Measurements of nitrogen fixation in the oligotrophic North Pacific Subtropical Gyre using a free-drifting submersible incubation device
J. Plankton Res. (
Measurements of nitrogen fixation in the oligotrophic North Pacific Subtropical Gyre using a free-drifting submersible incubation device
DENIZ BOMBAR 0 1
CRAIG D. TAYLOR 0
SAMUEL T. WILSON 0 3
JULIE C. ROBIDART 0 1 2
KENDRA A. TURK-KUBO 0 1
JOHN N. KEMP
DAVID M. KARL 0 3
JONATHAN P. ZEHR 0 1
0 CENTER FOR MICROBIAL OCEANOGRAPHY
1 OCEAN SCIENCES DEPARTMENT
2 NATIONAL OCEANOGRAPHY CENTRE
3 DEPARTMENT OF OCEANOGRAPHY
One challenge in field-based marine microbial ecology is to achieve sufficient spatial resolution to obtain representative information about microbial distributions and biogeochemical processes. The challenges are exacerbated when conducting rate measurements of biological processes due to potential perturbations during sampling and incubation. Here we present the first application of a robotic microlaboratory, the 4 L-submersible incubation device (SID), for conducting in situ measurements of the rates of biological nitrogen (N2) fixation (BNF). The free-drifting autonomous instrument obtains samples from the water column that are incubated in situ after the addition of 15N2 tracer. After each of up to four consecutive incubation experiments, the 4-L sample is filtered and chemically preserved. Measured BNF rates from two deployments of the SID in the oligotrophic North Pacific ranged from 0.8 to 2.8 nmol N L21 day21, values comparable with simultaneous rate measurements obtained using traditional conductivity - temperature - depth (CTD) - rosette sampling followed by on-deck or in situ incubation. Future deployments of the SID will help to better resolve spatial variability of oceanic BNF, particularly in areas where recovery of seawater samples by CTD compromises their integrity, e.g. anoxic habitats.
nitrogen fixation; in situ device; Pacific Ocean; diazotroph
I N T RO D U C T I O N
Biological nitrogen (N2) fixation (BNF), the conversion of
N2 gas to ammonia (NH3), is performed by a select
group of microorganisms, termed diazotrophs. BNF is a
key component of the oceanic nitrogen cycle, with
estimates of up to 200 Tg nitrogen (N) being fixed per year on
a global scale (Karl et al., 2002; Capone et al., 2005;
Gruber and Galloway, 2008). However, such estimates
have large uncertainties, partly due to an incomplete
understanding of the full diversity and ecology of marine
diazotrophs (Goebel et al., 2010; Moisander et al., 2010;
Farnelid et al., 2011). In the ocean, the major groups of
diazotrophs include: (i) the filamentous, non-heterocystous
cyanobacterium Trichodesmium (Mague et al., 1977; Capone
et al., 1997; Capone et al., 2005; Laroche and Breitbarth,
2005), (ii) unicellular, free-living cyanobacteria such as
Crocosphaera watsonii (UCYN-B) (Hewson et al., 2009; Webb
et al., 2009; Bench et al., 2011; Foster et al., 2013) and (iii)
cyanobacteria that form symbioses with eukaryotic algae,
e.g. the heterocystous genera Richelia and Calothrix that are
associated with diatoms (Villareal, 1990; Janson et al.,
1995; Foster and Zehr, 2006) and unicellular Candidatus
Atelocyanobacterium thalassa (UCYN-A) associated with
prymnesiophytes (Zehr et al., 2001; Moisander et al., 2010;
Thompson et al., 2012; Bombar et al., 2014).
Although the major abiotic, nutrient and internal
controls of BNF activity and its distribution appear to
have been identified (Sohm et al., 2011; Voss et al., 2013),
the relative strengths with which these different factors
govern BNF under different environmental settings
remain elusive. These variable controls are likely
responsible for the unexplained large spatiotemporal variability
in the abundances of diazotrophs in the surface ocean,
recently highlighted by high-resolution sampling using a
drifting robotic gene sensor (Robidart et al., 2014). It is
currently unknown how these abundance fluctuations
affect the variability in BNF. Thus, conducting
corresponding rate measurements at similarly high
spatiotemporal resolution is critical for better understanding
the role of diazotrophs in oceanic N cycling.
BNF field measurements are typically conducted using
versions of the 15N2 tracer gas technique (Montoya et al.,
1996). This method requires the collection of seawater
using a conductivity – temperature – depth (CTD) – Niskin
rosette sampling system, and subsequent on-deck or in
situ incubations of tracer-amended seawater, lasting
anywhere from a few hours to a few days. Seawater samples
in traditional BNF studies experience changing pressures,
light levels and temperatures upon recovery and seawater
transfer, while characterization from cultured
representatives and sorted cells suggest that many of these processes
affect diazotroph populations (e.g. Thompson et al.,
2012). Further, methodological improvements on the
15N2 tracer technique have demonstrated that BNF rates
have been underestimated in most studies, with the
addition of 15N2 tracer in a dissolved form more
representative of the actual rates (Grosskopf et al., 2012).
Automated sampling devices capable of conducting
sampling and incubation in situ are a promising approach
for resolving variability associated with biogeochemical
cycling in the marine environment. Such devices have
been developed and successfully employed to increase
the spatial and temporal resolution of planktonic primary
production rates (Dandonneau and Bouteiller, 1992;
Taylor and Howes, 1994), as well as for comparing
manipulations performed in situ and under “simulated”
in situ conditions (Gundersen, 1973; Lohrenz et al., 1992).
The submersible incubation device (SID) was originally
designed for primary productivity measurements (Taylor
and Doherty, 1990) and has since been adapted for a
range of oceanographic measurements (Lohrenz et al.,
1992; Taylor and Howes, 1994; Albert et al., 1995;
Edgcomb et al., 2011, 2014; Pachiadaki et al., 2014).
Integrating the BNF method protocol with a large
capacity submersible in situ device (4L-submersible incubation
device, 4L-SID) became more feasible with recent
developments of the 15N2 assimilation technique, which
requires dissolving the 15N2 gaseous tracer in sterile
seawater prior to its addition to the samples (Mohr et al.,
2010; Grosskopf et al., 2012). In the present study, a
modified version of the 4L-SID was deployed, which is
capable of autonomously executing the entire sampling,
tracer amendment ( predissolved 15N2, 13C-bicarbonate),
incubation and filtration processes associated with BNF
and primary production measurements in situ. By
conducting the entire sampling and incubation procedure
directly in the water column, delays in the onset of the
incubations and perturbations of the microbial
community assemblages during sampling are minimized.
Further, such devices have the potential to help overcome
the major hurdle of achieving higher sampling
resolution, which could reveal currently unknown
heterogeneity in BNF rates and the key environmental factors that
M E T H O D
The BioLINCS cruise (Biosensing Lagrangian
Instrumentation and Nitrogen Cycling Systems) was
conducted in the North Pacific Subtropical Gyre (NPSG)
(24.39 –25.138N, 158.20 –157.298W) in September 2011,
aboard the R/V Kilo Moana (Fig. 1). The overall goal of
Fig. 1. Sampling stations north of Station ALOHA visited during the
BioLINCS cruise (gray dots), locations of the SID on the mornings of
all incubation starts (black crosses) and stations at which the in situ array
was deployed (black circles). The gray contour lines (mesoscale altimetry)
indicate the presence of two anticyclonic eddies that influenced the drift
paths of the SID (described in the text). Color-coded near-surface
chlorophyll concentrations are averages of satellite data from Archiving,
Validation and Interpretation of Satellite Oceanographic data (AVISO)
and Moderate Resolution Imaging Spectroradiometer (MODIS) Aqua,
for 6–20 September 2011.
the scientific cruise was to examine microbial
biogeochemical cycling associated with the nitrogen cycle and was an
ideal context for implementing the 4L-SID test. To
characterize the hydrographic and biogeochemical conditions of
the upper water column, vertical profiles were conducted
daily using a CTD system coupled to a rosette consisting of
24 12 L Niskin bottles. Oxygen (O2) and fluorescence
sensors were calibrated against discrete measurements of
dissolved O2 (Carritt and Carpenter, 1966) and chlorophyll
extracted and analyzed by fluorometry (Strickland and
Parsons, 1972). Seawater for determination of nutrient
concentrations was sampled and analyzed as documented
in the online manual for “HOT Laboratory Protocols”
.html). Regional ocean color and sea-level anomaly for
the NPSG were analyzed using satellite-derived images
from the Moderate Resolution Imaging Spectroradiometer
Operation of the submersible incubation
Since the original description of the SID in the 1990’s
(Taylor and Doherty, 1990), there have been several
subsequent versions of the SID concept which have adapted
the instrumentation (Lohrenz et al., 1992; Taylor and
Howes, 1994; Albert et al., 1995; Edgcomb et al., 2014;
Pachiadaki et al., 2014). This study is the first time that
the 4L-SID has been used for conducting 15N2 rate
measurements and therefore the entire instrument
configuration relevant to quantifying N2 fixation is outlined here.
The SID, as configured for this study, consists of a
hydraulically driven, syringe-like 4-L incubation chamber,
an 18-port fluidic distribution valve (FDV) for directing
fluid flows, an array of 8 “version 1” Fixation Filter units
(FF1s, Taylor et al., 2013) for collection and preservation
of incubated particulate samples and a controlling
electronics/battery pack (Fig. 2A). The incubation chamber
consists of a precision bore borosilicate glass chamber
[interior silane treated with SurfaSilTM siliconizing fluid
(Thermo Scientific) for biological inertness]. Each end of
the incubation chamber is capped with silicone
O-ring-sealed polycarbonate end caps and it contains a
silicone O-ring-sealed polycarbonate floating piston. The
rotor/stator components of the FDV in contact with
sample are made of PVC and Teflonw and interfacing
tubing between the incubation chamber, FDV and FF1s
are made of Teflonw. All interiors were acid-washed and
rinsed with deionized water prior to deployment.
Communication with the instrument prior to and after
each deployment for programming and data retrieval
was via a serial RS-232 link with a laptop PC. The
4L-SID was mounted to a free-drifting spar float system
(Fig. 2B) for deployments at a fixed depth of 25 m.
During the instrument operation in situ, the location of
the spar float system was constantly monitored via two
iridium GPS transponders. The FF1s are unique in-line
filter units that each contain an appropriate chemical
preservative that is delivered (with no moving parts) through
the filter by density-driven laminar convection after
completion of filtration (Supplementary data, Fig. S1).
The SID was configured for deployment with the
hydraulically driven floating piston flushed against the
check valve-containing end cap (Fig. 3A). The space
behind the floating piston was filled with deionized water.
After deployments in the afternoon, the incubation
sequence was programmed to automatically commence
the next morning at 0530. To condition the interior of
the incubation chamber with environmental sample,
500 mL of seawater from the depth of deployment was
drawn into the chamber via the inlet check valve (ICV,
red inset, Fig. 3B) and expelled back into the
environment via the FDV “Waste” outlet (Fig. 3A). A total of two
flushes were executed. The flushing operation was
immediately followed by complete filling of the incubation
chamber with sample via the ICV, advancement of the
FDV rotor to the first FF1 filter unit and immediate
filtration of the entire 4 L to obtain a natural abundance time
zero (T0) particulate sample. The ICV has a large
enough internal spacing that will not select against larger
organisms (21 mm diameter annulus with a 1.63-mm
spacing, through six 2.38-mm diameter holes, ultimately
into the chamber via a 4.76-mm diameter orifice; see
Fig. 3B inset). During filling, the ICV exerts low shear
stress of 1.2 pascals (Pa); max. 1.7 Pa, at a flow rate of
200 mL/min (Taylor et al., 2015). The FDV advanced
to the next valve port connected to the first bag of tracer
and the chamber then re-filled as described above. The
slight negative pressure that developed within the chamber
during filling also quantitatively draws the entire 15N2
contents from the flexible tracer bag, which also
quantitatively sweeps the 13C-bicarbonate contained within the
in-line injector coil into the chamber as shown in Fig. 3B
(tracer details described below). The gentle turbulence
generated from the main bulk of the sample entering the
chamber via the ICV completely mixes the tracer with
the sample as it enters the chamber (confirmed by dye
studies). The 4-L sample was then incubated for a
preprogrammed 23.5 h, followed by direction of sample to the
next FF1 to obtain the Tincub sample (Fig. 3C). Upon
completion of the incubation the chamber was flushed
4 as described above to remove tracer. The taper of
floating piston and front end cap were machined to the
same angle, minimizing the dead volume remaining when
the piston meets the front endcap. Assuming an interior
dead volume of 4 mL when the chamber is empty, the 4
flushing cycle dilutes the tracer contents by 5.6 orders of
magnitude, which is well below background
concentrations. A given incubation cycle consumes three ports of the
FDV and two FF1s (Fig. 3A). The 4L-SID, as configured,
was thus able to conduct four in situ incubations.
All filtrations were collected onto 47-mm diameter
precombusted glass fiber filters (GFF) and chemically
preserved in a pH 2 acid buffer inside the FF1s (Taylor and
Doherty, 1990; Taylor and Howes, 1994; Supplementary
data, Fig. S1), which terminates biological activity and
preserves the sample for at least 1.5 month in warm water
(confirmed by a Bermuda Test Bed mooring SID
deployment where data agreed well with Bermuda Atlantic
Time-Series measurements made at the same depth
[C. D. Taylor, unpublished data]). Once the SID was
recovered aboard ship, the filters were immediately
recovered and dried for 48 h at 608C in a drying oven and then
stored at room temperature until analyzed. In the
laboratory, the filters were pelleted and sent for isotopic analysis
at the University of California, Davis Stable Isotope
Facility. BNF rate calculations followed the protocol of
Montoya et al. (Montoya et al., 1996). To test for leakage
of low molecular weight (LMW) metabolites into the
acidic preservative, the remaining preservative contents
of the FF1s were also recovered, evaporated onto GFF
filters (soaking the GFFs with the preservative and
putting them in a drying oven), and these filters were
treated as described above for the particulate filters. We
found only very low or even undetectable amounts of
carbon and nitrogen on these filters, and more
importantly, the d13C and d15N values were equal or even lower
than those of the respective non-tracered T0 samples.
Thus, the SID-derived rates were not underestimated
due to loss of tracer to the LMW fraction.
The two tracers added to the SID incubations were
15N2 gas to obtain estimates of BNF and 13C-bicarbonate
for measurements of primary productivity. 15N2 gas (98
atom%; Sigma-Aldrich) was added to seawater samples
as “15N2-enriched seawater” which was prepared on land
prior to the cruise using sterile-filtered surface seawater
from Station ALOHA (10 mL 15N2 per liter of seawater;
Wilson et al., 2012). The 15N2 gas used in this study was
from a batch manufactured from 2008 to 2009 by
Sigma-Aldrich, and we identified it as not causing severe
contamination with other bioavailable inorganic N
species (Dabundo et al., 2014). After enrichment, the
tracer water was stored in 200-mL gas-tight tri-layer
aluminized polyethylene bags (http://www.pmcbag.com/).
The bags were individually connected to the 18-port
FDV via Luer locks and 1.6-mm I.D. Teflon tubing and
coiled in-line tracer loops made of Teflonw tubing, as
illustrated in Fig. 3A. A complete 200-mL bag was added to
each 4-L incubation, providing a final atom enrichment
of 5%. For the 13C additions, 400 mL of a 0.1-M solution
of H13CO32 were stored within a coiled section of the
Teflon tubing (see inset in Fig. 3A). To facilitate loading
of the 13C-tracer into the coil using a syringe, small
bubbles (volume 50 mL) were introduced at the
beginning and end of the injection. The leading bubble isolates
the tracer from the water contained within the tubing
leading to the FDV, allowing it to be introduced as a
“plug flow” instead of the spreading of tracer by the
parabolic laminar flows that would otherwise occur. The
trailing bubble provides isolation from the 15N2-enriched
seawater in the bag. The surface tension of the small
bubbles quite effectively confines the tracer within the
loop and resists modest vibration.
Complementary 15N2 measurements
conducted during the cruise
Measurements of BNF were also conducted during the
cruise by sampling the water column using the CTD–rosette
Fig. 3. Diagrammatic illustration of 4L-SID functions. (A) Deployed
configuration. FDV: 18 port fluidic distribution valve. The inset shows an
enlarged view of the 13C-bicarbonate stored in the tracer coil. (B)
Procurement of sample and introduction of tracer. The inset illustrates the
water flow through the inlet check valve (ICV). The poppet is normally
closed by a light duty, Teflonw coated spring, except for when sample is
drawn into the chamber. When the piston reaches the full extent of its travel
in either direction, it is “lugged down” and the reduction in pump rounds
per minute sensed by the electronics turns off the pump. (C) Delivery of
incubated sample through an FF1 filter holder. The inset illustrates the
seating of the Poppet O-ring against an annular “Knife Edge” (KE) to
prevent loss of sample through the check valve during emptying of the
and incubating the seawater samples either using an
in situ array or on-deck incubators which simulated in situ
conditions. The in situ incubations were used to obtain
vertical profiles of BNF. Seawater was collected from
depths of 5, 25, 45, 75, 100 and 125 m into replicate 4.3- L
polycarbonate bottles, amended with 15N2-enriched water
and attached to a free-floating in situ array at the
appropriate depth for a 24-h period (Church et al., 2009). The
on-deck incubations were performed for 24 h using blue
shaded incubators cooled with running surface seawater
and additional neutral mesh shading to mimic the
corresponding light irradiances for each depth. Both sets
of BNF measurements were also amended with 400 mL
of a 0.1-M solution of H13CO32 injected through the
septum cap with a syringe. Upon termination of the
incubations, the seawater samples were gently filtered
through precombusted Whatman GFF filters (0.7-mm
nominal pore size) and processed as described for the
Different diazotrophs present in the water column were
quantified using quantitative PCR (qPCR) enumeration
of specific nifH gene copies. Water column samples were
collected from between 5 and 175 m depth. Once the
CTD was recovered, the seawater was immediately
drained from the Niskin bottles into acid-washed 4-L
polycarbonate bottles. Using peristaltic pumps, 2 L from
each depth was filtered in-line through 10 mm (Polyester,
Sterlitech, Kent, WA, USA) and 0.2 mm (Supor; Pall Life
Sciences, Ann Arbor, MI, USA) pore-size filters, held in
25-mm-diameter Swinnex filter holders (Millipore,
Billerca, MA, USA). The filters were placed into sterile
1.5-mL cryotubes containing 0.1 g autoclaved glass
beads, frozen in liquid nitrogen and stored at 2808C
until processing in the laboratory. DNA extractions were
carried out as described previously (Bombar et al., 2013).
We used previously designed Taq-Manw primer-probe
sets, including cyanobacterial phylotypes UCYN-A and
UCYN-B (Moisander et al., 2010), Trichodesmium (Church
et al., 2005) and two Diatom–Diazotroph Associations
(DDAs) (Foster et al., 2007) termed het-1 (Rhizosolenia-Richelia)
and het-2 (Hemiaulus-Richelia). Additionally, we quantified
presumed heterotroph diazotroph phylotypes HM210397
(g-Proteobacteria) and KC013231 (Cluster 3), described in
Bombar et al. (Bombar et al., 2013). qPCR optimizations
and quantifications have been described in detail in
Moisander et al. (Moisander et al., 2010), Halm et al. (Halm
et al., 2012) (specifications for phylotype HM210397) and
Bombar et al. (Bombar et al. 2013) (specifications for
R E S U LT S A N D D I S C U S S I O N
The application of in situ devices for observing physical,
chemical and biological parameters in the ocean is an
important approach for better understanding the complex
relationships between the physical and chemical
environment and microbial distributions and activities (Taylor and
Howes, 1994; Johnson et al., 2010; Ottesen et al., 2011;
Robidart et al., 2014). This study reports the first successful
deployment of an autonomous device capable of
conducting sampling, incubation and filtration processes for BNF
measurements in situ. We evaluate the operation of the SID
as an in situ instrument for BNF measurement, compare
the SID-derived BNF rates with commonly applied
sampling and incubation methods and discuss the SID rates in
the context of physical, chemical and microbial data
obtained during the BioLINCS cruise.
SID operation and measurements of BNF
The SID was deployed twice during the 2011 BioLINCS
cruise (Fig. 1). During the first deployment (4L-SID
Deployment 1) from 9 to 14 September 2011, it followed
a 47-km drift path in a north-easterly direction, and
during the second deployment (4L-SID Deployment 2)
from 16 to 20 September 2011, it drifted 28 km westwards
(Fig. 1). During both deployments, the 4L-SID performed
four autonomous tracer incubations over a 4-day period at
a depth of 25 m in the water column. With respect to the
proximity of each incubation event, the 4L-SID sampled
approximately every 16 km during Deployment 1 and
every 10 km during Deployment 2.
The 4L-SID-derived BNF measurements ranged from
0.8 to 1.9 nmol N L21 day21 (average 1.4 + 0.5 nmol N
L21 day21) during Deployment 1 and from 1.4 to 2.8 nmol
N L21 day21 (average 2.0 + 0.6 nmol N L21 day21)
during Deployment 2 (Table I). Thus, during each 4-day
deployment, a 2-fold variation of BNF was recorded, and
BNF rates were overall higher during Deployment 2. The
simultaneous rate measurements of 13C primary production
by the 4L-SID also revealed higher values for Deployment
2 (344 + 40 nmol C L21 day21) compared with
Deployment 1 (207 + 48 nmol C L21 day21). The higher
rates of both 13C primary production and BNF suggest that
the difference in BNF rates between the two deployments
was not due to methodological errors that were only specific
to the BNF measurements (Table I, Fig. 4).
Compared with 4L-SID-derived rates of 15N2
assimilation, BNF rates obtained from incubations in the on-deck
incubator (with water from 25 m) were higher during
Deployment 1 (3.3 + 0.2 nmol N L21 day21) and more
variable (ranging from 1.4 + 0.1 to 6.6 + 1.9 nmol N
day) in 2011
September 12–13 September 13–14
N2 fixation production
(nmol N L21 (nmol C L21
September 10–11 September 16–17
Parallel to the first incubation of each deployment, a comparison incubation in the shipboard incubator was carried out, using water sampled from 25 m at
stations close to the SID location. These comparisons, even though they are not perfect control measurements taken in closest proximity to the SID,
suggest that the SID-derived rates are of realistic magnitude.
Fig. 4. (A) BNF and (B) primary production rates obtained from both SID deployments in the context of vertical rate profiles obtained from in situ
array deployments (n ¼ 2). Incubations at Station 11 are shown as well, but were done in the shipboard incubator due to the loss of the in situ array.
No primary production was measured on Station 11.
L21 day21) during Deployment 2 (Table I). There are
several factors which might contribute to the difference
in 4L-SID and on-deck measurements of 15N2
assimilation. The first is the potential for sampling different
populations of diazotrophs by the SID and the shipboard
CTD – rosette. The distance between the ship and the
SID when the comparative samples were taken was 5.4
and 3.9 km for 4L-SID Deployment 1 and Deployment
2, respectively (Fig. 1). Another potential source of
variability is the method of incubation itself. The on-deck
incubators are a best-effort to mimic temperature and
light levels equivalent to a depth of 25 m in the water
column, but have a few limitations. For example,
nearsurface seawater intake is used for cooling which has a
slightly higher temperature than 25 m water, and the
existence of variability in light intensity due to shading
among incubation bottles or from ship structures.
Potential perturbations of the natural abiotic
conditions during sampling and incubations are a
wellknown problem (Feike et al., 2012), and highlight the
necessity for using in situ devices especially when longer
incubations are required. This theory seems to be
supported by the BNF measurements obtained using an in
situ array which performed incubations at 25 m, the
results of which are better aligned with the 4L-SID
Deployment 1 values (Table I, 0.9–2.9 nmol N L21
day21). Unfortunately, corresponding in situ array
incubations to match Deployment 2 were not obtained due to
the loss of the array at sea. In contrast, the available rates
of 13C primary production (Table I) do not mirror the
high similarity of BNF observed between the SID and the
in situ array measurements, with higher rates obtained by
the in situ array compared with the SID. However, overall
both the 4L-SID-derived BNF and primary production
rates compare favorably with values obtained using
traditional methods during the same oceanographic
expedition and during time-series measurements conducted at
nearby Station ALOHA. BNF and 14C primary
production (in situ array incubations) rate measurements are
conducted on a nearly monthly basis at Station ALOHA,
situated 100 km to the south of the BioLINCS
expedition region [Hawaii Ocean Time-series (HOT) and
other funding programs (Grabowski et al., 2008; Church
et al., 2009)]. BNF rates at 25 m depth ranged from 1 to
5 nmol N L21 day21 during the summer months (July –
September) between 2004 and 2007 (Church et al., 2009,
15N2 bubble addition protocol). During a more recent
cruise in June 2014, BNF rates were measured daily for 7
days and produced values of 5.9 + 1.1 nmol N L21
day21 (range 3.7 – 7.2 nmol N L21 day21) (S. T. Wilson,
unpublished data, 15N2 added as predissolved in
sterilefiltered seawater). The average rate of 14C primary
production for the month of September during the years
1989 – 2011 is 463 + 126 nmol C L21 day21, which is
quite similar to values obtained using different incubation
techniques during our cruise (Table I). We can only
speculate on the reasons for the comparably high primary
production from in situ array 2 (11 September 2011, 594 +
68 nmol C L21 day21) or the relatively low values from
SID deployment 1 (207 + 48 nmol C L21 day21).
Interestingly, CTD chlorophyll fluorescence measurements
were relatively high at 25 m during sampling for in situ
array 2 (0.20 + 0.00 mg L21) compared with any
measurements at 25 m at four stations in proximity to 4L-SID
samplings during deployment 1 (0.13 + 0.04 mg L21). Given
the high spatiotemporal variability in prokaryotic
distributions and abiotic conditions for our particular expedition
(Fig. 5, also see Robidart et al., 2014), the observed range in
primary production appears realistic. Without more
specific comparison experiments, at this point it is difficult to
claim that one or the other method delivers more
trustworthy rate measurements. Most importantly, the overall
similar range of BNF and primary production values
obtained using the different methods supports the efficacy
of the SID for in situ rate measurements.
Hydrographic and biogeochemical
4L-SID Deployment 1 was on the northern edge of an
anticyclonic eddy (Fig. 1). During the second
deployment, the SID drifted west between the primary eddy
and a smaller adjoined anticyclonic eddy (Fig. 1). The
drift paths for both 4L-SID deployments followed the
clockwise circulation patterns of the two anticyclonic
eddies revealed by sea surface altimetry (Fig. 1) and
shipboard acoustic doppler current profiler measurements
(Robidart et al., 2014). These mesoscale eddies introduced
small-scale physical and chemical heterogeneity in the
area, which clearly affected microbial distributions,
especially of diazotroph cyanobacteria (Robidart et al., 2014).
Several hydrographic and biogeochemical conditions
may have influenced the 4L-SID measurements. Low
average wind speeds of 4.8 m s21 for Deployment 1 and
6.9 m s21 for Deployment 2 contributed to a shallow mixed
layer depth (MLD) during both deployments, but with an
average of 17 m (range 10 – 27 m) during Deployment 1
and 30 m (range 14 – 57 m) for Deployment 2 (based on
0.03 density offset from 10 m criterion) (De Boyer
Monte´gut et al., 2004). Accordingly, along the 25-m
depth horizon seawater temperatures averaged 25.98C
(range 25.7 – 26.18C) during Deployment 1 and 26.28C
(range 26.1 – 26.28C) during Deployment 2 (Fig. 5).
Therefore, possibly the majority of sampling conducted
by the 4L-SID during Deployment 1 was beneath the
mixed layer, and thus below the main accumulation of
diazotrophs, which could partly explain why lower rates
were found compared with Deployment 2. However,
with the available qPCR data (samples from 5, 25, 45 m
etc.) we cannot resolve whether the MLD had an
influence on the vertical distribution of diazotrophs that
would explain the variations in BNF.
Nutrient concentrations in the upper 100 m were mostly
low, which is common in oligotrophic oceanic gyres
(Fig. 5). However, some of the NO22 þ NO32 (low-level
nitrogen, LLN) and phosphate concentrations near the
surface equaled concentrations much deeper in the water
column at around 125 m (Fig. 5D, near the “apex”), which
is atypical for NPSG waters (Robidart et al., 2014). Along
the 25-m-depth horizon where the 4L-SID was situated,
LLN concentrations ranged from 2 – 6 nmol L21 and
phosphorus concentrations ranged from 20 to 137 nmol
L21 for the complete 12-day oceanographic expedition
(Fig. 5). During Deployment 1, the 4L-SID encountered
remarkably steep gradients in salinity and phosphate
concentrations (Fig. 5B and D). In turn, waters sampled
during Deployment 2 were relatively rich in chlorophyll
(Fig. 1). Overall, these data suggest that different water
types were sampled during 4L-SID Deployments 1 and 2.
The highest BNF and primary production rates were
measured nearest to the “apex” of the cruise transit
(Fig. 5), where nutrient concentrations were elevated.
Possibly, diazotrophs in this region were stimulated by the
nutrients and were able to respond with higher BNF,
although it is not clear whether nutrient concentrations
were elevated due to influx from depth, atmospheric
deposition (Kim et al., 2014), or whether there was low
demand in the microbial community which led to its
accumulation within the surface layer. The 4L-SID drifts in
a Lagrangian manner and can provide unbiased samples
describing the variability in BNF within a complex
setting of small spatiotemporal fluctuations in abiotic
parameters. In order to pinpoint specific abiotic/biotic
parameters responsible for observed variations in BNF
rates, future versions of the SID need to include
additional oceanographic sensors, like a CTD package including
O2, nitrate and optical sensors (see next section).
The qPCR quantifications of nifH gene copies suggest
that unicellular cyanobacterial diazotrophs (Candidatus
Atelocyanobacterium thalassa, “UCYN-A,” and Crocosphaera
watsonii, “UCYN-B”) were the most abundant diazotrophs
present in the water column throughout the sampled area
(Fig. 6A). In the region near the apex, nifH genes of these
organisms attained concentrations of 2.0 108 copies
m22, which was 96% of all quantified nifH gene copies.
Trichodesmium sp. was the next most abundant diazotroph
(up to 3.0 107 nifH gene copies m22); heterocystous
symbionts of diatoms (het-1 and het-2) as well as
heterotrophic bacteria were present at much lower abundances
(Fig. 6B). While nifH gene copy inventories appeared to
co-vary for unicellular cyanobacterial diazotrophs (UCYN-A
and UCYN-B, Fig. 6A), a different pattern was observed for
the remaining five phylotypes, i.e. Trichodesmium,
heterocystous cyanobacterial symbionts het-1 and het-2, and the two
heterotroph diazotrophs (Fig. 6B). The nifH gene
abundances in this latter group also co-varied, but appeared to
be relatively more abundant at stations parallel to 4L-SID
deployment 1 (Fig. 6B). This group includes the
diazotrophs typically assigned to the .10-mm size fraction
(Trichodesmium and heterocystous cyanobacteria), and was
up to 42% of nifH inventories in the “Transit 1” area
(Fig. 6C). These data suggest that the eddy-induced
advection and mixing in the area had clear effects on the
distribution of different diazotrophs. While such data
cannot be used to infer which diazotrophs were
responsible for the measured BNF rates, it is noteworthy that
the overall higher rates obtained during 4L-SID
deployment 2 coincided with a generally lower abundance of
diazotrophs in the .10-mm size fraction.
Recommendations for future SID
In its current configuration, the SID was successfully
deployed and recovered on two occasions, providing
daily BNF and primary production measurements in the
surface waters of the oligotrophic open ocean. The SID
concept can contribute more environmentally relevant
rates of BNF to inform global flux calculations, and this
technology has been validated in this study and others
(Taylor and Howes, 1994; Pachiadaki et al., 2014;
Edgcomb et al., 2014) for expeditions in various marine
provinces. The SID could be especially helpful for studies
in “delicate” habitats (e.g. anoxic habitats), where the
seawater samples are severely compromised when they are
brought onboard the research vessel for on-deck
incubations (Feike et al., 2012; Edgcomb et al., 2014).
Fig. 6. Depth-integrated inventories of nifH gene copies of diazotroph
microorganisms (quantified by qPCR) on ship stations in close
proximity to the SID drift paths. Section distance on the x-axis is the
total distance covered by shipboard stations along the northeast and
following westward transit. (A) nifH inventories of UCYN-A (Candidatus
Atelocyanobacterium thalassa) and UCYN-B (Crocosphaera watsonii). (B)
nifH inventories of Trichodesmium sp., het-1 and het-2 [Diatom –
Diazotroph associations between the diazotroph cyanobacteria Richelia
intracellularis and two different diatom species, Rhizosolenia-Richelia (RR)
and Hemiaulus-Richelia (HR)] and heterotroph phylotypes HM210397
(Halm et al., 2012) and KC013231 (Bombar et al., 2013). (C) Percentage
of the total nifH gene copies m22 of diazotrophs that are usually found
in the .10 mm size fraction (Trichodesmium, het-1, het-2). The triangles
at the top axis indicate where the SID took samples along the section
distance during its two consecutive deployments (four samplings per
The current SID technology would be improved by
conducting simultaneous replicate measurements. Furthermore,
an increased number of FF1s would permit more samples
to be processed and longer deployment periods. The
ability to also collect replicate samples for metagenomic
and metatranscriptomic analysis will enable investigators
to link biological function with the identity and activity of
the prokaryotic key players present in the water column
at the exact time of sampling (Robidart et al., 2014). To
this end, a new SID-implemented fixation filter (FF3,
Taylor et al., 2015) capable of chemically preserving
particulate microbial samples in a manner compatible with
subsequent metagenomic and metatranscriptomic study
(Edgcomb et al., 2014) has just been developed. Finally,
the addition of further oceanographic sensors, like an
ISUS for NO32 measurements (Johnson and Coletti,
2002), and the ability for adaptive sampling in response
to thresholds in environmental parameters, would allow
the SID to sample along environmental gradients. A
relatively new version microbial sampling-SID (MS-SID,
Edgcomb et al., 2014; Pachiadaki et al., 2014; with a host
of sensors [CTD, turbidity sensors, oxygen optode]), that
was not available during this study, is now in hand and
possesses the ability of collecting/in situ chemically
preserving up to 48 incubated samples and/or larger
volume microbial samples as well as 24 samples in
gastight bags. At present, a dual incubation chamber SID is
in advanced development (Vent-SID, laboratories of
C. Taylor and S. Sievert, WHOI).
C O N C L U S I O N S
The development of a device for conducting the entire
sampling, incubation and filtration processes of 15N2 rate
measurements in situ could help in the future to obtain
higher resolution coverage of direct estimates of oceanic
BNF. The two 4-day deployments conducted during the
September 2011 BioLINCS cruise were successful with
respect to instrument operation and obtaining BNF rates
comparable with those achieved by traditional CTD –
rosette sampling and incubation. Overall, the SID offers
increased sampling resolution of BNF measurements and
a platform for conducting in situ sampling of oceanic
water columns where on-deck incubations are not
feasible. This device will help in identifying driving factors
of BNF in situ, and could be used to test important
hypotheses about the regulation of BNF within the oceanic
S U P P L E M E N TA RY D ATA
Supplementary data can be found online at http://plankt.
AC K N O W L E D G E M E N T S
We thank M. Hogan and T. Cote for help in cruise
organization, the personnel onboard R/V Kilo Moana
for technical assistance and John Ryan for help with Fig. 1.
F U N D I N G
This study was supported by Gordon and Betty Moore
Foundation (GBMF) Marine Investigator Awards (J.P.Z.
and D.M.K.), the MEGAMER facility (supported by
GBMF), the grant from the NSF Emerging Frontiers
Program (Center for Microbial Oceanography: Research
and Education, grant DBI-0424599) and grants NSF
OCE-1061774 to V. Edgcomb and C.T.
R E F E R E N C E S
and consequences for nitrogen fixation measurements. PLoS ONE, 9,
analysis of the globally important keystone
Crocosphaera watsonii. ISME J., 3, 618 – 631.
Dandonneau, Y. and Bouteiller, A. L. (1992) A simple and rapid
device for measuring planktonic primary production by in situ
sampling, and 14C injection and incubation. Deep Sea Res. Part I, 39,
795 – 803.
Albert , D. B. , Taylor , C. and Martens , C. S. ( 1995 ) Sulfate reduction rates and low molecular weight fatty acid concentrations in the water column and surficial sediments of the Black Sea . Deep Sea Res. Part I , 42 , 1239 - 1260 .
Bench , S. R. , Ilikchyan , I. N. , Tripp , H. J. and Zehr , J. P. ( 2011 ) Two strains of Crocosphaera watsonii with highly conserved genomes are distinguished by strain-specific features . Front. Microbiol ., 2 , 261 .
Bombar , D. , Heller , P. , Sanchez-Baracaldo , P. , Carter , B. J. and Zehr , J. P. ( 2014 ) Comparative genomics reveals surprising divergence of two closely related strains of uncultivated UCYN-A cyanobacteria . ISME J ., 8 , 2530 - 2542 .
Bombar , D. , Turk-Kubo , K. A. , Robidart , J. , Carter , B. J. and Zehr , J. P. ( 2013 ) Non-cyanobacterial nifH phylotypes in the North Pacific subtropical gyre detected by flow-cytometry cell sorting . Environ. Microbiol . Rep., 5 , 705 - 715 .
Capone , D. G. , Burns , J. A. , Montoya , J. P. , Subramaniam , A. , Mahaffey , C. , Gunderson , T. , Michaels , A. F. and Carpenter , E. J. ( 2005 ) Nitrogen fixation by Trichodesmium spp.: an important source of new nitrogen to the tropical and subtropical North Atlantic Ocean . Global Biogeochem. Cycles , 19 , GB2024: 1 - 17 .
Capone , D. G. , Zehr , J. P. , Paerl , H. W. , Bergman , B. and Carpenter , E. J. ( 1997 ) Trichodesmium, a globally significant marine cyanobacterium . Science , 276 , 1221 - 1229 .
Carritt , D. E. and Carpenter , J. H. ( 1966 ) Comparison and evaluation of currently employed modifications of the Winkler method for determining dissolved oxygen in seawater; a NASCO report . J. Mar. Res ., 24 , 286 - 318 .
Church , M. J. , Jenkins , B. D. , Karl , D. M. and Zehr , J. P. ( 2005 ) Vertical distributions of nitrogen-fixing phylotypes at Stn ALOHA in the oligotrophic North Pacific Ocean . Aquat. Microb. Ecol., 38 , 3 - 14 .
Church , M. J. , Mahaffey , C. , Letelier , R. M. , Lukas , R. , Zehr , J. P. and Karl , D. M. ( 2009 ) Physical forcing of nitrogen fixation and diazotroph community structure in the North Pacific subtropical gyre . Global Biogeochem. Cycles , 23 , GB2020.
Dabundo , R. , Lehmann , M. F. , Treibergs , L. , Tobias , C. R. , Altabet , M. A. , Moisander , P. H. and Granger , J. ( 2014 ) The contamination of commercial 15N2 gas stocks with 15N-labeled nitrate and ammonium De Boyer Monte´gut , C., Madec , G. , Fischer , A. S. , Lazar , A. and Iudicone , D. ( 2004 ) Mixed layer depth over the global ocean: an examination of profile data and a profile-based climatology . J. Geophys. Res. Oceans , 109 , C12003.
Edgcomb , V. , Orsi , W. , Taylor , G. T. , Vdacny , P. , Taylor , C. , Suarez , P. and Epstein , S. ( 2011 ) Accessing marine protists from the anoxic Cariaco Basin . ISME J ., 5 , 1237 - 1241 .
Edgcomb , V. P. , Taylor , C. , Pachiadaki , M. G. , Honjo , S. , Engstrom, I. and Yakimov , M. ( 2014 ) Comparison of Niskin vs. in situ approaches for analysis of gene expression in deep Mediterranean Sea water samples . Deep Sea Res. Part II . doi: 10.1016/j.dsr2.2014.10.020
Farnelid , H. , Andersson , A. F. , Bertilsson , S. , Al-Soud , W. A. , Hansen , L. H. , Sorensen , S. , Steward , G. F. , Hagstrom , A. et al. ( 2011 ) Nitrogenase gene amplicons from global marine surface waters are dominated by genes of non-cyanobacteria . PLoS One , 6 , e19223 .
Feike , J. , Juergens , K. , Hollibaugh , J. T. , Krueger , S. , Jost , G. and Labrenz , M. ( 2012 ) Measuring unbiased metatranscriptomics in suboxic waters of the central Baltic Sea using a new in situ fixation system . ISME J ., 6 , 461 - 470 .
Foster , R. A. , Subramaniam , A. , Mahaffey , C. , Carpenter , E. J. , Capone , D. G. and Zehr , J. P. ( 2007 ) Influence of the Amazon River plume on distributions of free-living and symbiotic cyanobacteria in the western tropical north Atlantic Ocean . Limnol. Oceanogr., 52 , 517 - 532 .
Foster , R. A. , Sztejrenszus , S. and Kuypers , M. M. M. ( 2013 ) Measuring carbon and N2 fixation in field populations of colonial and free-living unicellular cyanobacteria using nanometer-scale secondary ion mass spectrometry1 . J. Phycol., 49 , 502 - 516 .
Foster , R. A. and Zehr , J. P. ( 2006 ) Characterization of diatomcyanobacteria symbioses on the basis of nifH, hetR, and 16S rRNA sequences . Environ. Microbiol ., 8 , 1913 - 1925 .
Goebel , N. L. , Turk , K. A. , Achilles , K. M. , Paerl , R. W. , Hewson , I. , Morrison , A. E. , Montoya , J. P. , Edwards , C. A. et al. ( 2010 ) Abundance and distribution of major groups of diazotrophic cyanobacteria and their potential contribution to N2 fixation in the tropical Atlantic Ocean . Environ. Microbiol., 12 , 3272 - 3289 .
Grabowski , M. N. W. , Church , M. J. and Karl , D. M. ( 2008 ) Nitrogen fixation rates and controls at Stn ALOHA . Aquat . Microb. Ecol.., 52 , 175 - 183 .
Grosskopf , T. , Mohr , W. , Baustian , T. , Schunck , H. , Gill , D. , Kuypers , M. M. , Lavik , G. , Schmitz , R. A. et al. ( 2012 ) Doubling of marine dinitrogenfixation rates based on direct measurements . Nature , 488 , 361 - 364 .
Gruber , N. and Galloway , J. N. ( 2008 ) An earth-system perspective of the global nitrogen cycle . Nature , 451 , 293 - 296 .
Gundersen , K. ( 1973 ) In situ determination of primary production by means of a new incubator ISIS . Helgol. Wiss. Meeresunters., 24 , 465 - 475 .
Halm , H. , Lam , P. , Ferdelman , T. G. , Lavik , G. , Dittmar , T. , Laroche , J., D'hondt, S. and Kuypers , M. M. M. ( 2012 ) Heterotrophic organisms dominate nitrogen fixation in the South Pacific Gyre . ISME J ., 6 , 1238 - 1249 .
Hewson , I. , Poretsky , R. S. , Beinart , R. A. , White , A. E. , Shi , T. , Bench , S. R. , Moisander , P. H. , Paerl , R. W. et al. ( 2009 ) In situ transcriptomic Janson , S. , Rai , A. N. and Bergman , B. ( 1995 ) Intracellular cyanobiont richelia-intracellularis-ultrastructure and immuno-localization of phycoerythrin, nitrogenase, Rubisco and glutamine-synthetase . Mar. Biol ., 124 , 1 - 8 .
Johnson , K. S. and Coletti , L. J. ( 2002 ) In situ ultraviolet spectrophotometry for high resolution and long-term monitoring of nitrate, bromide and bisulfide in the ocean . Deep Sea Res. Part I , 49 , 1291 - 1305 .
Johnson , K. S. , Riser , S. C. and Karl , D. M. ( 2010 ) Nitrate supply from deep to near-surface waters of the North Pacific subtropical gyre . Nature , 465 , 1062 - 1065 .
Karl , D. , Michaels , A. , Bergman , B. , Capone , D. , Carpenter , E. , Letelier , R. , Lipschultz , F. , Paerl , H. et al. ( 2002 ) Dinitrogen fixation in the world's oceans . Biogeochemistry , 57 /58, 47 - 98 .
Kim , I.-N. , Lee , K. , Gruber , N. , Karl , D. M. , Bullister , J. L. , Yang , S. and Kim , T.-W. ( 2014 ) Increasing anthropogenic nitrogen in the North Pacific Ocean . Science , 346 , 1102 - 1106 .
Laroche , J. and Breitbarth , E. ( 2005 ) Importance of the diazotrophs as a source of new nitrogen in the ocean . J. Sea Res ., 53 , 67 - 91 .
Lohrenz , S. E. , Wiesenburg , D. A. , Rein , C. R. , Arnone , R. A. , Taylor , C. D. , Knauer , G. A. and Knap , A. H. ( 1992 ) A comparison of in situ and simulated in situ methods for estimating oceanic primary production . J. Plankton Res ., 14 , 201 - 221 .
Mague , T. H. , Mague , F. C. and Holm-Hansen , O. ( 1977 ) Physiology and chemical composition of nitrogen fixing phytoplankton in the central North Pacific Ocean . Mar. Biol., 41 , 213 - 227 .
Mohr , W. , Grosskopf , T. , Wallace , D. W. R. and Laroche , J. ( 2010 ) Methodological underestimation of oceanic nitrogen fixation rates . PLoS One , 5 , e12583. doi:10.1371/journal.pone.001258.
Moisander , P. H. , Beinart , R. A. , Hewson, I. , White , A. E. , Johnson , K. S. , Carlson , C. A. , Montoya , J. P. and Zehr , J. P. ( 2010 ) Unicellular cyanobacterial distributions broaden the oceanic N2 fixation domain . Science , 327 , 1512 - 1514 .
Montoya , J. P. , Voss , M. , Kahler , P. and Capone , D. G. ( 1996 ) A simple, high-precision, high-sensitivity tracer assay for N2 fixation . Appl. Environ. Microbiol ., 62 , 986 - 993 .
Ottesen , E. A. , Marin , R. , Preston , C. , Young , C. R. , Ryan , J. P. , Scholin , C. and Delong , E. ( 2011 ) Metatranscriptomic analysis of autonomously collected and preserved marine bacterioplankton . ISME J ., 5 , 1881 - 1895 .
Pachiadaki , M. G. , Taylor , C. , Oikonomou , A. , Yakimov , M. M. , Stoeck , T. and Edgcomb , V. ( 2014 ) In situ grazing experiments apply new technology to gain insights into deep-sea microbial food webs . Deep Sea Res. Part II . doi:10.1016/j.dsr2.2014.10.019.
Robidart , J. C. , Church , M. J. , Ryan , J. P. , Ascani , F. , Wilson , S. T. , Bombar , D. , Marin , R. , Richards , K. J. et al. ( 2014 ) Ecogenomic sensor reveals controls on N2-fixing microorganisms in the North Pacific Ocean . ISME J ., 8 , 1175 - 1185 .
Sohm , J. A. , Webb , E. A. and Capone , D. G. ( 2011 ) Emerging patterns of marine nitrogen fixation . Nat. Rev. Microbiol. , 9 , 499 - 508 .
Strickland , J. D. H. and Parsons , T. R. ( 1972 ) A Practical Handbook of Seawater Analysis . Vol., Fisheries Research Board of Canada , Ottawa.
Taylor , C. D. and Doherty , K. W. ( 1990 ) Submersible Incubation Device (SID), autonomous instrumentation for the in situ measurement of primary production and other microbial rate processes . Deep Sea Res. Part I , 37 , 343 - 358 .
Taylor , C. D. , Doherty , K. W. and Honjo , S. ( 2013 ). Fixation Filter Assembly . US Patent # 8 ,426, 218 .
Taylor , C. D. , Edgcomb , V. P. , Doherty , K. W. , Engstrom , I. , Shanahan , T. , Pachiadaki , M. G. , Molyneaux , S. J. and Honjo , S. ( 2015 ) Fixation filter, device for the rapid in situ preservation of particulate samples . Deep Sea Res. Part I , 96 , 69 - 79 .
Taylor , C. D. and Howes , B. L. ( 1994 ) Effect of sampling frequency on measurements of seasonal primary production and oxygen status in near-shore coastal ecosystems . Mar. Ecol. Prog. Ser. , 108 , 193 - 203 .
Thompson , A. W. , Foster , R. A. , Krupke , A. , Carter , B. J. , Musat , N. , Vaulot , D. , Kuypers , M. M. M. and Zehr , J. P. ( 2012 ) Unicellular cyanobacterium symbiotic with a single-celled eukaryotic alga . Science , 337 , 1546 - 1550 .
Villareal , T. A. ( 1990 ) Laboratory culture and preliminary characterization of the nitrogen-fixing Rhizosolenia-Richelia symbiosis . Mar. Ecol., 11 , 117 - 132 .
Voss , M. , Bange , H. W. , Dippner , J. W. , Middelburg , J. J. , Montoya , J. P. and Ward , B. ( 2013 ) The marine nitrogen cycle: recent discoveries, uncertainties and the potential relevance of climate change . Philos. Trans. R. Soc. B , 368 , 20130121 .
Webb , E. A. , Ehrenreich, I. M. , Brown , S. L. , Valois , F. W. and Waterbury , J. B. ( 2009 ) Phenotypic and genotypic characterization of multiple strains of the diazotrophic cyanobacterium, Crocosphaera watsonii, isolated from the open ocean . Environ. Microbiol ., 11 , 338 - 348 .
Wilson , S. T. , Bo¨ttjer , D. , Church , M. J. and Karl , D. M. ( 2012 ) Comparative assessment of nitrogen fixation methodologies, conducted in the oligotrophic North Pacific Ocean . Appl. Environ. Microbiol ., 78 , 6516 - 6523 .
Zehr , J. P. , Waterbury , J. B. , Turner , P. J. , Montoya , J. P. , Omoregie , E. , Steward , G. F. , Hansen , A. and Karl , D. M. ( 2001 ) Unicellular cyanobacteria fix N2 in the subtropical North Pacific Ocean . Nature , 412 , 635 - 638 .