 Limnological P arameter s The proximity and dependence of phytoplankton on their environment suggest the importance of limnological parameters in the study of N2 fixation (Havens, 2003). Thus the concentration and relative proportions of nutrients in the water column shape phytoplankton community compos ition, growth and N2 fixation rate per unit of limiting nutrient. Nutrients that affect N2 fixation are most generally associated with N limitation, thus promoting a community shift toward diazotrophs and increased rates of N2 fixation Among these nutrients are P (DRP and TP), N ( DIN, NOx, NH4 +,T N) and C (DOC, TC) (Hecky and Kilham 1988; Howarth et al., 1988(a) ). Important limn o logical parameters also showed variation between sites ( Table 2 1), and many of their mean concentrations in the water showed temporal patterns similar to AR pattern. The bimodal pattern was predominant among several parameters including chl -a dissolved oxygen (DO) POC:PON ratio, p H, and conductivity, while DRP had an inverse pattern (Figure 2 5 ). S alinity and silica (SiO2) showed same trend and together with water temper ature had lowest values during the first weeks of study but reached their highest points in early June and remained high until end of August ( Figure 2 5 ). I n contrast total dissolved Kjeldahl N (TDKN), dissolved organic carbon (DOC), and the DIN:DRP ratio showed no clear patterns Among possible explanations to observed seasonal trend are higher precipitation rate observed during the second sub-period (data not included). Dissolved organic N (DON), which like other organic forms is usually connected to non -point sources was almost constant during the entire study period with an average of 0.7 mg l1 (ranged from 0.6 to 0.83 mg1). Inter e st ing ly, during the low AR activity measured around the middle of study period ( start of June to mid July), nitrate/nitr ite ( NOx) levels peaked and showed a bimodal peak (Figure 2 7 ). A mmonia (NH4 +) remained low, yet at the first week of July it peaked to its second highest level (0.016 mg l1) (Figure 2 7 ). U nlike the first and last peaks t he second peak of NOx appeared
37 d uring low A R activity Among possible reasons that may explain the trends of NH4 + and NOx are point sources, flux from sediment, inflow of seawater (Morris 2000; Malecki et al. 2004) and nearby environmental sources (e.g. agricultural land ). Ammonium con centrations were low in the aerobic surface waters probably due to biotic uptake or oxidation except for a peak on 07/01 that coincided with a decrease in NOx (decreased by 0.007 mg l1 followed by an increase of 0.006 mg l1) (Figure 2 7 ). In addition, it may be a typical mid -summer peak in NH4 +, and DRP concentrations in surface waters, from imbalances between assimilation and mineralization, that has been described in other systems ( Kemp 1989). DOP and DRP did not show a clear trend during our study bu t increased afterwards (after September) in agreement with observation from previous years (SJRWMD report, 2002). Levels of DRP were consistently low and did not show significant correlation with AR A, yet high levels of chl -a might suggest th at P was taken by phytoplankton and thus was not as limiting as N. In their work, Tilman et al., (1982) suggested us ing the resource competition theory as an explanation to species dominance by looking at the type and levels of resources required by each species. Thus, a focus on seasonal population dynamics and succession might be an approach that explains the competition for nutrients ( e.g., diatoms that compete better for P given that N and silicon are not limiting). Measured nutrient flux (DRP and NH4 +) from sediment to water column at LSJR showed a strong negative relationship with O2 availability in the water column and suggested that phytoplankton blooms will increase nutrient flux from the sediment (Malecki et al ., 2004). For the above reasons, it is more likely t hat P was released during bloom events and dark periods when DO levels were lower. As a result of released P, nutrient limitation shifts to N and N2 fixation increases.
38 Evaluation of N utrient L imitation Probably the most common method to evaluate nutrient limitation was through examination of the nutrient levels and their ratio in the water column and phytoplankton biomass. In his paper, Redfield (1934) identified a constant relationship between nutrient levels in the marine environment and their phytoplan kton biomass. He identified the stoichiometric ratio of C:N:P to be 106:16:1. Since its publication, the Redfield ratio h as been extensively used to estimate phytoplankton nutrient limitation in both marine and fresh water systems. The Redfield ratio of mo lar TN:TP suggested that Lake George was limited mostly by P with several periods of N limitation ( Figure 2 6) P articulate C:N can largely be looked upon as the C:N ratio of algal biomass, and thus, can also be used to identify relative patterns of C and N availability. In this study, the w eight ratio of POC:PON ranged from 1.2 to 10.3 with an average of 4.4 (Figure 2 5). In this approach, values lower than 7 can be used to in fer N limitation (Redfield, 1934) indicating that Lake George was mostly N limit ed during the period of this study. In contrast to the Redfield ratio, a n alternative metric using DIN:TP was found by Morris and Lewis (1988) t o more accurate ly predict nutrient limitation of phytoplankton. This met ric classified lakes with a DIN:TP ratio (by weight) of <0.6 as N limited and lakes with a ratio > 4 as P limited, while those with ratios between 0.6 and 4 were considered to be under both N and P limitation. When using this approach in Lake George the averaged DIN:DRP molar ratio was lower th an 0.6 during most of the study period indicating a N -limited system (Figure 2 6). Thus, the two metrics suggested different nutrient limitation: While a ccording to the Redfield ratio, N limitation was short and began the first week of June until the end o f July (7/22), according to the DIN:TP weight ratio of Morris and Lewis (1 988) N was limited most of the study period (lower than 0.6) except for one month around the beginning of July and during the last date of sampling ( Figure 2 6). The high N2 fixation rates measured in the lakes water indicated a N
39 limited system, suggesting that the metric of Morris and Lewis (DIN:TP weight ratio ) was more accurate than Redfield ratio in this system. Additional support for this conclusion is found in, the fact that c lose association of AR A with the DIN:TP weight ratio (Figure 2 6). In natural environments i dentification of nutrient limitation by coupling concentrations to primary producers growth i s difficult due to system complexity and dynamics (Hecky, 1988). In o rder to identify and rank the parameters most closely associated to N2 fixation a s tepwise multiple regression models were developed for both sub-periods ( Table 2 4). In the first subperiod, the stepwise model ranked DO as the most correlated parameter t o ARA followed by NOx (negatively correlated to ARA) (R2 = 0.48 (P 0.0005)). On the other hand, the model of the second sub-period ranked POC:PON ratio (negatively correlated with ARA) as the most correlated parameter to ARA (P=0.001) followed by DON (negatively correlated), TDKN, NH4 + and water temperature (negatively correlated) (R2 = 0.9 8 ); the negative correlations implied that the se parameters were either inhibitory to or resulted from the outcome of N2 fixation while N (TKN and NH4 + during second sub-period) was the most strongly related parameter to N2 fixation The positive correlation of DO with ARA is likely the result of the importance of heterocystous cyanobacteria in the diazotrophic community. The reason P limitation was less apparent in the water column might be due to internal storing of P by cyanobacter ia (SJRWMD, 2002) and P deposits in sediment that can be used when N enters the system. In addition, readily available forms of P (mostly as DRP) are not expected to remain in the water column for long periods of time (ca. days) due to fast uptake by cyano bacteria. The result from predictors of nutrient limitation supported previous studies in Lake George (SJRWMD, 2002 Annual Project Report) which demonstrated a community that was limited mostly by N but also co -limited by both N and P. It was also noticed that although
40 diazotrophic activity constantly introduced new N into the water column, N was generally deficient relative to P and therefore P was not the sole limiting nutrient. The ability of N2 fixation to shift limitation from N to P may vary on diff erent locations or time periods but may be reflected by different community during bloom events P ossible explanations for the variation between nutrient limitation and diazotrophic activity in Lake George could include differences in the stoichiometric re quirements of the phytoplankton (Arrigo 2005) possible N storage mechanisms in periods of high N availability nutrient distribution/sources, and seasonal chang es (e.g., formation of thermocline followed by depletion of the initial levels of N as seen in other eutrophic lakes and estuaries (Schindler et al., 2008) The importance of distribution of environmental parameters can be seen in the resource competition theory and may explain species dominance and the spatial/temporal factors that play an import ant role. Due to heterogeneity of aquatic systems, three factors relating algal communities must be considered : a ) all phytoplankton are motile (the non-motile are affected by water currents and Brownian motion) ; b ) nutrient uptake is not directly coupled to reproduction; and that c ) the water column is heterogeneous (Tilman et al. 1982). According to t his theory, various niches and individual algal blooms can be limited by a single nutrient (e .g., dia zotrophs by P ), yet preventing the lake from being comp letely described as limited by a single nutrient. Thus as mentioned earlier, an approach that focuses on seasonal population dynamics and their succession offer a better explana tion for the competition for nutrients inside the system or between trophic le vels (like diatoms). The spatial distribution of environmental factors can be used to asses their importance on the distribution of N2 fixatio n. Using measured limnological parameters, canonical analysis was performed to identify the distributions of envi ronmental parameters that distinguish ed between
41 the three transects in order to identify pattern seen by ARA (Figure 2 1 0 ). The analysis showed that DOC, log NH4 +, log dissolved reactive P (DRP), DON, log TN:TP, and conductivity were the most different bet ween the east and west transects during the entire study period; on the other hand, during the first sub-period the central transect was similar to the east while on the second sub -period it shifted to be more like the west. Th e fact that there were differ ences in the distribution of water chemistry parameters between the two sub-periods suggested that an environmental change took place. Among possible explanations to observed seasonal trend are higher precipitation rate observed during the second sub-perio d (data not included) wave action and spring inputs that can form a gradient of decreasing conductivity out from the west shoreline depending on flow rate (Stewart et al., 2006). Although the similarity in nutrient distribution was not identical to ARA di stribution, it demonstrated their association and support ed the assumption that nutrient distribution could be affect ing N2 fixation rates. Our results indicated that water by the shorelines (east and west) of the lake ha d different composition. Several of these parameters (mostly N and P) are known to affect AR and their distribution is expected to influence N2 fix ation rate and phytoplankton abundance. During the first sub-period, the similarity in water composition of the west and center sites compared t o the east was also observed in AR rates however, while t he AR trends were kept during the second sub-period unlike nutrient distribution that shifted and could not be related to N2 fix ing. Microscopic A nalysis M icroscopy identified biovolume and density of Cylindrospermopsis raciborskii as the dominant genus in Lake George during our study period followed by Anabaena (Figure 2 9 ). While the first bloom was formed by both Anabaena and Cylindrospermopsi s the remaining blooms were formed mostly by Cylindro spermopsi s.
42 Both Anabaena and Cylindrospermopsis are considered to be opportunist s and are fitted for such systems for several reasons including: a ) they exhibit N2 fixation under N limited conditions b ) they form heterocysts, c) they have the ability to store P under saturated conditions and efficient buoyancy regulation, d) they have a r elatively low light requirement, e) they have the ability to uptake DIN and f) they have a high temperature and salinity resistance Cylindrospermopsis was more fitted t han Anabaena in this environment probably due to its ability to scavenge P rapidly under low concentration smaller cells and higher number of heterocysts. Thus, it is possible that after the first bloom P levels were very low selecting for the scavenger Cylindrospermopsi s. Due to their ability to uptake DIN and to fix N2, diazotrophs were expected to be mostly P limited; thus, P probably permits their blooms and higher N2 fixation rate in Lake George. These factors increase the difficulty in understanding and predicting phytoplankton blooms (namely cyanobacteria and diazotrophs abundance) in subtropical aquatic systems by looking at environmental conditions mostly in the water column, and their influence on N2 fixation. In addition, m icroscopic analysis re vealed that cyanobacteria dominated the phytoplankton community (E.J. Phlips, personal communication) During the period of study, bloom events measured by microscopy (densities and biovolume s ) showed temporal trends which seemed to follow AR rates and nut rient availability (DON, DIN, NOx, NH4 +, POC:PON ratio) ( Figures 2 8 and 2 9) The lowest concentration of cyanobacterial biovolume was measured at the first and last days of study (4/08/08 and 9/02/08) at all except for the East sites where biovolume s inc reased at the end of study, and the highest biovolume was between 5/20/08 5/27/08. In agreement with AR peaks, biovolume s of heterocyst ou s cyanobacteria yielded four peaks (the first by Anaba ena and Cylindrospermopsis and the others by Cylindrospermopsi s )
43 supporting the assumption that most of the N2 fixation is done by heterocystous cyanobacteria. Total c yanobacteria biovolume followed similar trend however with a lag compared to heterocystous cyanobacteria. The lag and fact that total cyanobacteria bio volume were low between the second and third blooms when heterocystous cyanobacteria biovolume were very low may be due to their dependence on fixed N (Figure 2 8). The result support ed the assumption that heterocystou s c yanobacteria were closely associate d with the entire cyanobacterial community and supplied them with fixed N (Schindler et al., 2008). The heterocystou s Anabae na and Cylindrospermopsis were the dominant diazotrophs and their association with AR rate demonstrated their important role in biol ogical N input in to the lake (Figure 2 9 ). The period around the middle of study (start of June to mid July), was between blooms of heterocystous cyanobacteria (their biovolume levels were at their lowest) and as expected showed low rates of ARA ( Figures 2 3, 2 8 and 2 9) In order to avoid the comparison of AR rates that were attributable to differences in heterocyst ou s cya nobacteria biovolume, AR rates were normalized1 by dividing each AR rate by the heterocystous cyanobacteria biovolume at that day. Norm alized AR rate s were expressed as AR rate s in terms of moles of C2H4 generated per unit of heterocyst ou s cya nobacteria biovolume ( l ) and time (h ), a comparison was done between samples. In addition to remov ing variability due to difference in numbers of he terocystous cyanobacteria normalized rates had the potential to identify the importance of individual heterocystous cyanobacteria within the cyanobacterial community (because different gen era of heterocystous cyanobacteria have different sizes and numbers of heterocysts per biovolume, thus resulting in different AR rates). 1 Also normalized to cyanobacteria biovolume expressed as N fixed per 3 cyanobacteria, the center on 4/22 was low (1.4) and ranged between 10 -17 mol to 10 -21 mol N per 3 cyanobacteria.
44 According to results, the highest AR rates were at center of lake and the lowest were at the west sites. Normalized AR rates ranged from 2.6 1010 to 3.9 108 N m3 heterocystous cy anobacteria h1. In addition, the centra l transect had one N2 peak between the second and third bloom s ( 6/24 to 7/ 15), when biovolume s of heterocyst ous cyanobacteria were at their lowest This pattern of cent ral stations correlated with NOx peaks, suggesti ng that N2 fixation in heterocyst ou s cyanobacteria may not have been inhibited by NOx or that they released fixed N. The increased efficien c y of fixation per biovolume may be due to low biomass or environmental conditions (e.g., water temperature) during t hat period Conclusions Based on the seasonal measurements of AR, we can conclude that N2 fixation was an important driver of cyanobacterial bloom formation ( perhaps through supplied N) and resulted in a significant contribution to the lakes N budge t (730 MT). This amount was within the range measured in other systems (e.g., Clear Lake (oak arm), CA ( Table 2 3)) and could explain a significant portion of the ~600 MT N increase observed in the outflow of Lake George I n addition to this N, there are many ot her N sources affecting the amount of N export from Lake George, including other sources of N2 fixation (e.g., from sediments), atmospheric deposition, groundwater discharges, and surface runoff from the areas surrounding the lake T he combination of these N sources with the amount of watercolumn N2 fixation estimated in this study is potentially much larger than the total amount of measured N export from the lake. However, it is important to remember that other N cycle processes such as sediment accumula tion and denitrification can also remove N from the lake I nclusion of these processes could balance much of the N inputs, and therefore, it is quite possible that the amount of N derived from watercolumn N2 fixation in this study is reasonable.
45 Light ha d a significant effect on AR rates. While ARA was used to measure nitrogenase activity, incubation under different light levels de m onstrated the degree to which N2 fixation was stimulated by light and that N2 was fixed mostly by phot o trophs requir ing low l ight levels. This conclusion was supported by the measured abundance of heterocystous cyanobacteria, and as well, by the positive correlat ion of AR with DO. These heterocystous cyanobacteria formed the first bloom and appeared to promote proliferation of t he entire cyanobacterial community as noticed by similar patterns of the two biovolumes. The lag seen in biovolume of total cyanobacteria compared to heterocystous cyanobacteria supported the idea of nutrient exchange within this community and the importance of diazotrophs to bloom formations. The overall patterns of nitrogenase activity indicated the importance of location and seasonality as regulator s of N2 fixation and diazotrophic growth in Lake George Patterns of algal blooms and selected nutrient s v aried seasonal l y and spatially as well, showing similar patterns to AR and indicating their association with observed rates of N2 fixation ( Figure 2 8). Moreover 3 -way ANOVA supported this conclusion with significant effects of both site and date on AR ra tes ( Table 2 2 ). In particular forms of N and P (DIN NH4 + and DRP) somewhat explain the differences observed between the measurements of the East West and Central portions of the lake. Heterogeneity may be explained by different water sources, comprisi ng mostly of the main river water source ( from the S t. J ohns) and other inputs from lake shorelines. The difference in parameters (conductivity, pH, DIN, DRP and NH4 +) seen in the canonical analysis between the first sub -period and the second subperiod po tentially showed the importance of external nutrient inputs which could include water inflow from upstream (SJR) and the west shoreline (by Artesian springs), natural events (such as rain events and water m ixing by wind). The apparent change in the distribution of water parameters (especially,
46 conductivity and pH) agreed with higher precipitation rate recorded during the second sub-period (data not included). Nutrient distribution or nutrient levels are known to control N2 fixation and algal bloom formation, and t he effect of environmental conditions on the diazotrophic community clearly indicated a shift from Anabaena to Cylindrospermopsis. The ability of Anabaena to uptake P under moderate levels faster than Cylindrospermopsis can explain its abundance in the first bloom, while the domination of Cylindrospermopsis during the rest of the year is probably due to its ability to scavenge P rapidly under low concentration Never the less, it was not precisely clear how N and P, or which mechanisms control diazo trophic composition and domination of single species. Due to low DRP concentration in the water, this parameter did not seem to correlate to N2 fixation, yet the N:P r a tio suggested a N limited system that agreed with higher AR rates. H igher P, and lower N:P ratio (DIN:DRP or TN:TP ratio) were expected to increase demand for fixed N. The indicator DIN:DRP seemed to better r e flect nutrient limitation in Lake George more accurately than Redfield ratio when compared with AR rates. Acetylene Reduction in late summer was lower than early summer probably due to the higher ratio of non heterocystous species of cyanobacteria that were seen by microscopy and can be explained by higher ratio of N:P (DIN:TP). It is also possible that P was deposited in the sediment or biomass of nondiazotrophic phytoplankton several months before bloom formation, or that there was a relatively constant supply of P obtained from the sediments. Apart from P, h e t erocystous cyanobacteria are also known to be affected by factors such as sa lt, high temperatures and intense sunlight that have adverse effects on many biochemical processes and might decrease N2 fixation rate. As mentioned above, various nutrient sources and a nthropogenic activities likely have change d the balance of both N an d P, in Lake George and affect ed phytoplankton community
47 structure (composition and biomass) and function (primary productivity and N2 fixation) during the study period. These alterations in the biogeochemical cycl es in Lake George supported conditions of N limitation (as indicated by N:P ratios) and diazotrophic proliferation. Fixed N acts to alleviate the demand for N in Lake George, and also turned the lake into a source of N to downstream systems These results clearly demonstrated that role which inland waterbodies like Lake George and processes such as N2 fixation can play in the health and function of entire watersheds Some of the limitations of this work include failure to correlate AR with nutrient distribution, use of few measurements to capture spatial and temporal patterns, understanding of bloom dynamics (shift from Anabaena to Cylindrospermopsis) and interaction within the phytoplankton community. For this reason, a dditional work should be focused on increas ing the ability to predict, explain, and control nutrients/environmental conditions that lead to algal blooms in Lake George and similar systems. Some improvements for future studies should potentially include the identification of specific nutrients that may be transient or pulsed and act t o initiate bloom formation T racing N sources and fixed N will help to identify external N supply as well as potential N transfer between cyanobacteria and other phytoplankton groups. Moreover, several ARA measurements during fall and winter months and bet ter understanding of light intensity or diurnal patterns would help improve the calculation of system N budget. A ccording to this study, the restoration and control of the phytoplankton community is complex tasks that cannot be achieved solely by reducing single nutrient inputs I n fact, such a reduction has been done in the last years without much success and even leading to accele rated eutrophication (Conley et al., 2009; Howarth and Marino, 2006). For this reason, other remedi a tion approaches (e.g., int roduction of submerged macrophytes to reduce sediment
48 resuspension and wave action thus limi t ing internal nutrient sources) may also be cautiously combined with nutrient reduction to reduce growth of cyanobacteria and other diazotrophs in the Lake.
49 Tabl e 2 1. Characteristic values of selected chemical parameters in Lake George water. Values represent the means of all samples collected for each transect during the study period (Analyzed by SJRWMD). Dissolved oxygen (DO) ; total nitrogen (TN); total phosphorus (TP); chlorophyll a (chl -a ); dissolved inorganic nitrogen (DIN); dissolved organic nitrogen (DON); dissolved reactive phosphorus (DRP); dissolved organic phosphorus (DOP). Table 2 2 Type 3 test of fixed effects based on nitrogenase activity rate (ARA) for the two sub periods (n=99). Effect Time Period -------------P -------------April June July September Site 0.01* <.0001* Date <.0001* <.0001* Light <.0001* <.0001* Site date <.0001 <.0001* Date light <0.002* <.0001* Site light 0.49 <.0004* indicates a statistically significant difference ( P Site DO (mg l 1) TN (mg l 1) TP (mg l 1) pH Conduct. (S cm 1) Chl a (mg l 1) DIN (mg l 1) DON (mg l 1) DRP (mg l 1) DOP (mg l 1) East 6.97 1.29 0.06 8.4 1561 32.5 0.03 0.6 0.003 0.008 West 7.15 1.20 0.05 8.5 1605 32.3 0.03 0.6 0.003 0.009 Center 7.02 1.25 0.06 8.5 1565 35.2 0.02 0.6 0.003 0.008
50 Table 2 3 Nitrogenase activity (measured using ARA) at various types of aquatic ecosystems. Unit System 3 h 1 nmol N l 1 h 1 g N m 2 y 1 nmole C 2 H 4 l 1 h 1 1 h 1 Reference Lake George (potential) 8760 52 4 156 1.5 This Study Lake George 78 Paerl et al., 2000. Eutrophic lake Valencia, Venezuela 1.3 Levine and Lewis, 1987. Eutrophic lake Lyngby, Denmark 80 Ahmad, 1981. Eutrophic Clear lake (oak arm) CA 310 (total 500 MT N y1) Horne and goldman, 1972. Eutrophic Reservoir Waco, TX 0.6 Scott et al., 2009. Meso trophic Lake Washington, WA. 0.013 Tison et al., 1977. Temperate Amazon flood plain, Lake Calado 30 Doyle and Fisher, 1994. North Sea in Europe 1450 Howarth et al. 1996.
51 Table 2 4 Model parameters for stepwise multiple regression analysis of log transformed AR rates of sub period 1 and 2. Models are presented for combined transect samplings. Sampling Date Model R 2 (n) Variables Included Estimate SE Pr> F Sub period 1 4/08 6/17/08 0.617 (23) Constant 238.49 51.4991 <0.0001* DO 38.7184 6.65975 NO x 95.291 43.1937 Sub period 2 6/249/02/08 0.9786 (29) Constant 21.6709 1.06814 <0.0001* POC:PON ratio 338.53 0.99719 DON 436.35 21.0301 TKN T 19.9688 1.13567 NH 4 + 67.3339 15.0517 Water temperature 110.49 41.8737 indicates a statistically significant difference ( P
52 Figure 2 1 Map showing the location of Lake George in Florida including its main water source and output (SJR).
53 0 50 100 150 200 250 300 0 4.5 9 13.5 18 22.5 Dark Low light High lightARA rate ( nmol C2H4l1h1) April May June July August September Subperiod 1 Subperiod 2 0 50 100 150 200 250 300 0 4.5 9 13.5 18 22.5 Dark Low light High light 0 50 100 150 200 250 300 0 4.5 9 13.5 18 22.5 Dark Low light High lightARA rate ( nmol C2H4l1h1) April May June July August September Subperiod 1 Subperiod 2 Figure 2 2 Temporal distribution of nitrogenase activity ( measured by acetylene reduction assay ) under three light levels (115, 66 and 0 M m2 S1 PAR) for Lake George water. Dotted line separates between the two sub periods. Values represent the mean of nine r eplicates ( 1 SE) obtained for each treatment during study period in 2008. Figure 2 3 Spatial and temporal distribution of nitrogenase activity (measured using acetylene reduction assay) under high light incubation (115 M m2 S1 PAR) for Lake Georg e water during study period (2008). Values represent the mean of three replicates ( 1 SE) obtained from each transect.
54 Dark 0 ( M photon m-2s-1PAR) Low 66 ( M photon m-2s-1PAR) High 115 ( M photon m-2s-1 PAR) Normalized AR ( nmol l1h1) 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Dark 0 ( M photon m-2s-1PAR) Low 66 ( M photon m-2s-1PAR) High 115 ( M photon m-2s-1 PAR) Normalized AR ( nmol l1h1) 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 a b b Figure 2 4. The effect of light intensity on acetylene reduction assay rates in Lake George water during the study period. Each ra te was normalized as the percent of maximum AR value for a given date (see text for description).
55 0 10 20 30 40 50 60 70 80 6.5 7.0 7.5 8.0 8.5 9.0 9.5 0.000 0.005 0.010 0.015 0.020 0.025 17 22 27 32 0 2 4 6 8 10 12 600 800 1000 1200 1400 1600 1800 0 1 2 3 4 5 6 7 8 April May June July August April May June July AugustDRP (mg l1) POC:PON wt. ratio chl a (mg l1) DO (mg l1) Conductivity ( S cm l1)oCSiO2(mg l1) 0 2 4 6 8 10 April May June July August April May June July AugustpH Figure 2 5 Temporal patterns of environmental parameters Points represent the means of nine replicates ( 1 SE) obtained from Lake George w ater -column A) DRP (dissolved reactive P) B) chlorophyll a (chl -a ). C) dissolved oxygen (DO) D) particulate organic carbon: particulate organic nitrogen ratio (POC:PON) E) conductivity F) dissolved silica (SiO2). G) pH H) water temperature (Celsius) All analyses were done by the SJRWMD.
56 Figure 2 6 Temporal patterns of N:P ratio in Lake George water -column. Points represent means of nine replicates. A) molar t otal N to total P ratio (TN:TP). B) dissolved inorganic N to dissolved reactive P ratio by weight (DIN:DRP). Redfield ratio and Morris and Lewis ratio are marked by dotted red line.
57 0 0.02 0.04 0.06 NH4 D NOx Dmg l1 April May June Ju ly August September 0 0.02 0.04 0.06 NH4 D NOx D 0 0.02 0.04 0.06 NH4 D NOx Dmg l1 April May June Ju ly August September Figure 2 7. Temporal patterns of labile N: ammonium (NH4 +) and nitrite/nitrate (NOx) in Lake George water -column during study period. Points represent means of nine replicates.
58 Figure 2 8 Spatial and temporal patterns of cyanobacteria biovolume (m3 ml1). Points represent the means of three replicates ( 1 SE) obtained from Lake George water column A ) heterocystous cyanobacteria B) total cyanobacteria. Microscopic analysis by Cichra and Phlips.
59 Figure 2 9 Spatial and temporal patterns of heterocystous cyanobacteria biovolume (m3 ml1). Points represent the means of three replicates ( 1 SE) obtained fr om Lake George water -column. A ) Cylindrospermopsis .B ) Anabaena. Microscopic analysis by Cichra and Phlips.
60 12 15 18 21 C E W DOC SiO2-D Water_Temp pH-Field DON log_Nox Conduct_norm log_DRP log_DOP log_NH4 log_TN:TP log_DIN -3 0 3 6 9 -96 -94 -92 -90 -88 C E W DOC SiO2-D Water_Temp pH-Field DON log_Nox Conduct_norm log_DRP log_DOP log_NH4 log_TN:TP log_DIN -24 -22 -20 -18 -16 Canonical 1 (78%)Canonical 2 (22%) Canonical 2 (10%)Canonical 1 (90%) West West East East Center Center Figure 2 10. Separation of lake transect locations by Canonical analysis based on limnlogical parameters in Lake George water column A) fir st sub -period (4/08 6/17) B) second sub-period (6/24 9/02). The m ost influential parameters were dissolved organic C (DOC), log NH4 +, log dissolved reactive P (DRP), dissolved organic N (DON), log TN:TP, and conductivity.
61 CHAPTER 3 MOLECULAR ANALYSIS Dinitrogen gas (N2) is the most abundant gas in the atmosphere comprising about 78% by volume, yet its consumption is limited to an exclusive group of organisms called diazotrophs. The nitrogenase enzyme is a common protein to all diazotrophs due to its ro le in breaking the triple bond of N2 to fix (reduces) it into a bioavailable form, NH4 +. Almost all known nitrogenase enzymes are irreversibly inhibited by oxygen (O2), yet they are the only known group of enzymes that reduces N2. Due to nitrogenase import ance, N availability is often coupled to biogeochemical processes such as photosynthesis and mineralization, thus shaping ecosystem ecology. Under normal conditions, primary producers are not limited by carbon (C) for photosynthesis, but as an ecosystem be comes more limited by available N, restricted primary production occurs, and N2 fixation by diazotrophs has the potential to supply the demanded N into the system ( Flett et al., 1980; Smith 1983). As N limitation increases, higher rates of N2 fixation ar e generally observed (Vitousek and Howarth, 1991), and the ability to fix N2 gives a competitive advantage to diazotrophs over other organisms (Karl et al., 1997). Such conditions promote dominance of diazotrophs, and may result in aquatic systems being so urces of N to connected systems (Schindler, 1977; Tilman et al., 1982). Nutrient levels (especially N) in the water can be correlated to N2 fixation rate or to limitation of other essential nutrients (such as phosphorus (P) and iron (Fe)) probably due t o m icrobial activity (Vitousek and Howarth 1991; Mills et al., 2004). Temporal changes in aquatic systems have both a direct and an indirect impact on nutrient abundance and distributions (Thayer, 1971), thus resulting in seasonal patterns (Baird and Ulanowic z, 1989). These temporal changes are controlled and regulated functions of microbes that are defined by their genetic potential.
62 Nitrogenase Gene Nitrogenase is a complex consisting of two enzymes : dinitrogenase ; and dinitrogen reductase. Dinitrogenase is a tetramer with molybdenum iron ( MoFe ) co -factor composed of two identical subunits encoded by nifD (Lammers and Haselkorn, 1984) and the other two identical subunits are encoded by nifK ( Mazur and Chui, 1982). Dinitrogenase reductase contains an iron cof actor and is composed of two identical subunits encoded by nifH (Mevarech et al., 1980). Although the phylogeny of diazotrophs using nifD 16S rRNA genes (rDNA) and nif H is conserved, some discrepancies are present suggesting that both 16S rDNA and nifH se quences may be ascribed to a lower level of conservation than nifD High level s of conservation make nifD and nifH good candidate genes for monitoring potential shifts in diazotrophs populations Although more research was done using nifH and its GeneBank l ibrary is larger nifD is expected to be more appropriate for systems in which most of the diazotrophic community is composed of cyanobacteria. Moreover, previous studies have indicated that the divergence found in nif D is intermediate between nifH and nif K (Henson, 2005); thus, it is believed that when used as a phylogenetic marker, its resolution among closelyrelated diazotrophic microorganisms is higher and might suffice to distinguish nifD gene family members as well as alternative nitrogenases such as the vanadium -containing enzyme ( Zehr et al., 2003; Henson et al., 2004a ). The n itrogenase g ene operon may be located on plasmid s (for example, Rhizobium meliloti carries two mega plasmids (Honeycutt et al., 1993)) or on the chromosome (for example, in Azor hizobium caulinodans ). From an evolutionary point of view, it is not clear if the symbiotic genes were originally located on the chromosome and evolved through excision, or if they were originally on a plasmid that became incorporated into the chromosome. For example, gene analysis of Klebsiella pneumoniae identified 20 adjacent nif genes that were organized in eight operons within ca. 24 kb (kilo base pairs) of DNA (Arnold et al., 1988). The most common
63 nitrogenase complex employs a MoFe cofactor (the non-protein part of nitrogenase) that exhibits a low efficiency level when fixing N2. As expected, some diazotrophs are able to adapt to different condition s by expressing alternative nitrogenase systems that can result in similar structures of nitrogenase. Although similar in structure, the alternative proteins are less efficient than the Mo protein (Robson et al., 1986) Both vn f and anf are two similar sets of genes which modif ied their N2 fixation mechanism ( vnf assemble d a (vanadium) VFe cofactor to produc e a haloenzyme with similar structure to nitrogenase). For example, when Mo is not available Azobacter use d V to create a less efficient form of the nitrogenase protein; when Mo bec a me available, it inhibit ed the vnf mechanism and resume d its Mo utilizatio n (Robson et al., 1986) In addition, other environmental conditions and nutrient (among them N, P and O2) concentration s are critical signals for the regulation of nif gene expression in most of the diazotrophs. The extremely high energy demand of N2 fixa tion (about 40 moles of ATP for the reduction of one mole of N2 to ammonia) requires efficient r egulation of nif (Witz et al., 1966). The expression of N2 fixation genes i s controlled by cascades of hierarchically organized regulatory genes which enable th e diazotrophs to sense environmental conditions required for N2 fixation and transmit this information to their gene expression. For example, Azospirillum lipoferum is able to regulate its nitrogenase activity at the posttranslational level in response to the presence of combined N (Ludden et al., 1989). While such a control is based on the reversible ADP -ribosylation of nitrogenase reductas e, a similar mechanism that can switch on and off the nitrogenase gene was identified in A. caulinodans (Kush et al., 1985). Diazotrophs that cannot adjust to new conditions would disappear causing a community shift toward more fit
64 organisms; thus, their community composition may infer something about their environmental conditions. In the past, the Polymerase Chain React ion (PCR) was used extensively as a molecular approach to amplify the 16S rDNA of many bacteria and helped characterization of their composition and evolutionary relationships. Although this approach established their relative diversity in general, in most cases its use as a marker did not identify a specific diazotrophic cyanobacteria. In his works, Zehr used nifH genes to identify several probes and primers that were able to target the subunits of the nitrogenase gene (Zehr and McReynolds, 1989). Although nifH of cyanobacteria clustered closely together and were expected to facilitate the design of specific PCR primers, it was not possible to design primers for specific cyanobacterial nifH genes (Zehr and McReynold 1989 ). Later, Henson et al., (2002) used complete nifD sequences to evaluate differentiation between two genera of heterocystous cyanobacteria ( Nostoc and Anabaena) and concluded that they were indeed separated. Later, during the development of molecular techniques that would detect N2 fixing cyanobacteria in environmental samples (complex microbial community), it was found that gene sequences of nifD contained conserved regions that permited the design of PCR primers specific for cyanobacterial nifD (Roselers et al., 2007). For the above reason, we decided to use nifD primers to specifically identify spatial and temporal distribution of diazotrophs in complex microbial communities that may be dominated by cyanobacteria. Cyanobacteria Cyanobacteria are gram negative bacteria that have a thicker p ept i doglycan layer although their polysaccharides are similar to those of gram positive cells (Weckesser and Jurgens, 1998). Although unicellular organisms, t he majority of cyanobacteria are aerobic photoautotrophs and their morphology covers unicellular, colonial and multicellular filamentous forms. Even though
65 photosynthesis is their principal mode of energy metabolism and C source some species we re able to survive long periods in complete darkness and others exhibit diverse heterotrophic nutrition pa tterns ( Fay, 1965). Their oxygenic photosynthesis activity is responsible for the majority of dissolved oxygen (DO) in their water column (especially in the absence of macrophytes) due to low diffusion rate of atmospheric O2 into water. Cyanobacteria are s uperior competitors under eutrophic conditions for several reasons and frequently thrive in nutrient -enriched waters (Vincent 1987). Among the major reasons that could explain their success are : a ) capable of optimal growth under high light conditions; b ) prefer vertically stratified (via either temperature or salinity) conditions (Paerl et al ., 1985); c ) many genera (including pest s ) are buoyant (Reynolds 1987; Reynolds and Walsby 1975); d) th eir growth rate is relatively slow making them favorable unde r longresidence time conditions (Paerl 1998). Although extensively studied, classification of many members of the cyanobacteria has not been completed (Litvaitis, 2002). The phylum Cyanobacteria branches into five taxonomic subsections based mostly on mo rphological characteristics. The simplest forms are unicellular while the more complex are highly structured and may exhibit morphology of branching filaments and even differentiation into specialized cells. Out of the five major subsections, two are those having vegetative cells versus those that form heterocysts. Subsection I is composed of unicellular cyanobacteria that reproduce by budding or binary fission. Subsection II is composed of unicellular cyanobacteria as well but reproduce by internal multipl e fissions. Subsection III is composed of nonheterocystous diazotrophs that form filaments and reproduce by binary fission with unbranched trichomes. Both subsection IV and V are composed of filamentous strains of heterocystous diazotrophs that reproduce by binary fission. Subsection IV reproduce in one plane without true
66 branching of trichomes, while subsection V reproduce in more than one plane, forming true branching (Rippka et al., 1979; Castenholz & Waterbury, 1989). Both subsections IV and V have the ability to differentiate their cells into specialized N2 compartments called heterocysts. These specialized cells enabled N2 fixation during the day and under aerobic conditions. Heterocyst ous Cyanobacteria Th e heterocyst ous structure protect s the O2-s ensitive nitrogenase from O2 that may enter from the environment or from nearby vegetative cells that produce O2 under photosystem II In general, heterocysts are larger than vegetative cells and although their differentiation results from change in gene expression, their frequency may not be solely related to nutrient limitation (Vintila and El -Shehawy, 2007) Outside their cell wall, the heterocyst structure consists of three layers which form a hydrophobic barrier that prevent O2 diffusion. Like other v egetative cells, heterocystous cells can store N in specialized polymers (cyanophycin), yet in addition they contain a honeycomb membrane that exhibits high respiratory activity (Murry et al., 1981); as a result of higher respiration, O2 is depleted rapidly, and more ATP is synthesized for N reduction (Fay, 1992). Their polar region, where they join vegetative cells, is narrow relative to other cells, and the outer cell membrane surrounds the entire filament to permit access of other cells to the periplasm space (Wolk, 1968; Drews and Weckesser, 1982); by using this structure, other members of the filament (vegetative cells) supply needed substrate generated by photosynthesis. It is possible such transfer between cells of the same filament was the evolution ary path to a multicellular organism. Heterocystous cells cannot photoreduce CO2 and use RPP (reductive pentose phosphate) pathway to supply C skeletons for assimilation of fixed N, because they lack photosystem II activity and ribulose bisphosphate carbox ylase (Wolk, 1982). Different levels of labile N (most common are, ammonium, nitrate, nitrite, and urea (Florres and Herrero, 1994))
67 inhibit heterocyst differentiation, although concentration thresholds vary by species F or example, differentiation of Anabaena sp PCC 7120 under ammonium concentration of 3 to 7 M is inhibited. Nostocaceae Anabaena sp. PCC 7120 is probably the most well -studied heterocys t -forming cyanobacteria. Other important heterocys tous cyanobacteria are: Anabaenopsis, Nodularia, Cylindrospermum, Cylindrospermopsis, Scytonema, Calothrix and Fischrella (Rippka et al., 1979). Cylindrospermopsis appeared in many aquatic systems in the past 30 years, and their increased production was associated with long periods of growth in high light intensity conditions (Dyble et al., 2006). Microcystis and t he genus Anabaena are among the most important cyanobacteria in many other eutrophic Florida lakes (e.g., Lake Apopka, Griffin and Okeechobee (Chapman and Schelske 1997; Cichra 1995)); in these systems, heterocysts were the predominant site s for N2 fi xat ion and their establishment wa s probably regulated by several variables including external fixed N2. As mentioned above, cyanobacteria are not the only diazotrophs in aquatic systems. Proteobacteria Several members of the phylum Proteobacteria that carry the nitrogenase enzyme were found in aquatic systems. The species Rhodospirillum centenum (also known as Rhodocista centenaria ) is a member of the class alphaproteobacteria that is capable to fix N2 under aerobic growth conditions We chose R. centenum as the root in the construction of phylogetic analysis, because it clustered outside of the cyanobacterial sequences. As heterotrophic, they could metabolize a unique set of C sources, yet could not use C 4 dicar boxylic acids as a C source. On the other hand, Fogg (1969) found that most c yanobacteria that form ed blooms prefer organic matter (OM) enriched conditions, including OM that comes from agricultural runoff, urban
68 wastewater and soil erosion products (Paerl 1988a). Many of the above mentioned condition s are related to increased rates of urbanization, and affected many aquatic systems in Florida including the St. Johns River ( SJR ) watershed (Hendrickson and Konwinski 1998). In fact, US Southeast riverine and estuar ine waters (including the SJR) are cons idered to be both nutrient enriched and N limited (due to low N:P ratio), which favor cyanobacterial dominance in the phytoplankton community (Smith 1983). Lake George Historically, around 1900, macrophytes (water hyacinth) dominated and covered great are as of the S t. Johns River (SJR) in Florida. Later, the US Ar my Corp s of Engineers used concentrated spraying of herbicides around 1940 to control them (Simberloff and Schmitz 1997). The adver se effect s of decayed macrophytes and introduced herbicide s into the system resulted in the release of nutrients back into the water column that promoted alga l proliferation (also, due to less competition with macrophytes for nutrient acquisition) (Moody, 1970). In addition, without macrophytes, more heat and light cou ld penetrate greater areas of the SJR providing better conditions and energy to cyanobacteria that became dominant. This was seen in many waterbodies including Lake George, which is the biggest water body on the SJR. The composition of diazotrophs in Lake George is of importance due to the coupling of their activity to nutrient levels (mostly P) and their impact on the N budget. In addition, their distribution may imply different inflow sources and niches in the lake and possible relationships with other m embers of the lake community. Dominance of heterocystous cyanobacteria reinforced expected conditions of relatively high dissolved oxygen ( DO) levels and light penetration through most of the water column. Distribution of non-heterocystous diazotrophs is m ore effected by phototrophic activity that introduced DO into water and possible association with primary producers (especially true for heterotrophs).
6 9 Studies conducted to understand the composition and function of phytoplankton community at Lake George h ave focused on measuring limnological parameters, enzyme assays (e,g. photosynthesis), and characterizing community composition using microscopic analysis. Because cyanobacteria are located at the base of the food web, changes in nutrient content and taxon omic composition may be used as an early indicator of trophic state and nutrient limitation. In addition, documenting the compositional changes in diazotrophic assemblages in response to temporal and varying nutrient concentrations is important in understa nding similar systems and enables their comparison. Thus far, only microscopic methods have been used to identify the diazotroph ic community in Lake George and correlate it to seasonal and environmental conditions. Use of molecular approach es is expected t o permit greater resolution inside families or genera and identify organisms not identified by microscop y This work identifies the genetic potential and variation of diazotrophs in an attempt to investigate how environmental factors a ffect their community composition and how their composition determine s N2 fixation at Lake George We hypothesized that nif D diversity will vary with nutrient levels that a re mostl y effected by temporal changes such as precipitation temperature, wind, salinity lake water res idence time and light intensity. Materials and Methods Site Description and Sample Collection Lake George is approximately 21,000 hectares in size, making it the second-largest freshwater lake in Florida, and the largest on the SJR (Figure 3 1). Lake Geor ge is shallow relative to its size, and a thermocline seldom forms in the lake, and its water dynamics appear to be dominated by subtidal variability of the Atlantic Ocean water level (Morris, 1995). Between the years 1996 to 2005, the SJR Water Management District (SJRWMD) estimated the mean
70 turnover rate of the lake to be 84 days. They concluded that peaks in algal biomass were partially controlled by flushing, and that water quality was significantly influenced by local groundwater sources entering by springs. Integrated water samples were collected weekly from nine locations in Lake George during the summer of 2008 (22 weeks starting April 8 to September 2). The sites represented the range of water chemistry parameters at Lake George. Water samples were transferred in three 1 L carboys to the lab where they were homogenized and split into three samples, based on sites, and vacuum filtered through 0.7 m glass fibre prefilters (Millipore CAT No. APFF04700) to concentrate water colmn biomass Filters were frozen at 20oC until DNA extraction. DNA Extraction Each filter was first thawed on ice, then its contents were washed with double distilled water o r Tris acetateEDTA (TAE) buffer into an Eppendorf tube, and centrifuged to keep precipitate. DNA was extracted from approximately 0.05 g of precipitate using an UltraClean Plant DNA Isolation Kit (MoBio, Solana Beach, CA; catalog # 13000 50) and followin g the kit instructions. Extracted DNA was divided into two; one half was stored at 20oC and the second half was combined with 50% ethanol and stored in 80oC until further analysis. Amplification of nif D by Polymerase Chain Reaction (PCR) PCR amplificatio n was conducted using a degenerated primer set that flanks a conserved region of the nifD gene (from position 552 to 861 in the nifD sequence of Anabaena cylindrica PCC 7122, AF442506) and designed by Henson et al., (2002). The nucleotide sequence s for the primers we used were: forward primer nifD552 -F 5 ; and reverse primer nifD861R 5 (MWG Biotech, Huntsville, AL).
71 The reaction mixture used for PCR amplification was composed of 25 l GoTaq Green Master Mix (Promega, Medison, WI), 1 l of each primer (100 pmo l l1), 13 l of distilled water, and 10 l of diluted DNA solution. PCR amplification was done using an iCycler thermal cycler (BIORAD, Hercules, CA) with the following conditions: initial enzyme activation and DNA denaturation of 10 min at 95oC, followe d by 35 cycles of 1 min at 95oC for denaturation, 1 min at 52oC for annealing, and 1 min of extension at 72oC, with a final extension of 72oC for 7 min. After PCR, products were evaluated on 1.5% (wt/vol) agarose gel made in TAE buffer (Sambrock et al., 1989) to confirm the expected fragment size (~ 310 bp). For cloning pGEMT and pGEMT Easy Vector Systems (Promega WI ) were used with a small modification: ligation reaction mix volume was 12 l ( instead of 10) and contained 6 l of 2X Rapid Ligati on Buffer (instead of 5) and 4 l of fresh PCR amplicons (instead of 3). Reaction mix was ligated into pCRII TOPO cloning vector and transformed into chemically competent XL10-Gold Ultracompetent Cells (Stratagene CA ) according to th e manufactures protocol. Inserts within white colonies were evaluated by PCR amplification using the same primer set and PCR protocol described earlier, and their size was confirmed by agarose gel electrophoresis. Sequencing and Phylogenetic Analysis Four sampling dates (4/22, 7/22, 8/19 and 9/02/08) were chosen for phylogenetic analysis based on measured nitrogenase activity (NA see Chapter 2 for details) A total of 12 clones (three per day) were sequenced at the DNA Sequencing Core Laboratory at the Uni versity of Florida using internal vector primers (M13). Each DNA sequence of nifD was compared to recorded sequences (GeneBank) from previous studies using BLAST queries (http://www.ncbi.nlm.nih.gov) and revealed that only 11 clone libraries were acceptable. Next,
72 the sequences were aligned with related sequences and one outgroup using ClustalX2 (Larkin et al., 2007). For community analyses, operational taxonomic units (OTUs) were generated using DOTUR under furthest neighbor algorithm and a threshold of 10% difference in nucleic acid sequences. The frequency of each OTU was used to construct a rarefraction analysis by comparing obtained versus cumulative expected phylotypes and was used to evaluate its diversity a nd richness. Thus, nonparametric estimates of richness and diversity were evaluated using DOTUR (Schloss and Handelsman, 2005) and included OTU, Chao1, Shannon index, and Simpson index (calculated using default parameters of the program DOTUR). Phylogene tic trees were conducted using MEGA version 4 (Tamura et al., 2007) w ith a neighbor joining analysis using a Tamura 3 parameter method for distance estimation of bootstrap analysis. Our selected outgroup was Rhodospirillum centenum, which is a non-sulfur p urple photosynthetic bacterium that prefers to grow in anoxic zones. Libshuff (Schloss et al., 2004) was employed to evaluate whether large differences observed between diazotrophic communities in chosen dates represent statistically different populations. The program -Libshuff evaluates community relationship (the exist ence of individuals in each community), and is considered to be less sensitive to library size than similar tests (Schloss et al., 2004). Well aligned nifD sequences were used to construct a distance matrix, using Jukes -Cantor corrected pairwise distance, by PHYLIP suite (Felsenstein, 2004). Next, the analysis was done using Libshuff (Schloss et al., 2004) with the Monte Carlo method and 10,000 permutations to calculate the integral form of the Cramrvon Mise statistic by constructing random sub-set populations from the entire data set and comparing the coverage of the gener ated populations to coverage in the experimentally obtained data set. Populations were considered significantly
73 different with P value below 0.01 after a Bonferroni correction for multiple pairwise comparisons ( =0.05, n = 132). E valuat ion of environmenta l parameters that may affect community composition was performed using the Mantel test (Mantel, 1967; Mantel and Valand, 1970). This test estimates correlations between observed differences in nifD diversity between transects and measured environmental par ameters of chosen dates to acquire an understanding of factors that may control community composition in Lake George. Mantel test is executed in R (R Development Core Team, 2008) using the package Vegan (Oksanen et al., 2010) and is based on a nonparametri c general regression model which employs squared Euclidean distance matrices between variables to test significance of and degree of predictability one variable has on another (Dutilleul et al., 2000). Eleven nifD clone libraries with 198 clones were chos en to represent the diazotrophic community composition during the summer of 2008 ( Figures 3 2 to 3 12; Table 3 1). On the third week of study (4/22), 22 clones were selected from the eastern and central transects and 21 from the western transect and were s ent to sequence ( Figure 3 14). Unifrac PCA analysis of clone libraries ( Lozupone and Knight, 2005) was used to evaluate diazotrophic community composition at different dates and sites in order to establish their relationships ( Figure 3 15). The clone libra ry that was constructed from the central sites on 9/02 contained only several inserts that were doubtable sequences (possibly chimeras) and were discarded. Results The number of groups of operational taxonomic units (OTU) varied between sites but was not s ignificantly different and showed no trends. Based on rarefaction analysis that used the numbers of clones per OTU, a nearly complete coverage of diazotroph diversity in all clone libraries was reached when defining 90% similarity of sequences as comprisin g an OTU ( Figure
74 3 13). All inde xes including the number of OTU s varied between sites and dates ( Table 3 1). Three indexes were used to rank community diversity. Shannon diversity index was chosen to estimate the diversity and evenness of the species (incr ease d by unique species, or greater species evenness) Simpson evenness estimation predict the probability that two randomly selected clones belong to the same species (accentuates abundant diazotrophs thus less affected by sample size) The Chao1 richness estimator is particularly useful for data sets that are skewed toward low abundance classes (Chao, 1984) (Table 1). Phylogenetic analysis of individual sequences cloned from PCR products identified distinct lineages of bacteria that possessed a nifD The majority of clones were related to cyanobacterial species (mostly filamentous and few u nicellular ) while the most diverse were heterocystous clustered within the family Nostocaceae (subsection IV). Only several free living proteobacteria that might be asso ciated with phytoplankton or with invertebrates and sediments were identified. Our p hylogenetic analysis of nifD was consistent with nifD literature that was relatively similar to cyanobacterial nifH and 16S rDNA gene sequences ( Henson et al., 2003; Givoan noni et al ., 1988; Roeselers et al., 2007). During the entire study period, nifD diversity was the highest in the east transect on the beginning (4/22) and end (9/02) of study. Although both dates were relatively highly diverse (8 and 7 OTUs respectively) than in the rest of the lake, they were quite different: on April the entire clone library clustered within heterocystous cyanobacteria, while in September it clustered also within Rhodospirillum (bootstrap value of 99) and the gram positive order Actino mycetales As mentioned, most clones clustered within cyanobacteria in two major clades: the bigger cluster was highly similar to Anabaena cylindrical PCC 7122 (93% similarity to known sequences in the
75 database ), while the other to Calothrix sp. 7101 (84% similarity to known sequences in the database ). In some samples like the one taken from east transect on 8/19, one clone was believed to be unicellular cyanobacteria from the family Pleurocapsales (subsection II) and more specifically the order Chroococci diopsis (bootstrap value 99); on the other hand, on 7/29, five highly related (bootstrap 100) clones were sequenced from the central transect, and branched inside Chroococcidiopsis (bootstrap 98) Chroococcidiopsis PCC 7293, closely related here, is a nonh eterocystous strain capable of forming specialized survival cells (Fewer et al 2002) that are similar to the akinetes found only in the heterocystous cyanobacteria (Rippka et al., 1979). These clones might be Microcystis (belonging to Chroococcidiopsi s ), even though t heir sequences showed the highest similarity (84%) to the genus Leptolyngbya sp. PCC 7004 which is in the order Oscillatoriales (subsection III). This order does not form heterocysts but rather uses different mechanism to fix N2 under aero bic condition. In fact, in Lake Griffin Florida, the filamentous Oscillatoriales outcompeted other cyanobacteria and dominated the diaz otrophic community between 20012002 ( Frost 2005 ). On 9/02, t hree clones from the west transect clustered close to the g enus Trichodesmium (also from the order Oscillatoriales ) but with low bootstrap value (29). These sequences showed high similarity to Calothrix sp. 7101 and Anabaena cylindrical PCC 7122 (84% and 83% similarity to known sequences in the database ). In addit ion, one clone clustered outside the c yanobacteria (bootstrap 87), probably within the Proteobacteria or within the order of the gram positive Actinomycetales (due to its clustering in proximity to Frankia ). Another interesting clone was found on the west transect on 7/29; which despite having only 5 OTUs, they belonged to different groups : in addition to the three clusters of cyanobacteria (one appeared as a
76 unicellular) and the Actinomycetales, one clone was highly related to our outgroup root Rhodospiril lum Surprisingly at this day, the center transect had four OTUs (heterocystous and unicellular cyanobacteria), but the east transect had only three OTUs which were composed solely of heterocystous cyanobacteria; it was the lowest number calculated (assum ing the diversity in the east transect on the 7/29, which was also equal to three, was higher than captured). According to the Mantel Test, the following environmental parameters were positively correlated to community composition particulate organic N (PO N), total N (TN), total Kjeldahl N (TKN) and chlorophyll a (chl -a ) (Table 3 3 ). Results from Unifrac PCA analysis of nifD clone libraries suggests that community composition is more affected by seasonality than location (Figure 3 15). Comparisons of nifD D NA gene libraries done by -Libshuff suggest that diazotrophic communities in Lake George were not significantly different from each other, except in two cases: the first in the west transect on 4/22 and 8/19 and the second between 8/19 west and 9/02 west (Table 3 2 ). Discussion DNA-based molecular characterization of nifD showed that the majority of sequences were distributed amongst cyanobacterial clades. More specifically, most clones showed similarity to filamentous heterocystous cyanobacteria and clust ered in two main groups that appeared in all samples. A distinct group of cyanobacteria clustered within unicellular but were found only in several samples According to nucleotide sequence, the potential diazotrophic cyanobacterial genera that were the mo st related were Anabaena PCC 7120, Calothrix sp., Chroococcidiopsis sp. and Leptoly ngb y a sp. This domination of cyanobacteria was expected in eutrophic systems like Lake George, due to nutrient levels and subtropical climate (Vincent 1987). In addition, t he
77 lakes shallow water, which permitted light penetration through most of its water column, and the small amount of macrophytes, favored cyanobacterial growth. Several clusters remained unidentified but the fact they clustered within the branch of cyanoba cteria suggest they may be novel groups of diazotrophic cyanobacteria. A possible source for this variation can be due to gene duplication and a secondary nitrogenase system that some cyanobacterial generas carry (Young, 1992). These additional systems evo lved possibly due to a lateral transfer or gene duplication I t was not clear how these systems are distributed, but such a system was found in vegetative cells of Anabaena variabilis strain ATCC 29413 and showed similarity to nif of Anabaena PCC 7120 or t o non -heterocyctous cyanobacteria (Thiel, 1993; Thiel et al., 1995). Fo r this reason we can not conclusiv e ly identity those cyanobacterial clones, especially the ones show ing low similarity to known sequences The other diazotrophs were not observed using microscopy, but most of their clones clustered close to Actinobacteria and Alphaproteobacteria and showed high similarity to their specie level (Rhodosperillium sp. and Frankia sp.). Rhodospirillum forms colonies that migrate toward or away from light, de pending on the wavelength, by using surface induced lateral flagella, chemotaxis, and a photosynthetic apparatus (Jiang et al., 1997). Both Rhodosperillium and Frankia may be inhabitants of the lake that ass ociate with other phytoplankton. A mong them are h eterotrophs that might be important to nutrient cycling and readily available organic C sources. On the other hand, if some entered the system recently they may be used to trace water sources or give information about their origin type (ex, leaching) or m icro niches (ex, anoxic zone for anaerobic diazotrophs). Because of their lower energetic levels compared to primary producers, and their sensitivity to UV radiation and DO in the water, heterotrophic organisms probably do not
78 account for a major portion o f the fixed N2 in the lake Thus, an ideal time for their fixation is during night (dark conditions) after photosynthesis, the main source of DO in the water, has cease d as respiration and bacterial decomposition have lowered remaining DO levels. Due to La ke George shallow water and its possible mixing by wind, some heterotrophs may be connected to sediment or benthic layer that can provide them with UV protection and possibly a DO gradient. Regardless of their quantitatively small contribution to Lake Geor ge N budget, they should be further studied, because they may be more sensitive to environmental changes than cyanobacteria; thus, understanding of these organisms might reveal their role as indicators of the lake condition or its phytoplankton dynamics. T he fact that they were harder to cultivate and identify using microscopy, favored a molecular approach for their study. After the examination of the diazotrphoic community composition at all clone libraries, the results confirmed the presence of an active phototrophic and heterotrophic diazotrophic community. This suggested a heterogeneous community that evolved to fit Lake George and some of its different niches. It is likely that heterotrophic metabolic processes done by diazotrophs were strongly tied to phototropic activities that supplied them with photosynthate. Nevertheless, classification of some nifD clones into a strong phylogeny were not supported by high bootstrap values, mostly because they showed low similarity to sequences associated with prev iously characterized organisms. This limited the confidence of our characterization, but since in most cases the clones formed strong clusters, their existence could not be ignored. Similar to other ecological studies, the amplified n ifD sequences supplie d high taxonomic resolution in our study of Lakes George diazotrophic community. Although nifD was more suited to characterization of cyanobacteria, GeneBank was lacking Cylindrospermopsis raciborskii sequences T his fact hindered our ability to different iate between the two members of
79 the family Nostocaceae (Cylindrospermopsis and Anabaena) that were the most abundant according to microscopic analysis (Chapter 2). Clones that clustered under cyanobacteria showed varying degree of similarity to known organ isms.The most diverse clade was related to Anabaena PCC 7120 and, in few samples, was divided into several related sub -clades, which was believed to include both Anabaena and Cylindrospermopsis The second largest clade branched outside Nostocaceae and sho wed similarity with Calothrix The third cyanobacterial clade was observed only twice and clustered within the unicellular group that was related to Chroococcidiopsis but might actually be M icrocystous (according to microscopy). Several studies of unicell ular non-heterocystous cyanobacteria demonstrated that some can fix N2 under fully oxic conditions and while oxygenic photosynthesis was taking place by employing different strategies to protect their nitrogenase enzyme during daytime. For example, some di azotrophs use temporal separation, as their main strategy, to fix N2 during the day (Stal, 1995; Fay, 1992). The results demonstrated the genetic dynamics and potential of diazotrophic diversity was not very high as supported by Libshuff analysis that demonstrated that most communities were not significantly different. Currently, we could not relate actual N2 fixation rates to each organism or clone. Such measurements can be beneficial and might explain the shift in community an d role of its members. This analysis can be performed by Real Time Polymerase Chain Reaction (RT PCR) using nif D primers, preferably combined with nif H ; by comparing both results under different conditions, our analysis would show the organisms that have t he potential to fix N2 (i.e., posses nitrogenase genes), while RT PCR would identify the ones that were active (i.e., expressing nitrogenase)
80 Table 3 1 Values of nifD diversity and richness in Lake George water, as estimated by Shannon diversity index, Simpson index, and Chao1 richness calculated using DOTUR (Schloss and Handelsman, 2005). Date Site No. of clones sequenced No. of OUT's Shannon index Diversity Richness 4/22/2008 East 22 8 1.3 (0.9, 1.7) 0.4 7.5 (6, 21) 4/22/2008 West 21 5 1.3 (1, 1.6) 0.3 6 (5, 19) 4/22/2008 Center 22 4 1.1 (0.7, 1.5) 0.5 5 (4, 17) 7/29/2008 East 8 3 0.7 (0.1, 1.4) 0.5 4 (3, 16) 7/29/2008 West 17 5 1.2 (0.6,2)* 0.4 6.5 (5, 20) 7/29/2008 Center 15 4 1.1 (0.7, 1.5) 0.4 5 (4,17) 8/19/2008 East 21 3 0.9 (0.6, 1.2) 0.5 3 (3, 3) 8/19/2008 West 17 6 1.6 (1.3, 1.9) 0.2 7 (6, 20) 8/19/2008 Center 19 6 1.3 (0.8, 1.8) 0.3 8 (6, 21) 9/2/2008 West 19 5 1.4 (1.1, 1.7) 0. 8 5 (5,NA) 9/2/2008 East 17 7 1.7 (1.3, 2.1) 0. 8 8 (7, 18)
81 Table 3 2 Population simila rity P values for comparison of nifD clone libraries determined using Cramer -von Mises test statistic, implemented in Libshuff (Schloss et al., 2004). Values in bold indicate significant P values ( P < 0.004) after Bonfe rroni correction for multiple pairwise comparisons. Libraries are distinct from one another if both comparisons ( X versus Y and Y versus X ) are significant. Comparisons were made using -Libshuff (Schloss et al., 2004) with 10,000 randomizations. The margi n of error for the P values 95% confidence interval for the P values near 0.05 was 0.004. Homologous Library (X) P Values comparison of heterologous library 4/22 East 4/22 Center 4/22 West 7/29 East 7/29 Center 7/29 West 8/19 East 8/19 Center 8/19 West 9/02 East 9/02 West 4/22 East 0.5363 0.3228 0.0894 0.0155 0.4943 0.0443 0.8441 0.0000 0.0051 0.0150 4/22 Center 0.2536 0.8145 0.1207 0.0224 0.1458 0.0328 0.8615 0.0000 0.0031 0.0084 4/22 West 0.1278 0.0045 0.1229 0.0065 0.0 380 0.3659 0.9647 0.0016 0.0002 0.0055 7/29 East 0.1876 0.4110 0.1103 0.2699 0.2878 0.7621 0.8534 0.8851 0.8682 0.7757 7/29 Center 0.0092 0.0216 0.0723 0.3475 0.0600 0.9313 0.5032 0.0057 0.0018 0.0165 7/29 West 0.3743 0.0251 0.1589 0.1749 0.1931 0.14 04 0.2473 0.0059 0.0023 0.0280 8/19 East 0.0040 0.0000 0.0628 0.2010 0.0257 0.0059 0.8429 0.0048 0.0001 0.0000 8/19 Center 0.0006 0.0000 0.0021 0.0428 0.0006 0.0000 0.1901 0.0057 0.0000 0.0000 8/19 West 0.0000 0.0000 0.0000 0.0106 0.0000 0.0000 0.0084 0.2076 0.0957 0.0001 9/02 East 0.0000 0.0000 0.0002 0.0040 0.0000 0.0000 0.0039 0.1526 0.0077 0.0022 9/02 West 0.0000 0.0000 0.0001 0.0055 0.0005 0.0000 0.0005 0.0000 0.0000 0.0002
82 Table 3 3 Correlation between diazotrophic composition and environmental parameters (u sing M a nt e l Test). Parameter Mantel statistics (R) Significance TKN T 0.35 23 0.018* Chl a 0.3993 0.022* TN 0.3223 0.023* PON 0.3113 0.029* Total Kjeldahl nitrogen (TKN T ); chlorophyll a (chl -a ); total nitrogen (TN); particulate organic nitrogen (PON) indicates a statistically significant difference ( P Figure 3 1 Selected dates where samples were obtained for molecular analysis (marked by arrows) based on acetylene reduction assay rate. Three samples were used at each date except to last date (used only the east and west transects).
83 0 1 2 3 4 5 0 5 10 15 20 25 0 1 2 3 4 0 5 10 15 20 25 0 1 2 3 4 0 5 10 15 20 25 0 1 2 3 4 5 0 5 10 15 20 25 0 1 2 3 0 5 10 15 20 25 0 1 2 3 4 0 5 10 15 20 25 0 2 4 6 0 5 10 15 20 25 0 1 2 3 0 5 10 15 20 25 0 2 4 6 0 5 10 15 20 25 0 2 4 6 8 0 5 10 15 2025 a e f g h j i b c d East Center WestOperational Taxonomic UnitsNo. of clones analyzed 0 1 2 3 4 5 0 5 10 15 20 25 k 0 1 2 3 4 5 0 5 10 15 20 25 0 1 2 3 4 0 5 10 15 20 25 0 1 2 3 4 0 5 10 15 20 25 0 1 2 3 4 5 0 5 10 15 20 25 0 1 2 3 0 5 10 15 20 25 0 1 2 3 4 0 5 10 15 20 25 0 2 4 6 0 5 10 15 20 25 0 1 2 3 0 5 10 15 20 25 0 2 4 6 0 5 10 15 20 25 0 2 4 6 8 0 5 10 15 2025 a e f g h j i b c d East Center WestOperational Taxonomic UnitsNo. of clones analyzed 0 1 2 3 4 5 0 5 10 15 20 25 k Figure 3 2 Rarefraction analysis for nifD collected from Lake George water A ) to C) sampled on 4/22. D to F) sampled on 7/29. G to I) sampled on 8/12. J to k) sampled on 9/22/08. Based on 10% difference.
84 Figure 3 3 Phylogenetic tree of genomic DNA ni fD from 4/22/08 east transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling.
85 Chroococcidiopsis Trichodesmium sp LG12 LG5 Anabaena cylindrica PCC 7122 D122 Anabaena sp PCC 7108 D101 D107 Cylindrospermum stagnale Nodularia spumigena D104 D121 Nostoc sp Calothrix sp PCC 7101 Cylindrospermum majus Chlorogloeopsis fritschii Chlorogloeopsis fritschii Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Nostoc sp PCC 7423 Anabaena 7120 Nostoc PCC 7120 Mastigocladus laminosus Fischerella sp UTEX 1903 Fischerella muscicola Calothrix desertica Fischerella UTEX1931 Fischerella sp. Scytonema hofmanni PCC 7110 Scytonema Leptolyngbya sp PCC 7104 Cyanothece sp PCC 7424 Pseudanabaena sp PCC 6802 Oscillatoria sancta Xenococcus sp PCC 7305 Synechococcus Pseudanabaena sp PCC Leptolyngbya sp PCC 7375 Lyngbya aestuarii PCC 7419 Gloeothece Pleurocapsa Frankia sp Bradyrhizobium sp Bradyrhizobium sp Bradyrhizobium spp Methylosinus Rhizobiales bacterium Rhizobium Methylobacterium Mesorhizobium Pseudomonas stutzeri A.faecalis Pseudomonas azotifigens Bradyrhizobium sp Herbaspirillum sp. H.seropedicae Rhodospirillum centenum 99 99 98 98 96 96 94 93 90 88 63 58 84 62 50 60 60 45 35 28 27 23 13 12 11 2 6 12 5 8 28 78 51 47 66 52 30 17 17 44 52 36 0.1 Cluster 1 Cluster 3 Cluster 2
86 Figure 3 4. Phylogenetic tree of genomic DNA nifD from 4/22/08 center transect. Numbers at branch points refer to bootstrap analysis bas ed on 1000 resampling.
87 LG11 LG2 D320 Anabaena cylindrica PCC 7122 D319 Anabaena sp PCC 7108 Nostoc sp 1189P Cyanothece sp CCY 0110 Chroococcidiopsis Trichodesmium sp IMS101 Cylindrospermum stagnale Nodularia spumigena PCC Nodularia spumigena D303 Calothrix sp PCC 7101 LG2(2) LG3 Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Chlorogloeopsis fritschii PCC 6718 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Fischerella sp UTEX 1903 Mastigocladus laminosus Fischerella UTEX1931 Scytonema hofmanni PCC 7110 Fischerella sp. Fischerella muscicola Scytonema Calothrix desertica Pseudanabaena sp PCC 6802 Oscillatoria sancta Leptolyngbya sp PCC 7375 Synechococcus Pseudanabaena sp PCC Lyngbya aestuarii PCC 7419 Pleurocapsa Cyanothece sp PCC 7424 Gloeothece Leptolyngbya sp PCC 7104 Xenococcus sp PCC 7305 Frankia sp. Bradyrhizobium sp U11 Bradyrhizobium spp Bradyrhizobium sp 1808 Herbaspirillum sp. H.seropedicae Methylobacterium Rhizobium Rhizobiales bacterium Mesorhizobium Pseudomonas azotifigens Pseudomonas stutzeri A.faecalis Methylosinus Bradyrhizobium sp Rhodospirillum centenum 65 62 99 99 99 99 76 99 17 99 72 98 98 98 93 97 91 88 69 85 62 72 58 54 38 36 30 26 31 27 23 17 18 16 46 93 60 43 84 80 58 56 43 37 19 27 32 0.1 Cluster 1 Cluster 2
88 Figure 3 5 Phylogenetic tree of genomic DNA nifD from 4/22/08 west transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling
89 LG7 F20-A01 F20-A12 Calothrix F20-A07 Nodularia Cylindrospermum Nodularia Trichodesmium F20-A11 LG2 Nostoc Anabaena F20-B03 Anabaena LG7(2) Cyanothece Chroococcidiopsis Nostoc Anabaena Nostoc Nostoc Nostoc Anabaena Chlorogloeopsis Cylindrospermum1 Chlorogloeopsis Fischerella6 Mastigocladus Mastigocladus2 Mastigocladus1 Fischerella Scytonema Scytonema1 Fischerella1 Calothrix Fischerella3 Synechococcus Cyanothece Leptolyngbya Gloeothece Xenococcus Pseudanabaena Pseudanabaena2 Oscillatoria Pleurocapsa Lyngbya Leptolyngbya1 Frankia H.seropedicae Herbaspirillum Bradyrhizobium Pseudomonas A.faecalis Pseudomonas Methylosinus Bradyrhizobium Methylobacterium Rhizobium Rhizobium Mesorhizobium Rhizobiales Rhizobiales Rhodospirillum 100 100 100 77 99 99 99 99 99 18 99 98 98 91 53 98 97 92 96 67 78 95 92 86 72 72 44 75 57 48 27 26 25 23 23 20 19 3 1 3 14 37 82 66 79 69 59 21 29 32 0.1 Cluster 1 Cluster 2 Cluster 3
90 Figure 3 6 Phylogenetic tree of genomi c DNA nifD from 7/29/08 east transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling.
91 LG7 5 LG7 7 LG7 2 LG7 1 LG7 4 LG7 3 LG7 8 Calothrix sp PCC 7101 Anabaena cylindrica PCC 7122 Nostoc sp 1189P Anabaena sp PCC 7108 LG7 6 Nodularia spumigena Cylindrospermum stagnale Nodularia spumigena PCC Chroococcidiopsis Trichodesmium sp IMS101 Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Chlorogloeopsis fritschii PCC 6718 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Mastigocladus laminosus Fischerella sp UTEX 1903 Calothrix desertica Fischerella sp. Fischerella UTEX1931 Fischerella muscicola Scytonema Scytonema hofmanni PCC 7110 Leptolyngbya sp PCC 7104 Pseudanabaena sp PCC Oscillatoria sancta Pseudanabaena sp PCC 6802 Lyngbya aestuarii PCC 7419 Leptolyngbya sp PCC 7375 Synechococcus Cyanothece sp PCC 7424 Xenococcus sp PCC 7305 Gloeothece Pleurocapsa Frankia sp. Bradyrhizobium spp Bradyrhizobium sp 1808 Bradyrhizobium sp U11 Bradyrhizobium sp Methylosinus Pseudomonas stutzeri A.faecalis Pseudomonas azotifigens Methylobacterium Rhizobium Rhizobiales bacterium Mesorhizobium Herbaspirillum sp. H.seropedicae Rhodospirillum centenum 67 51 64 100 99 59 100 9 2 99 92 99 99 99 99 45 99 97 96 79 49 36 79 92 85 67 47 33 32 19 17 27 22 41 84 80 60 68 44 35 27 16 33 0.1 Cluster 1 Cluster 2 Cluster 3
92 Figure 3 7. Phylogenetic tree of genomic DNA nifD from 7/29/08 center transect. Numbers at branch points refer to bootstrap analy sis based on 1000 resampling.
93 LG5 LG3 9 LG3 2 LG3 8 Calothrix sp PCC 7101 LG3 15 Nostoc sp 1189P Anabaena sp PCC 7108 Anabaena cylindrica PCC 7122 Nodularia spumigena Cylindrospermum stagnale Nodularia spumigena PCC Trichodesmium sp IMS101 Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Chlorogloeopsis fritschii PCC 6718 Mastigocladus laminosus Fischerella sp UTEX 1903 Scytonema Fischerella muscicola Scytonema hofmanni PCC 7110 Fischerella UTEX1931 Fischerella sp. Calothrix desertica Chroococcidiopsis LG5 Leptolyngbya sp PCC 7375 Oscillatoria sancta Leptolyngbya sp PCC 7104 Xenococcus sp PCC 7305 Synechococcus Cyanothece sp PCC 7424 Gloeothece Pseudanabaena sp PCC Lyngbya aestuarii PCC 7419 Pseudanabaena sp PCC 6802 Pleurocapsa Frankia sp. Rhizobiales bacterium Mesorhizobium Rhizobium Methylobacterium Bradyrhizobium sp 1808 Bradyrhizobium sp U11 Bradyrhizobium spp Methylosinus Bradyrhizobium sp Herbaspirillum sp. H.seropedicae Pseudomonas stutzeri A.faecalis Pseudomonas azotifigens Rhodospirillum centenum 100 99 17 97 52 99 99 43 55 99 9 3 0 99 98 98 98 47 98 46 97 96 95 55 36 38 73 91 84 40 42 31 11 21 28 28 34 82 65 67 51 25 16 18 26 26 0.1 Cluster 1 Cluster 3 Cluster 2
94 Figure 3 8 Phylogenetic tree of genomic DNA nifD from 7/29/08 west transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling
95 LG11 LG5 17 LG5 16 Calothrix sp PCC 7101 LG5 13 Anabaena cylindrica PCC 7122 Anabaena sp PCC 7108 Nostoc sp 1189P Nodularia spumigena Nodularia spumigena PCC Cylindrospermum stagnale Gloeothece Leptolyngbya sp PCC 7104 Leptolyngbya sp PCC 7375 Xenococcus sp PCC 7305 Pseudanabaena sp PCC 6802 Cyanothece sp PCC 7424 Pseudanabaena sp PCC Synechococcus Oscillatoria sancta Pleurocapsa Lyngbya aestuarii PCC 7419 Trichodesmium sp IMS101 Chroococcidiopsis LG5 8 Calothrix desertica Fischerella sp. Fischerella muscicola Fischerella UTEX1931 Scytonema hofmanni PCC 7110 Scytonema Fischerella sp UTEX 1903 Mastigocladus laminosus Chlorogloeopsis fritschii PCC 6912 Cylindrospermum majus PCC Nostoc sp PCC 7423 Chlorogloeopsis fritschii PCC 6718 Anabaena variabilis ATCC 29413 Nostoc PCC 7120 Anabaena 7120 Nostoc PCC 6720 Rhizobium LG5 9 LG5 1 Frankia sp. Mesorhizobium Rhizobiales bacterium Methylobacterium Bradyrhizobium sp 1808 Bradyrhizobium spp Bradyrhizobium sp U11 Bradyrhizobium sp Methylosinus H.seropedicae Herbaspirillum sp. Pseudomonas stutzeri Pseudomonas azotifigens A.faecalis LG5 10 Rhodospirillum centenum 100 21 10 8 16 78 43 100 99 10 6 0 2 99 45 97 54 97 92 82 59 80 68 78 57 59 72 72 68 51 20 18 10 6 9 20 58 57 44 44 39 36 35 32 35 25 19 25 26 98 0.1 Cluster 1 Alphaproteobacteria Actinomycetales Cluster 2
96 Figure 3 9 Phylogenetic tree of genomic DNA nifD f rom 8/19/08 east transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling.
97 LG14 F20-D11 F20-D05 LG2 LG3 Calothrix sp PCC 7101 Nodularia spumigena Cylindrospermum stagnale Nodularia spumigena PCC Chroococcidiopsis Trichodesmium sp IMS101 Cyanothece sp CCY 0110 Anabaena sp PCC 7108 Anabaena cylindrica PCC 7122 Nostoc sp 1189P Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Chlorogloeopsis fritschii PCC 6718 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Mastigocladus laminosus Fischerella sp UTEX 1903 Fischerella UTEX1931 Scytonema Fischerella sp. Scytonema hofmanni PCC 7110 Fischerella muscicola Calothrix desertica Xenococcus sp PCC 7305 Lyngbya aestuarii PCC 7419 Pleurocapsa Synechococcus Cyanothece sp PCC 7424 Pseudanabaena sp PCC 6802 Leptolyngbya sp PCC 7104 Gloeothece Leptolyngbya sp PCC 7375 Oscillatoria sancta Pseudanabaena sp PCC Frankia sp. Bradyrhizobium sp U11 Bradyrhizobium spp Bradyrhizobium sp 1808 Herbaspirillum sp. H.seropedicae Bradyrhizobium sp Methylosinus Pseudomonas azotifigens Pseudomonas stutzeri A.faecalis Methylobacterium Rhizobium Rhizobiales bacterium Mesorhizobium Rhodospirillum centenum 76 99 99 99 18 99 99 99 99 99 30 27 99 47 97 92 93 91 89 67 69 38 71 45 39 37 33 27 17 9 15 14 10 29 88 66 42 33 64 59 54 45 27 36 0.1 Cluster 1 Cluster 2
98 Figure 3 10. Phylogenetic tree of genomic DNA nifD from 8/19/08 center transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling.
99 LG7 LG4 F20-H01 LG2 Calothrix sp PCC 7101 Nodularia spumigena Cylindrospermum stagnale Nodularia spumigena PCC Trichodesmium sp IMS101 Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Chlorogloeopsis fritschii PCC 6718 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Mastigocladus laminosus Fischerella sp UTEX 1903 Scytonema Fischerella UTEX1931 Scytonema hofmanni PCC 7110 Fischerella sp. Fischerella muscicola Calothrix desertica Anabaena sp PCC 7108 Anabaena cylindrica PCC 7122 Nostoc sp 1189P Cyanothece sp CCY 0110 Chroococcidiopsis LG2(2) Pseudanabaena sp PCC 6802 Xenococcus sp PCC 7305 Pleurocapsa Leptolyngbya sp PCC 7104 Leptolyngbya sp PCC 7375 Lyngbya aestuarii PCC 7419 Oscillatoria sancta Gloeothece Pseudanabaena sp PCC Synechococcus Cyanothece sp PCC 7424 F20-H12 Frankia sp. LG2(3) Bradyrhizobium sp U11 Bradyrhizobium spp Bradyrhizobium sp 1808 Herbaspirillum sp. H.seropedicae Methylosinus Bradyrhizobium sp Pseudomonas azotifigens Pseudomonas stutzeri A.faecalis Rhizobium Rhizobiales bacterium Mesorhizobium Methylobacterium Rhodospirillum centenum 99 99 99 76 99 63 32 99 82 99 99 20 98 97 97 96 46 95 94 94 91 79 61 27 42 39 35 30 24 22 15 14 8 17 36 95 85 82 43 40 24 23 17 19 19 70 59 50 85 0.1 Cluster 1 Cluster 2 Actinomycetales
100 Figure 3 11. Phylogenetic tree of genomic DNA nifD from 8/19/08 west transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling.
101 LG6 F20-F09 LG2 Anabaena cylindrica PCC 7122 Anabaena sp PCC 7108 Nostoc sp 1189P Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Chlorogloeopsis fritschii PCC 6718 Fischerella sp UTEX 1903 Mastigocladus laminosus Scytonema Scytonema hofmanni PCC 7110 Fischerella UTEX1931 Fischerella sp. Fischerella muscicola Calothrix desertica Trichodesmium sp IMS101 Cylindrospermum stagnale Nodularia spumigena PCC Nodularia spumigena Calothrix sp PCC 7101 LG2(2) F20-E04 F20-E12 LG2(3) Cyanothece sp CCY 0110 F20-F11 Chroococcidiopsis Leptolyngbya sp PCC 7104 Gloeothece Pseudanabaena sp PCC 6802 Cyanothece sp PCC 7424 Synechococcus Oscillatoria sancta Xenococcus sp PCC 7305 Lyngbya aestuarii PCC 7419 Pseudanabaena sp PCC Leptolyngbya sp PCC 7375 Pleurocapsa F20-E03 Frankia sp. Bradyrhizobium sp U11 Bradyrhizobium spp Bradyrhizobium sp 1808 Herbaspirillum sp. H.seropedicae Bradyrhizobium sp Methylosinus Pseudomonas azotifigens Pseudomonas stutzeri A.faecalis Rhizobium Rhizobiales bacterium Mesorhizobium Methylobacterium Rhodospirillum centenum 99 76 99 99 74 99 73 99 99 16 7 97 66 97 97 97 96 91 95 86 71 80 88 70 70 70 56 48 28 28 24 22 21 13 4 9 41 87 69 85 79 61 47 21 21 17 13 12 34 53 0.1 Cluster 1 Cluster 2
102 Figure 3 12. Phylogenetic tree of genomic DNA nifD from 9/02/08 east tr ansect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling.
103 Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 GC2 GC5 GC1 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Chlorogloeopsis fritschii PCC 6718 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Mastigocladus laminosus Fischerella sp UTEX 1903 Calothrix desertica Scytonema Fischerella UTEX1931 Scytonema hofmanni PCC 7110 Fischerella sp. Fischerella muscicola Chroococcidiopsis Trichodesmium sp IMS101 Cyanothece sp CCY 0110 Anabaena cylindrica PCC 7122 Anabaena sp PCC 7108 Cylindrospermum stagnale Nodularia spumigena PCC Nodularia spumigena Nostoc sp 1189P Calothrix sp PCC 7101 GC18 GC19 GC26 GC4 GC30 Lyngbya aestuarii PCC 7419 Leptolyngbya sp PCC 7375 Leptolyngbya sp PCC 7104 Gloeothece Pseudanabaena sp PCC 6802 Cyanothece sp PCC 7424 Synechococcus Oscillatoria sancta Xenococcus sp PCC 7305 Pseudanabaena sp PCC Pleurocapsa Frankia sp. GC22 LG6 Bradyrhizobium sp U11 Bradyrhizobium spp Bradyrhizobium sp 1808 Herbaspirillum sp. H.seropedicae Pseudomonas azotifigens Pseudomonas stutzeri A.faecalis Methylosinus Bradyrhizobium sp Rhizobium Rhizobiales bacterium Mesorhizobium Methylobacterium Rhodospirillum centenum GC29 GC36 99 99 82 99 99 99 99 48 99 98 98 62 98 98 94 97 52 96 42 58 74 47 43 35 26 12 12 10 3 8 20 42 97 97 81 57 47 23 54 44 84 99 72 0.2 Cluster 1 Actinomycetales Alphaproteobacteria Cluster 2
104 Figure 3 13. Phylogenetic tree of genomic DNA nifD from 9/02/08 west transect. Numbers at branch points refer to bootstrap analysis based on 1000 resampling.
105 LG6 Anabaena cylindrica PCC 7122 Anabaena sp PCC 7108 Nostoc sp 1189P Calothrix sp PCC 7101 Nodularia spumigena Cylindrospermum stagnale Nodularia spumigena PCC Trichodesmium sp IMS101 LG3 Cyanothece sp CCY 0110 Synechococcus Leptolyngbya sp PCC 7104 Leptolyngbya sp PCC 7375 Lyngbya aestuarii PCC 7419 Pleurocapsa Xenococcus sp PCC 7305 Gloeothece Pseudanabaena sp PCC Cyanothece sp PCC 7424 Pseudanabaena sp PCC 6802 Oscillatoria sancta LG2 6 LG2 2 Chroococcidiopsis LG2 13 Anabaena 7120 Nostoc PCC 7120 Nostoc sp PCC 7423 Nostoc PCC 6720 Anabaena variabilis ATCC 29413 Chlorogloeopsis fritschii PCC 6718 Cylindrospermum majus PCC Chlorogloeopsis fritschii PCC 6912 Fischerella sp UTEX 1903 Mastigocladus laminosus Scytonema hofmanni PCC 7110 Scytonema Fischerella muscicola Fischerella sp. Fischerella UTEX1931 Calothrix desertica Frankia sp. Bradyrhizobium sp 1808 Bradyrhizobium spp Bradyrhizobium sp U11 Herbaspirillum sp. H.seropedicae Mesorhizobium Rhizobiales bacterium Rhizobium Methylobacterium Bradyrhizobium sp Methylosinus Pseudomonas azotifigens A.faecalis Pseudomonas stutzeri Rhodospirillum centenum LG7 99 99 99 99 99 99 98 98 97 97 97 50 96 93 95 90 69 67 84 81 80 70 63 30 38 44 62 71 49 68 60 55 37 32 29 19 19 16 12 6 10 51 92 99 0.1 Cluster 2 Cluster 1 Cluster 3 Alphaproteobacter ia
106 7/29/08 9/02/08 8/19/08 P1 22.9 % P2 12.5 %4/22/08 7/29/08 9/02/08 8/19/08 P1 22.9 % P2 12.5 %4/22/08 Figure 3 14. Separation of samples by principle component analysis based on the diazotrophic community composition (analyzed genetic clades using Unifrac)
107 CHAPTER 4 SUMMARY AND CONCLUSION Eutrophication of aquatic systems is draw ing increasing amounts of interest, and one major reason of concern is their potential as pollution sources to other systems (especially downstream systems) that may be irreversibly impacted. Although eutrophic systems are rich in nutrients and experiencin g high productivity levels, primary production is usually limited by nutrients. Many previous studies have concluded that primary production in fresh water systems is mostly limited by phosphorus (P ) following nitrogen (N ) compared to marine systems and es tuaries that are primarily limited by N. This work focused on measuring biological N2 fixation in the watercolumn of a shallow, subtropical lake with the potential to export significant quantities of N to hydrologically connected systems downstream L ake G eorge is a shallow eutrophic lake system located on a geologically P rich region that has the potential to shift limitation of primary production to N. This nutrient shift causes increased demand for N that can be satisfied by biological N2 fixation and grants a competitive advantage to diazotrophic organisms. Downstream of Lake George, low le vels of dissolved oxygen (DO ) are measured annually with periods of hypoxi a that may collapse biological systems an d alter biogeochemical cycles. Despite regular moni toring of the surface import/export N balance of Lake George, it was not clear how significant biological N2 fixation is for the system N budget and what mechanisms may convert the lake into a pollution source. Therefore, the objectives of this research we re to: 1) characterize and q uantify N2 fixation spatially, seasonally and under different light levels, 2) i dentify environmental parameters related to N2 fixation and their patterns, 3) c haracterize diazotrophic community composition by examining the effe cts of location, date, nutrients and N2 fixation rate.
108 The acetylene reduction assay (ARA) was selected to measure nitrogenase activity in Lake George, identify its regulators, and estimate the quantity of N2 that was fixed including its contribution to the lakes N export. The importance of regulators of N2 fixation and N export in this system can be used for management purposes and also may be adapted to similar systems. By this method, the measured rate of N2 fixation was expressed in volume (550 g N m3 h1, or a mean of 20 nmol N2 l1 h1 and a range of 0 55), area (mean 1.4 mg N m2 h1), and heterocystous biovolume (ranged from 10 to 18 picomol N2 m3 heterocyst ou s h1). T hese potential rates agreed with previous measurement in S t. J ohns River Est uary and similar systems that showed that N2 fixation may be a significant component of the N budget (Pearl et al., 2003). The increasing rates and abundance of eutrophication seen globally and the ir importance to exported N which leads to eutrophication of other systems, have called for actions, including demands for their regulation and control. Eutrophication is a broad and complex process, but shallow lake s represent a relative simple system (because most experience little strafication and are easier t o sample and extrapolate ) that may be ideal for the study of eutrophication. With assumptions regarding light penetration and day length, it was also possible to use the rates recorded during this study in 2008 to roughly quantify the total amount of N tha t was fixed in Lake George water during our study period and examine its contribution to yearly N output Based on this calculation, it was estimated that during the study (late Spring through early Fall), N2 fixation rate in the dark amounted to an estimated 135 metric tones (MT, or 106 g) of N, while daytime N2 fixation accounted for a potential of 462 MT N to the lake F or the rest of the year (outside the study period) an additional 131 MT N was estimated ( based on an assumed constant rate of 28 n g l1h 1) Thus, the total N that was fixed during our study period is estimated to range
109 from 600730 MT during the entire year 2008. This quantity fit the lakes N budget and was a little larger than the estimated yearly amount of exported N (600 MT). It was evident that ARA had seasonal pattern that was consistent between sites and various light treatments. This pattern indicated the importance of seasonality as a regulator of N2 fixation and diazotrophic growth and also supported the mixing of water sources expected in shallow lakes. The pattern showed four peak events that were separated by low activity and coincided with cyanobacterial and heterocystous abundance and with measured patterns of watercolumn nutrient concentrations (N and P). Nitrogen fixation rate was positively influenced by light which reinforced the assumption that N2 fixation is photosynthetically derived through heterocystous cyanobacteria in Lake George; ARA showed insignificant difference between low (66 M photons m2 s1 PAR (photosynt hetically active radiation) and high light (115 M photons m2 s1 PAR ) incubation that supported the ability of these organisms to photosynthesis under low light levels S easonality and water source affected nutrient abundance yet N was the limiting nutri ent during most of the study period, as indicated by high rate of AR and by weight ratio of dissolved inorganic N: total P (DIN:TP). Several of the biggest difficulties were connecting between nutrient forms to cyanobacterial shifts and species domination and the ability to assess environmental impacts caused by algal blooms. Seasonal patterns of N and P, including ammoni um (NH4 +), nitrate/nitrite (NOx), particulate organic N (PON) and total N ( TN ):TP ratio, showed similarity to ARA pattern and reinforced t he assumption that nutrients lead to cyanobacterial dominance and thus, regulate N2 fixation rate s After dividing the study period into two sub-periods (based on ARA peaks and measured abundance of heterocystous cyanobacteria) s t atistical analyses reveal ed that several selected
110 nutrients (NH4 +, DO, dissolved organic carbon (DOC)) we re positively correlated to ARA, and st epwise regression analysis selected particulate organic C ( POC ): PON ratio (negatively correlated with AR A) as the strongest correlated pa rameter in a model that explained 98% of variance in AR A rate T he negative correlations implied that N level may be either inhibitory or may be the outcome of N2 fixation. Canonical correlation analysis was also used to identify environmental parameters that could explain spatial N2 fixation pattern within the lake The analysis showed the importance of the S t. J ohns River as well as land uses around the lake on the variability of lake water composition Results indicated that DOC, NH4 +, dissolved reactiv e P (DRP), dissolved organic N (DON), TN:TP, and conductivity were the most different between the east and west lake regions during the study period, while during the first sub-period the central transect was similar to the east, and shifted to be more lik e the west during the second sub-period S easonall y these shifts appeared to control nutrient patterns and support ed a N limited system in which P was an important regulator of N2 fixation. Patterns of algal blooms and nutrients composition were affected by seasonality and showed a spatial heterogeneity of nutrients and N2 fixation. Although, the effect of environmental conditions on the diazotrophic community clearly indicated a shift toward heterocystous cyanobacteria, it was not clear how N and P, or w hich mechanisms control diazotrophic composition and domination of single species. These blooms are common in shallow lakes and other freshwater systems and indicate N limitation, but their ability to act as sources of N to downstream ecosystems may be con trolled only after achieving greater understanding about their mechanisms.
111 The goal of the second part of the study was to characterize the diazotrophic community composition in Lake George using a non-culture based method. We chose to use the conserved su bunit of nitrogenase ( nifD sequences) to characterize the diazotrophic community composition in order to better understand their relationship between structure and function in Lake George This wa s the first ever attempt to characterize the phylogeny of di azotrophs in Lake George using nifD a preferred biomarker for characterizing cyanobacteria. After identifying all organisms that had the genetic potential to fix N2, phylogenetic analysis was used to infer their evolutionary relation and find distributional patterns. Several methods were chosen in our attempts to correlate our characterized diazotrophic community to ARA, location, date and environmental parameters. We found that seasonality was the only apparent factor controlling distribution of diazotrophs. And both N and P were controlling community composition. It is important to note that ARA measured the N2 fixation potential rate of the entire community without the ability to relate fixation rate to each group present and actively fixing N2. Never th e less, it demonstrated that the diazotrophic community at Lake George is active and diverse. Using this approach, the diazotrophic community in Lake George was found to be similar to communities of other subtropical, eutrophic systems, although novel stra ins were also recorded in this study. While most clones were unidentified or showed low percentage of similarity, the heterocystous family Nostocaceae was found in all samples and was clearly the most diverse group. This family includes two of Lake George s dominant diazotrophs from the genus Anabaena and Cylindrospermopsis that expected to dominate such systems as supported by microscopic analysis. Heterocystous cyanobacteria (especially Cylindrospermopsis ) invaded into many systems and became sources of c oncern (e.g. fix N2 and may release toxic) Other
112 diazotrophic members that were also identified included other cyanobacteria (both heterocystous and non -heterocystous), A lphaproteobacteria, and the gram positive Frankia (order Actinomycetales ). Knowledge about the diazotrophic composition gives essential information about the system that otherwise may be hard to ac hieve, including the possible origin, control s, and activities of these organisms ). Except in two cases that could not be explained, Libshuff analysis suggested that population membership in most clone libraries was related and did not differ significantly Principle component analysis of nifD sequences showed highest separation between clonelibraries based on date with a minor effect of site l ocation within the lake. Although more information is need about the mechanisms by which nutrients affected community shift s Mantel Test results suggested that N ( TN and PON) were the primary factors affecting diazotrophic community composition A s expect ed, this analysis demonstrated that as the concentration difference of each of these nutrients increased, the diversity of the diazotrophic community also increased, however, it was not clear how these nutrients affected community composition Although pre vious work concluded that Lake George is a N limited system controlled by P, AR rate and cyanobacterial biovolume were not highly correlated to P levels. Other limitations were connected to the chosen methods and assumptions that were taken and to the complexity of system. In future research, more sites should be used to potentially resolve the difficulties in identifying cause and effect of peaks Also, in order to encompass all diazotrophic activity at Lake George, other ecosystem components (e.g. sedimen t, macrophytes, etc) must be measured, and other N cycle processes (e.g. denitrification) should be measured in order to better constrain the N budget and increase its accuracy. Increased sampling during other seasons and better
113 estimates of diel patterns may give more information about the activities and diazotrophic community and increase accuracy of estimated yearly fixed N. For most samples the diazotrophic composition was not rich, compris ed mainly of cyanobacteria. In order to increase our confidence in the phylogenetic analysis, nifD taken from pure cultures of Anabaena and Cylindrospermopsis from Lake George should be sequenced and recorded in Genebank; in fact, using the combination of both nifD and nifH would offer higher resolution within cyanoba cteria, may identify additional diazotrophs and thus increase the confidence of phylogenetic analysis. Several molecular techniques that measure expression of nif can be employed to measure activity by each clone and so estimate the importance of each mem ber in the community. Phylogenetic analysis of the diazotrophic community at the lakes sediment can be used to estimate its environmental conditions (For example, presence of cyanobacteria indicates light penetration and relatively high DO levels during l ight periods or association with other organisms); in addition, it can shed light on phytoplankton migration and interaction with the sediments and their communities. Lake George, like many shallow lakes may act as a source of N to downstream systems and i ts destructive impact on the adjacent marine systems is a good example o f the importance of controlling algal blooms in inland waters Thus far, most research on Lake George focused on nutrient manipulation bioassays (identified nutrient limitation and the ir controls on activities), presence and controls of cyanobacteria toxins, and phytoplankton abundance including their predators. In the past, ARA was used in Lake George to measure N2 fixation under different nutrient additions and as a function of growth and salinity. This study is the first attempt to quantify N2 fixation in Lake George water and estimate its significance to the N budget. In addition, nifD was used in this system for the first time to characterize the diazotrophic
114 community composition i n the lake. Our results clearly showed that N and P affected N2 fixation in two levels: 1) directly affecting N2 fixation rate and 2) indirectly by affecting diazotrophic community composition. The molecular analysis work was important for the construction of genetic database of the diazotrophic community at Lake George and could be used as a preliminary data to support additional investigations. Our genetic database represented organisms that had the potential to fix N2 and may be used as a foundation in the study of the expression of nifD in Lake George B y combining information on potential gene expression with environmental parameters leading to expressed N2 fixation rates, it may be possible to identify the precise controls on N2 fixation during parti cular blooms F urthermore, it may also assist in the search for early warning indicators that can detect ecosystem changes in advance of bloom formation. Additional work will increase our ability to predict, explain, and control nutrients/environmental con ditions leading to algal blooms in Lake George and similar systems. Thus, this work can be viewed as a significant step toward the development of effective strategies to reduce N2 fixation and minimizing the environmental impacts of N export from lakes.
115 LIST OF REFERENCES Arnold, W., Rump, A., Klipp, W., Priefer, U. B., and Puhler, A. 1988. J. Mol. Biol. 203:715 738. Arrigo K. R. 2005. Marine Microorganisms and Global Nutrient Cycles. Nature: 437:349 355. Ashton P. and Hoare R. A. 1981. The sensitivity t o phosphorus and nitrogen loads, of Lake Rotorua, New Zealand. Water science and technology. 13:1349 1350. Bachmann, R. W., Hoyer, M. V., & Canfield, D. E., Jr. 2000. The potential for wave disturbance in shallow Florida lakes. Lake and Reservoir Managemen t, 16, 281 291. Baird D. and R. E. Ulanowicz. 1989. The Seasonal Dynamics of The Chespeake Bay Ecosystem. Eco. Monographs. 58:329 364. Beisner, B. E., Haydon, D. T., & Cuddington, K. 2003. Alternative stable states in ecology. Frontiers in Ecology and the Environment. 1:376 382. Blindow, I., Andersson, G., Hargeby, A., Johansson, S. 1993. Longterm pattern of alternative stable states in two shallow eutrophic lakes. Freshwater Biol. 30, 159 167. Bricker S. B., Clement C. G., Pirhalla D. E., Orland S. P. and Farrow D. G. G. 1999. National Estuarine Eutrophication Assessment: A Summary of Conditions, Historical Trends, and Future Outlook. National Ocean Service, National Oceanic and Atmospheric Administration, Silver Springs, MD. Castenholz R.W. and J.B. Water bury. 1989. Oxygenic photosynthetic bacteria. Group I. Cyanobacteria. In: Bergeys Manual of Systematic Bacteriology. 3:1710 1789. Chao, A. 1984. Non-parametric estimation of the number of classes in a population. Scand. J. Stat. 11:265 270. Chapman A. D a nd Schelske C. L. 1997. Recent appearance of Cylindrospermopsis (Cyanobacteria) in five hypereutrophic Florida lakes. J Phycol 33:191 195. Cloern, J.E., 2001. Our evolving conceptual model of the coastal eutrophication problem. Mar. Ecol. Prog. Ser. 210, 223 253. Conley D. J, H. W. Paerl, R. W. Howarth,D. F. Boesch, S. P. Seitzinger,K. E. Havens, C. Lancelot, G. E. Likens. 2009. Policy Forum, Controlling Eutrophication: Nitrogen and Phosphorus. Science. 323:1014 1015.
116 De, P. K. 1939. The role of blue green algae in nitrogen fixation in rice -fields. Proc. Roy. Soc. B, 127, 121. Dokulil M. T. and K. Teubner. 2003. Eutrophication and restoration of shallow lakes the concept of stable equilibria revisited. Hydrobiologia. 506 509: 29 35. Drewes G., and Weckesse r J. R.1982. Function, structure and composition of cell wall and external layers. In The Biology of Cyanobacteria, pp. 333 357. Blackwell Scientific Publications. Dutilleul P., J. D. Stockwell, D. Frigon, P. Legendre. 2000. The Mantel test versus Pearson s correlation analysis: assessment of the differences for biological and environmental studies. J Agric Biol Environ Stud 5:131 150. Dyble J., Tester P. A. and Litaker R.W. 2006. Effects of light intensity on cylindrospermopsin production in the cyanobacte rial HAB species Cylindrospermopsis raciborskii, Afr. J. Mar. Sci. 28:309 312. Environmental Protection Agency. 2001. National Coastal Condition Report. EPA 620/R 01/005, Office of Research and Development and Office of Water, U.S. Environmental Protection Agency, Washington, DC. Fay, P. 1965. Heterotrophy and nitrogen fixation in Chlorogloea fritschii. J. Gen. Microbiol. 39:11 20. Fay, P. 1992. Oxygen relations of nitrogen fixation in cyanobacteria. Microbiol Rev 56:340 373. Felsenstein, J. 2004. Inferring Phylogenies. Sinauer Associates, Sunderland, Mass. ) PHYLIP (phylogeny inference package) version 3.6. Fewer D., Friedl T. and Budel B. 2002. Chroococcidiopsis and heterocyst differentiating cyanobacteria are each others closest living relatives. Molecula r Phylogenetics and Evolution. 23:82 90. Flett R. J., Schindler D. W, Hamilton R. D. and Campbell N. E. R. 1980. Nitrogen fixation in Canadian Precambrian Shield lakes. Can.J Fish. Aquat. Sci. 37:494 505. Flores, E., and Herrero A. 1994. Assimilatory nit rogen metabolism and its regulation, p. 487 517. In D. A. Bryant (ed.), The molecular biology of cyanobacteria. Kluwer Academic Publishers, Dordrecht, The Netherlands. Fogg G. E. 1969. The physiology of an algal nuisance. Proc R Soc London B. 173:175 189.
117 Frost J. and E. J. Phlips. 2005. In preparation for submission to Archive Hydrobiologia. Longterm trends in the relationship between phytoplankton standing crops and nutrient compositions of Lake Griffin, Leesburg, Florida, USA. Gao X., 2005. DRAFT TMDL R eport Nutrient and Un-ionized Ammonia TMDLs for Lake Jesup (WBIDs 2981 and 2981A). Givoannoni S. J., S. Turner, G. J. Olsen, S. Barns, D. J. Lane and N. R. Pace. 1988. Evolutionary relationships among cyanobacteria and green chloroplasts. J Bacteriol 170:3584 3592. Havens K. 2003. N:P ratios, light limitation, and cyanobacterial dominance in a subtropical lake impacted by non -point source nutrient polluti on. Environmental Pollution. 122:379 390. Hecky R. E. and Kilham P. 1988. Nutrient limitation of phytop lankton in freshwater and marine environments: a review of recent evidence on the effects of enrichment. Limnology and Oceanography. 33:796 822 Hendrickson J., and Konwinski J. 1998. Seasonal nutrient import -export budgets for the Lower St. Johns River, F lorida. St. Johns River Water Manage. Dist., Palatka, FL. Henson B. J., L. E. Watson, S. R. Barnum. 2002. Molecular Differentiation of the Heterocystous Cyanobacteria, Nostoc and Anabaena, Based on Complete NifD Sequences. Current Micro. 45:161 164. Henson B. J., L. E. Watson, S. R. Barnum. 2003. The Evolutionary History of Nitrogen Fixation, as Assessed by NifD. Journal of Molecular Evo. 58:390 399. Henson B. J., S. M. Hesselbrock, L. E. Watson, S. R. Barnum. 2004. Molecular phylogeny of the heterocystous cyanobacteria (subsections IV and V) based on nifD International Journal of Systematic and Evolutionary Micro. 54:493 497. Honeycutt, R. J., M. McClelland, and B. W. S. Sobral. 1993. Physical map of the genome of Rhizobium meliloti 1021. J. Bacteriol. 175:6945 52. Howarth R. W. 1988. Nutrient limitation of net primary production in marine ecosystems. Ann. Rev. Ecol. Syst. 19 :89 110. Howarth R. W., R. Marino, and J. J. Cole. 1988(b) Nitrogen limitation in freshwater, estuarine, and marine ecosystems. 2. Bi ogeochemical controls. Limnol. Oceanogr. 33:668 701. Howarth R. W., R. Marino 2006. Nitrogen as the limiting nutrient for eutrophication in coastal marine ecosystems: Evolving views over three decades. Limnol. Oceanogr. 51: 364 376.
118 Howarth R.W., D.P. Sw a ney, E.W. Boyer, R. Marino, N. Jaworski and C. Goodale. 2006. The influence of climate on average nitrogen export from large watersheds in the Northeastern United States. Biogeochemistry. 79:163 186. Inglett P. W., E. M. DAngelo, K. R. Reddy, P. V. McCormick and S. E. Hagerthey. 2009. Periphyton nitrogenase activity as an indicator of wetland eutrophication: spatial patterns and response to phosphorus dosing in a northern Everglades ecosystem. Wetland Eco. And Mang. 17:131 144. Isv novics V, Shafik H. M., Presing M, Juhos S. V. 2000. Growth and phosphate uptake kinetics of the cyanobacterium, cylindrospermopsis raciborskii (cyanophyceae) in throughflow cultures. Freshwater biology.:43:256 275 Jensen, J. P, E. Jeppesen, K. Olrik, and P Kristensen. 1994. Imp act of nutrients and physical factors on the shift from cyanobacterial to chlorophyte dominance in shallow Danish lakes. Canadian Journal of Fisheries and Aquatic Sciences 51:1692 1699. Jeppesen E., M. Sndergaard, M. Meerhoff, T. L. Lauridsen J. P. Jensen 2007. Shallow lake restoration by nutrient loading reduction some recent findings and challenges ahead. Hydrobiologia. 584:239 252. Jiang Ze Yu, Brenda G. Rushing, Yong Bai, Howard Gest, and Carl E. Bauer. 1998. Isolation of Rhodospirillum centenum mutants defective in phototactic colony motility by transposon mutagenesis Journal of Bacteriology. 180:1248 1255. Karl D, R. Letelier, L. Tupas, J. Dore, J. Christian & D. Hebel.1997. The Role of Nitrogen Fixation in Biogeochemical Cycling in the Subtropical North Pacific Ocean. Nature 388:5 77 538 Keirn, M. A., and P. L. Brezonik. 1971. Nitrogen fixation in Lake Mize, Florida and in some lacustrine sediments. Limnol. Oceanogr. 16:720 731. Kush, A., C. Elmerich, and J. P. Aubert. 1985. Nitrogenase of Sesbania Rhizobium strain ORS571: purifica tion, properties and 'switch off' by ammonia. J. Gen. Microbiol. 1 31:1765 1777. Lammers P. J, R. Haselkorn. 1984. Sequence of the nifD geneogy coding for the alpha subunit of dinitrogenase from the cyanobacterium Anabaena. Cell 44: 905 911. Larkin M. A., B lackshields G., Brown N. P., Chenna R., McGettigan P. A., McWilliam H., Valentin F., Wallace I. M., Wilm A., Lopez R., Thompson J. D., Gibson T. J., Higgins D. G.. 2007. Clustal W and Clustal X version 2.0. Bionformatics. 23:2947 2948.
119 Litvaitis M. K. 200 2. A molecular test of cyanobacterial phylogeny: inferences from constraint analyses. Hydrobiologia 468:135 145. Lohman, D., J. R. Jones, and C. Baysinger Daniel. 1991. Experimental evidence for nitrogen limitation in a northern Ozark stream. Journal of th e North American Benthological Society 10 :14 23. Lozupone C. and Knight R. 2005. UniFrac: a new phylogenetic method for comparing microbial communities. Appl. Environ. Microbiol. 71:8228 35. Ludden, P. W., and G. P. Roberts. 1989. Regulation of nitrogenas e activity by reversible ADP ribosylation. Curr. Top. Cell. Regul. 30:23 56. Mantel, N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27:209 220. Mantel, N. and R. S. Valand. 1970. A technique of nonparametric multivariate analysis. Biometrics 26:547 558. Mazur B. J, C. F. Chui. 1982. Sequence of the gene coding for the beta -subunit of dinitrogenase from the blue green alga Anabaena. Proc Natl Acad Sci USA 79: 6782 6786. Mevarech M, D. Rice, R. Haselkorn. 1980. Nucleotide sequence of a cyanobacterial nifH gene coding for nitrogenase reductase. Proc Natl Acad Sci USA 77: 6476 6480. Mills M. M., C. Ridame, M. Davey, J. La Roche & R. J. Geider. 2004. Iron and phosphorus colimit nitrogen fixation in the eastern tro pical North Atlantic. Nature 429 :292 294. Moisander P. H, L. Shiue, G. F. Steward, B. D. Jenkins, B. M. Bebout and J. P. Zehr. 2006. Application of a nifH oligonucleotide microarray for profiling diversity of N2-fixing microorganisms in marine microbial ma ts. Environmental Microbiology 8:1721 1735. Moody, H. L. 1970. Factors in Decline of the St. Johns River, Florida. Florida Game and Fresh Water Fish Commision. Morris D. P. and W. M. Lewis, Jr. 1992. Phytoplankton nutrient limitation in Colorado mountain l akes. Fresh water Biology. 20:315 327 Morris, F. W. 1995. Lower St Johns River Basin Reconnaissance. Technical Publication SJ959, St Johns River Water Management District.
120 Moss, B. 1990. Engineering and biological approaches to the restoration from eutr ophication of shallow lakes in which aquatic plant communities are important components. Hydrobiologia 200/201: 367 377. Murry, M. A., A. G. Olafsen, and J. R. Benemann. 1981. Oxidation of diaminobenzidine in the heterocysts of Anabaena cylindrica. Curr. M icrobiol. 6:201 206. NRC 2000. Clean Coastal Waters: Understanding and Reducing the Effects of Nutrient Pollution. National Academies Press, Washington, DC. Odum H. T. 1953. Dissolved Phosphrus in Florida Waters. Report of Investigations No 9, Florida Geol ogical Survey, State of Florida. V:1 40. Oksanen J., R. Kindt, R. B. OHara. 2005. vegan: Community Ecology Package R. package Version 1.6 10. Paerl H. W, L. M. Valdes, J, L. Pinckney, M. F. Piehler, J. Dyble, P. H. Moisander. 2003. Phytoplankton photopigm ents as Indicators of Estuarine and Coastal Eutrophication. BioScience 53:953 964. Paerl H. W. 1988. Nuisance phytoplankton blooms in coastal, estuarine, and inland waters. Limnol Oceanogr. 33:823 847. Paerl, H. W., Crocker, K. M. and Prufert, L. E. 1987. Limitation of N2 fixation in coastal marine waters: Relative importance of molybdenum, iron, phosphorus, and organic matter availability. Limnol. Oceanogr. 31:525 536. Paerl, H. W., Fitzpatrick, M. and Bebout. B. M. 1996. Seasonal nitrogen fixation dynamic s in a marine microbial mat: Potential roles of cyanobacteria and microheterotrophs. Limnol. Oceanogr. 41:419 427. Paerl, H., W. 2009. Controlling eutrophication along the freshwater -marine continuum: Dual nutrient (N and P) reductions are essential. Estua ries and Coasts 32: 1559 2723. R Development Core Team. 2008. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Redfield A.C. 1934. On the proportions of organic derivations in sea water and their relation to the composition of plankton. In James Johnstone Memorial Volume. Reynolds C. S., A. E. Walsby. 1975. Water blooms, Biol. Rev., 50: 437 481.
121 Reynolds, C. S., R. L. Oliver and A. E. Walsby. 1987. Cynaobacterial dominance: the role of buoyan cy regulation in dynamic lake environments. New Zealand J. Mar. Freshwater Res. 21:379 390. Riegman, R. 1998. Species composition of harmful algal blooms in relation to macronutrient dynamics. In Anderson, D. M., Cembella, A. D. & Hallegraeff, G. M. [Eds.] Physiological Ecology of Harmful Algal Blooms. Springer, Berlin 475 88. Rippka R., J. Deruelles, J. B. Waterbury, M. Herdman and R. Y. Stanier. 1979. Generic Assignments, Strain Histories and Properties of Pure Cultures of Cyanobacteria. Journal of Gener al Microbiology. 111:1 61. Robson R. L., R. R. Eady, T. H. Richardson, R. W. Miller, M. Hawkins, J. R. Postgate. 1986. The alternative nitrogenase of Azotobacter chroococcum is a vanadium enzyme, Nature. 322, 388 390. Roeselers G., L. J. Stal, M. C. M. van Loosdrecht and G. Muyzer. 2007. Development of a PCR for the detection and identification of cyanobacterial nifD genes. Jour. of Micro. Method. 70:550 556 Rowan K. S. 1989. Photosynthetic Pigments of Algae. Cambridge, Cambridge University Press, 334 p. S AS Institute. 2001. SAS users guide of release version 8.2. SAS Inst., Cary, NC. Scheffer M., S. Rinaldi, A., Gragnani, L. R. Mur, E. H. van Nes. 1997. On the Dominance of Filamentous Cyanobacteria in Shallow, Turbid Lakes. Ecology. 78:272 282. Schindler D. W. 1977. Evolution of phosphorus limitation in lakes. Science 195:260 262 Schindler, D.W., R.E. Hecky, D.L. Findlay, M.P. Stainton, B.R. Parker, M. Paterson, K.G. Beaty, M. Lyng, and S.E.M. Kasian. 2008. Eutrophication of lakes cannot be controlled by reducing nitrogen input: Results of a 37 year whole ecosystem experiment. Proceedings of the National Academy of Science USA 105: 11254 11258. Schloss, P. D., and J. Handelsman. 2005. Introducing DOTUR, a computer program for defining operational taxonomic units and estimating species richness. Appl. Environ. Microbiol. 71:1501 1506. Schloss, P. D., B.R. Larget, and J. Handelsman. 2004. Integration of microbial ecology and statistics: a test to compare gene libraries. Appl. Environ. Microbiol. 70:5485 5492.
122 Simberloff D. and D. Schmitz. 1997. Biological invasions: A growing threat. Issues in Science and Technology 13:33 40. Sims, J.T., R.R. Simard, and B.C. Joern. 1998. Phosphorus loss in agricultural drainage: Historical perspective and current research. J. Environ. Qual. 27:277 293. SJRWMD report, 2002 Paerl H. W., M. F. Piehler, J. Dyble, P. H. Moisander, J. Leonard and A. Waggener. 2002. Phytoplankton and zooplankton in the St. Johns River System: Factors affecting community structure and function. SJRWMD Report. Smith V. H. 1983 Low nitrogen to phosphorus ratios favor dominance by blue green algae in la ke phytoplankton. Science 221:669 671 Stal, L.J. 1995. Physiological ecology of cyanobacteria in microbial mats and other communities. New Phytol. 131: 1 32. Stewart, S., P. Sucsy, and J. Hendrickson. 2006. Meteorological and Subsurface Factors Affecting Estuarine Conditions Within Lake George in The SJR, FL. Stracke S., C. Kistner, S. Yoshida, L. Mulder, S. Sato, T. Kaneko, S. Tabata, N, Sandal, J. Stougaard, K. Szczyglowski and M. Parniske. 2002. A plant receptor like kinase required for both bacterial a nd fungal symbiosis. Nature 417:959 962. Tamura K., J. Dudley, M. Nei and S. Kumar. 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Molecular Biology and Evolution 24:1596 1599. Thayer G. W. 1971. Phytoplankton productio n and the distribution of nutrients in a shallow unstratified estuarine system near Beaufort, N.C.,Chesapeake Sci. 12:240 253. Thiel T. 1993. Characterization of genes for an alternative nitrogenase in the cyanobacterium Anabaena variabilis. J. Bacteriolog y 175:6276 6286. Thiel T., E. M. Lyons, J. C. Erker and A. Ernst. 1995. A second nitrogenase in vegetative cells of a heterocyst -forming cyanobacterium. Proc. Natl. Acad. Sci. USA 92:9358 9362. Tilman D, Kilham S. S and Kilham P. 1982. Phytoplankton community ecology: The role of limiting nutrients. Ann. Rev. Ecol. Syst. 13 :349 372. Tilman D. 1982. Resource Competition and Community Structure, p. 86. Vincent W. F. 1987. Dominance of bloom forming cyanobacteria (Blue green algae). N. Z Jour Mar and Freshwat Res 21:361 542.
123 Vintila S. and R. El -Shehawy. 2007. Ammonium ions inhibit nitrogen fixation but do not affect heterocyst frequency in the bloom -forming cyanobacterium Nodularia spumigena strain AV1. Microbiology, 153:3704 3712. Vitousek P. M. and R. W. How arth. 1991. Nitrogen Limitation on Land and in the Sea: How Can It Occur?. Biogeochemistry. 13:87 115 Wang Z. C., Burns A., Watt G. D. 1985. Complex formation and O2 sensitivity of Azotobacter vinelandii nitrogenase and its component proteins. Biochemistr y. 24:214 21. Weckesser, J., and U. J. Jrgens. 1988. Cell walls and external layers. Methods Enzymol. 167:173 188. Witz, D. F, R. W. Detroy and P. W. Wilson. 1966. Nitrogen fixation by growing cells and cell free extracts of the bacillaceae. Wolk, C.P. 1968. Movement of carbon from vegetative cells to heterocysts in Anabaena cylindrical. Bacteriol. 96: 2138 2143. Wolk, C.P. 1982. Heterocysts. The biology of cyanobacteria (ed. N.G. Carr and B.A. Whitton), pp. 359 386. Blackwell Scientific Publishers. Young J. P. W. 1992. Phylogenetic classification of nitrogen-fixing organisms, p. 43 86. Zehr J. P. and McReynolds L. A. 1989. Use of degenerated oligonucleotides for amplification of the nifH gene from the marine cyanobacterium Trichodesmium thiebautii. Appl. Environ. Microbiol. 55 :2522 2526. Zehr J. P., Crumbliss L. L., Church M. J., Omoregie E. O., Jenkins B. D. 2003a. Nitrogenase genes in PCR and RT PCR reagents: implications for studies of diversity of functional genes. BioTechniques. 35:996 1002, 1004 1005. Zehr, J. P., M. T. Mellon and S. Zani. 1998. New nitrogen-fixing microorganisms detected in oligotrophic oceans by amplification of nitrogenase ( nifH ) genes. Appl. Environ. Microbiol. 64:3444 3450.
124 BIOGRAPHICAL SKETCH Moshe Doron was born in Tel Aviv, Israel. He received his bachelor degree in Microbiology and Cell Science on 2007 from the University of Florida, USA. He joined the Department of Soil and Water Science, which had an emphasis on ecology, because they gave him an opportunity to research genetics of microorganisms and their influences on the environment. During his volunteering time at Dr. Andrew V. Ogram lab, he was fascinated with the studies and research conducted in the lab. At first, he assisted the research on restoration of the Hole i n the -Donut (due to the invasion of the exotic Brazilian Pepper ( Schinus terebinthifolius )). The following semester, he received his own project about carbon as a final electron acceptor. Unfortunately, he had to stop it after a couple of months because t here were not enough funds for the project. Nevertheless, his hard work was noticed and he got accepted to UF University Scholars Program (USP). In this program, USP co -sponsored his research on biodegradation of naphthalene (naphthalene is a toxin that forms naturally from byproducts of organic matter decomposition and from petroleum contamination). Since he enjoyed working with professor Ogram and love d conducting research with microorganisms, he was looking to expand his scope of research to the entire s ystem in which they operate. Biogeochemistry attracted him b ecause it enable d him to see the entire picture from the smallest detail s ( e.g., genes) up to the biggest ones (the e ntire system). Furthermore, he believe d our planet has environmental problems t hat must be controlled or repaired through a better understanding of the biogeochemical processes On 2007, he joined the MS program at UF in the Soil and Water Science Department with Dr. W. Patrick Inglett as his advisor. He wanted to learn more about bi ogeochemistry and how to apply this knowledge to the management and recovery of our planet.