Possible association of diazotrophs with marine zooplankton in the Pacific Ocean

Abstract Dinitrogen fixation, the biological reduction in N2 gas to ammonia contributes to the supply of new nitrogen in the surface ocean. To understand the diversity and abundance of potentially diazotrophic (N2 fixing) microorganisms associated with marine zooplankton, especially copepods, the nifH gene was studied using zooplankton samples collected in the Pacific Ocean. In total, 257 nifH sequences were recovered from 23 nifH‐positive DNA extracts out of 90 copepod samples. The nifH genes derived from cyanobacteria related to Trichodesmium, α‐ and γ‐subdivisions of proteobacteria, and anaerobic euryarchaeota related to Methanosaeta concilii were detected. Our results indicated that Pleuromamma, Pontella, and Euchaeta were the major copepod genera hosting dinitrogen fixers, though we found no species‐specific association between copepods and dinitrogen fixers. Also, the digital PCR provided novel data on the number of copies of the nifH gene in individual copepods, which we report the range from 30 to 1666 copies per copepod. This study is the first systematic study of zooplankton‐associated diazotrophs, covering a large area of the open ocean, which provide a clue to further study of a possible new hotspot of N2 fixation.

Introduction into a form of nitrogen available to living organisms have been performed in studies examining the presence and diversity of N 2 -fixers in oceanic environments (Zehr et al. 1998). The most significant N 2 fixing microorganisms were thought to be filamentous cyanobacteria, such as Trichodesmium (Capone et al. 1997), symbiotic and freeliving unicellular cyanobacteria (UCYN, including Crocosphaera) (Zehr et al. 2001), and filamentous cyanobacteria of the order Nostocales that associate with diatoms (Carpenter et al. 1999). However, diverse nifH genes clustering into noncyanobacterial clades have also been reported from the surface waters down to below the euphotic zone in the open ocean (Zehr et al. 2003;Langlois et al. 2005;Riemann et al. 2010;Farnelid et al. 2011;Moisander et al. 2014). Although diazotrophic bacteria obtained from the open ocean have clustered into a wide range of bacterial groups, including α-, β-, γ-, and δproteobacteria and Firmicutes (Zehr et al. 2003), it has been speculated that many presently unidentified organisms may be active in ocean biological N 2 fixation.
As nifH genes have been previously reported from copepods, an association between them and N 2 fixing bacteria has already been demonstrated (Zehr et al. 1998;Braun et al. 1999;Scavotto et al. 2015). Our hypothesis built on this proposed association is that the guts of copepods provide a perfect anaerobic environment where N 2 -fixers could flourish. Whereas N 2 fixation by microbes in the alimentary (gut) tract of many terrestrial arthropods (insects) serves as a source of N (Ohkuma et al. 1996), few species of N 2 fixing organism have been identified or isolated from marine planktonic-arthropods (Proctor 1997;Zehr et al. 1998;Braun et al. 1999;Scavotto et al. 2015). Copepods, small marine arthropod crustaceans ranging in size from species <1 mm in length to those more than 10 mm, dominate mesozooplankton within the upper 600 m of the world oceans (Star and Mullin 1981), wherein they occur at densities typically ranging 1-100 individuals m −3 (Longhurst 1985). N 2 fixation in zooplankton-associated microbial communities could provide an important, little understood source of possible new N in the open ocean. Herein, we report novel data searching the presence, diversity, distribution, and abundance of possible N 2 -fixers associated with marine zooplankton over a wide area of the Pacific Ocean.

Study area and sample collection
In total, 140 zooplankton samples were collected from 12 locations in the Pacific Ocean during the cruises of R/V Mirai (MR-11-2), R/V Hakuho-maru (KH-11-10 and KH-13-7), and R/V Shinsei-maru (KS-13-T2) between December 2011 and January 2014 (Table 1; Fig. 1). The K2 and S1 stations were located in the subarctic and subtropical North Pacific, respectively. The ALOHA station was located in the subtropical North Pacific. The KT station was located in the Kuroshio Current area. The other eight stations were located in the tropical and subtropical areas of the South Pacific. Samples were collected by North Pacific Standard Net (NORPAC net, 100 μm mesh) towed at 0.7 m s −1 , vertically hauled from a depth of 200 m to the surface (200-0 m). Net contents were washed with 0.2 μm-filtered seawater. Selected copepod samples from station 5, hereafter referred to as "empty-gut" copepods, were incubated for 72 h in 0.2 μm filtered seawater to evacuate gut contents. From sample KT-8 only gut of the respective copepod was separated aseptically referred as only "gut sample." Zooplankton was sorted and identified using conventional light microscopy (Leica Wild MZ-8, Vashaw Scientific, Inc., Norcross, GA) and taxonomic guides, then stored individually at −30°C until analysis.

Hydrographic parameters and nutrients
Samples for salinity, dissolved oxygen (DO), and macronutrient analysis (NO − 3 , NO − 2 and PO 3− 4 ) were collected in acid-cleaned Teflon-coated 12-L Niskin-X bottles on a CTD (Conductivity, Temperature, Depth) carousel system attached at the end of titanium-armored cable. Temperature and salinity profiles were determined with SBE 911 (Sea-bird Electronics, Inc., Bellevue, WA). DO concentrations were measured by Winkler titration using an automatic titrator (806 Titrando, Metrohm AG, Herisau, Switzerland). During the R/V Hakuho-maru cruises, nanomolar macronutrient concentration levels were determined by liquid wave guide spectrophotometer (Hashihama et al. 2009), for which detection limits of NO − 3 + NO − 2 and PO 3− 4 were 3 nmol L −1 , whereas those made during the R/V Mirai cruise were determined by standard colorimetric methods.

Molecular analysis of nifH
DNA was directly extracted from a copepod and its associated microorganisms using a Qiagen Blood and Tissue Kit (Qiagen, Hilden, Germany), in accordance with manufacturer protocols. Six samples from KH-27 to KH-32 represent extracts from copepods with an empty-gut and the KT-8 represents an extract only from the gut of the respective copepod ( Table 2). The DNA concentration of samples was measured using a Quant-iT ™ Picogreen ® dsDNA Reagent and Kit (Invitrogen, Carlsbad, CA) and Microplate reader (SH-9000; Corona Electric, Ibraraki, Japan). Partial nifH fragments were amplified from the DNA of an individual sample by the nested PCR (Zehr et al. 2001). Two degenerate oligonucleotide PCR primer sets were used to amplify approximately 350-bp segments of the nifH gene (Zehr and McReynolds 1989;Zani et al. 2000). Each reaction contained 0.05 μL of 5U μL −1 EX Taq HS (TaKaRa, Tokyo, Japan), 1.0 μL of 10× ExTaq Buffer (TaKaRa), 0.8 μL of 0.2 mmol L −1 dNTP mixture (TaKaRa), 1.0 μmol L −1 each primer, 2 μg BSA (TaKaRa), and 1 μL of template DNA in a final volume of 10 μL. The first and second PCRs were run in triplicate; the first, 40 cycles at 95°C for 3 min, followed by the second, 35 cycles at 98°C for 10 sec, 54°C for 30 sec, 72°C for 30 sec, with a final extension at 72°C for 7 min.
Amplification of nifH was checked using 3 μL of the second PCR product using 1.5% agarose gel. No visible band was detected from the negative control in PCR reactions. Samples that produced a 350-bp visible product were extracted and gel purified (Qiagen PCR purification kit) and cloned using TOPO TA Cloning Kit (Invitrogen) for sequencing, transferred into E. coli DH5α competent cells (TaKaRa).
Clones were screened to identify those with the correct insert, and 24 colonies from each sample were selected for DNA sequencing with the BigDye v.3.1 Sequencing   Kit (Applied Biosystems, Foster City, CA). Sequencing of nifH was performed using an ABI 3130 genetic analyzer in accordance with manufacturer protocols. Vector and primer sequences were removed manually. The nifH sequences were analyzed by BLAST searches against the National Center for Biotechnology Information (NCBI) database. Clones sharing 100% similarity in amino acid sequences were clustered into the same phylotype using the CD-HIT program (Huang et al. 2010). Two hundred and fifty-seven representative sequences were aligned with MUSCLE (Edgar 2004) in the Molecular Evolutionary Genetics Analysis (MEGA ver. 5) software package (Tamura et al. 2011). The closest sequences of uncultured and cultured bacteria identified from BLAST comparisons were added to the dataset. A distance-based neighbor-joining phylogenetic tree was constructed. Bootstrap values were determined from 1000 iterations.

Quantification of nifH
For determining the copy number of nifH genes associating with an individual copepod, we used droplet digital PCR (ddPCR) system (Bio-Rad Laboratories, Hercules, CA). Six specific primer and probe sets were designed to detect the nifH gene from different clusters ( Fig. 2; Table 3) using GenScript Real-time PCR primer design software (https://www.genscript.com/ssl-bin/app/primer). Among γ proteobacterial nifH sequences, three subclades were designated (γ − 1, γ − 2 and γ − 3). From these three subclades, three reference sequences were selected for the design of primers and probes. From αproteobacteria, cyanobacteria and archaea, one reference sequence from each clade was selected ( Table 3). Specificity of these six primer and probe sets was tested by applying them to our cloned nifH sequences in qPCR (DNA Engine OPTICON2, Bio-Rad Laboratories). Standards for the different nifH clusters were achieved by cloning nifH amplicons using the TOPO TA Cloning Kit (Thermo Fisher Scientific Inc, Waltham, MA). Plasmid DNA was purified using PureLink Quick Plasmid Miniprep Kit (Thermo Fisher Scientific Inc, Waltham, MA) in accordance with manufacturer protocols. Plasmid concentration were determined by Quant-iT ™ Picogreen dsDNA Reagent and Kit (Thermo Fisher Scientific Inc) and a Microplate reader (SH-9000; Corona Electric). The number of nifH genes (copies μL −1 ) was determined using ddPCR system, in accordance with manufacturer protocols. Each PCR reaction contained 10 μL 2× ddPCR supermix, 1.8 μL of 10 μmol L −1 forward and reverse primer, 0.5 μL of 10 μmol L −1 probe, and 3 μL of template DNA in a final volume of 20 μL. Droplets then were prepared by QX100 Droplet Generator (Bio-Rad Laboratories). Duplicate runs were performed for all ddPCR. No template control was run in duplicate for each reaction. Cycling conditions of ddPCR were: 95°C for 10 m followed by 40 cycles of 94°C for 30 sec, 56°C for 60 sec and finally 98°C for 10 m. After PCR reaction, fluorescence measurement was performed by QX100 Droplet Reader (Bio-Rad Laboratories). The copy number of nifH genes (copies μL −1 ) was determined by QuantaSoft software (Bio-Rad Laboratories). The digital PCR detects one copy in one reaction. We added 3 μL of DNA extracts in one reaction solution. This is why the minimum copy number per a unit volume (μL) was less than one. Because the final volume of DNA extracts obtained from an individual copepod was 50 μL, the minimum copy number we could detect in this study was 17 copies in one copepod. The copy number per a unit volume (copies μL −1 ) was multiplied by 50 for determining the total nifH copy number of the individual copepod.

Nucleotide sequence accession numbers
Sequences recovered from this study have been deposited in the DNA Data Bank of Japan (DDBJ) with accession numbers LC012980-LC013236.
Totally 257 nifH clones obtained from 23 nifH-positive samples were sequenced. The sequences were clustered into 30 phylotypes (Fig. 2). Most (163 out of 257) nifH sequences were grouped with γproteobacterial ones. Many of them (82) were from the subtropical South Pacific samples (KH: 1, 4, 5, 11-14 and 27) and were closely related to nifH sequences of Vibrio natriegens and V. diazotrophicus (Fig. 2). Also, 24 γproteobacterial sequences from the Kuroshio region (sample KT-8) were related to the nifH of Teredinibacter turnerae, whereas 36 γproteobacterial sequences from the tropical North Pacific (samples from KH-19 to KH-21 and KH-35) were related to the nifH sequences of Thiocapsa bogorovii and T. marina (Tourova et al. 2009). All 50 αproteobacterial sequences were recovered from samples of empty-gut copepods from the tropical South Pacific (samples from KH-28 to KH-32) were closely related to the nifH of Azospirillum lipoferum (Doroshenko et al. 2007). Thirty-three nifH sequences obtained from three copepods (AL-26, KH-34 and KH-37) were grouped into a cluster of cyanobacteria (Zehr et al. 2003); 27 were related to the nifH of Katagnymene spiralis (Lundgren et al. 2001) and six were related to that of Trichodesmium (Zehr et al. 1998). Eleven sequences from the sample KH-22 were closely related to the nifH sequence of euryarchaeota, Methanosaeta concilii (Chien and Zinder 1996). No nifH sequences of unicellular cyanobacteria (UCYN-A, UCYN-B, and UCYN-C) were identified from any sample.

Discussion
Most of our sequences (213 of 257) were similar to those of heterotrophic bacteria, suggesting their dominance in copepod-associating N 2 fixing communities in oligotrophic tropical and subtropical Pacific Ocean waters (Fig. 2). βand γproteobacterial nifH sequences have been reported from calanoid copepods in the Gulf of Mexico (Zehr et al. 1998), γproteobacterial nifH have been reported from copepods in coastal waters of the North Atlantic Ocean (Scavotto et al. 2015) and α-, βand γproteobacterial nifH from copepods and euphausids in the Caribbean and western Bahama Sea (Braun et al. 1999). In this study, γproteobacterial phylotypes were more abundant than αproteobacterial phylotypes. Table 3. Primer and probe sets of ddPCR of nifH gene of γ-proteobacteria, α-proteobacteria, cyanobacteria, and archaeal clusters as described in Fig. 2. Some of our cyanobacterial nifH sequences clustered to those of Trichodesmium and K. spiralis, which have been previously recovered from the seawater samples of the Pacific (Lundgren et al. 2001 andOrcutt et al. 2002), in addition to copepods from a shallow eutrophic lake (Yan et al. 2011). All clones recovered from the ALOHA station sample were clustered to reported cyanobacterial nifH sequences, suggesting the type of N 2 -fixers associating with copepods are influenced by those in the surrounding water column.

Type
Eleven nifH sequences recovered from Pontella in the equatorial tropical Pacific Ocean were clustered with the nifH sequence of Methanosaeta concilii, an obligate anaerobic methanogenic euryarchaeota (Patel and Sprott 1990). These were aligned in the cluster II nifH sequences described by (Zehr et al. 2003). Sequences of this cluster have been recovered from anoxic environments including marine sediments, soils, and termite guts (Ueda et al. 1995;Zehr et al. 1995 andOhkuma et al. 1996). It has been well-known that methanogens are present in guts of copepods (Marty 1993;DeAngelis and Lee 1994). Also, some methanogens have reportedly possessed nifH genes and shown diazotrophic growth (Reeve 1992). Given these previously reported evidences, it is reasonably speculated that N 2 -fixing methanogens reside within the copepod gut. Our finding of nifH sequences related to methanogens supports this speculation.
All nifH sequences from empty-gut copepods at St-5 except KH-27 clustered with Azospirillum spp (Doroshenko et al. 2007) and related αproteobacteria (Fig. 2). Scavotto et al. (2015) found gamma proteobacterial nifH sequences and N 2 fixing activity in empty-gut copepods from coastal waters of the North Atlantic Ocean, suggesting these bacteria represent permanent diazotroph associations. Alpha proteobacteria found from empty-gut copepods may be more permanently associating diazotrophs with copepods than other bacteria.
Abundance of diazotrophs associated with marine copepods may be related to host body size, as more were found in the larger Pontella than Euchaeta, of body size 5.9-6.2 mm, and 2.5-3.5 mm, respectively (Bradford-Grieve 1999). Our results also suggest that the abundance of diazotrophs might vary within the copepod genus, given the highest and lowest number of copies of nifH genes were recovered from Pontella (Fig. 3A). A possible explanation is that the variation in nifH gene copy is related to the abundance of diazotrophs in waters in which copepods occur. The degree of diazotroph association may change depending on spatial variability in abundances of both copepods and diazotrophs in water.
Although marine diazotrophic abundance in water samples in the Pacific have been reported, ours is the first to quantify nifH gene copy number in individual copepods. Real-time PCR was reportedly used to quantify UCYN-A nifH copy number in copepod samples (Scavotto et al. 2015). Although 10 copepods (Acartia) were pooled to extract DNA or RNA, most of them were below detection and only one was detectable but not quantifiable. Generally speaking, ddPCR system provides orders of magnitude more precision and sensitivity than real-time PCR (Hindson 2011). We did not try to use a conventional real-time PCR in this study, however, ddPCR may be a better choice than the real-time PCR to quantify low abundance of nifH gene in copepod samples. Abundance of γ proteobacterial nifH gene was reportedly 2.5 × 10 3 in South Pacific waters (Mahaffey et al. 2005) and 6.8 × 10 3 copies L −1 in North Pacific waters (Bombar et al. 2013). Abundance of Trichodesmium nifH gene was 1.4 × 10 5 copies L −1 in South Pacific waters (Mahaffey et al. 2005). Abundance of nifH genes found in one copepod (30 to 1.7 × 10 3 copies ind −1 ) was lower than that found in 1 L of seawater.
Copepods are among the most abundant of invertebrates in marine mesozooplankton (Longhurst 1985), and are responsible for consuming most oceanic phytoplankton. Any gut microflora would be continuously supplied with energy-rich substrata for microbial metabolism. Although the nitrogenase enzyme is inhibited by oxygen, the anoxic conditions found within the copepod gut could prevent enzyme deactivation, rendering the gut an ideal environment for nitrogen fixation. In general, N 2 fixation in the ocean is limited by iron availability (Falkowski 1997). Copepod gut tracts undergo pH (Patel and Sprott 1990) and redox changes during feeding and digestion that could be important for increasing the bioavailability of trace elements (such as iron) for N 2 -fixers. The feeding habits of copepods solubilize phytoplankton cellular iron (Hutchins et al. 1995), therefore providing a suitable mechanism for making iron available to N 2 fixing gut microflora.
To our knowledge, this is a first systematic study of copepod or zooplankton-associated diazotrophs, covering a large area of the Pacific Ocean. All nifH-positive copepod samples with diverse N 2 fixing prokaryotes were collected from tropical and subtropical waters throughout the Pacific Ocean (Fig. 1). Although our study found no speciesspecific relationship between individual copepod and bacterial texa in our research, we do document presence, diversity and the abundance of open-ocean diazotrophs associated with marine copepods. In order to reveal the contribution of these zooplankton-associated diazotrophs to oceanic nitrogen cycles, the determination of N 2 fixation rates associating with individual copepods should be required in a further study.

Supporting Information
Additional supporting information may be found in the online version of this article: Table S1. Numbers of analyzed zooplankton samples. Note: only one nifH-positive copepod was found during cruises MR-11-2 and KH-11-10; only copepods were analyzed from cruises KS-13-T2 and KH-13-7. Table S2. Taxonomic identity of zooplankton samples from Pacific Ocean stations.*Copepod (genus). Figure S1. Photomicrographs of three major nifHpositive copepod genus: A, Pontella; B, Euchaeta; C, Pleuromamma.