Distribution Patterns and Phylogeny of Marine Stramenopiles in the North Pacific Ocean

Abstract

Marine stramenopiles (MASTs) are a diverse suite of eukaryotic microbes found in marine environments. Several MAST lineages are thought to contain heterotrophic nanoflagellates. However, MASTs remain uncultured and data on distributions and trophic modes are limited. We investigated MASTs in provinces on the west and east sides of the North Pacific Subtropical Gyre, specifically the East China Sea (ECS) and the California Current system (CALC). For each province, DNA was sampled from three zones: coastal, mesotrophic transitional, and more oligotrophic euphotic waters. Along with diatoms, chrysophytes, and other stramenopiles, sequences were recovered from nine MAST lineages in the six ECS and four CALC 18S rRNA gene clone libraries. All but one of these libraries were from surface samples. MAST clusters 1, 3, 7, 8, and 11 were identified in both provinces, with MAST cluster 3 (MAST-3) being found the most frequently. Additionally, MAST-2 was detected in the ECS and MAST-4, -9, and -12 were detected in the CALC. Phylogenetic analysis indicated that some subclades within these lineages differ along latitudinal gradients. MAST-1A, -1B, and -1C and MAST-4 size and abundance estimates obtained using fluorescence in situ hybridization on 79 spring and summer ECS samples showed a negative correlation between size of MAST-1B and MAST-4 cells and temperature. MAST-1A was rarely detected, but MAST-1B and -1C and MAST-4 were abundant in summer and MAST-1C and MAST-4 were more so at the coast, with maximum abundances of 543 and 1,896 cells ml⁻¹, respectively. MAST-4 and Synechococcus abundances were correlated, and experimental work showed that MAST-4 ingests Synechococcus. Together with previous studies, this study helps refine hypotheses on distribution and trophic modes of MAST lineages.

INTRODUCTION

Marine heterotrophic nanoflagellates (HNFs) are small unicellular eukaryotes typically ranging from 2 to 20 μm (13, 42, 46). They are thought to be important grazers of bacteria and picophytoplankton (diameter, ≤2 to 3 μm) and to contribute to nutrient cycling (1). Results from environmental 18S rRNA gene clone library studies indicate that many HNFs are novel marine stramenopiles (MASTs) (25, 26). MASTs are composed of 12 independent phylogenetic clusters (MAST-1 to MAST-12), none of which appear to be represented by cultured isolates. Among the MAST groups, MAST-1, -3, -4, and -7 have been found in both open-ocean and coastal systems, on the basis of 18S rRNA gene clone libraries (24, 26, 31). The trophic roles of each MAST group are not necessarily known. The MAST-3 cluster appears to contain parasites, based on the fact that Solenicola setigera, a parasite of diatoms (Leptocylindrus mediterraneus), groups with the MAST-3 clade (15, 16). MAST-1 and -4 contain HNFs that actively consume prey (27). Rates of MAST-1C and -4 consumption of fluorescently labeled bacteria (FLB) were estimated to be 3.6 and 1 to 1.5 bacteria per predator per hour, respectively, in coastal waters. Few studies have quantified MAST populations in the natural environment; combined counts of MAST-1 and MAST-4 together have ranged from 2 to 658 cells ml⁻¹ in the world oceans, accounting for 5 to 35% of HNFs in the <5-μm size fraction (26). MAST-1, -4, -6, and -7 have been shown to be active predators on bacteria and/or picophytoplankton, based on microscopic observations and isotope labeling (14, 27, 36).

Relatively few studies have investigated the abundance and 18S rRNA gene diversity of eukaryotes and, more specifically, MASTs in the North Pacific Ocean. The North Pacific is characterized by a number of different provinces (21) that are influenced by the largest of these, the North Pacific Subtropical Gyre (NPSG). Some of the most productive provinces are observed in regions heavily influenced by boundary currents in the eastern and western North Pacific Ocean. These regions can be highly complex. For example, the East China Sea (ECS) is the largest marginal sea in the western North Pacific and receives input from coastal and oligotrophic sources. Coastal regions of the ECS are influenced by inflow from the Changjiang River (also known as the Yangtze River), the fifth largest river in the world, and by anthropogenic activities that have resulted in the loading of inorganic nutrients and dissolved organic matter. This is thought to have led to eutrophication and an increase in microbial abundance and biomass in the microbial food web (45). The hydrographic characteristics of the ECS are also influenced by the Kuroshio (a western boundary current to NPSG), Yellow Sea coastal water, and Taiwan Warm Current. In contrast, the California Current (CC) system (CALC), a region which is less enclosed by landmasses than the ECS, is hydrographically dominated by the NPSG eastern boundary current, the California Current. This dynamic region experiences strong seasonal variation in upwelling, is susceptible to influences from El Niño Southern Oscillation events, and undergoes mesoscale variations in the form of fine filaments and eddies which occur between the hydrographic current boundaries of the California and Davidson currents (reviewed in reference 9). Taken together, this region is known as the California Current system (9) and is within the CALC province, as defined by Longhurst (21). With respect to the diversity of small eukaryotes, two studies have reported on the diversity of small eukaryotes in the South China Sea (5, 6). In the CALC, picoeukaryote 18S rRNA gene diversity has been investigated with a strong emphasis on photosynthetic taxa (10, 47).

Here, we investigate MAST groups in two North Pacific provinces. We quantified MAST-1A, -1B, -1C, and -4 cells with specific fluorescent in situ hybridization (FISH) probes and explored their relationship with environmental parameters and other microbes in data from two ECS cruises. Prey ingestion was tested in prey addition experiments. The diversity of MAST populations was investigated in euphotic zone waters of both the ECS and CALC provinces. Comparisons of 18S rRNA gene sequences with those from previous studies, along with enumeration, were used to explore the ecology and distribution of these uncultured eukaryotes.

MATERIALS AND METHODS

Study sites and general sampling.

Three cruise transects were sampled, one in the eastern North Pacific Ocean (CALC) and two in the western North Pacific Ocean (ECS) (Fig. 1). Contextual data were collected at 25 stations in the CALC along California Cooperative Fisheries Investigation (CalCOFI) Line 67 between 1 and 10 October 2007 on the R/V Western Flyer. Three of these stations, representing different water masses, were also sampled for DNA and other biological measurements (Fig. 1B). The ECS cruises were performed on the R/V Ocean Researcher I, and samples were taken along a grid over the continental shelf in the ECS during late spring (29 April to 13 May 2009) and summer (29 June to 15 July 2009), with data collected at a total of 32 and 47 stations, respectively (Fig. 1C and D). Samples from all cruises were collected with either 20-L Go-Flo or Niskin bottles. The vertical profile of the hydrographic conditions, including temperature, salinity, density, and dissolved oxygen, was recorded by instruments mounted on the rosette, including a conductivity, temperature, and depth (CTD) sensor using a CTD collector (SeaBird SBE9/11plus).

Fig 1.

Study regions and sampling sites in the Central California and eastern Pacific Ocean provinces. (A) Location of the study regions within the North Pacific Ocean; (B) CALC sites sampled in October 2007; main stations are represented by white dots. (C and D) ECS sites sampled from April to May 2009 (C) and June to July 2009 (D). The color background represents sea surface salinity according to the gradient bars provided. Note the different scales between the top and bottom panels.

Nucleic acid and microscopy sampling.

At CALC sites, water was filtered through a 20-μm-mesh-size nylon mesh net and then onto 3-μm-, 0.8-μm-, and 0.1-μm-pore-size 293-mm filters in series as described previously (11). Microbial biomass from three ECS sites (station [St] 12, St 19, and St 23) was collected for DNA extraction by filtering between 1.5 and 9 liters of seawater through a 20-μm-mesh-size nylon mesh net and subsequently prefiltering it through a 5-μm-pore-size filter and onto a 0.8-μm pore-size 47-mm filter (Whatman). Filters with microbial biomass from the ECS or the CALC were stored at −75°C and −80°C, respectively. FISH samples were collected without a prefiltration step from surface water (2 or 3 m) at all ECS stations, preserved with 37% formaldehyde (final concentration, 3.7%), and incubated at 4°C for 1 to 24 h. Subsequently, 100-ml (for coastal stations) or 200-ml (other stations) subsamples were filtered onto a 0.8-μm-pore-size polycarbonate membrane (47 mm) and then stored at −80°C.

Nutrient and Chl a measurements.

CALC nutrient and chlorophyll a (Chl a) data and the data collection methodology have been reported previously (32). In the ECS, macronutrients were measured according to previous studies (30, 33) and modified according to Gong et al. (17). One to 2 liters of seawater was filtered onto 25-mm-diameter GF/F membranes (Whatman) for determining ECS Chl a concentrations (35). For both sites, Chl a concentrations were measured with a fluorometer (Turner Designs) after extraction.

Flow cytometric analyses for picoplanktonic abundance.

Two-milliliter ECS samples were fixed using paraformaldehyde (final concentration, 0.2%) for 15 min in the dark, frozen in liquid nitrogen, and stored at −75°C for later analysis. Enumeration of Synechococcus and photosynthetic picoeukaryotes was performed on a FACSAria flow cytometer (Becton Dickinson). Heterotrophic bacteria were enumerated by staining a 1-ml subsample with SYBR green (Molecular Probes, Inc.) at a 1:10,000 dilution and incubated in the dark for 15 min. Synechococcus was determined by its characteristic orange fluorescence, while photosynthetic picoeukaryotes were counted using red fluorescence and scatter properties (23). Calibration beads (1-μm yellow-green fluorescence beads) were added to each sample as an internal reference. All flow cytometric data were acquired for 2 min, and the flow rate ranged from 0.020 to 0.031 ml min⁻¹.

Microscopy.

MAST-1 and -4 cells were labeled with specific FISH probes. Different probe sequences (published previously) were used for enumerating three distinct clades within the MAST-1 cluster: NS1A (5′-ATTACCTCGATCCGCAAA-3′), NS1B (5′-AACGCAAGTCTCCCCGCG-3′), and NS1C (5′-GTGTTCCCTAACCCCGAC-3′) (26). The NS4 probe (5′-TACTTCGGTCTGCAAACC-3′) was used to enumerate the MAST-4 cluster (25). Five ECS coastal stations (St 19 to 23) were tested using a negative-control probe based on the antisense sequence of NS4. For all hybridizations, a portion of each filter was incubated at 46°C for 3 h with hybridization buffer (30% formamide, 900 nM NaCl, 20 mM Tris-HCl, and 0.01% SDS) with oligonucleotide probes (final concentration, 5 ng μl⁻¹) containing a Cy3 moiety at the 5′ end. After hybridization, the filters were transferred into a wash buffer (110 mM NaCl, 20 mM Tris-HCl, 5 mM EDTA, and 0.01% SDS) and incubated at 48°C for 20 min (26). Finally, the filter was overlaid with a mixture of 1 μg ml⁻¹ diamidino-2-phenylindole (DAPI; final concentration) and antifading reagent (Citifluor Ltd., London, United Kingdom). Probe-positive cells were identified by their Cy3 fluorescence under green light excitation. By switching UV, blue, and green light excitation, MAST-1 and -4 cells could be differentiated from Chl a-containing eukaryotes. Enumeration of total HNFs was carried out with a different section of the same filter according to the method of Porter and Feig (37). To determine the size of MAST cells, we measured the length and width and then converted these measurements into equivalent spherical diameter (ESD). The biovolume-carbon conversion factor for MAST-1 and -4 was estimated to be 183 fg carbon μm⁻³ based on results presented elsewhere (3). Three ocular lines of each FISH slide were inspected under epifluorescence at ×1,000, and the average area of each line scanned was 1.32 ± 0.04 mm². Pearson's correlation was performed with SPSS software (SPSS Inc., Chicago, IL) to evaluate MAST abundances and relationships with environmental and biological factors.

Prey observation of MAST-4 by tracer addition.

To explore MAST-4 prey ingestion, a grazing experiment was conducted utilizing fluorescently labeled Synechococcus (FLS) at St 24 during the ECS summer cruise. FLS was prepared on the basis of methods described by Sherr and Sherr (43) and stored at −20°C until use at sea. Synechococcus (sp. strain WH7803) was grown at 25°C in f/2 medium (18). FLS was added to 1-liter polycarbonate bottles containing natural seawater and incubated for 40 min in an on-deck incubator with running seawater. Two hundred-milliliter and 400-ml FISH samples were collected at time zero (T₀) and after 40 min (T₄₀), respectively. The food vacuole content of 50 MAST-4 cells at T₀ and 100 MAST-4 cells at T₄₀ was inspected by epifluorescence microscopy. FLS could not be distinguished from natural Synechococcus cells; thus, counts represent both cell types.

DNA extraction and clone libraries.

DNA was collected from surface waters at three ECS stations (St 12, St 19, and St 23) and three CALC stations (St H3, St 67-70, and St 67-155) and from the deep chlorophyll maximum (DCM) at St 67-155, using the filtration methods detailed above. DNA was extracted from CALC samples using a sucrose-based procedure as described previously (11). ECS samples were treated as follows: cells on the membrane were disrupted using a lysis buffer (0.1 M EDTA [pH 8], 1 mM Tris HCl [pH 8], 0.25% SDS, and 0.1 mg ml⁻¹ proteinase K) with gentle shaking at 55°C overnight. Subsequently, polysaccharides were removed using cetyltrimethylammonium bromide for 15 min at 65°C. Genomic DNA was purified by phenol-chloroform extraction and finally dissolved in 50 μl Tris-EDTA (pH 7.5) (7, 8). The concentration and purity of DNA were measured using a spectrophotometer (Nanodrop).

The 18S rRNA gene-specific PCR primers used on ECS (28) and CALC (29) samples amplified the same nearly full-length gene product. Forward and reverse primer sequences for ECS samples (5′-AACCTGGTTGATCCTGCCAGTA-3′ and 5′-GATCCTTCTGCAGGTTCACCTAC-3′) were similar to those used on CALC samples (5′-ACCTGGTTGATCCTGCCAG-3′ and 5′-TGATCCTTCYGCAGGTTCAC-3′). PCR conditions for ECS samples were as follows: 95°C for 2 min, followed by 30 cycles of denaturing at 95°C for 45 s, annealing at 55°C for 1 min, and extension at 68°C for 2 min, with a final extension step at 68°C for 10 min with Advantage II DNA polymerase (Clontech). About 50 positive colonies were randomly picked from each library. ECS clones were sequenced using vector-targeted primers T7 and SP6, each rendering ca. 1,000-bp reads, which were then assembled. PCR conditions for CALC samples, cloning, and sequencing were performed as described in reference 11. Between 1,200 and 2,000 clones were sequenced at each site using vector-targeted primers M13F and M13R, as well as primers internal to the PCR product, 502F and 1174R (47). Sequencing was performed on a Prism 3100 or 3730xl genetic analyzer (Applied Biosystems).

Sequence and phylogenetic analyses.

Stramenopile sequences were identified and selected for further analysis. Sequences were subjected to NCBI BLAST analysis to screen for stramenopile-derived sequences, and BLAST analysis was used to compare the sequences against those in the SILVA database, which contains 18S rRNA gene sequences representing all known stramenopile groups. Sequences with ≤97% nucleotide identity to their closest relatives on NCBI BLASTn analysis were further checked for chimeras by breaking each into several short fragments and assessing the apparent origin of each fragment. Sequences were aligned using the SINA aligner tool and the SILVA SEED database (http://www.arb-silva.de/aligner/) (39). The alignment was manually adjusted according to secondary structure in the ARB software environment (22). Prior to phylogenetic analysis, gaps and regions where the alignment was ambiguous based on inspection by eye were removed. Trees were inferred using neighbor-joining (NJ) and maximum-likelihood (ML) methods using 100 bootstraps in the PHYLIP program, version 3.69 (12). The nucleotide substitution model for ML was selected with the jModelTest tool by the Bayesian information criteria (38).

Nucleotide sequence accession numbers.

Sequences from this study were deposited in GenBank under accession numbers JQ781881 to JQ782099.

RESULTS AND DISCUSSION

Characteristics of study sites.

MAST diversity was investigated at oceanic provinces on either side of the NPSG. The ECS lies over the Asian continental shelf, while the CALC region is off Central California but extends northwards to Alaska and south to Baja California. The latter includes the eastern boundary current of the NPSG and a highly productive inshore upwelling region. In each of these provinces, three sites were used to represent different habitats for MAST populations. Specifically, we sampled inshore at highly productive sites ECS St 19 and MBARI Time Series St H3, with the latter being a midbay site in Monterey Bay. We also sampled at transitional stations under the influence of intersecting water masses, ECS St 23 and CalCOFI Line 67 St 70 (St 67-70). Finally, we sampled a more oligotrophic site in each region: in the ECS, St 12 was on the shelf break and influenced by the Kuroshio, and in the CALC, the station (St 67-155) was located at the NPSG-California Current boundary.

The major stations in the two provinces had distinct characteristics and were subject to different seasonal influences. Surface water temperatures in the ECS ranged from 15.8 to 25.3°C in spring and 23.3 to 29.6°C in summer 2009, while salinity ranged from 27.9 to 34.7 and 23.8 to 34.1 in spring and summer, respectively (Fig. 1; see Fig. S1 in the supplemental material). Hydrographic features could be a summarized as follows. In spring, the China Coastal Current, low in both temperature and salinity (temperature, <20°C; salinity, ≤31), flowed along the coast of the mainland and warm, higher-salinity Kuroshio (>20°C, >34 psu) water entered onto the shelf from the southeastern shelf break. Cold, saline Yellow Sea waters (<20°C) intruded into ECS from the north. St 23 was located midshelf and was heavily influenced by Yellow Sea waters (Fig. 1C). Hydrographic features changed dramatically by the time of the summer cruise, when two water masses dominated, a warm stream of the Taiwan Warm Current (>26°C), which flowed northward from Taiwan Strait, and water diluted by Changjiang River inputs (≤31), resulting in mixed seawater and freshwater discharge (Fig. 1D) induced by the southwestern monsoon. Coastal St 19 was situated in this area and presumably subject to the fluctuations in surface (freshwater) runoff, especially during summer. Even St 23 was influenced by the lower-temperature, lower-salinity Changjiang River waters during the summer cruise (Table 1; Fig. 1D).

Table 1.

Environmental data corresponding to 18S rRNA gene clone library samples

Date (day-mo-yr)	Station	Depth (m)	Longitudea (°E)	Latitudea (°N)	Temp (°C)	Salinity	Chl a concn (mg m−3)	Concn (μM)				No. of MASTs/no. of stramenopilesc
								NH4b	NO2	NO3b	PO4
10-Oct-07	H3	5	−122.02	36.74	12.28	33.47	4.19	NMd	0.31	8.86	1.12	47/75
9-Oct-07	67–70	10	−123.49	36.13	15.57	33.12	2.72	NM	0.05	0.51	0.61	33/34
7-Oct-07	67–155	5	−129.43	33.29	19.02	33.19	0.10	0.02	0.07	0.01	0.66	33/34
6-Oct-07	67–155	86	−129.43	33.29	13.39	33.13	0.94	0.02	0.13	0.40	0.58	54/56
7-May-09	19	2	123.15	31.63	16.79	29.61	3.32	1.0	0.27	7.9	0.07	2/3
7-May-09	23	2	126.23	30.46	16.51	33.04	0.64	1.0	0.02	1.0	0.03	0/0
2-May-09	12	2	125.13	27.55	22.97	34.53	0.28	1.8	0.00	0.0	0.01	3/3
2-Jul-09	19	2	123.14	31.63	24.52	30.38	5.13	0.9	0.36	3.8	0.28	1/3
5-Jul-09	23	2	126.23	30.46	25.71	31.30	0.56	0.3	0.03	0.0	0.02	12/13
9-Jul-09	12	2	125.12	27.49	28.51	33.90	0.14	0.7	0.00	0.0	0.03	1/1

^aSpecified in decimal degrees.

^bThe numbers of significant digits were 1 decimal place in the ECS and 2 decimal places in the CALC.

^cThe number of MAST clones relative to the number of stramenopile clones.

^dNM, not measured.

Overall, sites along Line 67 had salinity within a tighter range than the ECS stations (Fig. 1). The warmest sites encountered over the 3 cruises were during the ECS summer cruise, and conversely, the St H3 and 67-70 surface waters were cooler than those at any of the ECS sites (see Fig. S1 in the supplemental material). Along Line 67, water temperatures were relatively low inshore and became warmer toward the open ocean (see Fig. S1B in the supplemental material). They were also more saline in Monterey Bay, freshened slightly over the transition into the California Current (CC), and were more saline again beyond the core of the CC (Fig. 1B). Conditions observed in the CALC during October 2007 were consistent with a relaxation in seasonal upwelling (9), including relatively warm temperatures (although they were still cooler than ECS sites) at coastal and midbay stations. Surface NO₃ concentrations were higher at the midbay station (St H3; 8.86 μM) than any other CALC or ECS stations for which clone libraries were constructed (Table 1). However, this was considerably lower than at some ECS sites where MAST cells were enumerated during cruises (e.g., ECS St 30, 25.3 μM; see Table S1 in the supplemental material). Surface NO₃ concentrations at St 67-155 (0.01 μM) were similar to values observed at Station ALOHA in the NPSG (20), and St 12 also had relatively low NO₃ concentrations. Although PO₄ concentrations were considerably higher at all CALC stations than those in the ECS, NH₄ concentrations were lower. Chl a concentrations were highest at inshore stations (St 19 and H3) and decreased moving off shore (Table 1).

Distribution of MAST-1 and MAST-4 cells in ECS.

Maximum abundances of MAST-1A, -1B, and -1C were 14 cells ml⁻¹, 114 cells ml⁻¹, and 543 cells ml⁻¹, respectively (Table 2). No cells were detected with the negative (antisense) probe when applied to environmental samples. Very few MAST-1A cells were detected at the ECS stations, and none were detected in the plume area (n = 18), with the exception of a single station (St 29) in spring (see Table S1 in the supplemental material). In contrast, MAST-1C cells were observed frequently and at greater abundances (Fig. 2C and D; Table 2; see Table S1 in the supplemental material). MAST-1C contributed 1% of the total HNFs detected in both ECS cruises. The maximum abundance of MAST-4 occurred in the Changjiang River plume during both cruises and extended to the middle of the continental shelf during spring (Fig. 2E and F). MAST-4 abundance ranged from 12 to 712 cells ml⁻¹ (average, 187 ± 29 cells ml⁻¹, n = 32) and 8 to 1,896 cells ml⁻¹ (average, 264 ± 60 cells ml⁻¹, n = 47) in spring and summer, respectively. A significantly positive relationship between the abundance of MAST-1B and MAST-4 was observed (r = 0.39, P < 0.001, n = 79) (Table 3). Notably, four 18S rRNA gene sequences in our libraries (represented by 155D1Ae4Eh in Fig. 7A below) had a mismatch with the MAST-4 probe. Likewise, the 3 probes targeting the 3 different MAST-1 subclades first described (MAST-1A, -1B, -1C) probably do not account for cells within a fourth clade defined here as an additional sublineage, MAST-1D (represented by clone 155S8Be86f; see below), due to mismatches. These mismatches may have resulted in underestimates of overall MAST-1 and MAST-4 abundance.

Table 2.

Abundance and size of MAST-1 and -4 cells in the ECS

Parameter	MAST-1A		MAST-1B		MAST-1C		MAST-4
	Spring	Summer	Spring	Summer	Spring	Summer	Spring	Summer
Mean diam (μm)	5.4 (0.1)a	4.5 (0.2)	4.0 (0.1)	3.7 (0.1)	4.9 (0.2)	4.9 (0.1)	2.2 (0.0)	2.2 (0.0)
Presenceb	12/32 (38)	7/47 (15)	18/32 (56)	23/47 (49)	27/32 (84)	40/47 (85)	32/32 (100)	47/47 (100)
Avg no. of cells ml−1	2 (0)a	0 (0)	11 (3)	7 (3)	16 (3)	40 (15)	187 (29)	264 (60)
Maximum no. of cells ml−1	14	6	72	114	47	543	712	1,896

^aData in parentheses represent standard errors.

^bThe numerator reflects the number of samples in which the group was detected, and the denominator represents the total number of samples interrogated. The number in parentheses reflects the percentage of samples in which the group was detected.

Fig 2.

Abundance during spring and summer of MAST-1B (A and B, respectively), MAST-1C (C and D, respectively), and MAST-4 (E and F, respectively) in the ECS, as determined by FISH (units are cells ml⁻¹). Note the differences in the color scales.

Table 3.

Significant Pearson correlations observed between MAST abundances and other factors^a

Data subset	Pearson correlation
	MAST-1A	MAST-1B	MAST-1C	MAST-4
All stations (n = 79)	Salinity, 0.29**	MAST-4, 0.39***		MAST-1B, 0.39***; Syn, 0.56***
>31 psu (n = 61)	Salinity, 0.31*	Temp, −0.28*	NO2, −0.33**; PO4, −0.27*	Syn, 0.55***
≤31 psu (n = 18)		HNF, 0.57*; PNF, 0.58*; MAST-4, 0.66**; Syn, 0.72**		HNF, 0.70**; PNF, 0.54*; MAST-1B, 0.66**; Syn, 0.77***

^aOnly factors that had a significant correlation are shown. Dissolved oxygen, NH₄, NO₂, NO₃, PO₄, Chl a, heterotrophic nanoflagellates (HNF), pigmented nanoflagellates (PNF), Synechococcus (Syn), photosynthetic picoeukaryotes, and heterotrophic bacteria were included in the analysis. r values are significant at P < 0.05 (*), P < 0.01 (**), or P < 0.001 (***).

Fig 7.

(A) Phylogenetic reconstruction of MAST-4 by NJ distance methods using partial-length 18S rRNA gene sequences. Nine representative CALC sequences (solid squares) were included, and 489 nucleotide positions were analyzed, after masking and gap removal. Bootstrap values were estimated using ML and NJ methods with 100 replicates and are shown for nodes with >70% support. Vertical red lines indicate probe mismatches with environmental sequences. (B) MAST-4 cells and ingested Synechococcus. The images are the result of an overlay of images from blue and green light excitation; red represents the MAST-4 cell, and yellow represents Synechococcus cells. Bar, 10 μm.

MAST-1B and -1C as well as MAST-4 had peak abundances in the plume area during summer (Fig. 2). Previous studies have suggested that discharge from the Changjiang River has an effect on the nanoflagellate community (45). Therefore, we compared MAST abundances between what we defined herein, using salinity measurements, as coastal waters (≤31) and more oceanic waters (>31) using correlations and t tests. MAST-1C abundance was higher in the coastal water (n = 18) than at the more oceanic sites (n = 61) (P < 0.002), and its abundance was negatively correlated with nitrite and phosphate in the oceanic samples (Table 3). Likewise, MAST-4 abundance was significantly higher in coastal waters than in the more oceanic waters (P < 0.02). Photosynthetic picoeukaryotes and bacteria also had higher abundances in the coastal water (P < 0.03 and P < 0.001, respectively). In addition, concentrations of NO₂, NO₃, and Chl a were significantly greater in coastal water than more oceanic water (P < 0.05, P < 0.001, and P < 0.001, respectively). Thus, it appears that MAST-1C and MAST-4 thrive in the inner shelf of the ECS, and this may be related to higher prey availability, higher productivity in general, or other environmental parameters associated with these sites.

MAST cell size, biomass, and prey.

Cell sizes of MAST-1 and -4 from smallest to largest were MAST-4, MAST-1B, MAST-1C, and finally, MAST-1A (Table 2). MAST-1A cells in the ECS were smaller (Table 2) than in data from high latitudes, where cell size was 7.7 ± 0.3 μm (26), but our values are based on very few cells. The average cell size of MAST-1C was about 4.9 μm (Table 2), although at two coastal stations, some cells were >8 μm in diameter (2% of measured cells; Fig. 3). In general, MAST-1 cells were larger than MAST-4 cells. Thus, although MAST-4 dominated numerically, composing 10% (spring) and 6% (summer) of the total HNFs, total biomass contributions of MAST-1C (1.7 ± 0.4 μg liter⁻¹) were equivalent to those of MAST-4 (1.7 ± 0.3 μg liter⁻¹) in spring. In summer, the total biomass of MAST-1C (3.8 ± 1.3 μg liter⁻¹) was greater than that of MAST-4 (2.4 ± 0.6 μg liter⁻¹). The cell sizes of MAST-1A, -1B, and -4 and HNFs were negatively correlated with temperature (Fig. 3), whereas in a previous study a significant (negative) relationship between MAST-1C and cell size was found (26).

Fig 3.

Relationship between water temperature and cell diameter of each of the MAST groups enumerated as well as HNFs as a whole. Significant relationships were determined for HNFs (r = −0.70, P < 0.001, n = 79), MAST-1B (r = −0.39, P < 0.05, n = 41), and MAST-4 (r = −0.25, P < 0.05, n = 79) but not MAST-1C (r = −0.21, P = 0.91, n = 67). For MAST-1A, the relationship was also (negatively) significant, but the analysis involved very few cells. MAST and HNF cell sizes were estimated using different methods, the former by FISH (effectively, the cytoplasm) and the latter by DAPI staining (effectively, the nucleus).

A grazing experiment using tracer addition was carried out at ECS St 24 during the summer cruise. FLS were added at a slightly higher concentration (final concentration, 14 × 10⁴ cells ml⁻¹) than the natural abundance at T₀ (12 × 10⁴ cells ml⁻¹). At T₀, 2% of MAST-4 cells contained Synechococcus in their vacuole. After T₄₀, FLS were observed in 11% of MAST-4 cells; 6% contained 3 FLS per cell, 3% contained 2 FLS per cell, and 2% contained 1 FLS per cell (see, e.g., Fig. 7B). A similar experiment using fluorescently labeled heterotrophic bacteria (FLB) appears to have failed due to the addition of too few FLB (less than 1% of the natural bacterial abundance). By this experiment alone, it is unclear whether the Synechococcus cells ingested during the grazing experiments were then digested and assimilated (19) or, alternatively, possibly later egested. However, there was also a statistically significant positive relationship between the abundance of MAST-4 cells and Synechococcus (Table 3). With respect to other potential picophytoplankton prey, the abundance of photosynthetic picoeukaryotes in the ECS was higher in coastal waters than in more oceanic waters and higher in spring than in summer (P < 0.05). Additionally, photosynthetic picoeukaryotes were occasionally observed within MAST-4 cells. Previously, MAST-4 has been detected by PCR (followed by cloning and sequencing) in a sample sorted by flow cytometry, based on chlorophyll-derived red fluorescence; the red fluorescence was attributed to prey contained in the MAST-4 food vacuole (44). Taken together, these results indicate that picophytoplankton may serve as important prey items for MAST-4.

Stramenopile 18S rRNA gene sequences and phylogenetics.

We obtained a total of 222 stramenopile sequences, 23 from the ECS (6 in spring, 17 in summer) and 199 from the CALC (Table 1). MAST-1, -3, -4, and -7 18S rRNA gene sequences have been found in a number marine environments in previous work (see the review in reference 24); many of the available sequences are partial length. Therefore, we cloned, sequenced, and analyzed nearly full-length 18S rRNA gene sequences. Two phylogenetic approaches were used to integrate these new sequences with data from past studies. In the first approach, only full-length sequences were analyzed, allowing us to analyze more nucleotide positions and differentiate lineages more clearly on the basis of stronger bootstrap support (Fig. 4). The second set of phylogenetic analyses used fewer nucleotide positions in order to analyze a broader suite of sequences, including published partial-length sequences from low to high latitudes as well as the deep sea (Fig. 5 to 8). This second set of analyses did not always retain the levels of bootstrap support seen in the analysis that used more phylogenetic information (Fig. 4). For each clone library generated, a single representative sequence for 98% identity groups was included in the alignment (Fig. 4 to 8). This resulted in 87 selected representatives, 17 and 70 from the ECS and CALC provinces, respectively.

Fig 4.

Stramenopile phylogeny by ML methods of ECS (open squares) and CALC (solid squares) sequences. Nearly full-length 18S rRNA gene sequences from cultured or described organisms (gray) or environmental sequences (black) were used, resulting in 1,399 analyzed positions. ML and NJ bootstrap values are shown at nodes retaining >70% support (100 replicates total). Solid circles, significant support by both methods; gray circles, supported only by ML; open circles, only NJ support. Actual bootstrap values (ML/NJ) are shown for MAST clusters. The TrN+I+G substitution model (I = 0.206 and α = 0.438) was used, and four dinoflagellate sequences served as an outgroup. ECS clone names contain the cruise (897, spring; 905, summer) and station information, while in CALC clone names, the first number after the station identifier, i.e., 1, 8, and 3, corresponds to the sequenced size fractions 0.1 to <0.8 μm, 0.8 to <3 μm, and 3 to <20 μm, respectively; S and D indicate surface and DCM samples, respectively. Numbers in parentheses beside sequence names from our study indicate the number of clones represented by that sequence (all having >98% identity).

Fig 5.

Phylogenetic analysis of MAST-1 using NJ distance methods with partial-length sequences and containing representative ECS (open squares) and CALC (solid squares) sequences. Bootstrap values estimated from ML and NJ methods (100 replicates) are shown for nodes with >70% support, as described in the legend to Fig. 4. A total of 499 positions were analyzed, after masking and gap removal. Latitude and depth at the site of sequence origin are provided in color bars. Samples from latitudes of <20° are from the Cariaco Basin, Caribbean (A95, AA, AB, BC, and CA), Indian Ocean (IND58, IND60, IND70, and IND72), and Pacific Ocean (OLI); those from latitudes from 20° to 45° are from the Atlantic Ocean (AMT15_33, ENI, N5, N10, SSRP, and numerical sequence identifiers), Indian Ocean (IND1, IND2, IND31, and IND33), Mediterranean (BL and ME), and Pacific Ocean (TH and this study); those from latitudes from between 45° and 60° are from the Antarctic (ANT and DH), Atlantic Ocean (NA, AMT15_1, and OR), English Channel (RA), Framvaren Fjord (FV and SIF), and Helgoland (HA); and those from latitudes of >60° are from the Arctic Ocean (MD and NW), adjacent seas of the Arctic Ocean (NOR), Franklin Bay (CS), and Norwegian Sea (AD and CD). All clones were derived from the water column, except those from sediments (BAQ, TAGIRI, and DSGM).

Fig 8.

Phylogenetic reconstruction of MAST-7 on the basis of NJ distance methods and partial-length sequences. Our North Pacific sequences included 1 representative sequence from the ECS (open square) and 6 from the CALC (solid squares). A total of 515 nucleotide positions were analyzed, after masking and gap removal. Bootstrap values were estimated using ML and NJ methods with 100 replicates and are shown for nodes with >70% support.

MAST-3 sequences were the most frequently detected among MAST sequences (13 in the ECS provinces and 77 in the CALC provinces) and were extremely diverse (Table 1 and Fig. 4 and 6). Our analysis supported the inference that MAST-3 organisms comprise an independent cluster with high bootstrap values. The data also indicated that MAST-3 contains two distinct, supported subclades (Fig. 4). One MAST-3 subclade contained sequences from more oligotrophic (St 12), DCM (St 67-155), and mesotrophic transitional (St 67-70) stations as well as sequences from the parasite Solenicola setigera (Fig. 4). The other MAST-3 subclade contained sequences from the ECS transitional station (St 23) and all CALC samples, of which 37 had 98% identity to each other (represented by clone H3S1Be4Kp from St H3; Fig. 4). Overall, MAST-3 18S rRNA gene sequence diversity was relatively high, as reported previously (24). Most MAST-3 sequences appeared to be from between latitudes 20° and 45°, and few have been retrieved from colder deep waters (>200 m) at high latitudes (>60°) (Fig. 6). Organisms that have a parasitic lifestyle, as reported for S. setigera, have been hypothesized to be characterized by extremely high diversity (41). Solenicola, which may alternatively be an epibiont, is broadly distributed, with proposed differences between coastal and oligotrophic communities (2, 15). Although the extent to which other MAST-3 organisms are parasites is not known and some are free-living HNFs (25), the presence of parasites in this lineage might contribute to its relatively high diversity.

Fig 6.

Phylogenetic reconstruction of MAST-3 by NJ distance methods with partial-length sequences, including 7 representative sequences from ECS (open squares) and 27 representative CALC sequences (solid squares). A total of 471 positions were analyzed, after masking and gap removal; bootstrap values were estimated using ML and NJ methods with 100 replicates and are shown for nodes with >70% support. Clonal libraries represented, latitudinal ranges, and depths are as described in the legend to Fig. 5.

We did not recover MAST-4 clones in the ECS but did find them at all CALC sites. Besides the biases inherent to PCR-clone library studies, the number of CALC clones sequenced was greater than the number of ECS clones sequenced and the size fractions sequenced differed in the two provinces, making comparison of numbers between provinces invalid. A relatively large number of MAST-4 sequences were retrieved from the CALC oceanic site; DCM clone 155D8Be8Za represented 24 sequences from that sample, and the clone 155S8Aeagi represented 15 sequences from the surface at St 67-155. At the mesotrophic transitional station, clone 70S1Be7jE represented 12 sequences. MAST-4 cells were also detected by FISH at all ECS stations (Table 2). In the Indian Ocean, MAST-4 sequences were more abundant in coastal sites than at open-ocean sites, although they also appeared to be more abundant in the DCM than at the surface, on the basis of quantitative PCR data (40). MAST-4 organisms have not been detected in polar regions (>60°), by either 18S rRNA gene clone libraries (Fig. 7) or FISH (26). The distribution of MAST-4 organisms may relate in part to prey availability or other factors associated with temperature. Overall, MAST-4 18S rRNA gene sequences appeared to have relatively less genetic distance and fewer subclades than MAST-3 or MAST-1 (Fig. 4).

In the CALC province, representatives of the 3 previously described MAST-1 subclades, MAST-1A, -1B, and -1C (26), were retrieved along with one from the ECS (from MAST-1C; Fig. 4 and 5). The MAST-1A clone 155D3Ae6ho was obtained from the DCM at the most oligotrophic station sampled. MAST-1A sequences have primarily been acquired from higher latitudes, and many of them composed a large supported subgroup (ECS and CALC sequences did not belong to this subgroup; Fig. 5). The fact that there are several phylogenetically distinct groups from high and low latitudes that are targeted by the same FISH probe (Fig. 5) could potentially explain the observed differences in the sizes of the ECS cells measured here and those measured in samples from higher latitudes (26).

MAST-1B sequences were retrieved from the DCM only at the most oligotrophic CALC station (represented by 155D3Ae6hF; Fig. 4 and Fig. 5). An additional sublineage, MAST-1D, was apparent. MAST-1B contained DCM clones and appeared to have a broader distribution from low to high latitudes, while MAST-1D was detected only in tropical and subtropical/temperate locations (latitudes < 45°) (Fig. 5). MAST-1C contained two distinct groups of sequences (Fig. 4), one with those from the transitional and oligotrophic CALC stations (70S3Be9KQ and 155D3Be4vg) and the other with sequences from coastal St H3 and St 19 (Fig. 4). The former group contained (within it) two additional supported clades which separated sequences from low and high latitudes (Fig. 5). Those from lower latitudes (including CALC sequences) seemed to be from more oceanic settings, such as the Sargasso Sea (Q2H11N10 [31]), although nutrient data were not available from all previous studies (Fig. 5). Sequences composing the group containing clones 897St19-43 and 14H3Te6QW from our most eutrophic sites have also been recovered from sediments, anoxic, and pelagic environments (Fig. 5). Although the ECS has been reported to be a low-oxygen area (4), hypoxia was not observed at St 19 in spring during this study. We hypothesize that this MAST-1C subclade can acclimate to dramatic variations in environmental conditions, including estuarine or anoxic environments.

In the ECS, only a single clone each was obtained from MAST-2 (clone 897St19-5) and MAST-7 (clone 905St19-25), and both were from the coastal station. In the CALC, although no MAST-2 sequences were recovered, MAST-7 sequences were found at all stations. The basal part of the MAST-7 tree was relatively diverse, and only the upper half of the tree contained sequences from between latitudes 20° and 60° (Fig. 8). Sequences from the CALC province were spread throughout the tree and primarily seemed to come from the most oligotrophic station, St 67-155 (Fig. 8). The MAST-2 sequence from St 19 was phylogenetically close to a sequence (MB04.36; 99% identity over 713 bp) from the Hong Kong coast in the subtropical western Pacific Ocean. These two sequences have a mismatch to the NS2 probe (26), which may explain why MAST-2 sequences were rarely observed by FISH (less than 10 cells detected in a total of 79 samples). There are still few published MAST-2 sequences, making it difficult to further inspect the relationship between the phylogeny and distribution.

MAST-8 and -11 sequences were retrieved only from St 67-70 and the DCM at St 67-155 (Fig. 4). All of these had low identities to deposited sequences (91 to 96%), except MAST-11 clone 905St23-34. Phylogenetic analysis supported MAST-8 being an independent stramenopile lineage. Most MAST-8 clones in the analysis were from our study because only one other nearly full-length sequence (that for clone OR000415.113) was available (24). Finally, the position of MAST-11 and most other MAST groups was not supported in the 18S rRNA gene tree (Fig. 4).

Other heterotrophic stramenopile sequences included quite divergent sequences from the most nutrient-rich stations sampled in both provinces, which grouped with the Bicosoecida and Oomycetes (Fig. 4). Labyrinthulida-like sequences were also detected, and one was similar to clone BL010625.31 from the coastal Mediterranean Sea. In addition, a colorless chrysophyte (Spumella) sequence was retrieved from the ECS transitional station (St 23) in summer. Chrysophytes were also detected in the CALC but belonged to uncultured taxa for which the trophic mode is unknown. A single pelagophyte sequence was identified (St 67-155). Twenty-two diatom sequences were found at the more eutrophic midbay station (St H3), and most belonged to a clade largely composed of Thalassiosira sequences. Only one was found at St 67-70, and none was found at St 67-155 or any of the ECS stations. Most diatom sequences were detected in the largest CALC size fraction (3 to 20 μm), and a similar size fraction was not sequenced in the ECS (where water was prefiltered through a 5-μm-pore-size filter).

In conclusion, our results indicate that many of the previous observations on the distribution and phylogeny of MAST lineages (6, 24, 25, 31, 26, 48) hold true in the North Pacific Ocean. Within the stramenopiles, many trophic modes exist, and this is presumably the case across MASTs as well. Their ecology, physiology, and diversity are still little understood, in part due to their uncultured status. The addition of sequences as well as abundance and size data from previously unexplored oceanographic provinces is helping to refine hypotheses on the distribution and activities of these organisms. We found that the cell size of MAST-1B and MAST-4 varied with temperature, indicating that different ecotypes may exist within these lineages. The analysis of full-length 18S rRNA genes also indicated that MAST-1 and MAST-7 contain distinct subclades, and groups within those were associated with specific latitudinal zones. Thus, perhaps not surprisingly, temperature or other environmental parameters associated with these zones appear to play a role in the 18S rRNA gene diversity and distribution of these eukaryotes.

Finally, MAST-4 organisms have been shown to consume heterotrophic bacteria. Here we observed their consumption of picophytoplankton, specifically, Synechococcus. Combined with the statistically significant relationship observed between the natural abundance of MAST-4 and Synechococcus, this may indicate that the latter serves as a food resource for MAST-4. Herbivory has previously been reported for MAST-1, MAST-6, and MAST-7, on the basis of label incorporation in nucleic acids from incubation experiments with cyanobacterial prey cells (Prochlorococcus or Synechococcus) in NPSG surface waters (14). MAST-6 organisms have been reported to ingest more algae than heterotrophic bacteria in brackish waters of the Baltic Sea, on the basis of visual assessment of food vacuole contents (36). The factors that shape prey selection by these MAST lineages and now MAST-4 influence their ecosystem roles, particularly with respect to control of primary producer communities.