To Fix or Not To Fix: Controls on Free-Living Nitrogen Fixation in the Rhizosphere

The diversity of free-living N fixers.The ability to synthesize nitrogenase and fix N is exclusively prokaryotic (16). While N-fixing organisms are predominantly bacteria, some methanogenic archaea have been observed to fix N (17). N-fixing organisms are found across a wide range of bacterial phyla, including, Alphaproteobacteria, Betaproteobacteria, Deltaproteobacteria, and Gammaproteobacteria, Firmicutes, Cyanobacteria, and green sulfur bacteria (17). Furthermore, soils are home to high diazotroph diversity, containing over 50% more operational taxonomic units (OTUs) than marine systems (17). This diversity can be observed even within rhizosphere communities. For example, diazotrophs isolated from the switchgrass rhizosphere represented at least 52 different bacterial phylotypes across multiple phyla, including Firmicutes, Alphaproteobacteria, Betaproteobacteria, Deltaproteobacteria, and Gammaproteobacteria (13). Overall, the diversity of diazotrophs actively fixing N in the rhizosphere at any given time is likely to be high.

Despite the high diversity of diazotrophs, nitrogenase, the enzyme involved in BNF, has only three known forms. Nitrogenase consists of two metalloproteins, an iron (Fe) protein responsible for ATP synthesis, and, most commonly, a molybdenum-iron (Mo-Fe) protein responsible for substrate (i.e., N₂) and proton reduction (18). Molybdenum nitrogenase (Mo-nitrogenase) is the most ubiquitous isozyme synthesized by organisms from bacterial phyla Alphaproteobacteria, Betaproteobacteria, Deltaproteobacteria, Gammaproteobacteria, Firmicutes, and Cyanobacteria. Many diazotrophs from Alphaproteobacteria, Betaproteobacteria, Gammaproteobacteria, and Firmicutes can also synthesize alternative forms of nitrogenase that substitute the Mo-Fe cofactor with vanadium-iron (V-nitrogenase) and/or iron-iron (Fe-nitrogenase) cofactors under Mo-limited conditions (18, 19). These slight variations in enzyme structure may influence FLNF and its responses to environmental conditions.

It has been shown that the different forms of nitrogenase vary in substrate affinity, efficiency, and temperature sensitivity, all of which influence FLNF rates. For example, Bellenger et al. demonstrate that alternative forms of nitrogenase exhibit lower R ratios (the ratio between FLNF rates measured by acetylene reduction, an indirect measure of N fixation, and rates measured via fixation of ¹⁵N₂) than Mo-nitrogenase (20). Azotobacter vinelandii R ratios for Mo-nitrogenase were found to 3.5 ± 1.1, while R ratios for V-nitrogenase and Fe-nitrogenase were 1.2 ± 0.4 and 0.5 ± 0.3, respectively (20). This indicates that alternative nitrogenase enzymes have a lower affinity for acetylene gas than the Mo-nitrogenase. V-nitrogenase also expresses higher isotopic discrimination against ¹⁵N₂ with a fractionation factor of –4‰ versus –1‰ for Mo-nitrogenase (21). Electron allocation varies among the different forms of nitrogenase as well (22). Mo-nitrogenase allocates the majority (∼75%) of its electrons to N₂ reduction, while Fe-nitrogenase allocates the majority of its electrons to proton reduction (22). V-nitrogenase electron allocation approaches a 50:50 exchange between N₂ reduction and proton reduction (22). These results seem to suggest that the Mo-nitrogenase is the most efficient nitrogenase at converting N₂ to NH₃; however, other work has shown that temperature influences the relative efficiencies of these isozymes, complicating this issue. V-nitrogenase has been shown to be more effective than Mo-nitrogenase at low temperatures (∼5°C), as illustrated by a 40-fold versus 400-fold decrease in activity, respectively, as temperature decreased from 30°C to 5°C (23). At higher temperatures (e.g., 30°C), Mo-nitrogenase is more efficient, most likely due to its higher affinity for N₂ than V-nitrogenase, as indicated by differences in K_m for the reduction of N₂ to NH₃ (19 and 29 kPa for Mo-nitrogenase and V-nitrogenase, respectively) (24). These functional differences illustrate the potential for different forms of nitrogenase to respond differentially to various environmental conditions, such as metal availability (see below) and temperature, and highlight the need for further research into how replacement of Mo with V or Fe influences nitrogenase function.

Carbon controls on free-living N fixation.It is well known that N fixation is an energetically and, therefore, carbon (C) expensive process. This was shown in early studies of FLNF, where the free-living N fixer Clostridium pasteurianum was incapable of fixing N unless supplied with adequate availability of C substrate (i.e., sucrose or pyruvate) (25). This substrate requirement is driven by the high demand for ATP by nitrogenase (26); 16 ATP and 8 electrons are required for the conversion of 1 N₂ molecule to 2 NH₃ molecules (27). Such high energy demands limit FLNF such that diazotrophs can only fix N when adequate supplies of C are available. In fact, FLNF was previously overlooked as a significant source of N because it was thought that soil organic matter could not provide enough energy, in the form of accessible C, to support N fixation (28). However, plant root exudates, C-rich secretions consisting of low-molecular-weight compounds such as sugars, organic acids, and mucilage (i.e., polysaccharides) (29), are a potential source of C capable of meeting diazotroph energy demands. Root exudation makes the rhizosphere a hot spot for microbial activity (30) and a key area for FLNF in the soil.

Plants allocate a significant portion of their fixed C belowground; on average, ∼40% of a plant’s photosynthate is translocated into belowground biomass (31). Of this translocated C, ∼12% is typically recovered in the soils as root exudates, root exudate-derived metabolites, and microbial biomass C (31). For some prairie grasses, the portion of fixed C recovered from soils can be as much as 15% (32). Switchgrass is known to allocate a significant portion of its fixed C belowground. Switchgrass C allocation to roots and soil was measured at 40% and 6%, respectively, 1 day after a ¹³C-CO₂ pulse-chase labeling (33). Of the 6% fixed C recovered from soils, 92% was found in microbial biomass (33), indicating that C recently fixed by switchgrass was quickly assimilated into the rhizosphere microbial community. This highlights the potential for switchgrass to support FLNF in its rhizosphere. Furthermore, as discussed in a recent review by Bowsher et al., both quality and quantity of root exudation responds to N availability, highlighting the interplay between plant C inputs and soil N availability (34). While we are unaware of any studies which have directly explored the response of FLNF to additions of root exudates, it is well established that C additions typically stimulate N fixation, with additions of C being an integral part of the methods used to measure rates of rhizosphere FLNF (35). There is great need for studies that elucidate the linkages and feedbacks between N availability, plant C exudation rates, and FLNF.

Although rhizosphere-focused studies are limited, work on FLNF in other regions, particularly bulk soil and litter, suggest that quality (i.e., form) of C substrates may be just as important as the quantity of C in regulating FLNF (1). Glucose has been used as a C source for methods assessing N fixation in the rhizosphere and bulk soil (35), but methods for isolating and culturing diazotrophs often use other forms of C, such as malate, mannitol, and sucrose (36). Other C compounds, including acetate, have been shown to inhibit nitrogenase activity; nitrogenase activity of Azotobacter paspali was completely inhibited in pure cultures grown on acetate and reduced by 50% on root surfaces exposed to acetate (37). Furthermore, there can be differential responses to C sources depending on environmental conditions. When in association with grass roots, N fixation by A. paspali increased with additions of citrate, but when in pure culture, citrate additions reduced N fixation rates by half (37). These studies provide some insight on the influence of specific C compounds on FLNF, but root exudates are a complex mixture of low-molecular-weight compounds (29). Switchgrass exudates, for example, were found to contain over 30 different compounds (L. K. Tiemann, A. W. Bowsher, M. L. Friesen, S. E. Evans, and D. W. Hoyt, unpublished data) (Table 1), suggesting that diazotrophs in the switchgrass rhizosphere have access to a diverse range of C compounds. Carbon form, particularly the diverse C forms in the rhizosphere, may be an important control on FLNF; however, it is very difficult to draw any concrete conclusions about the influence of C form on FLNF because of a lack of rhizosphere studies that explore this topic. This clear hole in our understanding of rhizosphere FLNF highlights the need for studies assessing how C compounds found in plant root exudates influence FLNF of individual diazotrophs as well as complex microbial communities (e.g., rhizosphere communities).

The form of C available to FLNF can also drive the efficiency and productivity of N fixation and determines the growth strategy of the organisms. Although little is known about the efficiency and productivity of rhizosphere FLNF specifically, we can make inferences based on general FLNF research. Surprisingly, FLNF may be more productive (i.e., greater N fixed per unit biomass) than symbiotic N fixation. Although symbiotic N fixers can fix more N per gram of cellular material (1.0 to 2.5 versus 0.1 g N₂ g⁻¹ cellular material for diazotrophs) (38), free-living diazotrophs live much shorter lives (on the order of hours versus weeks for symbionts) and have 10 times higher nitrogenase activity (25 to 50 mg N fixed g⁻¹ protein h⁻¹ versus 2 to 5 mg N fixed g⁻¹ protein h⁻¹ for symbiotic N fixation) (38). If these rates are averaged over the lifetime of the organism, rhizosphere diazotrophs may match or even exceed fixation by symbionts (38). Thus, if plants are supporting a large and active diazotroph community, they may be benefiting from highly productive N fixation, with N becoming available as the diazotroph biomass rapidly turns over. Although these data highlight the potential importance of FLNF as a plant N source, it is important to note that these FLNF rates are based on growth under optimal conditions. As discussed in subsequent sections, variations in other environmental conditions (e.g., oxygen availability and nutrient availability) are likely to influence the productivity of FLNF.

Oxygen controls on free-living N fixation.Oxygen concentrations in the rhizosphere are dynamic and extremely difficult to measure; therefore, most available research on rhizosphere oxygen concentration has been conducted in saturated systems (e.g., wetlands or sediments). For example, studies we could find were conducted using wetland plants grown in peat or sand (39, 40), agar (41), or using seagrass (Cymodocea rotundata) grown in saturated sediments (42). In these studies, rhizosphere oxygen concentration ranged from near 0% (anaerobic) to 20% (ambient). While these studies indicate a wide range of oxygen concentrations that occur in the rhizosphere, it is unclear if this variability in oxygen concentration is similar in nonsaturated rhizospheres such as that of switchgrass. We also note that soil texture (including particle size and compaction) is likely to impact rhizosphere oxygen concentrations; this topic is discussed further in the “Other environmental controls on free-living nitrogen fixation” section below.

Rhizosphere oxygen concentration also likely follows a diel pattern that could partially control the activity of rhizosphere-associated microbes (42). Modeling efforts suggest that diel changes in water flow through the rhizosphere (43) would have a big impact on oxygen concentration. Active root and microbial growth in the rhizosphere may also create oxygen depletion zones within the rhizosphere (44). Microsites of very low or relatively high oxygen concentration may also form in the rhizosphere, as occurs with roots of aquatic systems (45). We are unaware of any studies which have specifically examined oxygen as a control on FLNF in the rhizosphere; however, we know oxygen is a strong inhibitor of nitrogenase activity, and we can discuss physiological responses of diazotrophs to oxygen using research derived predominately from pure culture studies.

Oxygen irreversibly inhibits nitrogenase, even in aerobic organisms (46). Therefore, diazotrophs must employ protection mechanisms to maintain N fixation when oxygen is present. This includes avoidance of oxygen via growth strategy, spatial and/or temporal isolation of nitrogenase from oxygen, and production of biofilms as oxygen diffusion barriers (2, 19). Diazotrophs can also remove oxygen by increasing substrate utilization, which increases respiration rates, thereby decreasing oxygen concentrations (19, 38). This particular mechanism is likely at work in the switchgrass rhizosphere, as switchgrass has been shown to stimulate microbial growth in the rhizosphere via exudation and, thereby, substrate utilization (47). A rapidly growing rhizosphere community, regardless of diazotroph presence, is likely to reduce oxygen concentrations around the root to FLNF-favorable levels.

Oxygen management requires energy investment and so can greatly influence the efficiency of FLNF, or the amount of N fixed per unit C. Microaerophilic organisms are the most efficient N fixers, fixing an estimated 26 mg N₂ g⁻¹ C metabolized (27 and references therein). In contrast, anaerobes can fix 11 mg N₂ g⁻¹ C, while aerobes can only fix 7 mg N₂ g⁻¹ C (27 and references therein). One study found that upward of 60% of the energetic costs of N fixation are indirect costs associated with combating oxygen (48). This may be especially true when organisms use increased respiration to remove oxygen (49). Patra et al. found a negative relationship between substrate-induced respiration and rates of FLNF in both rhizosphere and bulk soil, suggesting that as diazotrophs increase respiration to combat oxygen damage, there is less C available for FLNF (50). However, even under high oxygen pressure, if carbohydrate availability is sufficiently high, diazotrophs may still carry out N fixation (49).

Under optimal oxygen concentrations, N fixation can actually be an energetically favorable mechanism for NH₃ acquisition, having a slight energetic advantage over assimilatory nitrate reduction (48). However, because of the diversity of diazotrophs present and potentially active in the rhizosphere, it is difficult to pinpoint one optimal oxygen concentration for FLNF. For example, Inomura et al. found the oxygen optima of Azotobacter vinelandii to be 3%, while Großkopf and LaRoche demonstrated that oxygen concentrations around 5% resulted in significantly higher nitrogenase activity and lower respiration by Crocosphaera watsonii, a marine cyanobacteria, than ambient oxygen concentrations (20%) (48, 49). It is likely that diazotrophs with different growth strategies, oxygen protection mechanisms, and C demands will also have different oxygen optima. Overall, work assessing the response of FLNF to different oxygen concentrations is sparse, and there is no information available about how whole communities of diazotrophs may respond to oxygen availability or if a community-level oxygen optimum exists.

Free-living N fixation and N availability and form.As discussed by Reed et al., diazotrophs can access N via N fixation or through uptake of externally available N sources (2), which can include both low- and high-molecular-weight organic N sources (51). Given the energy costs of FLNF, it is generally downregulated by increasing N availability, as diazotrophs use external N in favor of fixed N (2). This is corroborated by several field studies which demonstrate that environments with low N availability, whether soil, rhizosphere, or moss, typically have greater FLNF rates than sites with high N availability (50, 52, 53). The switchgrass rhizosphere is likely to be one such environment as root N uptake creates an N deplete zone in the rhizosphere (44) and switchgrass is thought to be particularly skilled at scavenging N (54). In fact, we confirm in our own work that unfertilized switchgrass rhizospheres exhibit greater FLNF rates than fertilized rhizospheres (Fig. 4). However, as different N sources will require variable amounts of energy for uptake and utilization, N form is likely to play an important role in how N availability influences FLNF.

Ammonium, the direct product of N fixation, is well known to inhibit N fixation (2) and has been shown to inhibit nitrogenase synthesis at the genetic level through the regulation of nifA gene transcription (19). However, it does not inhibit the activity of already synthesized nitrogenase in most organisms (38). In fact, there are only a few diazotrophs that regulate nitrogenase posttranslationally, including Azospirillum brasilense (19). This posttranslational regulation is carried out by dinitrogenase reductase ADP-ribosyltransferase (DraT) in response to ammonium and is reversed by dinitrogenase reductase-activating glycohydrolase (DraG) (19). Organisms with DraG-DraT regulation are likely to be more responsive to increases in ammonium, shutting down nitrogenase activity as soon as ammonium becomes available, while organisms without this posttranslational regulation may cease enzyme synthesis in response to ammonium but will continue to have functioning nitrogenase in their cells. Characterizing the presence of posttranslational regulation systems is important to understanding diazotroph response to changes in N availability, particularly ammonium. Huergo et al. suggest that DraT may be present in many diazotrophs (55), but to date, this regulatory system has only been well studied in the photosynthetic bacteria Rhodospirillum rubrum (56).

Glutamine also downregulates nitrogenase synthesis at the genetic level (nifA) via a pathway similar to ammonium regulation (19). In fact, glutamine may influence ammonium regulation of N fixation as intracellular glutamine levels regulate glnD modification of the P_II protein, an important cellular N sensor (55). Glutamine and glutamate were found to downregulate N fixation of Herbaspirillum seropedicae (57). However, in the same study, histidine, lysine, and arginine had no effect on nitrogenase activity. In a similar study, ammonium, glutamine, and nitrate reduced N fixation in A. brasilense (58). Conversely, nitrate concentrations of up to 800 µmol had no inhibitory effect on growth or N fixation by C. watsonii (48).

Overall, the availability of external N sources generally reduces rates of FLNF. Although, few studies have targeted rhizosphere diazotrophs to confirm these responses to N availability and the role of N form. Furthermore, the magnitude of this response in the rhizosphere (i.e., complete shutoff FLNF or reduced rates of fixation) which is likely to depend both on the concentration and the form of external N, is not known.

Controls of phosphorus and micronutrients on free-living N fixation.The availability of phosphorus (P) and micronutrients, including Fe, Mo, and V, is known to influence N fixation (2). Yet, to the best of our knowledge, there are no studies on the controls of P, Fe, Mo, and/or V specifically targeted to rhizosphere FLNF. There have been many studies examining the importance of these nutrients on FLNF, mostly summarized by Reed et al., but these have been almost exclusively conducted in bulk soil and leaf litter (2). Here, we summarize more recent findings that may help us better understand the nutrient constraints on rhizosphere FLNF.

Phosphorus is a key nutrient in energy production and has been frequently documented as a control on N fixation (59, 60). FLNF associated with the nonlegume tree Eucalyptus urophylla was ∼27% higher in soils receiving additions of P than in the “no-added P” control (61). Furthermore, additions of N and P resulted in similarly boosted FLNF rates (61), suggesting that P limitation was a stronger driver of FLNF than N availability in this system. A meta-analysis examining FLNF responses to nutrient additions found P fertilization to significantly increase FLNF but only in tropical forest systems (62). We may expect a similar response in the rhizosphere where root uptake of P can result in a P-depleted rhizosphere (44); thus, P may be a particularly important limiter of rhizosphere FLNF.

The response of FLNF to P availability is highly variable and can be further complicated when other nutrients, including Fe, Mo, and V, which are all essential components of nitrogenase, are also limiting. For example, Wurzburger et al. found that the limitation of FLNF by Mo and P varied along a P gradient of Panamanian soils (63). In P-rich soils, Mo was most limiting, but this shifted to a colimitation of Mo and P in P-poor soils (63). However, P alone never limited N fixation in this system (63). Conversely, a study of Costa Rican soils found that P availability was the dominant control on N fixation, while Mo concentrations did not correlate with soil FLNF rates (60). These contrasting findings may be the result of differences in soil organic matter. Wichard et al. describe how binding of Mo to organic matter can prevent Mo limitation (64). This suggests that Mo may be more available in organic soils and, therefore, a less important control on FLNF in organic-rich soils. Overall, the majority of studies indicate that increased P, Fe, Mo, and V availability generally have positive effects on FLNF. However, the most limiting nutrient and, therefore, the dominant control on FLNF is variable. This could be particularly relevant in the rhizosphere where recent advances in two-dimensional (2D) and three-dimensional (3D) element mapping have revealed connections between root growth and exudation and micronutrient concentrations, particularly metals, that can control microbial community composition and physiology (65). For example, high concentrations of available iron found at root tips (66) could be important for diazotrophs, as iron plays such a crucial role in nitrogenase construction and functioning. This highlights the need to explore how availability of these nutrients, specifically in the rhizosphere, may influence diazotroph community composition and the potential for FLNF.

Other environmental controls on free-living N fixation.The rhizosphere habitat poses an extra set of challenges to FLNF not experienced by symbiotic N fixation, as diazotrophs are more directly influenced by variation in the soil environment. For example, soil texture may influence how diazotrophs manage oxygen because of the relationship between texture and diffusion of substrate (i.e., C) and oxygen. Increasing clay content of soils can create microaerophilic and anaerobic microsites where bacteria can be protected from oxygen exposure (67), thus, potentially supporting larger populations of N fixers and/or more efficient N fixers. Indeed, Gupta and Roper found more rhizosphere N fixation in soils with greater clay content, which shifted to greater N fixation along root surfaces as clay content decreased (67). While soil texture is known to influence soil microbial community activity, there is little research exploring how soil texture may influence FLNF in the rhizosphere.

Soil pH, which is also likely to be highly variable in the rhizosphere, may also be an important environmental control on FLNF. The rhizosphere is a dynamic environment in which root growth is continuously altering the pH of the surrounding soil (39); therefore, pH may have a different effect on rhizosphere FLNF. For example, pH can change 0.5 to 1 pH units when moving just 1 mm away from the root surface (68). We also know that at the field scale, acidic soil pH has been shown to decrease N-fixation rates in aerobic soils (1). Furthermore, a study of alpine meadow soils, with pH values ranging from 5 to 8, found lower richness and diversity of diazotroph communities at acidic pH (69) suggesting that lower N fixation in acidic soils may be due to reduced community redundancy. Wang et al. also examined the relative abundances of the three dominant genera in their soils (Azospirillum, Bradyrhizobium, and Mesorhizobium) across the pH gradient (69). Azospirillum (a free-living N fixer) abundance did not vary significantly with pH, but Bradyrhizobium (a free-living and symbiotic N fixer) abundance increased with decreasing pH and Mesorhizobium (a symbiotic N fixer) abundance was reduced at acidic pH (69). This work provides some insight into response of rhizosphere FLNF to soil pH and suggests it may depend heavily on the dominant diazotroph in the community. But, with the lack of in situ rhizosphere studies, it is difficult to draw any strong conclusions about the magnitude or direction of rhizosphere FLNF response to changes in soil pH.

FLNF is also known to be temperature and moisture sensitive, increasing as both temperature and soil moisture increase, and these responses are well summarized by Reed et al. (2). However, rhizosphere-focused studies are lacking, highlighting a need for both mechanistic and field-scale studies aimed at addressing this knowledge gap. Soil moisture may be of particular interest, as the roots are likely to exhibit strong control over rhizosphere water availability, creating water accumulation and depletion zones according to uptake and consumption (44), which could dramatically alter oxygen dynamics.

Methodological considerations for studying free-living N fixation.The assessment of FLNF rates and diazotroph diversity across environmental gradients and ecosystems is crucial to furthering our understanding of FLNF. However, there are methodological issues that should be considered in the assessment of both rates and diversity. FLNF rates are most commonly measured either via acetylene reduction or ¹⁵N isotope enrichment. Acetylene reduction takes advantage of the ability of nitrogenase to reduce triple bonded molecules other than N₂. In this method, diazotrophs are supplied acetylene (C₂H₂) which is reduced to ethylene (C₂H₄) via nitrogenase (70). Concentrations of ethylene can then be measured over time to obtain a proxy for FLNF rates. Unfortunately, there are multiple issues that should be accounted for when using this method. First, this is an indirect measure of FLNF and, as such, requires a conversion constant that relates acetylene reduction rates to N₂ reduction rates. While a conversion factor of 3 is commonly used to transform acetylene reduction rates to FLNF rates (70), the actual conversion value can vary from 30 to 1 (20). As discussed above (see “The diversity of free-living nitrogen fixers”), the different forms of nitrogenase have different affinities for acetylene which can result in highly variable ratios between acetylene reduction rates and FLNF rates. Moreover, this method relies on the measurement of ethylene production over time. It is assumed that this ethylene production is the direct result of acetylene reduction; however, ethylene has been shown to evolve from soils independent of acetylene (71). Ethylene is an important compound in plant growth and signaling and is produced by both plants and bacteria (72). This production is not associated with the presence of acetylene. Controls that account for background ethylene production can be included with the analysis; however, they have been shown to result in misrepresentative final rates of acetylene reduction (71). Therefore, we recommend that acetylene reduction not be used for the assessment of rhizosphere FLNF rates.

The ¹⁵N isotope incorporation method represents a good alternative to acetylene reduction. This method supplies diazotrophs with ¹⁵N₂ in place of atmospheric N₂, which is dominated by ¹⁴N (35). As diazotrophs fix N, the ¹⁵N label is incorporated into the soil and can be measured to determine FLNF rates (35). A reference sample which provides ¹⁵N content in a given sample prior to ¹⁵N₂ fixation is required to obtain FLNF rates. ¹⁵N isotope enrichment is a direct measure of FLNF rates and is, therefore, more accurate than acetylene reduction. However, this method is costlier in time and funds than acetylene reduction. The contribution of N via N fixation has also been frequently assessed isotopically by assuming all fixed N found in soils will have a value of 0‰ (73, 74). In other words, it is assumed that there is no fractionation during N fixation. Although symbiotic N fixation does not result in any fractionation, this is not true of FLNF which has an average fractionation factor of –2.5‰ (21). However, the true fractionation factor of a sample will depend on the relative abundance of different nitrogenase forms and may actually range from –1‰ if Mo-nitrogenase dominated to –4‰ if V-nitrogenase dominated (21). When trying to assess the relative contribution of N fixation to soil N pools, this fractionation must be accounted for in order to accurately assess the contribution of FLNF as well as symbiotic N fixation.

The assessment of diazotroph diversity also comes with challenges and we direct you to Gaby and Buckley for a detailed description of some of these challenges (75). In brief, Gaby and Buckley discuss two major concerns surrounding diazotroph diversity assessment (75). First, PCR primer selection for the nifH gene can lead to bias in measures of both diversity and relative abundance of amplified organisms (75). If using universal PCR primers, it is important to select primer pairs with high coverage but low phylogenetic bias (76). Second, paralogs of nifH can lead to a false-positive detection of the nifH gene (75). Lastly, as discussed earlier, there are several forms of nitrogenase. These isozymes are regulated by different genes, namely, nifH, anfH, and vnfH for Mo-nitrogenase, Fe-nitrogenase, and V-nitrogenase, respectively (77). Thus, it may also be prudent to assess the diversity of the anfH and vnfH genes alongside nifH.