Comparative Assessment of Nitrogen Fixation Methodologies, Conducted in the Oligotrophic North Pacific Ocean

Abstract

Resolution of the nitrogen (N) cycle in the marine environment requires an accurate assessment of dinitrogen (N₂) fixation. We present here an update on progress in conducting field measurements of acetylene reduction (AR) and ¹⁵N₂ tracer assimilation in the oligotrophic North Pacific Subtropical Gyre (NPSG). The AR assay was conducted on discrete seawater samples using a headspace analysis system, followed by quantification of ethylene (C₂H₄) with a reducing compound photodetector. The rates of C₂H₄ production were measurable for nonconcentrated seawater samples after an incubation period of 3 to 4 h. The ¹⁵N₂ tracer measurements compared the addition of ¹⁵N₂ as a gas bubble and dissolved as ¹⁵N₂ enriched seawater. On all sampling occasions and at all depths, a 2- to 6-fold increase in the rate of ¹⁵N₂ assimilation was measured when ¹⁵N₂-enriched seawater was added to the seawater sample compared to the addition of ¹⁵N₂ as a gas bubble. In addition, we show that the ¹⁵N₂-enriched seawater can be prepared prior to its use with no detectable loss (<1.7%) of dissolved ¹⁵N₂ during 4 weeks of storage, facilitating its use in the field. The ratio of C₂H₄ production to ¹⁵N₂ assimilation varied from 7 to 27 when measured simultaneously in surface seawater samples. Collectively, the modifications to the AR assay and the ¹⁵N₂ assimilation technique present opportunities for more accurate and high frequency measurements (e.g., diel scale) of N₂ fixation, providing further insight into the contribution of different groups of diazotrophs to the input of N in the global oceans.

INTRODUCTION

The biological conversion of dinitrogen (N₂) gas into ammonia (NH₃) is performed by a select group of organisms, termed diazotrophs. Examining the identity of diazotrophs present in the marine environment, determining their abundance, and understanding their physiological characteristics is vital to improving our understanding of the marine nitrogen (N) cycle. Equally important is quantifying the contribution of diazotrophs to the fixed pool of N in the global ocean, which is estimated to range from 100 to 200 Tg of N year⁻¹ (24). Calculating global N₂ fixation via field-based measurements (5, 25) results in lower estimates of global N₂ fixation compared to geochemical evidence such as nutrient stoichiometry and isotopic ratios (10, 20, 27). In part, this may derive from the methodologies used to measure N₂ fixation, a topic that forms the basis of the present study. Recently, there have been several model-based estimates on the specific contributions of individual diazotroph groups to the fixed pool of N (18, 29). The model outputs support field-based observations that the contribution of fixed N by unicellular N₂-fixing cyanobacteria, including group A (termed UCYN-A) and Crocosphaera spp., can equal or even exceed the contribution of larger N₂-fixing microorganisms, such as Trichodesmium and various heterocystous cyanobacteria (12, 31, 41).

The commonly applied methods to measure the rates of marine N₂ fixation include the acetylene reduction (AR) assay (4, 7, 12, 36) and the ¹⁵N₂ assimilation technique (30). The AR assay relies on the preferential reduction of acetylene (C₂H₂) to ethylene (C₂H₄) by nitrogenase, instead of reducing N₂ to NH₃ (34). The most extensive applications of the AR assay at sea include the measurement of C₂H₄ production by Trichodesmium colonies via gas chromatography (6, 7) and the use of an online system incorporating a laser photoacoustic detector (37, 42). However, the AR assay has not escaped criticism during the past 4 decades of its use, predominantly due to potential indirect effects of C₂H₂ on microbial metabolism and the reliability of the factor used to extrapolate rates of C₂H₄ production to rates of N₂ fixation (14, 16, 21, 36). An additional concern, especially for oligotrophic seawater samples, is the need to concentrate the microbial biomass, either by shipboard filtration or net tows, in order to obtain a detectable signal of C₂H₄ production (6, 12, 37). The filtration process can impose undesired effects on the microbial population due to cell damage (1, 13). In comparison, the ¹⁵N₂ assimilation technique follows the net incorporation of the ¹⁵N₂ tracer into cellular biomass after a predetermined incubation period (2, 17, 26). Because the ¹⁵N₂ technique has a lower detection limit than the AR assay, it is the preferred method for the oligotrophic marine environment (30). Indeed, ¹⁵N₂ assimilation has been used to describe rates of N₂ fixation at Station (Stn) ALOHA (A Long-term Oligotrophic Habitat Assessment) located at 22°45′N, 158°W in the North Pacific Subtropical Gyre (NPSG) (9, 11). However, it has recently been demonstrated that the method used to add ¹⁵N₂ to the seawater sample can affect the measured rate of N₂ fixation (28). The authors of that study found that the addition of ¹⁵N₂ as a gas bubble underestimated N₂ fixation because the gas bubble introduced does not attain equilibrium within the incubation period, causing an unknown and time-dependent ¹⁵N/¹⁴N ratio for the N pool.

We describe methodological developments to improve both the AR assay and the ¹⁵N₂ assimilation technique. We demonstrate that the rates of C₂H₄ production are quantifiable after incubations of 3 to 4 h by measuring the C₂H₄ concentration in equilibrated headspace gas using a reduced gas analyzer that incorporates a reducing compound photodetector. Importantly, no preconcentration of microbial biomass is required. Additional key methodological aspects of the AR assay are also documented, including the saturation of nitrogenase with C₂H₂ and the importance of blank control treatments. The AR assay measurements were conducted alongside rates of ¹⁵N₂ assimilation, whereby ¹⁵N₂ was added to seawater samples in both gaseous form and as ¹⁵N₂-enriched seawater. We describe the preparation and storage of ¹⁵N₂-enriched seawater for use on oceanographic cruises, making its use in the field more time efficient. The resulting advances to the AR assay and ¹⁵N₂ assimilation method described here allow more accurate quantification of N₂ fixation and increase the ability to conduct time-resolved rate measurements in the oligotrophic ocean.

MATERIALS AND METHODS

The fieldwork was conducted on multiple expeditions to the NPSG between October 2010 and December 2011. Seawater samples were collected during Hawaii Ocean Time-series (HOT) cruises to Stn ALOHA and an additional oceanographic cruise located ∼50 nautical miles to the north of Stn ALOHA (25°N, 157°30′W).

AR assay.

The AR assay described here differs from the “typical” AR assay (see, for example, reference 4) in two main aspects. The first is with regard to sample preparation with C₂H₂ gas added in dissolved form to the seawater sample, followed by incubation of the seawater sample with no headspace, and subsequent quantification of the ensuing C₂H₄ gas in the whole sample. This differs from the routinely reported applications of the AR assay (4, 7, 12), whereby C₂H₂ gas is added to serum vials containing approximately two-thirds sample seawater and one-third headspace, and aliquots of the gas phase are analyzed for C₂H₄ concentrations at selected intervals during the sample incubation period or a predetermined endpoint. The second major methodological difference is that C₂H₄ concentrations are quantified by using a reducing compound photodetector in contrast with the more routinely utilized gas chromatography-flame ionization detector (GC-FID). Below, we describe the five steps of the modified AR assay: (i) preparation of the C₂H₂-enriched seawater, (ii) sample collection and incubation, (iii) extraction of the dissolved C₂H₄ gas, (iv) quantification of the C₂H₄ concentrations, and (v) calculation of the C₂H₄ concentration.

(i) Preparation of C₂H₂ enriched seawater.

The C₂H₂-saturated seawater for the AR assay was prepared ∼1 h prior to use. To generate the C₂H₂ gas, 6 g of calcium carbide (CaC₂; Sigma) was added to 150 ml of deionized water in a 250-ml side-arm glass flask according to the following equation:

$$Ca{C}_2+2H_{2}O=C_2{H}_2+Ca{(OH)}_2\mathrm{.}$$ (1)

The resulting C₂H₂ gas was transferred via the side-arm of the flask and 1/8-in. polytetrafluoroethylene (PTFE) tubing to the base of a secondary 1-liter glass flask that contained 300 ml of filtered (0.2-μm pore size) seawater collected from the sampling location. To enhance the mass transfer of C₂H₂ gas into the filtered seawater, the outlet tubing was fitted with a ceramic air stone diffuser resulting in the production of microbubbles. After purging the filtered seawater with C₂H₂ and shaking vigorously for 5 min, the C₂H₂-enriched seawater was stored in the dark until use. No measurements were made of C₂H₂ concentrations using the reducing compound photodetector since this would rapidly deplete the bed of mercuric oxide. Rather, we relied on empirical solubility studies showing that seawater saturated with C₂H₂ (1.6 ml of C₂H₂ per ml of water) contains 65 mM C₂H₂ (35). It should also be noted that C₂H₂ produced from calcium carbide does not result in pure C₂H₂ and contaminants are present, including C₂H₄ (22). These can potentially be decreased by scrubbing the C₂H₂ gas stream through an additional water flask (8); however, the presence of any level of contaminant C₂H₄ necessitates careful time zero measurements and blank controls, discussed in detail below.

(ii) Sample collection and incubation.

Seawater samples were collected using a CTD-rosette and transferred to acid-washed, combusted, glass-stoppered 300-ml Wheaton bottles that were filled to two times overflowing. Subsequently, 20 ml of the seawater sample was removed from the Wheaton bottle using a pipettor and replaced with 20 ml of C₂H₂-enriched seawater, with a final C₂H₂ concentration of 4 mM. Replicate (n = 3) samples were immediately analyzed after the introduction of C₂H₂ enriched seawater to provide a time zero C₂H₄ concentration. Control samples, consisting of 0.2-μm-pore-size filtered seawater, were inoculated with 20 ml of C₂H₂-enriched seawater and analyzed in triplicate whenever measurements were conducted on seawater samples. Experimental and control treatments were incubated in deckboard incubators plumbed with flowing surface seawater and shaded using blue Plexiglas to 20% light intensity, approximately equivalent to a depth of 25 m in the water column. The typical incubation period was 3 to 4 h, although on a few occasions the incubation time was extended up to 8 h.

(iii) Extraction of dissolved C₂H₄ gas from incubated samples.

To measure the dissolved C₂H₄ concentrations, a headspace analysis method was used to extract the C₂H₄ from seawater (Fig. 1). Samples of seawater from the Wheaton bottle were withdrawn into a 50-ml gas-tight, glass syringe (Perfektum). Syringes were rinsed at least twice with sample seawater, prior to withdrawing the final bubble-free sample for analysis. Accurate and reproducible volumes of sample seawater, typically 30 ml, are achieved by housing the 50-ml syringe within a custom-built syringe actuator (Fig. 1). The headspace gas, typically 10 ml, is added to the syringe via the bypass outlet of the analyzer, which passes the carrier gas (ultra-high-purity air) through a combustor providing a C₂H₄-free (<10 parts per trillion) source of air. The water and gas phases in the syringe are then equilibrated at room temperature by vigorous shaking of the syringe for 3 min. The equilibrated headspace is subsequently injected into the gas analyzer inlet which incorporates a 0.2-μm-pore-size PTFE membrane hydrophobic filter (Acrodisc; Pall Life Sciences) to prevent the accidental injection of seawater. An internal 10-port switching valve is subsequently activated that injects 100 μl of the sample gas stream onto the chromatography columns described below. The total analytical time per sample ranges from 7 to 8 min.

Fig 1.

Schematic of the setup and analysis of dissolved ethylene (C₂H₄) concentrations as part of the AR assay.

(iv) Quantification of C₂H₄ concentrations.

C₂H₄ concentrations were measured using a reduced gas analyzer (RGA; Peak Laboratories). The RGA has previously been used to quantify other reduced gases such as hydrogen and carbon monoxide (33, 39, 40) and was modified for the quantification of C₂H₄ by installing two chromatographic columns—a precolumn packed with a Unibeads 60/80 mesh to trap water vapor and a 1/8-in. Unibeads 1S 60/80 analytical column to separate C₂H₂ and C₂H₄. Furthermore, the analytical time program of the RGA was altered to divert the gas flow from the column after C₂H₄ had eluted to ensure that the large pulse of C₂H₂ did not reach the detector. One of the main advantages of using the RGA compared to a GC-FID is the high sensitivity to C₂H₄ with a detection limit of 30 ppb. This high sensitivity is achieved by passing the sample gas stream over a bed of heated mercuric oxide and quantifying the resulting mercury vapor with a UV photodetector (Fig. 1). Calibration of the RGA was routinely conducted using serial dilutions of a 10.3 ± 0.1 ppm of C₂H₄ standard in N₂ (Scott-Marrin, Riverside, CA).

(v) Calculation of C₂H₄ concentrations.

The measured concentration of C₂H₄ in the equilibrated headspace was used to calculate the total dissolved C₂H₄ concentration ({C₂H₄}_w in ml of C₂H₄/ml of H₂O) remaining in the water after equilibration by

$${\left\{{\text{C}}_2{\text{H}}_4\right\}}_{\text{w}}={10}^{-6}\hspace{0.17em}\beta \hspace{0.17em}{m}_{\text{a}}\hspace{0.17em}p$$ (2)

where β (ml of C₂H₄/ml of H₂O/atm) represents the Bunsen solubility coefficient of C₂H₄ (3), m_a is the measured concentration of C₂H₄ in the equilibrated headspace (in parts per million by volume), and p is atmospheric pressure (atm) of dry air. The C₂H₄ concentration in the initial seawater ({C₂H₄}_aq in ml of C₂H₄/ml of H₂O) was calculated, assuming mass balance, as follows:

$${\left\{{\text{C}}_2{\text{H}}_4\right\}}_{\text{aq}}=\hspace{0.17em}\hspace{0.17em}\left({\left\{{\text{C}}_2{\text{H}}_4\right\}}_{\text{w}}\hspace{0.17em}{V}_{\text{w}}\hspace{0.17em}+\hspace{0.17em}{10}^{-6\hspace{0.17em}}{m}_{\text{a}}\hspace{0.17em}{V}_{\text{a}}\right)/V_{\text{w}}={10}^{-6}\hspace{0.17em}{m}_{\text{a}}\hspace{0.17em}\left(\beta p{V}_{\text{w}}+\hspace{0.17em}{V}_{\text{a}}\right)/V_{\text{w}}$$ (3)

where V_w is the water sample size (ml) and V_a is the volume of headspace air (ml), followed by the conversion of {C₂H₄}_aq to units in nM ([C₂H₄]_aq) as follows:

$${\left[{\text{C}}_2{\text{H}}_4\right]}_{\text{aq}}\hspace{0.17em}\hspace{0.17em}{10}^9\hspace{0.17em}\times \hspace{0.17em}p{\left\{{\text{C}}_2{\text{H}}_4\right\}}_{\text{aq}}/\left(RT\right)$$ (4)

where R is the gas constant (0.08206 atm liter mol⁻¹ K⁻¹) and T is temperature (K). After the C₂H₄ concentrations in sample and control treatments were measured at both the beginning and end of the incubation period, the rate of C₂H₄ production was calculated as follows:

$$({Cs}^{final}-{Cc}^{final})-({Cs}^{t=0}-{Cs}^{t=0})/\mathrm{\Delta }t$$ (5)

where C represents the concentration of C₂H₄ in the sample (s) and the control treatment (c), as calculated using equations 1 to 3, at the end (final) or beginning (t=0) of the incubation period.

¹⁵N₂ assimilation technique.

In conjunction with the AR assay, we also measured N₂ fixation using the ¹⁵N₂ assimilation technique. It was recently shown that the addition of ¹⁵N₂ gas as a bubble resulted in the underestimation of rates of N₂ fixation in cultures of Crocosphaera (28). Therefore, the present work compared rates of ¹⁵N₂ fixation when ¹⁵N₂ gas was added either as a bubble or dissolved in seawater samples collected from discrete depths in the upper 50 m of the water column at Stn ALOHA. We describe below the preparation of the ¹⁵N₂-enriched seawater and the sample collection, preparation, and analysis.

Preparation of ¹⁵N₂-enriched seawater.

During a first comparative test, the ¹⁵N₂-enriched seawater was prepared at sea by adding filtered (0.2-μm pore size) degassed (250 mbar for 30 min) seawater to a crimp-sealed glass vial and injecting an appropriate quantity (1 ml of ¹⁵N₂ gas per 100 ml of seawater) of ¹⁵N₂ gas through the septum. Vigorous shaking for ∼30 min aided the complete dissolution of the bubble, and the ¹⁵N₂-enriched seawater was then added to the seawater samples (10 ml of ¹⁵N₂-enriched seawater per liter of seawater sample). Subsequent to this preliminary comparison, an alternative procedure was adopted whereby the ¹⁵N₂-enriched seawater was prepared at the shore-based laboratory prior to the oceanographic cruise (Fig. 2). To prepare the ¹⁵N₂-enriched seawater, 3.5 liters of filtered (0.2-μm pore size) seawater was added to a 4-liter acid-washed vacuum flask (Fig. 2). Ambient air was removed from the analytical apparatus by purging the system with helium at 5 lb/in² (∼500 ml min⁻¹) for 15 min. The seawater was placed under vacuum (250 mbar) for 60 min and transferred to a 3-liter gas-tight PTFE bag (Welch Fluorocarbon), which was cooled in a 4°C water bath. Subsequently, 40 ml of ¹⁵N₂ gas (98 atom%; Sigma-Aldrich) was injected via the sampling port and the gas-tight bag was physically agitated until the gas bubble was completely dissolved into the seawater (∼10 min). Afterward, 70-ml borosilicate glass vials were filled from the gas-tight bag to two times overflowing using nylon tubing (4-mm outer diameter by 2-mm inner diameter; Legris) and immediately crimp sealed with no headspace using Teflon-lined septa. The glass vials containing ¹⁵N₂-enriched seawater were stored in the dark at 4°C until required for experimental purposes.

Fig 2.

Schematic of the experimental setup for the preparation of seawater enriched with ¹⁵N₂ gas. The experimental apparatus, including the filtered seawater, was purged with helium prior to vacuum degasification. The ¹⁵N₂ gas was added to the filtered seawater after it was transferred to a gas-tight bag and cooled in an ice bath, and the ¹⁵N₂-enriched seawater was stored in the dark at 4°C in 70-ml crimp-sealed vials.

Validation of ¹⁵N₂-enriched seawater.

One of the major considerations for any analytical method is the ease of its application and although the ¹⁵N₂-enriched seawater was prepared at sea on one occasion, it was found to be more time efficient to prepare the ¹⁵N₂-enriched seawater on land prior to its use at sea. We therefore investigated the feasibility of storing the ¹⁵N₂-enriched seawater in crimp-sealed 70-ml borosilicate glass vials over a 4-week period by analyzing the dissolved ¹⁵N₂ content at weekly intervals using a membrane inlet mass spectrometer (MIMS) (23). In brief, the MIMS provides rapid and accurate measurements of gas ratios by coupling semipermeable, microbore tubing with the inlet vacuum line of a quadrupole mass spectrometer. Reference measurements consisted of a 1-liter reservoir of filtered (0.2-μm pore size) surface seawater collected from Stn ALOHA. Instrument drift in the N₂/Ar ratio over the 4-week period was 0.6%. The analytical temperature for reference seawater and samples was kept constant at 25°C by immersing 1/16-in. stainless steel inlet tubing inside a water bath. The gases analyzed included those with a mass to charge ratio of 28, 30, and 32 (corresponding to ¹⁴N₂, ¹⁵N₂, and O₂), and they were detected sequentially using a repetitive cycle of 1.5 Hz. Replicate samples of prepared ¹⁵N₂-enriched seawater were analyzed for the loss of ¹⁵N₂ at weekly intervals over a 1-month period, comparing the ratio of mass 30 in ¹⁵N₂-enriched seawater/reference seawater.

Sample collection, inoculation, and incubation.

Seawater samples were collected using a CTD-rosette from depths of 5, 25, and 45 m and subsampled into acid-washed and seawater-rinsed 4.3-liter polycarbonate bottles. Once the bottles were completely filled, 50 ml of seawater was removed and replaced with 50 ml of ¹⁵N₂-enriched seawater from the 70-ml crimp-sealed vials using a 50-ml glass syringe, resulting in a final ¹⁵N₂ enrichment of 1.5 atom%. The syringe was attached to a 15-cm length of 1/8-in. PTFE tubing which enabled the ¹⁵N₂-enriched seawater to be added below the neck of the 4.3-liter polycarbonate bottle. Bottles were carefully closed using septum closure caps (Thermo Scientific) with no headspace and inverted 20 times. In addition, replicate seawater samples which had been filled and capped with no headspace were injected with 3 ml of ¹⁵N₂ gas (98 atom%; Sigma-Aldrich) using a gas-tight syringe (SGE Analytical Sciences) through the septum cap into the bottle. The 3-ml ¹⁵N₂ gas-injected sample bottles were gently shaken, and then all of the bottles were incubated using either (i) a free-floating in situ array at three depths consisting of 5, 25, and 45 m (as described in reference 9) or (ii) deckboard incubators plumbed with surface seawater and shaded to 50, 25, and 10% of full sunlight to represent 5, 25, and 45 m.

Sample analysis and calculation of ¹⁵N enrichment.

At the end of the incubation period, the entire content of the 4.3-liter polycarbonate bottles was filtered onto combusted (450°C for 5 h) 25-mm glass fiber filters that were subsequently placed on combusted foil pieces in polystyrene petri dishes and stored frozen at −20°C for transport. To quantify the ¹⁵N₂ enrichment of particulate material, samples were dried at 60°C for ∼24 h and pelleted before analysis with an elemental analyzer-isotope ratio mass spectrometer (EA-IRMA Carlo Erba NC2500 coupled to a Thermo-Finnigan Delta S) by the Stable Isotope Facility, University of Hawaii.

N₂ fixation rates were determined from the particulate N (PN) and δ¹⁵N-PN content of incubated samples, the ¹⁵N atom% enrichment, ambient dissolved N₂ concentration, and the δ¹⁵N-PN for control (unamended) samples. The δ¹⁵N-PN of control samples (described as the AP_PN-initial in equation 6 below) was provided by natural abundance PN concentration and δ¹⁵N-PN at a depth of 25 m in the water column at Stn ALOHA. Atmospheric N₂ gas is used as the reference standard (¹⁵N/¹⁴N = 0.0036765, δ¹⁵N = 0‰). After the isotope mass balance calculation, the rates of N₂ fixation rates were expressed as PN by including a factor of 2 (as described in reference 30):

$$V(t^{-1})=({AP}_{PN\mathrm{-}final}-{AP}_{PN\mathrm{-}initial})/({AP}_{N2}-{AP}_{PN\mathrm{-}initial})\cdot 1/\mathrm{\Delta }t\cdot {PN}_{final}/2$$ (6)

where AP is the atom% ¹⁵N of the sample (e.g., PN_final) or substrate (i.e., N₂) pool and Δt is the length of the sample incubation period.

RESULTS

AR assay samples and control treatments.

A typical AR measurement from near-surface (25 m) seawater collected from the oligotrophic NPSG (Fig. 3) highlights several important considerations when conducting the assay. First, at the beginning of the incubation period (time zero) dissolved C₂H₄ concentrations were always measurable, exceeding the analytical detection limit of 5 pmol liter⁻¹ by several orders of magnitude. In this particular example, at time zero the dissolved C₂H₄ concentrations measured 37 nmol of C₂H₄ liter⁻¹ (Fig. 3). At the end of the 4-h incubation period (time final), the blank/biological signal ratio decreased to 80%, when a significant difference was measured between the control and sample treatment (one-way analysis of variance [ANOVA], P = 0.036 and P < 0.05). The biological C₂H₄ production associated with this particular seawater sample, as calculated per equation 5, was 2.7 nmol of C₂H₄ liter⁻¹ h⁻¹. An additional important observation is the 12% increase in C₂H₄ concentrations in the control treatment during the 4 h of incubation period (Fig. 3). The nonbiological production of C₂H₄ was investigated further, and an increase was repeatedly observed, even when the filtered seawater controls were amended with mercuric chloride (HgCl₂; 200 μl of saturated HgCl₂ in 300 ml of seawater sample) and regardless of whether samples were incubated in the light or dark. The abiotic production of C₂H₄ during the AR assay has previously been reported, due to reactions with the serum cap liners often used to crimp-seal glass vials (38). The observation of increasing C₂H₄ concentrations in glass Wheaton bottles highlights the importance of quantifying the C₂H₄ concentration at time zero in both the samples and control treatments and also reporting the blank to biological signal ratio for the appropriate time points of the experimental period (as previously highlighted by reference 14).

Fig 3.

C₂H₄ concentrations in whole seawater (white bars) and control treatments (0.2-μm-pore-size filtered seawater) (gray bars) at time zero and after 4 h of incubation in deckboard incubators.

Following the analytical procedures described above, the AR assay was conducted on surface seawater samples collected at 25 m in the oligotrophic NPSG on multiple occasions between March and October 2011 (Table 1). Concentrations of C₂H₄ at the end of the incubation period ranged from 42 to 67 nmol liter⁻¹, with an overall analytical precision of ± 0.05 nmol liter⁻¹ and the blank/signal ratio at the end of the incubation period ranged from 67 to 82%. The corresponding rates of C₂H₄ production varied by an order of magnitude ranging from 0.46 nmol liter⁻¹ h⁻¹ (incubation period: 2100 to 0600, 19 September) to 4.7 nmol liter⁻¹ h⁻¹ (incubation period: 1000 to 1400, 20 July).

Table 1.

Measurements of AR and ¹⁵N₂ assimilation, determined using ¹⁵N₂-enriched seawater, for water column samples collected from 25 m at Stn ALOHA between July and September 2011^a

Date	Incubation period (hourly range)	Mean ± SD		C2H4/N2 ratio
		C2H4 production (nmol of C2H4 liter−1 h−1)	15N2 assimilation rate (nmol of N liter−1 h−1)
20 July 2011	Day (1000–1400)	4.7 ± 0.3	0.15 ± 0.01	18
	Night (2200–0500)	0.7 ± 0.2
30 Aug 2011	Day (1200–1500)	2.0 ± 0.1	0.22 ± 0.09	7
	Night (2200–0600)	1.0 ± 0.2
9 Sept 2011	Day (1030–1430)	2.1 ± 0.1	0.08 ± 0.02	27
	Night (2200–0600)	3.4 ± 0.1	0.22 ± 0.10	16
13 Sept 2011	Day (1000–1400)	0.9 ± 0.2	0.08 ± 0.02	12
	Night (2200–0600)	1.3 ± 0.4	0.10 ± 0.01	13
19 Sept 2011	Day (1100–1330)	4.0 ± 0.5	0.26 ± 0.05	16
	Night (2100–0600)	0.5 ± 0.2	0.04 ± 0.02	13

^aThe length of the sample incubation period is reported for the AR assay in parentheses, whereas it was either 12 or 24 h for ¹⁵N₂ assimilation. Both C₂H₄ production and ¹⁵N₂ assimilation are reported as hourly rates (n = 3).

Determining the quantity of C₂H₂ added to seawater sample.

A major consideration of the AR assay is the quantity of C₂H₂ added to the seawater samples. The addition of low C₂H₂ concentrations will insufficiently saturate nitrogenase (42); however, complete saturation of nitrogenase will inhibit N₂ fixation, resulting in N starvation and production of new nitrogenase (36). Balancing these considerations has resulted in C₂H₂ typically added in gaseous form to the headspace of a sample at 10 to 20% (vol/vol) (4). We measured the effect of C₂H₂ concentrations on the rate of C₂H₄ production in seawater samples by increasing the volume of C₂H₂ added from 2 to 10%, equivalent to C₂H₂ concentrations of 1.3 to 6.5 mM, respectively. Overall, a nearly 2-fold increase was observed in the rate of C₂H₄ production, from 2.7 nmol liter⁻¹ h⁻¹ at 2% (vol/vol) to 4.5 nmol liter⁻¹ h⁻¹ at 10% (vol/vol) (Fig. 4). All subsequent measurements of C₂H₄ production were conducted with a 10% addition of C₂H₂-saturated seawater.

Fig 4.

Rates of C₂H₄ production in oligotrophic, near-surface (25 m) seawater samples incubated for 3 h at different C₂H₂ concentrations.

Incubation period.

During both day and night sample incubation periods, the increase in C₂H₄ production rate was linear over an 8-h period (Fig. 5). The minimum length of the incubation period is dictated by the quantity of time required to measure a significant increase in the quantity of C₂H₄ produced relative to the control treatments. We advise restricting the AR assay incubation period to a few hours since the presence of C₂H₂ can adversely affect other metabolic pathways and the wider microbial community (32).

Fig 5.

Linear response of C₂H₄ concentrations measured over an 8-h incubation period in surface seawater incubations (open circles) compared to filtered seawater controls (gray triangles).

¹⁵N₂ assimilation.

In addition to conducting the field assessment of the ¹⁵N₂ tracer technique, the methods used to prepare and store the ¹⁵N₂-enriched seawater were also assessed. A mass-to-charge ratio of 30, corresponding to ¹⁵N₂, was measured in ¹⁵N₂-enriched seawater/reference seawater over a 4-week period to determine any loss of ¹⁵N₂ in the ¹⁵N₂-enriched seawater. Over the storage period, the coefficients of variation averaged 1.7%, and no decrease was detected within this range of variation (Fig. 6).

Fig 6.

Analysis of the storage capability of ¹⁵N₂-enriched seawater. The mass-to-charge (m/z) 28, 39, 30, and 32 peak areas as measured by membrane inlet mass spectrometry (MIMS) are shown for reference seawater (A) and samples of ¹⁵N₂-enriched seawater (B) with a ¹⁵N₂ atom% enrichment of 12 ml liter⁻¹. (C) m/z ratio corresponding to ¹⁵N₂ in ¹⁵N₂-enriched seawater/reference seawater over a 4-week period.

On five occasions between October 2010 and September 2011, comparisons of the ¹⁵N₂ assimilation methods were conducted, adding ¹⁵N₂ tracer to sampled seawater incubations either as a gas bubble or in dissolved form. On two occasions (20 July and 30 August), samples were incubated using an in situ floating array (Fig. 7). Rates of ¹⁵N₂ assimilation were significantly greater (one-way ANOVA, P = <0.001) when ¹⁵N₂ tracer was added to the seawater sample in dissolved form, averaging 4.6 ± 0.9 nmol of N liter⁻¹ day⁻¹ (±represents the standard deviation, n = 18) compared to an overall average of 2.7 ± 0.8 nmol of N liter⁻¹ day⁻¹ when ¹⁵N₂ tracer was added as a gas bubble. On two occasions (20 July and 29 September), samples were incubated in deckboard incubators (Fig. 8). Similar to the seawater samples incubated using the in situ array, the rates of ¹⁵N₂ assimilation were significantly greater (one-way ANOVA, P = <0.001) when ¹⁵N₂ tracer was added to the seawater sample in dissolved form (overall average of 5 ± 1.3 nmol of N liter⁻¹ day⁻¹) compared to the addition of a gas bubble (overall average of 1.9 ± 1.1 nmol of N liter⁻¹ day⁻¹).

Fig 7.

Depth profiles of N₂ fixation rates at Stn ALOHA incubated in situ for 24 h using a free-drifting array on two separate occasions: 20 July 2011 (A) and 31 August 2011 (B). The ¹⁵N₂ tracer was either added as a gas bubble (open bars) or in dissolved form (gray bars). Error bars where shown represent standard deviation (n = 3).

Fig 8.

N₂ fixation rates on surface seawater samples at Stn ALOHA incubated in deckboard incubators at 3 different light intensities for 24 h on two separate occasions: 20 July 2011 (A) and 27 September 2011 (B). The ¹⁵N₂ tracer was either added as a gas bubble (open bars) or in dissolved form (gray bars). Error bars represent the standard deviation (n = 3).

On 20 July 2011, replicate samples collected from 5, 25, and 45 m, were incubated simultaneously using the in situ array and deckboard incubators. When the ¹⁵N₂ tracer was added as a bubble, there was no significant difference (one-way ANOVA, P = 0.98) between samples incubated using the in situ array and deckboard incubator. In contrast, when the ¹⁵N₂ tracer was added in dissolved form, there was a weak significant difference (one-way ANOVA, P = 0.044 and P < 0.05) between the in situ array and deckboard incubator, with a 30% increase in the depth-integrated (0 to 45 m) ¹⁵N₂ assimilation rates in the deckboard-incubated samples (Table 2). We speculate that the difference between deckboard and in situ incubations was primarily due to a daytime increase (ca. 2 to 3°C) in seawater temperature in the deckboard incubators, which was measured at the start of the cruise, together with irradiance, but not continuously logged. We therefore recommend continuously monitoring deckboard incubator conditions with a temperature/light logger (e.g., HOBO Pendant, Onset, MA) or conducting in situ analyses.

Table 2.

Depth-integrated (0 to 45 m) rates of N₂ fixation as measured by ¹⁵N₂ assimilation by adding the ¹⁵N₂ gas in dissolved and bubble form^a

Date	Mode of incubation	15N2 assimilation rate (μmol of N m−2 day−1)
		Bubble	Dissolved
20 July 2011	Array	82	150
20 July 2011	Incubator	87	198
30 Aug 2011	Array	104	195
29 Sept 2011	Incubator	59	206

^aThe seawater samples were incubated either using an in situ array or in a shipboard incubator.

An additional experiment was performed to ensure the validity of comparing the addition of ¹⁵N₂ in gaseous or dissolved phase. It was considered whether the addition of 50 ml of exogenous seawater to a 4.3-liter polycarbonate bottle (which represents a 1.2% [vol/vol] addition) could have stimulated N₂ fixation. Therefore, on 7 November 2011, a third treatment was included in addition to the comparison of ¹⁵N₂ in gaseous or dissolved phase. The third treatment consisted of 50 ml of degassed, but non-¹⁵N₂-enriched seawater that was added to seawater samples prior to the injection of the ¹⁵N₂ gas bubble. No significant difference was observed in the rate of ¹⁵N assimilation when amended with a bubble (2.1 ± 0.6 nmol of N liter⁻¹ day⁻¹) compared to the bubble plus degassed, nonenriched seawater (1.3 ± 0.2 nmol of N liter⁻¹ day⁻¹) (one-way ANOVA, P = 0.096 and P > 0.05). In contrast, the rate of ¹⁵N assimilation in seawater samples inoculated with dissolved ¹⁵N₂ averaged 6.5 ± 0.4 nmol of N liter⁻¹ day⁻¹ and were significantly higher than the ¹⁵N₂ gas bubble injected treatments (one-way ANOVA, P = 0.003 and P < 0.05).

Comparison of AR assay and ¹⁵N₂ assimilation.

On eight separate occasions, the AR assay was conducted alongside rate measurements of ¹⁵N₂ assimilation conducted using ¹⁵N₂-enriched seawater (Table 1). The incubation period differed for the two analyses, ranging from 3 to 8 h for the AR assay and either 12 or 24 h for the ¹⁵N₂ tracer. Expression of C₂H₄ production and ¹⁵N₂ assimilation incubated with ¹⁵N₂-enriched seawater as an hourly rate measurement revealed a positive relationship (Pearson correlation, r = 0.83, P = 0.017) with the ratio of C₂H₄ production to ¹⁵N₂ assimilation ranging from 7 to 27.

DISCUSSION

Measuring N₂ fixation in the oligotrophic open ocean is essential for a comprehensive understanding of the marine N cycle and the role of diazotrophs in upper-water-column biogeochemistry. This study reports on the specifics of two conventional methods frequently used to measure N₂ fixation in the marine environment: the ¹⁵N₂ assimilation technique and the AR assay. A common theme in the currently applied protocols of the AR assay and the ¹⁵N₂ assimilation is the issue of gas equilibration. It has previously been demonstrated that without sufficient agitation of the sample, dissolved C₂H₄ resulting from the reduction of C₂H₂ does not equilibrate with the gas phase (14). Similarly, it was recently shown that the addition of ¹⁵N₂ tracer in gas form does not completely equilibrate with the liquid phase, even over a 24-h incubation period (28). The loss of signal makes data comparison between separate studies difficult and presents a serious drawback when estimating the contribution of diazotrophs to the fixed pool of N over global ocean basins. To avoid the addition of ¹⁵N₂ and C₂H₂ in gaseous form, we present here a modified protocol for the AR assay and describe the preparation and use of ¹⁵N₂-enriched seawater. Both methods were tested on several oceanographic cruises in the oligotrophic open ocean environment of the NPSG between October 2010 and September 2011, with particular attention paid to the time efficiency and practicality of the protocols.

With respect to the AR assay, we found that the RGA provides accurate and reproducible measurements of dissolved C₂H₄. Two inherent features of the RGA which aid the analysis of C₂H₄ are the redirection of the carrier gas flow after C₂H₄ has eluted from the column avoiding the C₂H₂ reaching the detector bed and the installation of a primary chromatographic column to handle the high water vapor content of the sample. Installation of a water-impermeable filter prior to the analyzer inlet also prevents the accidental injection of seawater. Use of the RGA at sea is facilitated by the ease of transport and setup, requiring a single compressed gas cylinder to supply the carrier gas flow. The only custom-built feature of the analytical setup was the hand-held syringe actuator (Fig. 1), which may also be commercially available. When conducting the AR assay in the field, the most consistent measurements were obtained when analyzing three to four samples with a sample incubation period of 4 h. It is recommended to avoid having samples waiting to be processed due to the continuing biological and potentially abiotic production of C₂H₄ in the sample (15).

As concluded on previous occasions (4, 14, 30), the AR assay represents a valuable tool for obtaining a rapid and sensitive measurement of C₂H₄, particularly for field samples that contain mixed community assemblages of diazotrophs. This work demonstrates that no preconcentration of biological material is required when dissolved C₂H₄ concentrations are quantified using the RGA, excluding the potential of harming the N₂-fixing microorganisms during the filtration process. Regardless of whether C₂H₄ is added in gaseous or dissolved phase to the sample, it is recommended to report the blank to signal ratio for all time points when conducting the AR assay. Previous studies have been able to minimize the background C₂H₄ (<0.2 ppm) by using “clean C₂H₂,” which refers to gas cylinders of compressed C₂H₂ (37). To the best of our understanding, these C₂H₂ gas cylinders are no longer commercially available. Finally, we recommend before commencing extensive field measurements, an initial set of C₂H₄ rate measurements are conducted at a range of C₂H₂ concentrations, e.g., 5, 10, and 20% (vol/vol), to demonstrate the response of nitrogenase enzyme within the ambient microbial community to various levels of C₂H₂.

An additional issue regarding the AR assay concerns the use of a stoichiometric conversion factor to obtain an estimate of N₂ fixation. Since the development of the AR assay, there has been considerable research on the possibility of a universal conversion factor enabling the rate of C₂H₄ production to be converted into a rate of N₂ fixation. A theoretical ratio of 3:1 is often cited, based on the difference between two hydrogen ions required to reduce C₂H₂ to C₂H₄ and six hydrogen ions needed to reduce N₂ to 2NH₃. A ratio of 4:1 has also been suggested since this incorporates the H₂ evolution associated with the reduction of a N₂ molecule (4). A wide range of C₂H₄/¹⁵N₂ assimilation ratios have previously been observed, ranging from 0.93 to 7.26 (6), 6.7 to 11.6 (19), 3.3 to 56 (26), and 1.9 to 9.3 (30). The reasons for the discrepancies between the theoretical and observed ratios have previously been discussed at length (14, 16, 19) and demonstrate that it is unwise to assume a fixed conversion factor. Prior to converting C₂H₄ production to a rate of N₂ fixation, multiple simultaneous AR assay and ¹⁵N₂ assimilation measurements should be conducted, as previously recommended (4, 15, 19, 30).

With respect to the ¹⁵N₂ assimilation technique, the results demonstrate that introducing the ¹⁵N₂ gas into the seawater sample, either as a gas bubble or dissolved in sterile seawater, causes significant differences in the quantity of ¹⁵N recovered in particulate form (Fig. 7 and 8). The issue of ¹⁵N₂ gas solubility was recently demonstrated for laboratory cultures of the diazotroph Crocosphaera watsonii (28), and we now extend the observation of an underestimation of ¹⁵N₂ assimilation rates when adding ¹⁵N₂ as a gas bubble to open ocean seawater samples composed of mixed diazotrophs assemblages. We demonstrate here that the addition of ¹⁵N₂ gas in dissolved form increases the recovery of ¹⁵N compared to the addition of a gas bubble. Our findings help to resolve an existing conundrum in the upper oceanic N cycle at Stn ALOHA, whereby indirect estimates of N₂ fixation over an 11-year period, from 1989 to 2001, derived from the isotopic composition of sinking material exceeded rates of ¹⁵N₂ assimilation measured on five occasions by 40 to 80% (11). This previous study concluded that their ¹⁵N₂ assimilation tracer measurements, as conducted by the addition of N₂ gas in a bubble, were underrepresenting N₂ fixation rates in the upper 0 to 100 m of the water column. A more recent study of N₂ fixation at Stn ALOHA over a 3-year period measured depth-integrated (0 to 100 m) rates of ¹⁵N₂ assimilation of 150 μmol of N m⁻² day⁻¹ in the summer months (July to September) (9), which is within the range of surface N₂ fixation (100 to 400 μmol of N m⁻² day⁻¹) reported by Deutsch et al. (10). The rates of ¹⁵N₂ assimilation measured herey by adding ¹⁵N₂ in dissolved form are 2- to 3.5-fold higher than the rates of ¹⁵N₂ assimilation when ¹⁵N₂ is added as a gas bubble. Therefore, the present study supports previous findings (28) that conducting ¹⁵N₂ assimilation incubations using ¹⁵N₂-enriched seawater will reduce existing imbalances in the fixed pool of N in the surface seawater of the oligotrophic ocean.