Microbial processes are extremely important to the functioning of ecosystems, yet we know relatively little about the environmental controls on microbial community compositions. Research has typically focused on either process-based (i.e., rate measurements, flux rates) or, less-often, on microbial community characterization. To gain a true mechanistic understanding of ecosystem functioning, and ultimately predictive capabilities, research that couples biogeochemical process studies with characterization of the microbial community is required (Oremland et al. 2005; Gutknecht et al. 2006; Hobbie et al. 2007). Of particular importance is research that addresses the response of microbial communities to future climate change, in an effort to prepare for, and potentially to mitigate, the impacts of climate change (Doney et al 2004; Hobbie et al. 2007).

Climate change and freshwater withdrawal in watersheds can reduce the volume of freshwater delivered to the coastal zone. As freshwater discharge decreases and freshwater-saltwater mixing zones move upriver, saline waters intrude into previously freshwater regions (Hamilton 1990; Knowles 2002). Rising sea-levels (Fig. 1; Nakada & Inoue 2005; Wigley 2005; Church and White 2006), reduced precipitation in watersheds of rivers (Smith et al. 2005), resulting decline in streamflows (Milly et al. 2005), and global increases in water consumption (Gleick 2003) may result in widespread salinity intrusion into coastal ecosystems. Increased ionic strength and changes in concentrations of many substances (e.g., salts and nutrients) during the transition from freshwater to saltwater conditions may exert a powerful influence on animal (Kupschus & Tremain 2001), plant (Crain et al. 2004), and microbial communities (Rysgaard 1999; Mondrup 1999; Dincer & Kargi 1999; Pattnaik et al. 2000; Mishra et al. 2003).

The introduction of saline waters may present a particular threat to freshwater marshes in coastal zones undergoing salinity intrusion. Marshes must accrete to keep pace with rising sea levels to avoid inundation and eventual wetland loss (Reed 1995). Accretion in marshes depends on allochthonous sources such as land-derived sediment and autochthonous organic matter production and peat formation (Reed 1995). In freshwater marshes, organic matter accumulation is a major mechanism of marsh accretion, and freshwater marshes therefore contain large reservoirs of organic carbon (C) in the form of peat. Increasing salinity may alter pathways of microbial organic matter degradation in freshwater marshes (Weston et al. 2006), potentially increasing mineralization rates of stored marsh organic C. Salinity-induced stress on the freshwater plant community is also likely to decrease production and therefore accretion rates (Pezeshki et al. 1987; McKee & Mendelssohn 1989). The ability of freshwater marshes to keep up with rising sea levels during these periods of elevated C mineralization and reduced accretion will determine the fate of freshwater marshes undergoing salinity intrusion (Weston et al. 2006).

Due to rapid consumption of oxygen (O2) close to the marsh sediment surface, much of the organic matter mineralization in marsh sediments occurs via anaerobic microbial pathways. In general, metabolic processes mediating the oxidation of organic matter in sediments are coupled to the reduction of electron acceptors, the use of which depends on thermodynamic (relative energy yield) and kinetic (reactivity and availability) constraints (Froelich et al. 1979). In freshwater marshes the availability of electron acceptors such as nitrate (NO3-), manganese oxides (MnO2), iron oxides (Fe(OH)3), and sulfate (SO42-) is often limiting, and therefore microbially-mediated methanogenesis is a major pathway of anaerobic organic matter mineralization (Capone & Kiene 1988). Methanogenesis is a terminal fermentation process in which C-1 compounds are disproportionated to methane (CH4) and carbon dioxide (CO2) or reduced with hydrogen (H2) to form CH4 (Ferry 1992; Thauer 1998). CH4 is a powerful greenhouse gas, and freshwater wetlands are a major natural source of CH4 to the atmosphere (Wuebbles & Hayhoe 2002).

Sulfate reduction replaces methanogenesis as the dominant anaerobic microbial terminal C mineralization process in marine sediments (Jørgensen 1982; Capone & Kiene 1988; Howarth 1993) due to the greater availability of SO42- in seawater (approximately 28 mM). The higher energy yield of organic C degradation coupled to sulfate reduction as compared to methanogenesis, in theory, results in the competitive inhibition of methanogens by sulfate reducers for organic matter substrates when SO42- is abundant (Capone & Kiene 1988; Mishra et al. 2003). The hydrogen sulfide (H2S) produced by sulfate reducers is toxic to methanogens and can further inhibit methanogenesis in marine sediments (Visser et al. 1993; O’Flaherty et al. 1998). The rates of CH4 production and emission from saline sediments are therefore often lower than those observed in freshwater sediments (Bartlett et al. 1987; Capone & Kiene 1988).

Previous studies demonstrated substantial differences in microbial community composition between marine and freshwater environments. Osmotic regulation along with the availability of electron donors (organic C) and electron acceptors (e.g. SO42-) are primary controlling determinants of microbial community composition (Csonka 1989). For instance, Gram-positive sulfate reducing bacteria such as Desufotomaculum appear to be dominant in low-SO42- freshwater systems (Castro et al. 2002; Detmers et al. 2004; Leloup et al. 2005a), whereas Gram-negative sulfate reducers in the d-Proteobacteria clade often dominate in saline environments (Ravenschlag et al. 1999; Purdy 2001; Leloup et al. 2005a; Bahr et al. 2005). Population size, as well as community composition, can differ between fresh and marine influenced sediments. For example, Leloup et al. (2005b) measured higher abundances of sulfate reducer functional genes in brackish water estuarine sediments than in freshwater sediments in the Seine Estuary.

Due to the dominance of methanogens in freshwater systems, and sulfate reducers in marine systems, it might be expected that sulfate reducers would simply replace methanogens in freshwater marsh sediments experiencing salinity intrusion, resulting in elevated CO2 production and reduced CH4 production (Weston et al. 2006). However, the measured response of freshwater tidal marsh microbial communities to salinity intrusion was not this straightforward. Ongoing research in the Delaware River estuary showed that both CO2 (reflecting total organic mineralization) and CH4 flux rates increased significantly from sediment cores transplanted from the tidal freshwater marshes (Rancocas) an intermediate salinity site (Salem; Fig. 2), and both of these rates decreased from cores transplanted to a higher salinity site (Stow; Fig. 2). These results (Fig. 3) indicate that, contrary to expectations methanogens are not only able to survive but to thrive during at least the initial phase of salinity intrusion when salinities remain low.

Results from a long-term (1 year) laboratory experiment in which tidal freshwater marsh sediments from the Delaware River were subjected to either freshwater or dilute saline (9 mS cm-1) conditions support the results from the field transplant experiment. As expected, sulfate reduction rates increased significantly following salinity amendments, resulting in an overall increase in CO2 flux for 8 months after salinity intrusion (Fig. 4). However, for 5 months following salinity intrusion CH4 emissions were also elevated (Fig. 4), suggesting that the methanogenic microbial community is stimulated by salinity intrusion rather than inhibited or out-competed, opposite to predictions based on prior work and on thermodynamics.

The results of our ongoing research indicate that a true understanding of the biogeochemistry can only be attained through an understanding of the response of the microbial community to salinity intrusion. The sulfate reducing and methanogenic microbial communities are clearly responding to salinity intrusion (Figs. 3 & 4), and the response of the methanogens is unexpected. Whether these responses are the product of simply an increase in metabolic rate, an increase in the population size, a shift in the community structure, or a combination of these responses is unclear. Given the increase in ionic strength and the relatively large accompanying changes in sediment biogeochemistry associated with salinity intrusion, it is likely that shifts in microbial community structure are responsible for a portion of the measured responses. Knowing the response of the sulfate reducing and methanogenic microbial populations specifically would be highly advantageous.

Commonly used molecular techniques targeting ribosomal RNA genes (i.e. 16S rDNA) are probably of limited value because it is often difficult to ascribe function to various parts of the microbial community. Alternately, analysis of functional genes targets specific members of diverse microbial communities capable of performing specific processes. Functional genes that encode key enzymes in the sulfate reducing (dissimilatory sulfite reductase; dsrAB) and methanogenic (methyl co-enzyme M; mcrA) metabolic pathways have been identified, and are present in all known sulfate reducers and methanogens.

Sulfate reducers are found in both eubacterial and archaeal lineages (Castro et al. 2000), making them difficult to target using 16S rRNA approaches. The dsrAB genes are sufficiently conserved to allow their detection by PCR, while heterogeneous enough to provide phylogenetic information (Wagner et al. 1998; Karkhoff-Schweizer et al. 1995; Zverlov et al. 2005). Indeed, the phylogeny of dsrAB functional genes closely matches the phylogeny of sulfate reducers deduced through 16S rRNA sequences (Wagner et al. 1998; Klein et al. 2001), although lateral gene transfer events for dsrAB have been noted (Klein et al. 2001; Zverlov et al. 2005).

Methanogen-specific functional genes (mcrA) have been used to determine the phylogeny of methanogens (Springer et al. 1995; Luton et al. 2002; Castro 2004; Shigematsu et al. 2004). As with sulfate reducers, the phylogeny of methanogens deduced using mcrA functional genes closely mirrors the relationship between methanogens calculated using 16S rRNA (Luton et al. 2002). Methanogens all belong to the archaeal kingdom Euryarchaeota, and they are found within the Methanomicrobiales, Methanococcales, Methanopyrales, Methanobacteriales, and Methanosarcinales orders, which are phylogenetically diverse (Thauer 1998). Fermentation of acetate and other organic compounds is limited to the Methanosarcinales, while all other methanogens utilize H2 exclusively (Thauer 1998). However, it is estimated that over 70% of total methanogenesis is acetoclastic (Jetten et al. 1992). Our ongoing research includes measurement of both acetoclastic and hydrogenotrophic methanogenesis in the field and laboratory experiments. Given the complex response of the methanogens to salinity intrusion in the field transplant experiments (Fig. 3) and the laboratory salinity amendment experiment (Fig. 4), knowing how the relative importance of the methanogenic pathways (hydrogenotrophic versus acetoclastic) shifts and how the methanogenic microbial community responds to salinity intrusion is key.

We conducted preliminary research targeting dsrAB and mcrA functional genes from freshwater (Rancocas) and brackish water (Stow) marsh sediments from the Delaware River Estuary (Fig. 2). Microbial DNA was extracted and dsrAB functional genes of the expected size were successfully amplified from the freshwater marsh and saltmarsh, and from a positive control (Desulfovibrio vulgaris culture) using the mixed degenerate primer set DSR1F/DSR1Fb/DSR4R (Zverlov et al. 2005). McrA functional genes from the marshes and from a Methanosarcina mazei culture of the expected size were also successfully amplified using the degenerate primer set ME1/ME2. This preliminary work suggests that saline sediments in the Delaware River contain qualitatively more sulfate reducer-specific functional genes than freshwater marsh sediments, as shown by Leloup (2005b). In addition, methanogen-specific functional genes appeared to be more abundant in freshwater marsh sediments than brackish sediments.

Cloning and sequencing of the dsrAB (Fig. 5) and mcrA (data not shown) functional gene amplicons in our preliminary work was also successful. There was relatively little overlap between freshwater and brackish water marsh dsrAB sequences, with three clusters (I, II and III) containing clones from a single site (Fig. 5). Four of the freshwater clones (cluster I) were affiliated with a dsrAB cluster within the d-Proteobacteria Desulfobacterales, composed of the Desulfosarcina, Desulfococcus, and Desulfonema genera. The five remaining freshwater dsrAB sequences (three in cluster II and two in cluster IV) and four sulfate reducers (Zverlov et al. 2005), although a more exhaustive search of uncultured clones may yield matches (Fig. 5). Three dsrAB sequences (cluster III) from the brackish Delaware River marsh were found in a cluster containing sulfate reducers from within the Firmicutes, including Desulfotomaculum, Thermodesulfobium, and Desulfosporosinus species, as well as several Archeaoglobus species within the Archaeal kingdom (Fig. 5).

A mechanistic understanding of the susceptibility of freshwater marshes to climate change-induced sea-level rise requires a comprehensive assessment of several controlling factors. A major challenge lies in coupling microbial population and community shifts with rates of biogeochemical cycling into a comprehensive model of ecosystem function. Shifts in the anaerobic microbial processes of methanogenesis and sulfate reduction will largely dictate how the stored C in freshwater marshes is affected by salinity intrusion. Results from both field transplant laboratory manipulation experiments suggest that the microbial community does not simply shift from dominance by sulfate reducers to dominance by methanogens. How shifts in the microbial community composition of methanogens and sulfate reducers are coupled to changes in the rates of these C mineralization pathways, as well as the timeframe of any shifts, is unknown. Understanding how the microbial community responds is key to elucidating how tidal freshwater marshes will respond to salt-water intrusion and the attendant ecosystem-level consequences.