Nitrogen Fixation and Diazotroph Community in the Subarctic Sea of Japan and Sea of Okhotsk
Abstract
Nitrogen-fixing microorganisms (diazotrophs) significantly influence marine productivity by converting N2 gas into bioavailable nitrogen. Recent studies revealed that nitrogen fixation occurs in both warm oligotroph and colder and/or N-rich water; however, little is known about the spatial variability of nitrogen fixation activity and diazotroph diversity in cold waters. In this study, we examined the nitrogen fixation activity and diazotroph community structure in the subarctic waters around Hokkaido in northern Japan, including the Sea of Japan, the Sea of Okhotsk, and the North Pacific Ocean. Nitrogen fixation activity was detected at temperatures ranging from −1.1 to 15.6°C and was significantly related to high salinity and high temperature, which characterize the Tsushima Warm Current. The highest recorded nitrogen fixation activity (5.42 nmol N L−1 d−1) was comparable to that of subtropical regions. Diazotrophs usually reported in subtropical regions dominated the diazotroph communities and were likely advected from the upstream area of the Tsushima Warm Current—a tributary of the Kuroshio Current. In particular, the nifH abundance of the most dominant diazotroph—the symbiotic cyanobacterium UCYN-A1—showed a significant positive relationship with nitrogen fixation rate. We also detected a diazotroph community and low but active nitrogen fixation in the N-rich waters sourced from the Sea of Okhotsk, which infers the influence of both advected and indigenous diazotrophs on the regional nitrogen fixation. Although further research is needed, our study points toward the widespread distribution of diazotrophy in the Sea of Okhotsk.
Key Points
-
Nitrogen fixation and diazotroph community structure were investigated for the first time in the subarctic Sea of Japan and Sea of Okhotsk
-
Nitrogen fixation was detectable at temperatures ranging from −1.1 to 15.6°C and mainly controlled by diazotrophs advected from the south
-
UCYN-A1 was ubiquitous and dominant in the diazotroph communities in most surface water
Plain Language Summary
The oceanic reservoir of bioavailable nitrogen is mainly controlled by the balance between the gains from nitrogen fixation and the losses from denitrification. However, direct measurements have identified significantly higher denitrification than nitrogen fixation, suggesting the lack of understanding of diazotroph diversity and abundance. Recent studies have shown that diazotrophs occur over a much wider habitat range than previously assumed warm oligotrophic waters. Here, we examined for the first time the nitrogen fixation activity and diazotroph community structure in the subarctic waters around Hokkaido in northern Japan, including the Sea of Japan and Sea of Okhotsk characterized as colder and/or N-rich environments. Nitrogen fixation activity was detected at temperatures ranging from −1.1 to 15.6°C; it was related with the warm water mass transported from the southern oligotrophic waters. We also found that cyanobacterial and non-cyanobacterial diazotrophs were distributed in a region influenced by the cold and/or N-rich subarctic water. Our results suggest that the extensive diazotrophy in the previously overlooked subarctic marginal seas, including the Sea of Okhotsk would contribute to the gain of bioavailable nitrogen.
1 Introduction
Biological nitrogen fixation by specialized microorganisms (diazotrophs) reduces N2 gas into ammonia and is a major source of bioavailable nitrogen in the ocean. The oceanic reservoir of bioavailable nitrogen is mainly controlled by the balance between the gains from nitrogen fixation and the losses from denitrification (Gruber & Galloway, 2008). Based on geochemical estimates, the gain and loss of nitrogen are considered roughly balanced (Brandes & Devol, 2002; Deutsch et al., 2004), while direct measurements identified significantly higher denitrification rates compared with nitrogen fixation (Codispoti, 2007). This discrepancy suggests a significant underestimation of marine nitrogen fixation in the field, which may be partly explained by methodological uncertainty. Recent studies have incorporated a revised dissolution method that attains equilibrium 15N2 gas to more accurately estimate in vitro seawater nitrogen fixation rates during a 12–24 h incubation period (Großkopf et al., 2012; Mohr et al., 2010). This change in methodologies has yielded an approximately two-fold increase in nitrogen fixation rate estimates. Another possible cause for the underestimation is the lack of understanding of diazotroph diversity and abundance, which are now known to occur over a much wider habitat range than previously assumed (Zehr & Capone, 2020).
Diazotrophs have the unique ability to utilize nitrogen from N2 gas and are therefore prolific in warm subtropical waters with low dissolved inorganic nitrogen (DIN) concentrations (Sohm et al., 2011). The distribution and biological activity of these diazotrophs showed strong temperature dependency (Stal, 2009). Although how much DIN may suppress diazotorphy is still controversial (Knapp, 2012), generally high DIN concentrations are suggested to inhibit diazotrophic activity (Holl & Montoya, 2005). Warm and oligotrophic environments—such as subtropical waters—have traditionally been thus considered prerequisites for nitrogen fixation (Luo et al., 2014; Sohm et al., 2011). In addition to the well-known nitrogen-fixing colonial cyanobacteria Trichodesmium spp., recent developments in molecular biology and nifH gene sequencing (which encodes the iron unit of nitrogenase) have elucidated the high diversity of unicellular cyanobacterial and non-cyanobacterial diazotrophs in subtropical oligotrophic waters (Bombar et al., 2016; Zehr et al., 2003). NifH-based studies have also identified the presence of diazotrophs in other environments, such as the colder and/or DIN-rich waters of the polar seas (Blais et al., 2012; Harding et al., 2018; Shiozaki et al., 2017, 2018, 2020), temperate coastal waters (Bentzon-Tilia et al., 2015; Rees et al., 2009; Shiozaki et al., 2015), and upwelling regions (Fernandez et al., 2011; Gradoville et al., 2017). Nitrogen fixation in these environments is variable (Mulholland et al., 2019; Tang et al., 2019) but may largely contribute to the global activity of marine nitrogen fixation. However, since studies on nitrogen fixation in subarctic and arctic waters are still recent and few, the effect of environmental conditions on the spatiotemporal variation of diazotrophic activity in cold and/or DIN-rich water is relatively unknown (von Friesen & Riemann, 2020). Emerging evidence has shown that UCYN-A1, a unicellular cyanobacterial diazotroph, is widely distributed from the lower latitude to polar regions and the key species to understand diazotrophy in cold waters (Harding et al., 2018; Shiozaki et al., 2020).
In the northwestern Pacific, studies on diazotrophy have been mostly restricted to subtropical waters due to the conventional assumption. These studies demonstrated that nitrogen fixation was widely active, and diazotroph communities were dominated by cyanobacteria (Cheung et al., 2017; Shiozaki et al., 2010). Although one recent study reported active nitrogen fixation and various diazotrophs in temperate waters off Japan (Shiozaki et al., 2015), knowledge in colder water is still needed for better understanding diazotrophy in the northwestern Pacific. This study examined the nitrogen fixation activity and diazotroph community structure in the subarctic Sea of Japan and the Sea of Okhotsk in early summer. The research area has spatiotemporally complex oceanography due to the confluence of a warm current (Tsushima Warm Current) and a cold current (East Sakhalin Current) (Takizawa, 1982). In summer, the warm current becomes predominant and flows into the coastal area around Hokkaido, meanwhile offshore waters transported by the cold current exist as Okhotsk Surface Water (OSW) and Intermediate Subsurface Cold Water (ISCW) at the surface and the subsurface, respectively (Itoh & Oshima, 2000; Kasai et al., 2010). Although most of the study stations were located in the subarctic region, the regional current system promotes the seasonal emergence of N-depleted waters from spring to autumn (Kasai & Hirakawa, 2015; Kasai et al., 2010), which suggests the occurrence of nitrogen fixation.
2 Materials and Methods
2.1 Water Sampling and Environmental Data
Sampling was conducted at 19 stations around Hokkaido during the HK1805 cruise onboard the R/V Hokko-maru from 31 May to 7 June 2018 (Figure 1). The study area includes the Sea of Japan (St. J1–J6), the Sea of Okhotsk (St. O1–O10), and the Pacific Ocean (St. P1–P3).

Current system and sampling stations around Hokkaido in northern Japan during the HK1805 cruise. Dotted and solid arrows represent warm and cold currents, respectively. The color of each plot corresponds to each water mass: Tsushima Warm Current originated Water (red), Okhotsk Surface Water (blue), and Mixed Water (black). Surface water samples were analyzed at all stations, and vertical water samples were additionally analyzed at three stations (indicated by diamonds). The first letter of each station name represents each sampling region (e.g., J of J1 represents the Sea of Japan).
Temperature and salinity profiles at each station were measured using a conductivity-temperature-pressure (CTD) system (Model-911 plus, SeaBird Electronics). Surface water samples were collected with an acid-cleaned bucket at all stations. Immediately after sampling, aliquots of surface water samples were taken for nutrient concentrations, chlorophyll a (Chl a) concentrations, incubation experiments, and DNA analysis. We were unable to collect Chl a samples at St. J1, J2, and P2 due to time constraints; Chl a for these stations were therefore derived via linear regression between the sampled Chl a concentrations and the CTD instrument fluorescence sensor (Figure S1). In conjunction with CTD observations, Niskin-X bottles were used to collect samples below the surface for incubation experiments and DNA analysis at depths (0–72 m) corresponding to 25%, 10%, 1%, and 0.1% of the surface light intensity at St. J2, O2, and O7. Nutrient concentration samples were also collected from several depths in the upper 200 m at St. O2 and O7, while only surface water samples were collected at St. J2. Light attenuation was determined using a Profiling Reflectance Radiometer PRR-600 (Biospherical Instruments Inc.) before CTD deployment.
2.2 Nitrogen Fixation and Primary Production
Nitrogen fixation and primary production rates were measured using the 15N2 gas dissolution (Mohr et al., 2010) and 13C methods (Hama et al., 1983), respectively. The incubation samples were filled into triplicate acid-cleaned 2 L polycarbonate (PC) bottles. At the beginning of incubation for each station and depth, samples (2.3 L) for determining the initial 15N and 13C enrichment of particulate organic matter were filtered onto pre-combusted (450°C for 6 h) GF/F filters (n = 1), and 15N2-enriched filtered seawater (50 mL) was then injected into each bottle. To prepare the 15N2-enriched filtered seawater, stored surface seawater in the subtropical North Pacific (see detail in Table S1) was filtered using a GF/F filter and degassed using a Sterapore membrane unit (Mitsubishi Rayon Co., Ltd.) at a flow rate of 500 mL min−1. Degassed seawater was stored in 1 L Tedlar bags without headspaces, and 15N2 gas (99.8 atom % 15N; SI Science) was dissolved at a ratio of 10 mL 15N2 per 1 L seawater. The enriched seawater was agitated to facilitate the dissolution and used within 48 h. 13C-labeled sodium bicarbonate (99 atom % 13C; Cambridge isotope Laboratories Inc.) was then added to the same triplicate bottles to a final concentration of 200 µM. For the sub-surface incubation samples, the light levels were adjusted to the corresponding levels using neutral-density screens. All bottles were sealed with a thermoplastic elastomer cap and incubated in an on-deck running surface seawater bath for 24 h. Incubations were terminated by gentle vacuum filtration (<200 mm Hg) through pre-combusted GF/F filters, which were kept frozen until onshore analysis. The filters were dried at 50°C overnight and then exposed to HCl fumes for 2 h to remove inorganic carbon before isotope analysis. Particulate N, particulate C, and isotopic ratios were analyzed using a Flash EA elemental analyzer (Thermo Electron) connected to a DELTAplus XP mass spectrometer (Thermo Electron). Nitrogen fixation and primary production rates were calculated as described in Montoya et al. (1996) and Hama et al. (1983), respectively. The volume of dissolved N2 was calculated using the equation given by Weiss (1970). Since we did not determine the isotope ratio of the 15N2-enriched seawater, calculated rates could contain uncertainty (White et al., 2020). We performed a sensitivity analysis for the nitrogen fixation measurements to determine the minimum quantifiable nitrogen fixation rates for each sample within the range of 0.12–4.22 nmol N L−1 d−1 (Table S2; Gradoville et al., 2017; Montoya et al., 1996). Depth-integrated nitrogen fixation was calculated for St. J2, O2, and O7 using trapezoidal integration. As some non-cyanobacterial diazotrophs are smaller than the GF/F pore size (0.6 µm), this study may underestimate the activity of smaller non-cyanobacterial diazotrophs of <0.6 µm diameter (Bombar et al., 2018). Moreover, the 15N2 gas stocks used in this study (purchased from SI Science) likely contain negligible contamination of 15N nitrate and ammonium (Shiozaki et al., 2015) and are therefore unlikely to cause overestimations of nitrogen fixation.
2.3 DNA Extraction, nifH Amplicon Sequencing, and Sequence Analysis
Water samples (1.2 L) for DNA analysis were filtered onto Whatman 0.2-μm pore size Nuclepore filters and stored at −20°C until analysis on land. DNA was extracted using a ChargeSwitch Forensic DNA Purification Kit (Invitrogen) following the manufacturer's protocol. Nested polymerase chain reaction (PCR) (Zehr & Turner, 2001) was performed to amplify nifH genes in all samples as described in Shiozaki et al. (2017) using PuRe Taq Ready-To-Go PCR Beads (GE Healthcare Life Science). In the first-round PCR, nifH3 and nifH4 primers (Zehr & Turner, 2001) were used as the forward and reverse primers, respectively. The forward and reverse primers (nifH1 and nifH2) (Zehr & Turner, 2001) with the attached Illumina adapter forward (5′-ACACTCTTTCCCTACACGACGCTCTTCCGATCT-3′) and reverse (5′-GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT-3′) primer pads were added to amplicons for the second-round PCR. Both the first and second PCRs were performed under the same conditions in the following order: 94°C for 2 min, 30 cycles of 94°C for 30 s, 52°C for 1 min, 72°C for 1 min, and finally 72°C for 7 min. Only samples with successful PCR amplifications (28 of 31 samples) confirmed by electrophoresis for the PCR products were used for further analysis. Pooled triplicate PCR products per sample were purified using an AMPure XP purification kit (Beckman Coulter). An index PCR was then performed to attach a 7-base barcode using a Nextera XT Index kit (Illumina) and a KAPA HiFi HotStart Ready Mix (KAPA Biosystem) under the following cycling conditions: 95C for 3 min, followed by 8 cycles of 95°C for 30 s, 55°C for 30 s, 72°C for 30 s, followed by a final extension at 72°C for 5 min. The PCR products were purified again using the AMPure XP purification kit and quantified using Quanu-iT Picogreen dsDNA Reagent and Kit (Invitrogen). All PCR amplicons were finally pooled at equimolar concentrations (∼14 ng PCR product per sample) and subjected to paired-end sequencing on the MiSeq platform (Illumina), where 300 bp fragments from each end of the libraries were sequenced using a MiSeq Reagent Kit v3 (600 cycles; Illumina) with Phix control v3 (Illumina). Sequencing reads were demultiplexed using the standard Illumina pipeline. We then performed a community analysis of the reads using the QIIME2 program (ver. 2020.2; Bolyen et al., 2019). Primer sequences were trimmed from the sequence using the Cutadapt plug-in (Martin, 2011). The reads were denoised and clustered based on sequence variants (SVs) at single-nucleotide resolution using the Deblur plug-in (Amir et al., 2017). The 3′ends of the forward and reverse sequences were merged and trimmed to 320 bp using the [–p-trim-length] command. The obtained SVs were searched against the publicly available nifH database (https://wwwzehr.pmc.ucsc.edu/nifH_Database_Public/) to discriminate diazotroph sequences. The SV sequences were translated into amino acid sequences and analyzed using BLAST searches against the National Center for Biotechnology Information databases; non-nifH SVs or SVs with a frame shift were removed. The nifH sequences were aligned using MUSCLE in the MEGA6 package (Tamura et al., 2013). A phylogenetic tree was constructed using the maximum likelihood method, and bootstrap values were determined using 1,000 iterations implemented in MEGA6. The nifH sequences were classified into canonical nifH clusters (Zehr et al., 2003) via BLAST search to a custom reference database of 879 full-length nifH sequence public databases (https://wwwzehr.pmc.ucsc.edu/Genome879) complemented by the decision tree statistical model, CART (Frank et al., 2016).
2.4 Quantitative Polymerase Chain Reaction
Since the nifH sequences of UCYN-A1 were retrieved as a dominant diazotroph from the Miseq results (see Section 3 section), we conducted a quantitative PCR (qPCR) assay using the MiniOpticon Real-Time PCR Detection System (Bio-Rad). The nifH standards were obtained by cloning the known UCYN-A1 nifH sequence, and previously designed primer and probe sets were used for the qPCR (Table S3; Shiozaki et al., 2014). The 20 µL qPCR mixture contained 10 µL of 2 × Premix Ex Taq (Probe qPCR; Takara Bio), 6.4 µL nuclease-free water, 0.6 µL each of 10 µM forward and reverse primers, 0.4 µL of 10 µM TaqMan probe, and 2 µL of template DNA. PCR amplification was performed in the following order: 95°C for 5 min, 49 cycles at 95°C for 4 s, and 60°C for 11 s; a detection step of 1 s was included at the end of each cycle. We ran the qPCR assays in duplicate or triplicate reactions for each sample. For qPCR of each environmental sample, duplicate standard curves ranging from 101 to 106 gene copies and the negative control were prepared for each reaction. The r2 value for the standard curves ranged between 0.99 and 1.00, and the PCR reaction efficiencies were between 96.6% and 103.8%. No signals were detected in the negative controls. Based on the extracted volumes, the detection limit was 75 nifH gene copies L−1 seawater.
2.5 Statistics
ANOVA and a post-hoc Tukey-Kramer test were used to compare the surface nitrogen fixation activity of the different water masses using the statsmodels library (Seabold & Perktold, 2010) in Python. A principal component analysis (PCA) was conducted to examine the relationships between nitrogen fixation activity, UCYN-A1 abundance, and environmental variables in the surface water using Python and the scikit-learn library (Pedregosa et al., 2011).
3 Results
3.1 Environmental Conditions
Surface temperature and salinity in the research area ranged from 5.0 to 15.6°C and 32.2 to 34.3, respectively (Table 1). The temperature (>14°C) and salinity (>34.1) were relatively high at the Sea of Japan stations (St. J1, J2, and J3) and relatively low (<10°C and <33.0, respectively) at most stations in the Sea of Okhotsk. The same trend was observed in the sub-surface layers; water temperature and salinity at St. J2 (Sea of Japan) ranged from 10.0 to 14.7°C and 34.1 to 34.4, respectively (Table S4), and those at St. O2 and O7 (Sea of Okhotsk) ranged from 0.8 to 5.5°C and 32.3 to 33.9, respectively. The water masses in the study area were divided into four groups (Figure S2; Tables 1 and S4) based on the definitions by Takizawa (1982) and Hanawa and Mitsudera (1986): Tsushima Warm Current originated Water (TWCW), OSW, ISCW, and Mixed Water (MW) of other unclassified water masses. Water masses affected by warm currents in the Sea of Okhotsk and Pacific Ocean have been defined as the Soya Warm Current Water and Tsugaru Warm Current Water, respectively (Hanawa & Mitsudera, 1986; Takizawa, 1982). Both currents were defined as TWCW in this study as they are tributaries of the Tsushima Warm Current (Figure 1). The surface waters at 9 out of the 19 stations were classified as TWCW and were characterized by high salinity. This group includes stations in a coastal area of the Sea of Okhotsk—such as St. O4 and O5—which indicates the inflow of the Tsushima Warm Current to the Sea of Okhotsk through the Soya Strait. Six stations in the Sea of Okhotsk were classified as OSW and were characterized by low salinity. The surface water of the four remaining stations in the Sea of Japan (J4), the Sea of Okhotsk (St. O1 and O2), and the Pacific Ocean (St. P3) were classified as MW. The ISCW is characterized by low salinity and temperature and was detected in the subsurface waters at 25% and 10% light depth at St. O2 and 10%, 1%, and 0.1% light depth at St. O7 (Table S4).
Stationa | Water massb | Latitude | Longitude | Bottom depth (m) | SST (°C) | Salinity | Chl a (µg L−1) | NO3−(µM) | PO43−(µM) | N/P ratiod | N2 fixation (n mol N L−1 d−1) | Primary production (n mol C L−1 d−1) | UCYN-A1 (copy L−1) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
J1 | TWCW | 41° 30N | 139° 50E | 1,364 | 15.6 | 34.3 | 0.5c | 0.1< | 0.05 | 1.87 | 5.42 ± 0.89 | 965 | 8.83 × 105 |
J2 | TWCW | 42° 31N | 139° 42E | 1,254 | 15.2 | 34.1 | 0.3c | 0.1< | 0.03 | 3.94 | 0.74 ± 0.34 | 1,205 | 3.02 × 103 |
J3 | TWCW | 43° 30N | 140° 20E | 642 | 14.0 | 34.2 | 0.6 | 0.1 | 0.1 | 1.00 | 1.21 ± 0.34 | 662 | 4.07 × 105 |
J4 | MW | 44° 30N | 141° 30E | 76.2 | 12.3 | 32.9 | 0.8 | 0.1< | 0.04 | 2.50 | 0.73 ± 0.06 | 967 | 1.48 × 103 |
J5 | TWCW | 45° 30N | 141° 25E | 111 | 10.1 | 33.8 | 0.6 | 0.1< | 0.04 | 2.50 | 0.58 ± 0.05 | 213 | 5.13 × 102 |
J6 | TWCW | 45° 33N | 142° 05E | 40 | 11.6 | 33.8 | 1.6 | 0.1< | 0.03 | 3.33 | 1.54 ± 0.49 | 2,398 | 1.35 × 105 |
O1 | MW | 45° 30N | 142° 30E | 82 | 5.0 | 33.5 | 1.3 | 7.2 | 0.72 | 10.0 | 1.83 ± 0.39 | 8,380 | 3.02 × 103 |
O2 | MW | 45° 19N | 142° 41E | 98 | 6.9 | 33.0 | 0.8 | 1.8 | 0.43 | 4.30 | n.d. | 3,477 | 1.48 × 103 |
O3 | OSW | 45° 01N | 143° 10E | 132 | 9.9 | 32.2 | 0.6 | 0.1< | 0.29 | 0.34 | n.d. | 1,258 | 1.51 × 103 |
O4 | TWCW | 44° 53N | 142° 53E | 88 | 7.8 | 34.1 | 0.7 | 2.4 | 0.27 | 8.76 | 1.53 ± 0.75 | 1,770 | 4.30 × 105 |
O5 | TWCW | 44° 30N | 143° 20E | 58 | 9.1 | 33.8 | 1.4 | 0.6 | 0.2 | 3.00 | 1.34 ± 0.43 | 3,103 | 7.50 × 103 |
O6 | OSW | 44° 14N | 144° 08E | 109 | 8.4 | 32.5 | 1.2 | 0.1< | 0.3 | 0.33 | n.d. | 2,026 | 4.44 × 102 |
O7 | OSW | 44° 30N | 144° 20E | 841 | 7.2 | 32.3 | 0.8 | 0.1< | 0.3 | 0.33 | n.d. | 1,066 | n.d. |
O8 | OSW | 44° 60N | 144° 45E | 1,562 | 7.6 | 32.3 | 0.8 | 0.1 | 0.21 | 0.48 | n.d. | 1,421 | n.d. |
O9 | OSW | 45° 25N | 145° 10E | 3,010 | 6.4 | 32.3 | 3.0 | 0.1 < | 0.15 | 0.67 | n.d. | 3,087 | n.d. |
O10 | OSW | 44° 38N | 145° 22E | 2300 | 8.4 | 32.5 | 1.9 | 0.2 | 0.22 | 0.91 | n.d. | 2,439 | 4.21 × 104 |
P1 | TWCW | 42° 10N | 141° 15E | 217 | 11.4 | 33.8 | 0.6 | 0.1< | 0.16 | 0.62 | 1.66 ± 1.16 | 1,611 | 8.73 × 104 |
P2 | TWCW | 41° 53N | 142° 44E | 417 | 10.5 | 33.7 | 0.8d | 0.1< | 0.04 | 2.54 | 1.52 ± 0.91 | 2,526 | 8.88 × 103 |
P3 | MW | 43° 00N | 145° 50E | 717 | 8.4 | 33.4 | 1.2 | 0.1< | 0.31 | 0.32 | 0.91 ± 0.44 | 2825 | n.d. |
- a The first letter of each station name represents the region (e.g., J of J1 represents the Sea of Japan).
- b TWCW, Tsushima Warm originated current water; MW, Mixed Water; OSW, Okhotsk Surface Water (Takizawa, 1982).
- c Chl a at St. J1, J2, and P2 were calculated via a regression equation between Chl a obtained by the fluorescence sensor and by water sampling (Figure S1).
- d When the nutrient concentration was below the detection limit, the value of the detection limit was used for the analysis. n.d., not detected.
Surface nitrate concentrations were below 1 µM at most stations—except for those around the Soya Strait (St. O1, O2, and O4), with concentrations as high as 7.2 µM (Figure 2a). Surface phosphate concentrations were low in the Sea of Japan and the Pacific Ocean (0.03–0.31 µM) and predominantly high (0.15–0.72 µM) in the Sea of Okhotsk (Figure 2b). The nitrate to phosphate (N/P) ratio ranged from 0.32 to 10.0, generally lower than the Redfield ratio (Table 1). Both nutrient concentrations and the N/P ratio were especially high at stations around the Soya Strait. Similarly, surface Chl a concentrations were low (0.3–1.6 µg L−1) in the Sea of Japan and the Pacific Ocean and high (up to 3.0 µg L−1) in the Sea of Okhotsk (Figure 2c).

Horizontal distribution of surface (a) nitrate, (b) phosphate, (c) chlorophyll a, (d) primary production, (e) nitrogen fixation, and (f) UCYN-A1 abundance. The crosses indicate values below the detection limit. Chl a at St. J1, J2, and P2 was calculated using the values from the CTD fluorescence sensor.
3.2 Primary Production and Nitrogen Fixation
Surface primary production showed a similar distribution pattern to nutrients and Chl a, with generally higher values in the Sea of Okhotsk and the Pacific Ocean and lower values in the Sea of Japan (Figure 2d). The average volumetric primary production was 1,068 ± 671, 2,321 ± 516, and 2,803 ± 2,021 nmol C L−1 d−1 in the Sea of Japan, Pacific Ocean, and Sea of Okhotsk, respectively. Among all stations, the highest primary production rate (8,380 nmol C L−1 d−1) and phosphate and nitrate concentrations were recorded at St. O1.
Nitrogen fixation activity was detected in the surface waters of 12 stations and ranged from 0.58 to 5.42 nmol N L−1 d−1 (average: 1.00 nmol N L−1 d−1); the remaining seven stations where nitrogen fixation was not detected were located in the Sea of Okhotsk (Figure 2e). The highest value (5.42 nmol N L−1 d−1) was observed at the southern-most station (J1), which was also characterized by higher temperature and salinity (Table 1). Nitrogen fixation was detected at all stations of the TWCW but was undetected at all stations in the OSW. Nitrogen fixation in the remaining four MW stations showed large variability, ranging from undetectable to 1.83 nmol N L−1 d−1. The magnitudes of nitrogen fixation were significantly different between the TWCW and OSW (ANOVA, p < 0.05, Turkey Kramer test, p < 0.05), and the highest nitrogen fixation rates were observed in the TWCW stations (Figure S3).
The vertical profiles of nitrogen fixation showed distinct patterns at three stations (St. J2, O2, and O7; Figure 3). The water column at St. J2 in the Sea of Japan is consistently covered by TWCW, and low but active nitrogen fixation was only detected at the surface and 0.1% light depth (0.74 and 0.27 nmol N L−1 d−1, respectively). Nitrogen fixation was not detected at the coastal station (St. O2) in the Sea of Okhotsk, which is covered by both MW and ISCW. In contrast, low nitrogen fixation was detected at 10%, 1%, and 0.1% light depth (0.59, 0.85, and 0.65 nmol N L−1 d−1, respectively) at the offshore station (St. O7), where the corresponding subsurface water mass was identified as the low temperature ISCW (−1.1 to 2.0°C). The depth-integrated nitrogen fixation at St. J2, O2, and O7 was 6.35, undetectable, and 42.6 µmol N m−2 d−1, respectively.

Vertical profiles of nitrogen fixation (red), temperature (black), salinity (black dotted), and nitrate (blue) at St. J2 (Japan Sea), O2 (Okhotsk coast), and O7 (Okhotsk offshore). The nitrogen fixation error bars denote the standard deviation. Nitrate concentration at St. J2 was only obtained for the surface water.
3.3 nifH Gene Composition and UCYN-A1 Abundance
In total, 94,803 reads were retrieved from 28 DNA samples, yielding 140 SVs. The number of reads varied between 237 and 19,126. The representative SVs were affiliated with cyanobacterial diazotrophs (Cluster 1B; 22 SVs), non-cyanobacterial diazotrophs of nifH Cluster 1A (putative δ-proteobacteria, 39 SVs), Cluster 1J/K (putative α-proteobacteria, 4 SV), Cluster 1G (putative γ-proteobacteria, 31 SVs), and Cluster III (putative δ-proteobacteria, 41 SVs) (Zehr et al., 2003) (Table S5; Figure 4). SV001—which is omnipresent throughout the study region—and representative SVs accounting for ≥3% of the total nifH sequences at the study stations were selected for further analysis (25 SVs accounted for >83% of all sequences). Overall, cyanobacterial diazotrophs (Cluster 1B) were the most abundant diazotroph group. SV001 was the most highly recovered SV and displayed 100% nucleotide identity to the symbiotic unicellular cyanobacteria, UCYN-A1, which accounted for 51% of the total sequences. SV002 showed the second highest recovery, displaying 99% similarity at the nucleotide level to the free-living unicellular cyanobacteria Crocosphaera watsonii. SV007 and SV020 showed 99% and 100% nucleotide similarity to the cyanobacterial diazotrophs UCYN-A3 and Trichodesmium, respectively. All other SVs retrieved from DNA samples were classified as non-cyanobacterial diazotrophs. Among the non-cyanobacterial SVs, six were affiliated with the Cluster 1G diazotroph, and SV004 and SV005 showed high nucleotide similarity (>97%) to the subtropical heterotrophic diazotroph γ-2477A11 (NCBI accession No. EU052413). SV003 in Cluster 1G was identical to the γ-proteobacteria retrieved from the South Pacific subtropical gyre (NCBI accession No. HM210403). Seven SVs fell into Cluster 1A, several of which were closely related to nifH sequences found in the North Sea water column and/or bottom sediment (Fan et al., 2015; Figure 4). Seven SVs belonged to Cluster III which were dominated by sequences from putative anaerobic bacteria and archaea (Zehr et al., 2003). Two SVs (SV064 and SV067) in Cluster III also displayed high similarity (>98%) to the North Sea nifH sequences (Fan et al., 2015; Figure 4).

Maximum likelihood phylogenetic tree of nifH gene sequences. The 25 representative amplicon sequence variants (SVs) in this study are shown in red. Canonical nifH clusters according to Zehr et al. (2003) are displayed by colored external strips with the cluster names. The areas of the purple circles are proportional to the Bootstrap value (>50%) determined from 1,000 iterations. The tree was produced using the Interactive Tree of Life (http://itol.embl.de/; Letunic & Bork, 2019).
Overall, UCYN-A1 was ubiquitous and dominant in the diazotroph communities in most surface waters, except at St. O8 and O10 (Figure 5a). Crocosphaera, γ-2477A11, and SV003 were also broadly distributed in most surface waters. Trichodesmium was only detected at the eastern-most station in the Pacific Ocean (St. P3). In contrast, the community structures at St. O8 and O10 in the Sea of Okhotsk were distinct from other stations. Cluster III (SV063) dominated at St. O8, followed by Cluster III (SV064), and then Cluster 1A (SV065). Cluster 1J (SV034) dominated at St. O10.

(a) Horizontal distribution of diazotroph community around Hokkaido, northern Japan during HK1805. (b) Vertical distribution of diazotroph community (pie chart) and UCYN-A1 abundance (black line) at St. J2 (Sea of Japan), O2 (Okhotsk coast), and O7(Okhotsk offshore). n.d. denotes water in which nifH was not amplified by nested PCR.
The vertical profiles of the diazotroph community significantly differed between stations (Figure 5b). At St. J2 in the Sea of Japan, UCYN-A1 occupied more than 60% of the total diazotroph sequences from the surface to 1% light depth, while Cluster 1A (SV104) occupied ∼40% of the community at 0.1% light depth. Despite the absence of nitrogen fixation throughout the water column at St. O2 (coastal area of the Sea of Okhotsk), various diazotrophs were still detected at different depths: UCYN-A1, Cluster 1A (SV112), Cluster 1G (SV028), and Cluster 1A (SV031) dominated the surface, 10%, 1%, and 0.1% light depth, respectively. At the offshore station in the Sea of Okhotsk (St. O7), UCYN-A1 was dominant throughout the entire water column, while another community dominated by Cluster 1G (SV028) accounted for 37% of the total diazotrophs at 1% light depth.
The diazotroph communities in MW, OSW, and ISCW—which are less influenced by the Tsushima Warm Current—were notably distinct from those in TWCW. Diazotrophs in these waters were dominated by not only Cluster 1B (cyanobacteria) and Cluster 1G (γ-proteobacteria) but also Clusters 1A, III (δ-proteobacteria), and Cluster 1J (α-proteobacteria) (Figure S4). Cluster 1A and Cluster III were the dominant or semi-dominant diazotrophs in several stations of the Sea of Okhotsk, including at 0.1% light depth in St. O2 (MW), the surface in St. O8 (ISCW), and 10% light depth in St. O7 (ISCW).
The results of the qPCR analysis revealed the extensive detection of UCYN-A1 across the study region (20 of 31 samples), ranging from 4.5 × 102 to 8.9 × 105 copies L−1 (Figures 2f and 5b). UCYN-A1 was detected from the surface to 0.1% light depth and covered a wide range of temperatures (−1.1–15.6°C) and salinities (32.2–34.3). The maximum abundance was observed at the southern-most station (St. J1) with the highest temperature and salinity measurements.
3.4 Relationships Between Nitrogen Fixation Rates, UCYN-A1 Abundance, and Environmental Variables
The result of the PCA of the surface water variables suggests that salinity and UCYN-A1 abundance were the dominant influence on nitrogen fixation rate (Figure 6) in the study region. This relationship was further supported by the significant correlation between nitrogen fixation and UCYN-A1 (p < 0.01), temperature (p < 0.05), and salinity (p < 0.05) (Table 2). The highest UCYN-A1 abundance was observed at the station with the maximum nitrogen fixation activity (St. J1); however, a high abundance was also detected in samples with no nitrogen fixation activity, regardless of the sampling depth (surface: up to 4.2 × 104, subsurface: 7.1 × 105 copies L−1). We observed no significant correlation between nitrogen fixation and nutrient concentrations (nitrate and phosphate) or N/P ratio.

Relationship between each surface environment variable and each station via principal component analysis (T, temperature; S, salinity; N, nitrate; P, phosphate; N:P, N/P ratio; PP, primary production; N2, nitrogen fixation; and UC, UCYN-A1 abundance). Red, black, and blue dots correspond to the TWCW, MW, and OSW, respectively.
SST | SSS | Nitrate | Phosphate | N/P ratio | Nitrogen fixation | Primary production | UCYN-A1 | |
---|---|---|---|---|---|---|---|---|
SST | 1 | |||||||
SSS | 0.674** | 1 | ||||||
Nitrate | −0.489* | −0.060 | 1 | |||||
Phosphate | −0.744** | −0.295 | 0.794** | 1 | ||||
N/P ratio | −0.225 | −0.001 | 0.841** | 0.444 | ||||
Nitrogen fixation | 0.526* | 0.522* | 0.136 | −0.192 | 0.253 | 1 | ||
Primary production | −0.593** | −0.074 | 0.853** | 0.772** | 0.600** | 0.032 | 1 | |
UCYN-A1 | 0.543* | 0.498* | 0.129 | −0.228 | 0.368 | 0.579** | −0.059794597 | 1 |
- *p < 0.05; **p < 0.01.
4 Discussion
4.1 Environmental Constraints on the Geographical Variability of Nitrogen Fixation
The Tsushima Warm Current strongly influenced the spatial variability of nitrogen fixation in our study. The Tsushima Warm Current is a Kuroshio offshoot transporting warm and saline water into the Sea of Japan and then flows into the Sea of Okhotsk and Sea of Japan as the Soya Warm Current and the Tsugaru Warm Current, respectively (Figure 1). Our knowledge of nitrogen fixation in the Tsushima Warm Current is limited to the study by Hashimoto et al. (2012) who reported diazotroph community and the dominance of UCYN-A1 in the Japan Sea. Active nitrogen fixation and various diazotrophs, however, have been often detected by several studies in its upstream, the Kuroshio region (Cheung et al., 2017; Shiozaki et al., 2010). These studies demonstrated that nitrogen fixation in the Kuroshio region is relatively high, and cyanobacterial diazotrophs tended to dominate diazotroph community. These current therefore likely transported active diazotrophs and enhanced nitrogen fixation to the study region. In agreement, the magnitude of nitrogen fixation in each water mass showed different trends. Surface nitrogen fixation was detected at all stations in the TWCW and showed a higher average compared with that of the other water masses in the region (Figure S3). In contrast, nitrogen fixation was not detected in the surface waters of the OSW. Shiozaki et al. (2015) reported high nitrogen fixation activities in the Tsugaru Warm Current—a tributary of the Tsushima Warm Current—in the northwestern coastal region of the North Pacific. This suggests that the Tsushima Warm Current is a dominant control on the spatial variation of nitrogen fixation in the coastal areas around Hokkaido. This study is the first to assess the spatial distribution of nitrogen fixation activity across the subarctic Sea of Japan and the Sea of Okhotsk. The measured nitrogen fixation activity in this study (0.58–5.42 nmol N L−1 d−1) was lower than in temperate regions, including the North Atlantic coast (up to 130 nmol N L−1 d−1; Mulholland et al., 2019) and the Danish Strait (47–83 nmol N L−1 d−1; Bentzon-Tilia et al., 2015); however, they were generally comparable to or exceeded the observed values in the Bering Sea (2.27–2.84 nmol N L−1 d−1; Shiozaki et al., 2017), Chukchi Sea (0.14–0.42 nmol N L−1 d−1; Harding et al., 2018), and the oligotrophic open ocean of the western and central North Pacific (0.17–3.62 nmol N L−1 d−1; Shiozaki et al., 2010, 2017). Further, the depth-integrated nitrogen fixation in the offshore station of the Sea of Okhotsk (42.6 µmol N m−2 d−1) was as high as that observed in the western and central subtropical North Pacific (39.2 ± 7.51 µmol N m−2 d−1; Shiozaki et al., 2010). Therefore, our findings infer the occurrence of substantial nitrogen fixation in the coastal area around Hokkaido.
Nitrogen fixation was detected at low rates in ISCW but not in OSW in the Sea of Okhotsk (Figures 2 and 3). The difference may be explained by the concentration of iron as a major limiting nutrient for nitrogen fixation, which is more enriched in the subsurface than in the surface in the Sea of Okhotsk (Nishioka et al., 2020). N-depleted OSW indicates that iron at the surface water could be used up by a spring bloom (Kasai et al., 2010, 2015), resulting in the lack of detectable nitrogen fixation rate in the OSW. In contrast, detectable nitrogen fixation rates were observed in the N-rich ISCW in which iron concentration was possibly enriched. A recent study identified the occurrence of nitrogen fixation in cold and N-rich water in the presence of an iron supply (Shiozaki et al., 2020), which infers the possibility of widespread diazotrophy in the Sea of Okhotsk. However, this cannot be confirmed by our study due to the low data coverage in the Sea of Okhotsk, though it is worth further investigation.
4.2 Geophysical Variation in Diazotroph Diversity
Although our study stations are located in the subarctic region, we identified many diazotroph species that had been reported to thrive in the subtropics such as UCYN-A1, Crocosphaera, Trichodesmium, and γ-2477A11. Warm currents likely transported some of these diazotrophs to the study area from lower latitudes, as inferred from the close relationship between nitrogen fixation and water masses. Interestingly, UCYN-A1 was also dominant in the OSW and ISW which were not affected by the warm currents, indicating that UCYN-A1 also inhabited with diazotrophic activity in the cold waters as recently reported (Harding et al., 2018; Shiozaki et al., 2020). Recent studies found that UCYN-A could grow in high nitrate waters (Shiozaki et al., 2020; Turk-Kubo et al., 2018) and that DIN enrichment did not inhibit its nitrogen fixation (Mills et al., 2020). UCYN-A1 would therefore likely adapt to the N-rich waters of the Sea of Okhotsk due to its unusual insensitivity to high DIN as well as low temperature.
Non-cyanobacterial diazotrophs occasionally dominated the diazotroph communities in cold and/or N-rich water, such as the ISW and the OSW (which were minimally influenced by warm water). It should be noted that the absolute abundance of the non-cyanobacterial diazotrophs could be very low because nifH gene sequencing can detect very rare organisms. This would be the reason for low or undetectable nitrogen fixation rates in those waters. However, it is still possible that diazotroph groups, especially Clusters 1A and III, play a certain role in the regional nitrogen cycle where subtropical diazotrophs are unable to flourish. These nifH clusters have been widely reported in both pelagic and sediment samples from temperate, cold, and/or N-rich regions, including the English Channel (Rees et al., 2009), Monterey Bay (Cabello et al., 2020), Danish Strait (Bentzon-Tilia et al., 2015), northwestern upwelling coast of Iberia (Moreira-Coello et al., 2019), and Chukchi Sea (Shiozaki et al., 2018). This is consistent with the detection of these groups in the cold waters (−0.1 to 10.0°C) of our study region. In contrast to previous assumptions of diazotroph distribution in the region, our findings suggest that the Sea of Okhotsk can accommodate variable diazotrophs.
5 Conclusions
Nitrogen fixation activity was detected at almost all stations in the study region within a temperature range of −1.1–15.6°C; however, its magnitude varied between the different water masses. The highest nitrogen fixation activity (comparable to subtropical water values) was detected in the TWCW, which is characterized by high salinity and high temperature. NifH sequences were omnipresent throughout the study region. The cyanobacterial diazotroph, UCYN-A1, was likely transported by the Tsushima Warm Current from the Kuroshio region and may be a major active diazotroph in the study region. Our results suggest that subarctic areas which are largely unrated in diazotrophy are potentially important areas for understanding the nitrogen cycle at regional and global scales. Furthermore, UCYN-A1 and non-cyanobacterial diazotrophs, such as Clusters 1A and III, were also detected in a region influenced by the cold and/or N-rich water of the OSW and ISCW; nitrogen fixation activity was also detected in these waters, indicating the possibility of diazotrophy in the Sea of Okhotsk.
Acknowledgments
We are grateful to the captain, crew, and all participants of the R/V Hokko-maru of the Hokkaido National Fisheries Research Agency (FRA) for their assistance with sample collection at sea and on-deck incubation. We also thank S. Horii for her laboratory support and advice with regards to the isotope and DNA analysis, K. Inomata, and T. Kitahashi for MiSeq sequencing, and M. Hirai for their helpful comments on the molecular analysis. We also thank T. Kodama for the nutrient analysis. We would like to thank Editage (www.editage.com) for English language editing. This work was supported by a JSPS KAKENHI Grant Numbers 20J22451, 20H03059, and 16H04959.
Open Research
Data Availability Statement
The recovered sequences have been deposited in the DNA Data Bank of Japan Sequence Read Archive under accession number DRA011126. The data used in this study are available in the UTokyo Repository (http://hdl.handle.net/2261/00079986).