SEARCH

SEARCH BY CITATION

Keywords:

  • DNA sequencing;
  • marine bacteria;
  • marine snow;
  • sediment traps

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Using marine sediment traps (named RESPIRE for REspiration of Sinking Particles In the subsuRface ocEan) designed to collect sinking particles and associated microbial communities in situ, we collected and incubated marine aggregates/particles in the southern Pacific Ocean from separate phytoplankton bloom events in situ. We determined the phylogenetic affiliation for the microorganisms growing on aggregates by pyrosequencing partial 16S rRNA gene amplicons. Water column samples were also collected and sequenced for comparison between sinking-particle-associated and planktonic bacterial communities. Statistically significant differences were found between the water column and sediment trap bacteria. Relative abundances of Pelagibacter sp. and multiple members of the Flavobacteria, Actinobacteria, and α-Proteobacteria were elevated in water column samples, while trap samples contained members of the Roseobacter clade of α-Proteobacteria in high relative abundances. Our findings indicated that rapid changes – within 24 h of collection – occurred to the microbial community associated with aggregates from either bloom type. There was a little change in the bacterial assemblage after the initial 24-h incubation period. The most abundant early colonizer was a Sulfitobacter sp. This study provides further evidence that Roseobacters are rapid colonizers of marine aggregates and that colonization can occur on short timescales. This study further demonstrates that particle origin may be insignificant regarding the heterotrophic bacterial population that degrades them.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Organic matter aggregates and sinking particles in marine systems (commonly known as marine snow) appear to provide an ecological niche for bacteria. Bacterial cell densities on marine snow are typically orders of magnitude higher than background measurements (Caron et al., 1986), and the processes of photosynthesis, decomposition, and nutrient regeneration on these particles also occur at rates much higher than on the surrounding water column (Alldredge & Silver, 1988; Smith et al., 1992). These observations can be explained, at least in part, by the fact that marine snow can be highly enriched in nutrients compared with the surrounding water column (Shanks & Trent, 1979). Increases in rates of the processes listed above often result in utilization of the labile material in a particle within the first 48 h after formation (Pomeroy et al., 1984; Smith et al., 1992).

The dynamic nature of marine snow microenvironments likely exerts strong selective pressures on associated microbial communities. During the rapid colonization and processing of this organic matter, complex successional changes occur in the microbial community across timescales of hours to days (Alldredge & Silver, 1988). Bacteria that rapidly colonize surfaces in marine systems (Dang & Lovell, 2000; Bearon, 2007) are thought to play a major role in elemental recycling on and within marine snow particles (Smith et al., 1992). Successional changes in the microbial community result in significant alterations in the chemical and biological properties of the colonized particle (Alldredge & Silver, 1988). This is evidenced by the fact that C/N ratios and the percentage of refractory matter in small aggregates (Riley, 1970) and large particles (Martin et al., 1987) increase with depth. The change in C/N ratios of sinking particles ultimately skews deepwater elemental distributions compared with surface waters, leading to pooling of recalcitrant dissolved organic matter (Jiao et al., 2010).

Traditional sediment traps have been used for decades to measure the quantity and quality of particles ‘raining’ down from the upper regions of the water column (Davis, 1967). In order for the contents of the trap to accurately reflect the original quantity, quality, and chemical nature of trapped particles, traps must inhibit microbial grazing and microbial decomposition. This is generally accomplished using a chemical poison or preservative to inhibit or terminate biological activity (i.e. sodium azide, formalin; Knauer et al., 1984; Taylor et al., 2009). A consequence of inhibiting all organisms in the trap is that it becomes impossible to determine the fate of the particles or their associated microbial assemblage over time. Thus, measurements of aggregate-/particle-associated rates of bacterial respiration, bacterial production, and microbial community compositional changes cannot be determined.

Biologically mediated transformations of sinking material play important roles in biogeochemical cycling and can influence climate (Denman et al., 1996; Kwon et al., 2009), yet our understanding of these processes and the factors influencing them is limited. It has been shown previously, in various marine environments, that the bacterial communities of free-living and attached bacteria are taxonomically distinct (Delong et al., 1993; Kellogg & Deming, 2009). What we know about particle-associated microbial communities and activities is based on experiments that required samples to be brought to the surface and incubated and/or analyzed ex situ (Alldredge & Youngbluth, 1985; Alldredge & Silver, 1988; Muller-Niklas et al., 1994; Zimmermann, 1997). Recently, new sediment traps (P.W. Boyd, J. Valdes, A. McDonnell, M.P. Gall, submitted) were constructed to collect particles and here are used to monitor the development of, and changes to, the marine-snow-associated bacterial community in situ. We describe the use of these traps to characterize changes in the particle-associated microbial communities compared with free-living bacterial communities in the water column during an initial diatom bloom and a secondary bloom codominated by cyanobacteria and diatoms, while the system transitioned from iron-replete to iron-limiting conditions (Wilhelm et al., 2013).

We hypothesized that over time, the trap communities would shift from a ‘water column-like’ assemblage to a more distinct community dominated by bacteria specializing in particle colonization and degradation. Our findings shed new light on the rapid changes occurring within particle-associated bacterial communities and provide insight on the phylogeny of microorganisms involved in processing this material during export from surface waters.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Samples were collected during the FeCycle II research cruise aboard the R/V Tangaroa. FeCycle II was conducted in subtropical waters east of the North Island of New Zealand and centered in a counterclockwise eddy (39°20′S 178°40′W). This location is the site of an annually recurring spring diatom bloom that persists for c. 2–4 weeks a year (Murphy et al., 2001; Nodder et al., 2005). FeCycle II was one of the first GEOTRACES process studies (http://www.geotraces.org) and as such provided a wealth of available metadata on in situ chemistry and processes (Boyd et al., 2012; King et al., 2012; Matteson et al., 2012, 2013).

Our experimental design and sampling efforts focused on two distinct environments (Fig. 1). For the first 10 days, our experiments were performed, and samples were collected within the anticyclonic eddy. This quasi-Lagrangian study of a singular patch of water was enabled through the use of a drogued radio drifter. After 10 days, we were blown to the outer edges of the eddy by high winds (> 15 m s−1; Boyd et al., 2012). The remaining 8 days of sampling efforts occurred on the outside edge of the eddy (Fig. 1).

image

Figure 1. (a) Heat map of chlorophyll concentrations. Sample collections are denoted with open circles for water samples and open diamonds for sediment trap samples. Sample numbers are noted within the circles. Samples were collected at 5 and 100 m for water column samples and 100 and 120 m for sediment traps. The orange-dashed line demarcates the sampling dates within the eddy vs. the outside edge of the eddy. (b) Sampling scheme for RESPIRE Traps. Chlorophyll data modified from Weller et al. (2013).

Download figure to PowerPoint

RESPIRE sediment traps (described in Boyd P.W. Boyd, J. Valdes, A. McDonnell, M.P. Gall, submitted) were deployed on floating arrays that were suspended at preset depths (100 and 120 m because of the minimal density changes at these depths (P. Boyd, unpublished data), standard practice for the deployment of free-drifting traps). The traps were programmed to collect particles for 24 h. During the collection period, particles falling from overlying water came to rest on the surface of a large (c. 15 cm diameter), dimpled, ‘golf ball-like’ mechanism within a collection cylinder. During this collection period, the particles were not enclosed in a bottle, and they were in constant contact with the surrounding marine environment. This allowed diffusion to occur and chemotactic bacteria and other motile bacteria not yet associated with the particles to potentially ‘find’ the settled aggregates.

After the 24-h collection period, some traps were closed and immediately brought to the surface (these samples are referred to as STL-initial). Other traps (still at depth) were remotely switched from ‘collect’ to ‘incubate’ mode. While switching to incubate mode, the ‘golf ball-like’ mechanism rotated 180° and trapped the particles inside the collection cylinder, sealing them from the surrounding environment. The traps, still at depth, incubated the collected particles and associated microorganisms for an additional 72 h before being brought to the surface (these samples are referred to as STL-72). For all RESPIRE traps, water and particles were carefully drained to avoid dislodging biofilms that may have been established on the walls of the incubation cylinder. Water and particles were then homogenized, and 50 mL of the water/particle suspension was filtered through 47-mm, 0.2-μM-pore-size polycarbonate filters (Millipore, Billerica, MA) to collect bacteria and particles. Filters were immediately frozen in liquid nitrogen for future analysis.

Water samples, serving as indicators of the free-living bacterial communities and as potential reservoirs of bacteria inoculating or migrating into the sediment traps, were also collected from 5 and 100 m. These samples were collected predawn using Niskin bottles on the ship's CTD rosette. Water (50–100 mL) was filtered onto 0.2-μM 47-mm polycarbonate filters. Filters were flash-frozen in liquid nitrogen for transport and then stored at −80 °C.

Chlorophyll a (c.f. Welschmeyer, 1994), enzymatic activities (Hoppe et al., 1993), and bacterial production rates (Kirchman, 2001) were determined for both water column and trap samples as comparisons of detrital material from phytoplankton and biological activity of the microbial community. Flow cytometric determination of water column phytoplankton and picoplankton was completed as previously described (Hall et al., 2006). Bacterial abundance in sediment traps was determined by quantitative PCR targeting bacterial 16S rRNA gene (primers 1055f and 1392r with TaqMan probe 16STaq1115 as described in Harms et al., 2003).

DNA was extracted from all filters using the MoBio PowerWater DNA Isolation Kit (MoBio Laboratories Inc., Carlsbad, CA) according to manufacturer's protocols. For amplification of bacterial 16S rRNA genes, bacteria-specific primers (Eurofins MWG Operon, Huntsville, AL) targeting bases 338–926 (E. coli numbering) of the 16S rRNA gene were used to amplify the V3–V5 region of the extracted DNA samples (Supporting Information, Table S1). We selected these primers based on a previous study by Wang & Qian (2009). This primer set has successfully been used in another study of 16S rRNA gene diversity coupled with 454 sequencing (Methe et al., 2012). PCRs were performed using Invitrogen Platinum Taq (Life Technologies, Grand Island, NY) using the following protocol: 95 °C for 5 min, followed by 30 rounds of 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s and then a final extension step at 72 °C for 10 min. Product amplification was verified on a 1% agarose gel stained with ethidium bromide and viewed on a UV transilluminator. Individual samples were then processed to remove unincorporated primers and nucleotides using the Qiaquick PCR cleanup kit (Qiagen, Valencia, CA). Amplicon concentrations were determined using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE).

Individual sample amplicons were barcoded (six additional PCR cycles: 95 °C for 30 s, 55 °C for 30 s, 72 °C for 30 s) with primers that contained a unique 8-bp barcode attached to the 454 fusion primers (Table S1). Our barcoding primers were designed for unidirectional sequencing on a 454 GSFLX sequencer (454 Life Sciences, Branford, CT). This strategy requires the use of the Lib-L kit (see Roche application brief 001-2009). We opted for unidirectional sequencing because our PCR product was larger than the average read length of the current 454 titanium sequencing chemistry. This approach ensured that sequences would overlap for the longest length possible. All barcoding reactions were prepared to have 0.5 ng μL−1 of amplicon DNA per reaction. After the barcoding reaction, we again verified our amplicons on an agarose gel and pooled all barcoded amplicons. Barcoded amplicons were processed to remove unincorporated primers and nucleotides using a single Qiagen Qiaquick column. Sequencing was completed at the University of Tennessee/Oak Ridge National Laboratory Joint institute of Biological Sciences (www.ceb.utk.edu/dnasequence.html), and information was deposited in the NCBI short-read archive (accession numbers SAM02318127SAM02318154).

We used the Mothur software package (version 1.24.1; Schloss et al., 2009) to screen sequences for sufficient length and quality. We processed our sequences similar to the Schloss SOP (http://www.mothur.org/wiki/Schloss_SOP) with some modifications in the shhh.flows command; we changed the number of flows value in the shhh.flows command to 360–720 from 450 (Quince et al., 2009). Mothur was also used to cluster sequences into operational taxonomic units (OTUs) and for phylogenetic classification. A 0.03 cutoff (97% identity) was chosen for OTU determination. We used Mothur to sort our sequences into different groups based on the source of the DNA, incubation time, and depth.

The primer-e software package (version 6; Clarke & Gorley, 2006) was used to interrogate the relationships between OTUs across samples and to also look for any correlations between OTU presence/abundance and environmental parameters. The ‘.shared’ file (a matrix file containing OTU abundances for each sample) created by Mothur was imported directly into primer-e. All OTUs were standardized to the total number of sequences per barcoded library (proportional abundances). The standardized relative abundances were then square-root-transformed to partially deemphasize more highly abundant OTUs. A Bray–Curtis similarity matrix was constructed and used to perform nonmetric multidimensional scaling analysis (NMDS) for visualization of community structure relationships between the different samples. We also investigated correlations between OTU relative abundances and environmental parameters including bacterial production rates, chlorophyll concentration, protease activity, hydrolase activity, peptidase activity, length of incubation, original source of the sample (sediment trap or water column), and location of sampling (inside or outer edge of the eddy) using the BIOENV program within primer-e (Clarke & Gorley, 2006). A principal components analysis (PCA) was performed on the square-root-transformed data to visualize community structure differences and identify OTUs driving the differences in the bacterial communities. Phylogenetic identities of all OTUs were determined using the Ribosomal Database Project (RDP) classification within Mothur. Both weighted and unweighted UniFrac (Lozupone et al., 2007) analyses were used to compare 16S libraries from the different sample sources.

Bacterial production rates were estimated from leucine incorporation rates measured by standard methods (Kirchman, 2001). Briefly, triplicate samples with 3H-leucine (20 nmol L−1) were incubated for 2 h at in situ temperature on deck. The incorporated 3H-leucine was precipitated by trichloroacetic acid (TCA), collected by centrifugation, and rinsed with TCA and ethanol. The samples were then dried and radioassayed to determine 3H-leucine uptake.

Results and discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Previous studies have established that free-living and particle-associated bacterial communities are different from one another (Delong et al., 1993). The factors that drive these changes and their impacts on the microbial community, however, are not completely known. We attempted to investigate, in unprecedented resolution, changes in the free-living and particle-associated microbial communities over the course of an initial diatom and subsequent secondary bloom dominated by cyanobacteria and diatoms (Fig. 1 and Supporting Information, Fig. S2). This was accomplished using newly developed RESPIRE traps that allowed for incubation at in situ temperature and pressure.

We had the opportunity to sample microbial communities on exported particles and within the water column during a succession of different phytoplankton populations during a southern Pacific Ocean spring bloom event. A detailed description of the bloom has been provided in another study (Boyd et al., 2012). In the current study, we have added water column depth resolution for phytoplankton biomass: our observations reveal a subsurface bloom (c. September 22, 2008) not resolved previously by satellite information as well as the presence of a significant population of Synechococcus c. February 28, 2008 (Figs 1, S1 and S2). While both diatoms (Asterionellopsis glacialis and Leptocylindrus sp.) and Synechococcus are present during these two periods, they occur at different proportionalities, implying that carbon and nutrients exported during these two periods should be compositionally different (Fig. S2).

In total, we obtained 83 582 raw sequences with a mean length of 529 bps from the sequencing facility. After processing, 54 723 sequences remained. These sequences formed a total of 766 OTUs at a distance cutoff of 0.03 (97% similarity). Heterotroph and photosynthetic/chloroplast sequences were then partitioned into separate files. After partitioning, there were 49 641 sequences in the heterotroph sequence file and 5082 sequences in the photosynthetic/chloroplast file. The number of OTUs in the heterotroph and chloroplast files was 674 and 92, respectively.

The NMDS plot (Fig. 2) demonstrated that sediment trap 16S libraries (referred to hereafter as STL) and water column 16S libraries (referred to hereafter as WCL) were distinctly different from one another (with the exception of NZ67). There was no clear difference between WCL collected from 5 or 100 m depth. The STL-initial trap samples were most similar to the STL-72 samples, despite the 72-h samples being incubated for a comparatively long period of time. Both weighted and unweighted UniFrac analyses, however, showed that all four sample types (WCL-5m, WCL-100m, STL-initial, and STL-72) were significantly different from one another (data not shown). The discrepancy between the NMDS plot and the UniFrac analyses arises from the fact that the NMDS analysis compared all samples individually, and the UniFrac analyses compared the four samples types to each other (STL-initial, STL-72, WCL-5m, and WCL-100m). Results from an anosim, performed within primer-e, showed that statistically significant differences (P < 0.001) only existed between the STL and WCL samples. Differences between STL-initial and STL-72 were statistically insignificant, and differences between WCL-5m and WCL-100m libraries were also statistically insignificant. In the anosim, all samples of the same type were pooled.

image

Figure 2. NMDS analysis plot comparing all sample libraries. Green triangles = sediment traps incubated for 72 h; brown triangles = STL-initial; blue squares = 5-m water samples collected with Niskin bottle; purple squares = 100-m water samples collected with Niskin bottle. I = inside eddy, O = outside eddy. Libraries with similarities > 20% or 40% are circled with yellow and blue lines, respectively. Similarity lines were automatically drawn based on results from a hierarchical cluster analysis performed within primer-e, version 6.

Download figure to PowerPoint

It appeared that differences between the heterotrophic bacterial communities of the WCL and STL were established quickly. Our sampling strategy prohibited us from knowing exactly when or how rapidly these changes occurred, but Riemann et al. (2000) and Mc Carren et al. (2010) both found that 24 h was sufficient time for measuring shifts in microbial communities. Another interesting finding was that the consortium of OTUs that comprised > 1% of the relative abundances in the STL was relatively constant between the 24-h and 72-h time points, indicating that once established, the community composition was relatively stable.

A heat map of the dominant OTUs (Fig. 3) and PCA plot based on these data (Fig. 4) both indicated that only a few OTUs were responsible for a large amount of the differences between communities. The OTUs responsible for these differences are listed in Table 1. The relative abundance of various OTUs was clearly dependent on the source of the sample (sediment trap or water column). For example, STL samples contained the genera Phaeobacter, Sulfitobacter, Ruegeria, and Mesonia in high relative abundances. Three of these four genera (Phaeobacter, Sulfitobacter, and Ruegeria) fall within the Roseobacter clade. Roseobacters have been demonstrated to colonize surfaces and form biofilms (Buchan et al., 2005; Dang et al., 2008). The fourth genus (Mesonia) is a member of the Flavobacteria. The strain of Mesonia most closely related to the OTU in our library was isolated from a marine alga (Nedashkovskaya et al., 2003). Flavobacteria are known to degrade complex organic molecules produced by algae (Williams et al., 2012) and are common during phytoplankton blooms (Delong et al., 1993; Teeling et al., 2012). WCLs, on the other hand, were mostly comprised of OTUs associated with Pelagibacter, Erythrobacter, Acidovorax, Nitriliruptor, Ilumatobacter, Gillisia, and Cloacibacterium (See Fig. 4 and Table 1). These genera are affiliated with the α-Proteobacteria (2), β-Proteobacteria (1), Actinobacteria (2), and Flavobacteria (2). A BIOENV analysis (performed using primer-e) indicated that 77.5% of the difference between sample types could be explained by whether the source DNA was from the WCL or STL.

Table 1. Accession numbers and phylogenetic hits of the closest relatives to the 11 most influential OTUS from water and sediment trap samples
OTU#Accession #Max identityE-valueScoreClosest relative (environmental samples excluded)
1 JQ806410.1 1007E-94351Sulfitobacter dubius strain HME8274 16S ribosomal RNA gene
2 JQ907337.1 1005E-105388Mesonia algae gene for 16S rRNA gene, partial sequence, strain: NBRC 100447
3 HE818197.1 1007E-94351Phaeobacter sp. Ph222 partial 16S rRNA gene, isolate Ph222
4 JQ963327.1 1007E-94351Erythrobacteraceae bacterium K-2-3 16S ribosomal RNA gene
6 JF488574.1 1007E-94351Alpha proteobacterium SCGC AAA160-I09 16S ribosomal RNA gene
11 JN688951.1 1005E-105388Cloacibacterium sp. RBC21 16S ribosomal RNA gene, partial sequence
13 HQ675191.1 1007E-94351Actinobacterium SCGC AAA015-M09 small subunit ribosomal RNA gene
16 JF488663.1 1007E-94351Actinobacterium SCGC AAA163-G08 16S ribosomal RNA gene
24 HQ675267.1 98.083E-97363Flavobacterium sp. SCGC AAA298-O21 small subunit ribosomal RNA gene
36 JQ975881.1 1008E-108398Acidovorax sp. ALI-INI32 16S ribosomal RNA gene
image

Figure 3. All OTUs in this figure comprise > 1% of sequences in at least two sample libraries. Taxa with a number in parentheses indicate the number of times separate OTUs were assigned to a specific genera (classifier assigned multiple OTUs to the same genera). Genera in red are Roseobacters. Blue circles = OTUs found only in water column samples; brown circles = OTUs found only in sediment samples. Color indicates percentage of library the OTU represents (see scale bar).

Download figure to PowerPoint

image

Figure 4. Principal components analysis of all 674 heterotrophic OTUs. The 10 OTUs with the largest sum of PC1 and PC2 eigenvectors are included in the figure. These 10 OTUs were labeled according to their phylogenetic affiliation. OTU numbers (assigned by Mothur) are listed in parentheses and correspond to numbers in Table 1. Principal components 1 and 2 accounted for 41.6% of variance between samples. blastn results for these OTUs can also be found in Table 1.

Download figure to PowerPoint

Some OTUs fluctuated wildly in relative abundance with regard to the sample source. For example, the OTU associated with Pelagibacter (Morris et al., 2002), a dominant member of the WCL (11 of 17 WCL had Pelagibacter relative abundances > 1%), was not present in any STL in relative abundances > 1%. In fact, only nine of the 2,314 (0.0039%) Pelagibacter-related 16S rRNA gene sequences were retrieved from our STL. Based on previous studies, it is not necessarily surprising that Pelagibacter relative abundances were drastically reduced in the incubated samples. Pelagibacter has been shown to be adapted to a pelagic lifestyle and not to one associated with colonization and decomposition (Giovannoni et al., 2005). Additionally, P. ubique abundances have been shown to be inversely proportional to Flavobacteria abundances and in situ fluorescence (Teeling et al., 2012). Known P. ubique growth issues [from previous culture-based studies (another confined, nutrient rich environment)] may also have played a role in the lack of sequences in our STL (i.e. slow growth rate and low cultured cell abundances; Morris et al., 2002; Rappe et al., 2002). Pelagibacter ubique were possibly overgrown by the faster replicating microorganisms that specialize in colonizing and degrading particles. Alternatively, the P. ubique cells may never have been present in the traps due to their lack of particle colonization abilities. Regardless of the reason, there were very few P. ubique cells retrieved from STL samples.

Despite the general relationship found between WCL and STL (WCL are distinct from STL), there was considerable variability between libraries within the same sample type. For example, no single OTU was found in relative abundances > 1% in every STL (Fig. 3) nor was any single OTU found in relative abundances > 1% in every WCL. The high degree of variation in relative abundances in both the STL and WCL may have resulted from an insufficient sampling effort (not enough sequences) or because this natural system was extremely dynamic and continually changing (See Figs 1 and 3). Rarefaction curves made of our sequencing efforts indicate that the STL were more fully sequenced than the WCL. All STL samples began to plateau (data not shown), and only four of the WCL samples were beginning to plateau (NZ70, NZ109, NZ21, and NZ79; data not shown).

Several OTUs with high relative sequence abundances were from the Roseobacter clade. These OTUs were related to Sulfitobacter, Jannaschia, Phaeobacter, and Ruegeria. All four of these OTUs were found in high relative abundances (> 1%) in at least two WCL and two STL. Finding Roseobacters in high relative abundance in WCL samples is consistent with previous studies that have measured Roseobacter abundances at > 10% in open ocean samples (Buchan et al., 2005). Roseobacters have also been shown to be adept at growth on particles due to their abilities to grow rapidly and make biofilms (Dang & Lovell, 2000, 2002; Buchan et al., 2005). Members of this lineage have also been shown to produce antimicrobial secondary metabolites (Brinkhoff et al., 2004; Wagner-Döbler et al., 2004; Martens et al., 2007; Cude et al., 2012) that might afford them a competitive advantage over other lineages, especially in the close proximity experienced during particle-associated growth.

Sulfitobacter appeared to be the dominant early colonizer of sinking particles in this environment. The Sulfitobacter OTU was found in extremely high relative abundances in the STL-initial samples (ranging from 47.94% to 83.80% of total sequences in each library, with an average percentage of 66.84). Additionally, one STL-72 sample was comprised of 95.67% Sulfitobacter sequences. Sulfitobacter was common in the other STL-72 samples as well, but not in as high relative abundance as the STL-initial samples (STL-72 relative abundances ranged from 0% to 95.67% with an average percentage of 20.79%). It appears that the Sulfitobacter OTU started to decline in relative abundance soon after the 24-h collecting phase as their relative abundances were substantially diminished in the STL-72 samples compared with the STL-initial samples. Other studies have shown that Sulfitobacter is an early colonizer of settling aggregates (Dang & Lovell, 2000; Jing et al., 2012), so it is not surprising that they were abundant in our STL. The Sulfitobacter OTU was not exclusive to the sediment traps. It was also present in seven of the 16 WCL at relative abundances higher than 1% (1.20–95.43%), but in only one water column sample (NZ67) did it actually exceed 5% relative abundance (95.43% of NZ67). NZ67 is the only WCL that clusters more closely with a STL than with other WCL (Fig. 2). The presence of Sulfitobacter in such high relative abundance is what caused NZ67 to cluster with the STL instead of the other WCL in Fig. 2. It is possible that during the collection of NZ67, we inadvertently collected a large aggregate from the water column. This would explain the presence of Sulfitobacter in such high relative abundance in the NZ67 sample.

It is not clear from our study whether the microorganisms associated with STL were present on the settling aggregates before falling into the sediment traps or whether the early colonizers swam towards and attached to the already settled aggregate. It is, however, likely that at least some of the Roseobacters were present on the sinking particles before entrapment because Roseobacters have been shown to be highly abundant in bacterial communities associated with marine algae, including natural phytoplankton blooms and algal cultures (Buchan et al., 2005; Grossart et al., 2005; Jing et al., 2012; Teeling et al., 2012). These phytoplankton-associated Roseobacters could have served as a seed stock for the massive increase in Roseobacter relative abundance we measured in the STL.

It is interesting to note that despite clearly sampling two different water masses (see Fig. 1), the WCL within the eddy and the other WCL on the outer edge of the eddy were not significantly different based on anosim, and they coclustered with one another (Fig. 2). This was also true for the STL. We initially expected to see two different communities associated with the water column and also in the sediment trap communities that collected particles from the two different blooms. We found the high degree of similarity between libraries to be a surprising and interesting occurrence as there were considerable differences in bloom type and status inside and on the outer edge of the eddy (Figs 1, 2 and S2). The differences between blooms likely resulted in particles of different types and sizes with varying nutrient ratios. Figures 2-4 hint at the presence of ‘core’ pelagic and particle-associated bacterial communities. Regardless of the bloom status of the overlying water, the heterotrophic community of the STL always reverted back to communities more similar to other STL than to any of the original overlying WCL. Likewise, regardless of whether the water column samples were collected within or outside of the eddy, the WCL were more similar to one another than to any STL (with the exception of NZ67).

We attempted to correlate relative abundances of all OTUs to other environmental parameters, including bacterial production rates, chlorophyll a concentrations, protease activity, phosphatase activity, β-glucosidase activity, incubation time, source of the sample, and location of sampling (Table S2). To our surprise, the only parameter that correlated with the OTU relative abundance patterns was the source of the sample (water column or sediment trap). This, too, points to the possibility of a core pelagic and particle-associated bacterial communities. Additionally, there was no consistent pattern between incubation time and bacterial relative abundances (Table S3).

A few things remain unknown, including whether the particles settled into the trap with an already established community or whether the aggregate bacterial assemblages shifted after the onset of the collection period. There is also the potential that the shifts in the bacterial assemblage are related to or caused by bottle effects. Regardless of whether the shift was due to bottle effects, the OTUs that grew up on the particles were consistent with previous phylogenetic studies. The phenotypic characteristics of the most abundant OTUs we observed are consistent with those of microorganisms that colonize particles.

In closing, our results confirm the observations of others that early colonization of particles occurred very rapidly by heterotrophic bacteria of the Roseobacter clade and some other bacteria. This study was unique in that it involved autochthonous particles that were incubated in situ without any on-deck manipulations, negating artifacts caused from samples being brought to the surface for experimentation. The possibility of separate ‘core pelagic’ and ‘core particulate’ communities was raised because the only driver of community structure, regardless of sample location, date, or bloom status was the source of the sample.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

We thank the captain and crew of the R/V Tangaroa and Alice Layton and Dan Williams for help with sequencing and Pat Schloss for running a fun and informative workshop on using Mothur. We also thank Ben Twining, David Hutchins, Scott Nodder, and the other members of the research team for support at sea and feedback. This research was funded by a grant from the New Zealand Ministry of Science and Innovation through the Coasts and Oceans OBI (to PWB) and a grant from the National Science Foundation (OCE-0825405) to SWW.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Acknowledgements
  7. References
  8. Supporting Information
  • Alldredge AL & Silver MW (1988) Characteristics, dynamics and significance of marine snow. Prog Oceanogr 20: 4182.
  • Alldredge AL & Youngbluth MJ (1985) The significance of macroscopic aggregates (marine snow) as sites for heterotrophic bacterial production in the mesopelagic zone of the subtropical Atlantic. Deep-Sea Res Part I Oceanogr Res Pap 32: 14451456.
  • Bearon RN (2007) A model for bacterial colonization of sinking aggregates. Bull Math Biol 69: 417431.
  • Boyd PW, Strzepek R, Chiswell S et al. (2012) Microbial control of diatom bloom dynamics in the open ocean. Geophys Res Lett 39: L18601.
  • Brinkhoff T, Bach G, Heidorn T, Liang LF, Schlingloff A & Simon M (2004) Antibiotic production by a Roseobacter clade-affiliated species from the German Wadden Sea and its antagonistic effects on indigenous isolates. Appl Environ Microbiol 70: 25602565.
  • Buchan A, Gonzalez JM & Moran MA (2005) Overview of the marine Roseobacter lineage. Appl Environ Microbiol 71: 56655677.
  • Caron DA, Davis PG, Madin LP & Sieburth JM (1986) Enrichment of microbial-populations in macroaggregates (marine snow) from surface waters of the North Atlantic. J Mar Res 44: 543565.
  • Clarke K & Gorley R (2006) PRIMER v6: User Manual/Tutorial. Bretonside Copy, Plymouth.
  • Cude WN, Mooney J, Tavanaei AA et al. (2012) Production of the antimicrobial secondary metabolite indigoidine contributes to competitive surface colonization by the marine roseobacter Phaeobacter sp strain Y4I. Appl Environ Microbiol 78: 47714780.
  • Dang HY & Lovell CR (2000) Bacterial primary colonization and early succession on surfaces in marine waters as determined by amplified rRNA gene restriction analysis and sequence analysis of 16S rRNA genes. Appl Environ Microbiol 66: 467475.
  • Dang HY & Lovell CR (2002) Seasonal dynamics of particle-associated and free-living marine Proteobacteria in a salt marsh tidal creek as determined using fluorescence in situ hybridization. Environ Microbiol 4: 287295.
  • Dang H, Li T, Chen M & Huang G (2008) Cross-ocean distribution of Rhodobacterales bacteria as primary surface colonizers in temperate coastal marine waters. Appl Environ Microbiol 74: 5260.
  • Davis MB (1967) Pollen deposition in lakes as measured by sediment traps. Geol Soc Am Bull 78: 849858.
  • Delong EF, Franks DG & Alldredge AL (1993) Phylogenetic diversity of aggregate-attached vs. free-living marine bacterial assemblages. Limnol Oceanogr 38: 924934.
  • Denman K, Hofmann E & Marchant H (1996) Marine biotic responses to environmental change and feedbacks to climate. Climate Change 1995: The Science of Climate Change (Houghton JT, Meira Filho LG, Callander BA, Harris N, Kattenberg A & Maskell K, eds), pp. 483516. Cambridge University press, Cambridge.
  • Giovannoni SJ, Tripp HJ, Givan S et al. (2005) Genome streamlining in a cosmopolitan oceanic bacterium. Science 309: 12421245.
  • Grossart HP, Levold F, Allgaier M, Simon M & Brinkhoff T (2005) Marine diatom species harbour distinct bacterial communities. Environ Microbiol 7: 860873.
  • Hall JA, Safi K, James MR, Zeldis J & Weatherhead M (2006) Microbial assemblage during the spring-summer transition on the northeast continental shelf of New Zealand. NZJ Mar Freshw Res 40: 195210.
  • Harms G, Layton AC, Dionisi HM et al. (2003) Real-Time PCR quantification of nitrifying bacteria in a municipal wastewater treatment plant. Environ Sci Technol 37: 343351.
  • Hoppe HG, Ducklow H & Karrasch B (1993) Evidence for dependency of bacterial growth on enzymatic hydrolysis of particulate organic matter in the mesopelagic ocean. Mar Ecol Prog Ser 93: 277283.
  • Jiao N, Herndl GJ, Hansell DA et al. (2010) Microbial production of recalcitrant dissolved organic matter: long-term carbon storage in the global ocean. Nat Rev Microbiol 8: 593599.
  • Jing HM, Shek L, Yung WY, Jin XW & Liu HB (2012) Dynamics of bacterial community composition during degradation of copepod fecal pellets. J Plankton Res 34: 700710.
  • Kellogg CTE & Deming JW (2009) Comparison of free-living, suspended particle, and aggregate-associated bacterial and archaeal communities in the Laptev Sea. Aquat Microb Ecol 57: 118.
  • King AL, Sanudo-Wilhelmy SA, Boyd PW et al. (2012) A comparison of biogenic iron quotas during a diatom spring bloom using multiple approaches. Biogeosciences 9: 667687.
  • Kirchman D (2001) Measuring bacterial biomass production and growth rates from leucine incorporation in natural aquatic environments. Methods Microbiol, Vol. 30 (Paul J, ed.), pp. 227237. Academic Press, San Diego, CA.
  • Knauer GA, Karl DM, Martin JH & Hunter CN (1984) In-situ effects of selected preservatives on total carbon, nitrogen and metals collected in sediment traps. J Mar Res 42: 445462.
  • Kwon EY, Primeau F & Sarmiento JL (2009) The impact of remineralization depth on the air-sea carbon balance. Nat Geosci 2: 630635.
  • Lozupone CA, Hamady M, Kelley ST & Knight R (2007) Quantitative and qualitative β diversity measures lead to different insights into factors that structure microbial communities. Appl Environ Microbiol 73: 15761585.
  • Martens T, Gram L, Grossart H-P et al. (2007) Bacteria of the Roseobacter clade show potential for secondary metabolite production. Microb Ecol 54: 3142.
  • Martin JH, Knauer GA, Karl DM & Broenkow WW (1987) Vertex - carbon cycling in the Northeast Pacific. Deep-Sea Res Part I Oceanogr Res Pap 34: 267285.
  • Matteson AR, Loar SN, Pickmere S et al. (2012) Production of viruses during a spring phytoplankton bloom in the South Pacific Ocean near New Zealand. FEMS Microbiol Ecol 79: 709719.
  • Matteson AR, Rowe JM, Ponsero AJ, Pimentel TM, Boyd PW & Wilhelm SW (2013) High abundances of cyanomyoviruses in marine ecosystems demonstrate ecological relevance. FEMS Microbiol Ecol 84: 223234.
  • Mc Carren J, Becker JW, Repeta DJ et al. (2010) Microbial community transcriptomes reveal microbes and metabolic pathways associated with dissolved organic matter turnover in the sea. P Natl Acad Sci USA 107: 1642016427.
  • Methe BA, Nelson KE, Pop M et al. (2012) A framework for human microbiome research. Nature 486: 215221.
  • Morris RM, Rappe MS, Connon SA, Vergin KL, Siebold WA, Carlson CA & Giovannoni SJ (2002) SAR11 clade dominates ocean surface bacterioplankton communities. Nature 420: 806810.
  • Muller-Niklas G, Schuster S, Kaltenbock E & Herndl GJ (1994) Organic content and bacterial metabolism in amorphous aggregations of the northern Adriatic Sea. Limnol Oceanogr 39: 5868.
  • Murphy RJ, Pinkerton MH, Richardson KM, Bradford-Grieve JM & Boyd PW (2001) Phytoplankton distributions around New Zealand derived from SeaWiFS remotely-sensed ocean colour data. NZJ Mar Freshw Res 35: 343362.
  • Nedashkovskaya OI, Kim SB, Han SK et al. (2003) Mesonia algae gen. nov., sp nov., a novel marine bacterium of the family Flavobacteriaceae isolated from the green alga Acrosiphonia sonderi (Kutz) Kornm. Int J Syst Evol Microbiol 53: 19671971.
  • Nodder SD, Boyd PW, Chiswell SM, Pinkerton MH, Bradford-Grieve JM & Greig MJN (2005) Temporal coupling between surface and deep ocean biogeochemical processes in contrasting subtropical and subantarctic water masses, southwest Pacific Ocean. J Geophys Res-Oceans 110: c12017.
  • Pomeroy LR, Hanson RB, McGillivary PA, Sherr BF, Kirchman D & Deibel D (1984) Microbiology and chemistry of fecal products of pelagic tunicates- rates and fates. Bull Mar Sci 35: 426439.
  • Quince C, Lanzen A, Curtis TP et al. (2009) Accurate determination of microbial diversity from 454 pyrosequencing data. Nat Methods 6: 639641.
  • Rappe MS, Connon SA, Vergin KL & Giovannoni SJ (2002) Cultivation of the ubiquitous SAR11 marine bacterioplankton clade. Nature 418: 630633.
  • Riemann L, Steward GF & Azam F (2000) Dynamics of bacterial community composition and activity during a mesocosm diatom bloom. Appl Environ Microbiol 66: 578587.
  • Riley GA (1970) Particulate matter in seawater. Adv Mar Biol 8: 1118.
  • Schloss PD, Westcott SL, Ryabin T et al. (2009) Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 75: 75377541.
  • Shanks AL & Trent JD (1979) Marine snow- microscale nutrient patches. Limnol Oceanogr 24: 850854.
  • Smith DC, Simon M, Alldredge AL & Azam F (1992) Intense hydrolytic enzyme-activity on marine aggregates and implications for rapid particle dissolution. Nature 359: 139142.
  • Taylor GT, Thunell R, Varela R, Benitez-Nelson C & Scranton MI (2009) Hydrolytic ectoenzyme activity associated with suspended and sinking organic particles within the anoxic Cariaco Basin. Deep-Sea Res Part I Oceanogr Res Pap 56: 12661283.
  • Teeling H, Fuchs BM, Becher D et al. (2012) Substrate-controlled succession of marine bacterioplankton populations induced by a phytoplankton bloom. Science 336: 608611.
  • Wagner-Döbler I, Rheims H, Felske A et al. (2004) Oceanibulbus indolifex gen. nov., sp. nov., a North Sea alphaproteobacterium that produces bioactive metabolites. Int J Syst Evol Microbiol 54: 11771184.
  • Wang Y & Qian P-Y (2009) Conservative fragments in bacterial 16S rRNA genes and primer design for 16S ribosomal DNA amplicons in metagenomic studies. PLoS ONE 4: e7401.
  • Weller DI, Law CS, Marriner A, Nodder SD, Chang FH, Stephens JA, Wilhelm SW, Boyd PW & Sutton PJH (2013) Temporal variation of dissolved methane in a subtropical mesoscale eddy during a phytoplankton bloom in the southwest Pacific Ocean. Progr Oceanogr 116: 193206.
  • Welschmeyer NA (1994) Fluorometric analysis of chlorophyll a in the presence of chlorophyll b and pheopigments. Limnol Oceanogr 39: 19851992.
  • Wilhelm SW, King AL, Twining BS et al. (2013) Elemental quotas and physiology of a southwestern Pacific Ocean plankton community as a function of iron availability. Aquat Microb Ecol 68: 185194.
  • Williams TJ, Wilkins D, Long E, Evans F, DeMaere MZ, Raftery MJ & Cavicchioli R (2012) The role of planktonic Flavobacteria in processing algal organic matter in coastal East Antarctica revealed using metagenomics and metaproteomics. Environ Microbiol 15: 13021317.
  • Zimmermann H (1997) The microbial community on aggregates in the Elbe Estuary, Germany. Aquat Microb Ecol 13: 3746.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results and discussion
  6. Acknowledgements
  7. References
  8. Supporting Information
FilenameFormatSizeDescription
fem12213-sup-0001-FigS1.TIFimage/tif2619KFig. S1. Flow cytometric analyses of A) phytoplankton and B) picoplankton during our observations.
fem12213-sup-0002-FigS2.tifimage/tif10469KFig. S2. Heat map of relative abundances of chloroplast sequences. Color indicates% of library the OTU represents (see scale bar).
fem12213-sup-0003-TableS1-S3.docxWord document22K

Table S1. PCR Primers for 16S amplification and bar-coding reactions.

Table S2. Summary of samples, environmental parameters and activities included in analyses. Bacterial productions data are given as 10-11 mmols leucine incorporated mL-1 hr-1.

Table S3. qPCR estimates of bacterial abundances for RESPIRE sediment trap samples. Sample NZ165 was not measured.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.