Viral and bacterial abundance and production in ... - Wiley Online Library

RESEARCH ARTICLE

Viral and bacterial abundance and production in the Western Pacific Ocean and the relation to other oceanic realms Janet M. Rowe1, Jennifer M. DeBruyn2, Leo Poorvin1, Gary R. LeCleir1, Zackary I. Johnson3, Erik R. Zinser1 & Steven W. Wilhelm1 1

Department of Microbiology, The University of Tennessee, Knoxville, TN, USA; 2Biosystems Engineering and Soil Science, The University of Tennessee, Knoxville, TN, USA; and 3Marine Laboratory, Nicholas School of the Environment, Duke University, Beaufort, NC, USA

Correspondence: Steven W. Wilhelm, Department of Microbiology, The University of Tennessee, Knoxville, TN 37996, USA. Tel.: +1 865 974 0665; fax:+1 865 974 0665; e-mail: [email protected] Present address: Janet M. Rowe, School of Biological Sciences, The University of Nebraska, Lincoln, NE, 68583-0900, USA. Received 5 April 2011; revised 26 September 2011; accepted 1 October 2011. Final version published online 1 November 2011.

MICROBIOLOGY ECOLOGY

DOI: 10.1111/j.1574-6941.2011.01223.x Editor: Gary King Keywords microbial ecology; marine viruses; spatial diversity.

Abstract We completed a transect through the Western Pacific Warm Pool to examine how environmental variables may influence viral and bacterial abundance and production rates in this globally important oceanic region. Of the variables analyzed, viral abundance and production had the most significant relationship to bacterial cell abundance: viral parameters were not significantly correlated to the measured environmental variables, including temperature. Bacterial production rates were significantly correlated to temperature in open ocean waters, but not in waters close to land masses. Analyses of 16S rRNA gene by pyrosequencing indicated only minor changes in eubacterial community structure across the transect, with a-proteobacteria dominating all sampled populations. Diversity within the prokaryotic community did not correlate directly with viral abundance or activity. Comparisons to two other ocean-scale transects (> 8000 km of open ocean in total) in the Atlantic Ocean indicated that correlations between viral and bacterial abundance and production relative to environmental variables are regime dependent. In particular, correlations to temperature showed remarkable differences across the three transects. Collectively, our observations suggest that seemingly similar oceanic regions may have very different microbial community responses to environmental variables. Our observations and analyses demonstrate that ocean-scale generalizations may not apply in the case of viral ecology.

Introduction Marine viruses contribute a diverse range of features to the aquatic microbial community. As agents of mortality (Thingstad & Lignell, 1997; Wommack & Colwell, 2000), nutrient recycling and regeneration (Poorvin et al., 2004; Middelboe & Jo¨rgensen, 2006) and potentially horizontal gene transfer (Lindell et al., 2004), viruses are integral members of all aquatic ecosystems and are able to affect community structure and dynamics (Brussaard et al., 2008; Wilhelm & Matteson, 2008). Given the global importance of their microbial prey in biogeochemical cycling (Fuhrman, 1992), viruses also have a large-scale influence as predators of these key players: a top-down effect (Wilhelm & Suttle, 1999). Typically, in situ information on marine viruses has been collected from coastal or short open ocean transects FEMS Microbiol Ecol 79 (2012) 359–370

(e.g. Wommack et al., 1992; Wilhelm et al., 2003; Brussaard et al., 2005). For example, examinations of viral and bacterial abundance in the Chesapeake Bay have focused on temporal (c. 1 year) variations (Wommack et al., 1992), while sampling in Sagami Bay was designed to compare benthic viral communities including those associated with cold seeps (Middelboe et al., 1996). Longer sampling tracks (c. 5000 km) such as those reported by Wilhelm et al. (2003) and Winter et al. (2008) are limited, but even in those cases, the studies represent primarily coastal communities. In one pelagic marine study, data collected from a transect of > 3000 km of open ocean were combined with data from shorter coastal studies as part of a large-scale evaluation of viral and bacterial variables such as abundance and production (Hewson & Fuhrman, 2007). While these studies have individual as well as collective value, there remains little ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

J.M. Rowe et al.

360

information concerning the constraints on virus activity spanning open ocean regimes. Of the variables that may be drivers of viral abundance and activity, temperature, nutrients and microbial diversity stand out as potentially the most important. Temperature may directly affect the physical structure of the viral particles, triggering particle decay or weakened infectiousness. Along with nutrient availability, temperature may also indirectly influence virus production by affecting the host’s growth and physiology. Changes in microbial diversity affect the availability of hosts, thus influencing the viruses that can survive in a particular environment. Previous observations in the Sargasso Sea (Rowe et al., 2008) demonstrated that sea surface temperature significantly correlated to viral abundance as well as other biological variables. Correlations between viral abundance and temperature have been observed in other environments, including the coastal waters of South Texas, Peconic Bay, NY (Suttle & Chan, 1993) and Tampa Bay, FL (Jiang & Paul, 1994). Additionally, increased temperatures have been shown to induce lytic activity from a latent virus infection of the zooxanthellae of Anemonia viridis (Wilson et al., 2001). Studies investigating virus– host interactions in higher plant systems demonstrated a wide range of results, including increases and decreases in infectivity (Boland et al., 2004). These observations led us to the hypothesis that increasing temperatures (potentially due to climate change) may have an influence on marine viruses and their activity. The Western Pacific Warm Pool contains the warmest pelagic ocean surface waters; temperatures can reach above 30 °C. This phenomenon is formed as part of the El Nino/Southern Oscillation (ENSO) event during the El Nino phase. The ENSO is the most well-known interannual variation in weather patterns as it has worldwide climatological and economic impacts (Halpert & Ropelewski, 1992; Goddard & Dilley, 2005; McPhaden et al., 2006). These high temperatures make the Western Pacific Warm Pool an ideal setting for studies on the effect of temperature on microbial communities; the global influence of this region make it an important subject for ecological analyses. Our goals in this study were twofold. First, we examined environmental and biological variables that influence viral activity in the Western Pacific while traversing the Western Pacific Warm Pool. Because data from our previous study showed significant correlations with temperature (Rowe et al., 2008), we had a sub-focus on its correlations to in situ viral and bacterial abundances and production rates across the Western Pacific transect. Secondly, we compared observations in the Western Pacific to those from the North Atlantic and Sargasso Sea (Rowe et al., 2008) to allow for a regional vs. global analyses of virus activity. ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

Methods Sample and metadata collection

Sampling was completed onboard the R/V Kilo Moana during the WP2 (Western Pacific Warm Pool) cruise (January 3–February 11, 2007). Daily CTD data and water samples were collected while traveling through the Western Pacific Warm Pool. The transect began at 19.0°N, 169.04°W, progressed through the Warm Pool down to 36.15°S, 161.79°E, and ended at 28.76°S, 155.37°E. The center of the Warm Pool (station 14), designated so by having the highest temperatures observed during the transect, was located at 9.25°S, 170°E. Stations 8 and 16 are the upper- and lowermost Warm Pool stations, respectively, based on the temperature demarcation (> 28.5 °C) used by Wang & Enfield (2001) for the Western Hemisphere Warm Pool. See Fig. 1 for sampling locations. For this study, all samples were collected at 5 m, unless

Fig. 1. Stations from the WP2 (Western Pacific Warm Pool) cruise, sampled from January to February in 2007. Location 14 corresponds to the center of the Western Pacific Warm Pool. Locations 8 and 16 are the northern and southern, respectively, boundaries of the Warm Pool.

FEMS Microbiol Ecol 79 (2012) 359–370

361

Viruses and bacteria in the Western Pacific

specified otherwise, during the morning (8 AM, local time). Water temperature and salinity were measured by the ship’s CTD package attached to the rosette sampler. Surface water samples for nutrient analyses were collected and analyzed as described by Hynes et al. (2009). Chlorophyll a

Size-fractionated chlorophyll a (chl a) concentrations were measured in duplicate for pico- (0.22–2.0), nano(2.0–20.0 µm), and microphytoplankton (> 20.0 µm) size classes. These measurements were taken from parallel filtrations using the corresponding nominal pore-size, 47-mm-diameter polycarbonate filters (Osmonics). Chl a was extracted in 90% acetone (24 h at + 4 °C) and the concentration determined fluorimetrically (Welschmeyer, 1994). Viral abundance and production

Samples for determining in situ virus particle abundance and abundances during production assays were prepared immediately after sample collection (Wen et al., 2004). Briefly, virus particle abundance was determined with epifluorescence microscopy using SYBR Green I (Noble & Fuhrman, 1998) on a Leica DMRXA microscope with a ‘wide blue’ filter set (kEx = 450–490 nm, kEm = 510 nm, with suppression filter at 510 nm). Single microscopic slides were prepared for each sample, and the variation of particles over the > 20 fields of view examined to ensure adequate representation of the particles present in each sample. The virus reduction assay (Wilhelm et al., 2002; Weinbauer et al., 2010) was used to measure virus particle production rates. For each 500-mL sample, viruses were reduced over a 0.22-µm filter, and then cells were resuspended in 500 mL of ultrafiltrate, generated from water at the same depth and station as previously described (Wilhelm & Poorvin, 2001). Triplicate 250-mL polycarbonate bottles, each containing 150 mL of the prepared sample, were incubated in an on-deck incubator covered with neutral density screening to reduce penetrating light to simulate in situ light and through which surface water was continually pumped to maintain temperature. Bottles were incubated for 10 h with samples for virus enumeration taken every 2.5 h. Time zero (T = 0) cell abundance was also determined with epifluorescence microscopy, using Acridine Orange stain (Hobbie et al., 1977). Production rates were calculated as the average of virus reoccurrence rates in the three independent incubations. Rates were corrected for host cell losses during processing by multiplying by the ratio of experimental T = 0 bacterial abundance to in situ bacterial abundance estimates. FEMS Microbiol Ecol 79 (2012) 359–370

Water column bacterial abundance, production, and viral–bacterial contact rates

Subsamples for bacterial abundance (including Archaea) were collected and frozen with 0.125% (final concentration) glutaraldehyde at 80 °C following Vaulot et al. (1989) for later analysis. These samples were stained with SYBR Green I prior to flow cytometric analysis, following Marie et al. (1997). All flow cytometry samples were run on a Becton Dickinson FACSCalibur flow cytometer modified with a syringe pump for quantitative sample delivery following Johnson et al. (2010). Heterotrophic bacterial production rates were measured using the microcentrifuge method of determining 3 H-leucine incorporation described by Kirchman (2001) using 2.0-mL screw cap copolymer microfuge tubes (Phenix, see Pace et al., 2004). Independent triplicate samples were corrected using killed (5% TCA) controls to provide average rates for each water sample. Virus–bacterium contact rates were estimated according to Murray & Jackson (1992) Contact rate ¼ ð2SpxDv ÞVB where S is the dimensionless Sherwood number, estimated at 1.06 (Wilhelm et al., 1998). Estimates for the diameter of a marine bacterium (x = 0.45 9 104 cm) were based on work by Lee & Fuhrman (1987). Dv, the diffusivity of viruses, is estimated at 3.456 9 103 cm2 day1. V and B are the abundances per milliliter of viruses and bacteria, respectively. Minimum burst size and percentage of visibly infected cells

Burst size and percentage of visibly infected cells were determined by transmission electron microscopy as described previously (Weinbauer & Suttle, 1996). Samples were preserved with 2% (final concentration) glutaraldehyde and stored at 4 °C until processing. Centrifugation was used to collect c. 30 mL of sample on top of 400mesh copper grids covered with collodian film and coated with carbon. Samples on top of the grids were stained with 0.75% uranyl formate. Prepared grids were then viewed with a Hitachi H-800 transmission electron microscope under various magnifications (c. 15 000– 20 000 for scanning and c. 30 000–40 000 for examining cells). A minimum of c. 200 cells were viewed per grid. Bacterial diversity estimates

The diversity of eubacterial populations across a subset of stations was investigated using PCR and high-throughput ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

J.M. Rowe et al.

362

pyrosequencing. DNA was extracted from particles in water samples collected onto 0.2-lm polycarbonate filters as previously described (Rinta-Kanto et al., 2005). DNA sequences (156-bp (after primer removal) of 16S rRNA genes (encoding the hypervariable V3 region) were amplified using Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA) and primers 338F and 533R (Huse et al., 2008). After the initial PCR amplification (35 cycles), PCR products were purified with the Qiaquick PCR cleanup kit (Qiagen, Valencia, CA). A second round of PCR (6 cycles) was performed on the PCR products to add forward and reverse barcodes to each sequence (Hamady et al., 2008, see Table S1, Supporting Information). Bar-coded products were then pooled and cleaned up using the Qiaquick PCR cleanup kit. The pooled products were prepared for pyrosequencing according to manufacturer’s FLX protocols (454 Life Sciences, Branford, CT) at the UTK/ORNL Joint Institute for Biological Sciences (Oak Ridge, TN). Pyrosequencing results were initially processed using the ROCHE software. Individual libraries were then processed through the Ribosomal Database Project pyrosequencing pipeline (Cole et al., 2007, 2009). For each read, the primers were trimmed from the beginning and the end, and low-quality sequences removed (Huse et al., 2007, 2008). Sequences were flagged as low quality when (i) they were < 50 nucleotides, (ii) the start of the raw sequence did not have an exact match to a primer sequence, (iii) the sequence contained one or more ambiguous nucleotides (Ns), or (iv) if the first nucleotides of a tag did not correspond to the expected nucleotide run key (used to sort the pyrosequencing reads). Sequences were binned into individual libraries based on the tags. Taxonomic assignments for each sequence were generated using the RDP Classifier program. Sequences were then processed by Complete Linkage Clustering and rarefaction estimated for 3% dissimilarities. Results from linkage clustering were also used to calculate diversity (Chao1, Shannon’s). See Tables S1 and S2 for primers and additional metrics, respectively. Principal component analyses (PCA) were completed on unique sequences (showing 100% identity) within individual libraries using the UNIFRAC service (Lozupone et al., 2006). A neighbor-joining tree was constructed using MEGA 4.0 (Tamura et al., 2007), bootstrapped with 5000 iterations and used as input for UNIFRAC. The environmental file for PCA input included weights for sequences (number of occurrences in each library) and was generated using the weighted UNIFRAC algorithm. Pairwise comparisons of environments using both the UNIFRAC and P tests were completed for 100 permutations. ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

Statistical correlations

To examine the relationships between pairs of data, both linear regressions and Spearman’s rank correlations were utilized. Linear regressions of select variables were performed using SIGMAPLOT (ver. 10.0) on untransformed data. Spearman’s rank correlations (a nonparametric analysis) were performed using the statistical software, NCSS and SPSS, to generate correlations between all pairs of variables. Significance was defined as P < 0.05. The multivariate analysis, canonical correspondence analysis (CCA), was used to assess the relationships among sets of data. For this, the program, CANOCO (ver 4.5; Plant Research International) was used, with the significance of the axes determined by Monte Carlo permutation tests. Relationships between bacterial community composition and production variables were examined in PRIMER-E (ver 6.1.1.3; Primer-E, Ivybridge, UK).

Results Biological and physicochemical variables in the Western Pacific transect are provided in Fig. 2 for samples collected at 5 m and include temperature, salinity, size-fractionated chl a, viral abundance, virus production rate, bacterial abundance, and bacterial biomass production rate. Temperatures increased and then decreased along the transect, with the maximum temperature (30.5 °C) at the Warm Pool center, station 14 (9.25°S, 170°E). Sizefractionated chl a concentrations show a consistent dominance of picophytoplankton with peak concentrations just before and just after the Warm Pool center, while nanophytoplankton and microphytoplankton remained at fairly low concentrations. Nanophytoplankton mirror the picophytoplankton trends with smaller but still significant increases immediately adjacent to the Warm Pool center. Estimates of viral abundance showed small but significant fluctuations across the transect, although no significant overall trends were noted. Virus production rates demonstrated few significant changes and no clear trend across this transect. Bacterial abundance also fluctuated without a clear pattern, but bacterial biomass production rates were highest at the center of the Warm Pool, decreasing to lower rates at stations further away. TEM analyses revealed a range of minimum burst sizes from 12 to 30 (Table 1), with an average of 16.6. Estimates of visibly infected cells ranged from 1.8% to 3.8%, with an average of 2.9%. Microbial diversity across the stations sampled during this cruise was relatively constant (Fig. S1, Table S2). The bacterial population was dominated by members of the phylum Proteobacteria at all stations, ranging from 70% to 92% of the total microbial community as identified by FEMS Microbiol Ecol 79 (2012) 359–370

363


32

6

(a)

(d) 5 –1

Viruses (x 10 mL )

28 26 24

20

2

2.5 –1

hr )

(b)

(e)

2.0

5

Virus production ( 10 mL

–1

36

3

1

22

Salinity (psu)

4

6

Temperature (°C)

30

35

34

33 1.0

1.5 1.0 0.5 0.0 1.4

(c)

(f)

1.2

0.4

0.2

10

15

20

Station #

25

30

–1 –1

5

0.6

0.2 12

(g)

10 8 6 4 2 0

5

10

15

20

25

30

Station #

Sargasso Sea

North Atlantic

Western Pacific

0.043–0.16

0.22–2.25

0.25–1.15

1.38–18.92

8.19–28.02

5.02–47.64

0.02–1.91

0–5.6

0–1.72

4.61–8.78

8.23–23.80

2.93–12.36

0.20–0.54

0.44–1.88

0.59–10.61

11.67 19.8–24.3 36.51–36.78

11–23 10.7–15.5 35.20–35.87

12–30 20.8–30.5 33.87–35.91


0.8

0.4

Table 1. Data ranges for selected measurements in the Sargasso Sea, North Atlantic, and Western Pacific. Sargasso and North Atlantic data are further described elsewhere (Rowe et al., 2008)

Chl a (lg L1, > 0.2 lm) Virus abundance (9 105 mL1) Virus production (9 105 mL1 h1) Bacterial abundance (9 105 mL1) Bacterial production (nmol leu L1 day1) Burst size (cell1) Temperature (°C) Salinity (psu)

1.0

6

0.6

Bacterial production (nmol leu L d )

Fig. 2. Virus–host variables measured across the Western Pacific. Points presented in chronological order of sampling date. (a) Temperature; (b) salinity; (c) size-fractionated chl a concentrations: circles (> 20.0 µm), squares (0.2–2.0 µm), and triangles (0.22– 2.0 µm). Error bars indicate the range of duplicate samples. (d) VLP abundance measurements. Error bars indicate the variation found after 31 fields of view were counted. (e) VLP production rate measurements. Error bars indicate the standard deviation from triplicate samples. (f) Bacterial abundance. (g) Bacterial biomass production rate measurements. Error bars indicate the standard deviation of triplicate samples.

Chla (µg L–1)

–1

Bacteria (x 10 mL )

0.8

the RDP II classifier program. An examination of diversity metrics relative to virus production and abundance revealed no significant relationships, although a PCA analysis of microbial community composition demonstrated that the stations with higher virus production rates also had the most similar microbial communities (Fig. S2). Figure 3 shows a comparison of viral abundance, virus production, bacterial abundance, and bacterial biomass production rate vs. temperature for the combined Atlantic and Pacific transects’ data. The North Atlantic viral abundances and production rates cover the same range as those in the Sargasso Sea and Western Pacific despite occurring at lower temperatures. Overall, while the lowest viral abundances and some of the lowest production rates were observed in the Sargasso Sea, the abundances and production rates of viruses were consistent across all temperatures as well as across the different regimes. Bacterial ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

364

J.M. Rowe et al.

Fig. 3. Viral and bacterial variables compared with temperature across all three transects. (a) VLP abundance; (b) VLP production rate; (c) bacterial abundance; (d) bacterial biomass production rate. Sargasso Sea = white triangles; North Atlantic = black circles, and Western Pacific = gray squares (in (d) stations north of the center of the Warm Pool are half black, and stations south of the center of the Warm Pool are crossed).

abundance was highest in the North Atlantic, with the Sargasso Sea and Western Pacific having lower abundances. With one exception, bacterial abundance above 24 °C appeared to hit a maximum just below 0.7 9 106 mL1, while samples below 24 °C showed a higher average (although n = 3). When bacterial biomass production rates in the Western Pacific are compared to temperature, it appears that there is a positive relationship; however, no significant correlations were observed. The majority of bacterial production rates in the Western Pacific were significantly higher than the highly productive North Atlantic and the Sargasso Sea. Bacterial production rates vary nearly 10-fold over the three transects. An abbreviated list of these and other variables commonly measured across all three transects is shown in Table 1. Using Spearman’s rank correlation analysis, all possible pairs of variables were examined for the Western Pacific, and then for all three transects combined. Previously in the Sargasso Sea, surface temperature was found to be significantly correlated with viral abundance, virus proª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

duction rate, size fractionations of chl a (0.22–2.0 µm, 2.0–20.0 µm, and > 20.0 µm), cell abundance, and bacterial biomass production rates (Rowe et al., 2008; Table S3). Despite being oligotrophic, like the Sargasso Sea, there were few significant correlations in the Western Pacific, none of which were with temperature. Abbreviated results are shown in Table 2, while the total Spearman’s rank correlations for the Sargasso Sea, North Atlantic, Western Pacific, and all three transects combined are shown in Tables S3–S6, respectively. To provide a multivariate view of the relationships among all measured variables, CCA was used to assess the similarity (shown by proximity) of data from the Western Pacific and from the stations of all three transects. Virus abundance, virus production rate, cell abundance, bacterial biomass production rate, burst size, and contact rate were used as the biological response variables, and temperature, salinity, and chl a were used as the environmental predictor variables. While the axes of the Western Pacific analysis were not significant (determined by the Monte Carlo permutation test), those of FEMS Microbiol Ecol 79 (2012) 359–370

365

0.536 0.728 0.575 0.371 0.034 0.662 0.004 0.562 0.791 0.180 0.362 N/D < 0.001

the overall analysis (including data from all three transects) were, meaning that the latter results were statistically significant. CCA was completed using all data (56 stations) from these three open ocean regimes (Fig. 4). As shown by the clustering of station points, each transect appears distinct from the others.

0.262 0.284 0.289 0.615 0.478 0.740 0.367 0.597 0.990 0.545 0.245 N/D 0.783

Discussion The current study allowed us to test the relationships between estimates of viral abundance and production rate with environmental variables and bacterial diversity in the Western Pacific. It further allowed us to build on previous observations in two distinct oceanic realms in the Atlantic Ocean (the Sargasso Sea and the North Atlantic) to determine whether similar variables constrained viral activity.

0.016 < 0.001 0.022 < 0.001 0.007 0.052 1.000 0.748 0.050 N/D N/D N/D N/D

0.923 0.730 0.818 0.967 0.945 0.789 0.000 0.511 0.890 N/D N/D N/D N/D

< 0.001 0.005 < 0.001 < 0.001 < 0.001 0.002 1.000 0.131 0.001 N/D N/D N/D N/D

0.288 0.276 0.273 0.132 0.191 0.087 0.233 0.155 0.004 0.158 0.298 N/D 0.072

0.156 0.088 0.142 0.225 0.516 0.111 0.646 0.163 0.072 0.331 0.229 N/D 0.808 0.245 0.478 0.398 0.370 0.774 0.874 0.509 0.951 0.251 – – – – 0.196 0.120 0.143 0.152 0.050 0.028 0.119 0.011 0.206 – – – –

0.423 0.226 0.319 0.453 0.066 0.272 0.046 0.409 0.087 – – – –

0.001 0.098 0.018 0.001 0.650 0.056 0.791 0.003 0.550 – – – –

0.152 0.127 0.079 0.164 0.552 0.067 0.200 0.188 0.006 0.176 0.103 0.527 0.042

0.676 0.726 0.829 0.651 0.098 0.855 0.704 0.603 0.987 0.627 0.777 0.117 0.907

0.291 0.308 0.042 0.312 0.417 0.195 0.600 0.103 0.165 0.038 0.271 0.267 0.025

0.167 0.143 0.845 0.138 0.068 0.397 0.285 0.630 0.440 0.865 0.211 0.218 0.911

0.653 0.852 0.625 0.686 0.702 0.572 0.000 0.117 0.667 N/D N/D N/D N/D

Across the Western Pacific


ND, not determined; rs, correlation coefficient; P, probability that correlation is because of chance.

9 0.842 0.725 0.995 0.001

Virus abundance Chlorophyll Total (>0.22lm) >20lm 2.0–20.0lm 0.22–2.0lm Cell abundance Bacterial production Burst size Salinity Temperature Orthophosphate Nitrate + nitrite Nitrite Silicate

9

9

0.517

0.154

9

9

rs rs rs P rs Parameter

Abundance

P

rs

P Production Production

Abundance

P

Production

P

0.005

9

P

9

0.052

9

P rs P rs rs

Abundance Production Abundance

North Atlantic Sargasso Sea Western Pacific All transects

Table 2. Spearman’s rank correlation analysis contrasting virus production rates and viral abundance with biotic and abiotic variables. Data from the Sargasso Sea and North Atlantic are reprinted from Rowe et al. (2008). Relationships considered significant in this study (P < 0.05) are given in bold


Discontinuities in temperature (Fig. 2a) and salinity (Fig. 2b) demonstrate that sampling occurred through multiple water masses during the transect through the Western Pacific. While it is not unexpected that picophytoplankton, primarily Prochlorococcus and Synechococcus, would dominate the Western Pacific (Fig. 2c), it is interesting to note that the highest picophytoplankton biomass concentrations (based on chl a) were found just before and just after the Warm Pool center at intermediate temperatures. To a lesser extent, this pattern is mirrored by the nanophytoplankton, but not the microphytoplankton. Although we observed a decrease in picophytoplankton abundance toward the Warm Pool center, the concentrations are within the range of those in other distant locations, and it must be considered that photobleaching of populations may introduce variance in flow cytometric estimates at some locations. All together, the results suggest that peak temperatures within the Warm Pool may not be optimal for phytoplankton biomass accumulation. In contrast to phytoplankton, viral abundance (Fig. 2d) and virus production (Fig. 2e) remained constant across the transect. The range of viral abundance and production estimates are similar to other open ocean data sets (Weinbauer et al., 2002; Winter et al., 2004; Rowe et al., 2008). However, despite these similarities, it is likely that the viral communities across these regions differ in composition and relative rates of production. This would be due to differing host populations and environmental factors and is also in agreement with the findings of Angly et al. (2006). Bacterial abundance (Fig. 2f) and biomass production (Fig. 2g) provide contrasting results from the viral ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

366

J.M. Rowe et al.

Fig. 4. CCA of Sargasso Sea, North Atlantic, and Western Pacific stations (see Fig. 3 for legend). Clustering of data points suggests that each transect is statistically distinct from the other two. The three points at the top of the image represent transitional stations between the Sargasso and North Atlantic. Predictor (environmental) variables = total, micro-, nano-, and pico-size-fractionated chl a, salinity, and temperature. Response (biological) variables = VA, virus abundance; VP, virus production rate; BA, bacterial abundance; BP, bacterial production rate; BS, burst size; and CR, contact rate are located at their optima. Where available, symbols have been color coded (log scale) for surface virus production rates.

metrics. Bacterial abundance over the transect displayed no clear trends, with surface samples within stations being within a 10% variance of each other. Bacterial production drastically and significantly peaked inside the Western Pacific Warm Pool, suggesting that bacterial production is not a simple function of bacterial abundance (or vice versa). In our survey, bacterial production rates were higher than reported in other oligotrophic open oceanic regimes (Noble & Fuhrman, 2000; Steinberg et al., 2001; Church et al., 2004; Rowe et al., 2008). It is possible that this regime is more conducive to higher production; though, it is also possible that the higher rates may be the result of more efficient detection (see Pace et al., 2004). Parametric analyses are not always ecologically relevant statistical methods because basic assumptions (e.g. normal distribution) may not be met, and such is the case with this data set. Spearman’s rank correlation is a nonparametric means of assessing significant correlations and is therefore better suited to the data we collected. For this reason, we undertook a Spearman’s rank correlation analysis for all possible pairs. In previous research, bacterial production correlated significantly to viral abundance in the Sargasso Sea (linear regression, r2 = 0.581, P = 0.004, and Spearman’s correlation, rs = 0.789; P = 0.002) but not in the North Atlantic (Table S3; Rowe et al., 2008). Given the similar trophic status of the Western Pacific transect waters to the Sargasso, we expected to observe a similar correlation between bacterial production and viral abundance; however, this was not the case. Similarly, ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

pairwise comparisons between temperature and viral and bacterial variables suggested no effects of temperature on virus production or abundance in the Western Pacific, implying that temperature is not a direct driver of in situ virus activity. It is equally surprising that neither bacterial abundance nor production was significantly correlated to temperature in spite of observable differences between bacterial production in the Warm Pool center relative to other stations. It is interesting to note that within the Western Pacific transect, stations north of the Warm Pool center (1–14, Fig. 3d, gray and black squares) did demonstrate a significant correlation between bacterial production and temperature (r2 = 0.645, P < 0.001), while the stations South of the Warm Pool center (15–30, Fig. 3d, crossed gray squares) did not. These results suggest that the transect through the Western Pacific crossed two very distinct oceanic realms. Indeed, the Northern stations are all open ocean in nature, while the Southern stations are in waters that in may be influenced by regional land masses (see Fig. 1). Microbial diversity during this cruise was estimated from 16S rRNA gene amplicon libraries (Fig. S1). In total, more than 20 000 sequences were generated for this study to determine whether major changes occurred along the transect. For surface water samples, all stations fell along a gradient in PCA analyses with the exception of Station 11 (Fig. S2). More than 19% of the sequences at that station were most closely related to GP IIa cyanobacteria (based on RDP II Classifier analyses). The samples collected at the deep chlorophyll maximum in the middle of FEMS Microbiol Ecol 79 (2012) 359–370


the Warm Pool (station 14-DCM) were predictably different from the surface samples (Figs S1 and S2). Unfortunately, the overlap in our molecular analyses and virus production assays did not contain sufficient replicates to find significant relationships in this data set. From the available data, an intriguing clustering of the stations (by diversity) with the highest virus production rate is apparent (Fig. S2). This observation and a growing body of literature suggest that in the near future, we will be able to build on this observation with higher resolution sampling to clarify the relationship(s) between virus production and cell diversity. Comparison with other regimes

Previously, we observed that biological variables were ‘predictable’ in the oligotrophic Sargasso Sea based on environmental measures, while in the North Atlantic, a threshold effect of environmental factors on viral production and abundance was evident (Rowe et al., 2008). Curiously, in the Western Pacific, we observed neither of these trends, and comparisons between temperature and viral variables across all three transects demonstrated no temperature effects. However, the ranges of most of the biological variables examined were overlapping across these three regimes (Table 1). A closer look at bacterial abundance and production compared to temperature across all three transects (Fig. 3c and d) revealed a dynamic relationship. Bacterial abundance in the cooler North Atlantic was higher than in the oligotrophic Sargasso Sea and Western Pacific, probably because of the high primary productivity associated with the North Atlantic Spring Bloom, which could support bacterial carbon demand. While it is logical that temperature could affect bacterial abundance and increase bacterial productivity, the actual standing stock is likely a more complicated function of various loss rates (e.g. viral lysis and grazing) that may also be temperature dependent. Although bacterial abundance and temperature were not significantly correlated in the Western Pacific, the combined data set (Western Pacific, Sargasso Sea, and North Atlantic transects) showed a significant second-order linear regression (r2 = 0.729, P < 0.001), with Spearman’s rank correlation analysis yielding further support (rs = 0.785, P < 0.001), suggesting that within the individual transects, the range and number of observations may be too limited for trends to be observed. This highlights the importance of geographic scale in understanding biological drivers and is consistent with other studies that show that ocean-scale changes in temperature may regulate the abundance and diversity of the bacterial community (Baldwin et al., 2005; Johnson et al., 2006). FEMS Microbiol Ecol 79 (2012) 359–370

367

Despite higher bacterial abundances and high primary productivity, the bacterial production rate estimates in the North Atlantic were lower than in the Western Pacific. While bacterial production was not significantly correlated to temperature in the Western Pacific, for all three transects both regression (r2 = 0.365, P < 0.001) and Spearman’s rank correlation analysis (rs = 0.451, P = 0.002) revealed positive, significant relationships. When overall analyses were again performed, but with Western Pacific data shortened to either the first or second half (Sargasso Sea, North Atlantic, and Western Pacific stations 1–14 or 15–30), the strength of correlation was higher, r2 = 0.518, P < 0.001 and r2 = 0.435, P < 0.001, respectively, than with the full Western Pacific transect included. This agrees with earlier analyses of the Western Pacific alone and could suggest that the impacts of temperature in near-land and open ocean environments are influenced by different factors. Future studies should aim to separate these influences along with providing better resolution of correlations between temperature and biological variables. Taken all together, the results indicate that constraints on the activity of viruses in one regime may not hold for other locations. Table 2 shows how an overall view of the data from all three transects tells a different story than each transect separately. For example, significant correlations observed in the combined analysis are not present in all three of the separate transects. This could be the result of a need for a larger sample size before a correlation reaches significance, as may be the case for viral abundance compared with salinity. However, another option is that the correlations found in one transect could be strong enough to still be seen in the overall analysis, albeit as weaker correlations (e.g. viral abundance compared to chl a measures from the Sargasso Sea transect). Because of these concerns, there was a need for multivariate analyses to examine these measured variables and to determine whether or not the data sets should be combined. CCA, a unimodal and constrained multivariate analysis, was chosen to compare all three complete data sets (Fig. 4). The unimodal component of this analysis allows one to detect optima along gradients rather than assuming linear relationships and thus may be of more ecological relevance than linear-based analyses. CCA for all three combined transects displayed distinct clusters corresponding to the different oceanic regions. Because sampling in the Sargasso Sea progressed into the North Atlantic, it was not unexpected that a few of the boundary sampling sites (the three stations at the top of Fig. 4) of the Sargasso Sea clustered closer to the North Atlantic sites, implying that this area of transition is both real and potentially unique. Overall, CCA demonstrates that these are clearly three statistically different open ocean regimes. ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

368

This is interesting to note as univariate analyses presented earlier could, by themselves, have suggested otherwise. As such, the results point to the need for both increased spatial and temporal resolution of viral and bacterial turnover and abundance, as well as the need for a better understanding of viral and host richness in the development of models for controls of viral activity in the global ocean. Monitoring microbial communities and their global-scale effects will require a delicate balance between assessing oceanic regions as individual environments and assessing their role in the greater marine ecosystem. Understanding how viruses interact with the environment is just the first step in our efforts to develop a clearer picture of global virus ecology, a necessary component of all models (and management strategies) that focus on anthropogenic influences and climate change.

Acknowledgements We thank the captain and crew of the R/V Kilo Moana as well as Audrey Matteson. We also thank Annette Hynes and Eric Webb for nutrient data. John Dunlap provided help and advice with transmission electron microscopy, and David Hutchins and Matt Cottrell provided invaluable assistance during the NASB 2005 cruise. This research was funded by grants for the National Science Foundation to ZIJ (OCE-0526462), ERZ (OCE-0526072), and SWW (OCE-0452409, OCE-0526159, OCE-0851113).

References Angly FE, Felts B, Salamon P et al. (2006) The marine viromes of four oceanic regions. PLoS Biol 4: e368. Baldwin AJ, Moss JA, Pakulski JD, Catala P, Joux F & Jeffrey WH (2005) Microbial diversity in a Pacific Ocean transect from the Arctic to Antarctic circles. Aquat Microb Ecol 41: 91–102. Boland GJ, Melzer MS, Hopkin A, Higgins V & Nassuth A (2004) Climate change and plant diseases in Ontario. Can J Plant Pathol 26: 335–350. Brussaard CPD, Kuipers B & Veldhuis MJW (2005) A mesocosm study of Phaeocystis globosa population dynamics – 1. Regulatory role of viruses in bloom. Harmful Algae 4: 859–874. Brussaard CPD, Wilhelm SW, Thingstad F et al. (2008) Global scale processes with a nanoscale drive – from viral genes to oceanic biogeochemical cycles. ISME J 2: 575–578. Church MJ, Ducklow HW & Karl DM (2004) Light dependence of [3H]leucine incorporation in the oligotropihc North Pacific Ocean. Appl Environ Microbiol 70: 4079–4087. Cole JR, Chai B, Farris RJ et al. (2007) The Ribosomal Database Project (RDP-II): introducing myRDP space and

ª 2011 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved

J.M. Rowe et al.

quality controlled public data. Nucleic Acids Res 35: D169– D172. Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS, McGarrell DM, Marsh TL, Barrity GM & Tiedje JM (2009) The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res 37: D141–D145. Fuhrman JA (1992) Bacterioplankton roles in cycling of organic matter: the microbial food web. Primary Productivity and Biogeochemical Cycles in the Sea (Falkowski PG & Woodhead A, eds), pp. 361–384. Plenum, New York, NY. Goddard L & Dilley M (2005) El Nino: catastrophe or opportunity. J Clim 18: 651–655. Halpert MS & Ropelewski CF (1992) Surface-temperature patterns associated with the Southern Oscillation. J Clim 5: 577–593. Hamady M, Walker JJ, Harris JK, Gold NJ & Knight R (2008) Error-correcting barcoded primers for pyrosequencing hundreds of samples in multiplex. Nature Methods 5: 235– 237. Hewson I & Fuhrman JA (2007) Covariation of viral parameters with bacterial assemblage richness and diversity in the water column and sediments. Deep Sea Res Part 1 Oceanogr Res Pap 54: 811–830. Hobbie JE, Daley RJ & Jasper S (1977) Use of nuclepore filters for counting bacteria by fluorescence microscopy. Appl Environ Microbiol 33: 1225–1228. Huse SM, Huber JA, Morrison HG, Sogin ML & Mark Welch D (2007) Accuracy and quality of massively parallel DNA pyrosequencing. Genome Biol 8: R143. Huse SM, Dethlefsen L, Huber JA, Welch DM, Relman DA & Sogin ML (2008) Exploring microbial diversity and taxonomy using SSU rRNA hypervariable tag sequencing. PLoS Genet 4: e1000255. Hynes AM, Chappell PD, Dyhrman ST, Doney SC & Webb EA (2009) Cross-basin comparison of phosphorus stress and nitrogen fixation in Trichodesmium. Limnol Oceanogr 54: 1438–1448. Jiang SC & Paul JH (1994) Seasonal and diel abundance of viruses and occurrence of lysogeny/bacteriocinogeny in the marine environment. Mar Ecol Prog Ser 104: 163–172. Johnson ZI, Zinser ER, Coe A, McNulty NP, Woodward EMS & Chisholm SW (2006) Niche partitioning among Prochlorococcus ecotypes along ocean-scale environmental gradients. Science 311: 1737–1740. Johnson ZI, Shyam R, Ritchie AE, Mioni C, Lance VP, Murray JW & Zinser ER (2010) The effect of iron- and light-limitation on phytoplankton communities of deep chlorophyll maxima of the western Pacific Ocean. J Mar Res 68: 283–308. Kirchman DL (2001) Measuring bacterial biomass production and growth rates from leucine incorporation in natural aquatic environments. Methods Microbiol 30: 227–236. Lee SH & Fuhrman JA (1987) Relationships between biovolume and biomass of naturally derived marine bacterioplankton. Appl Environ Microbiol 53: 1298–1303.



Lindell D, Sullivan MB, Johnson ZI, Tolonen AC, Rohwer F & Chisholm SW (2004) Transfer of photosynthesis genes to and from Prochlorococcus viruses. P Natl Acad Sci USA 101: 11013–11018. Lozupone C, Hamady M & Knight R (2006) UniFract – an online tool for comparing microbial community diversity in a phylogenetic context. BMC Bioinformatics 7: 371. Marie D, Partensky F, Jacquet S & Vaulot D (1997) Enumeration and cell cycle analysis of natural populations of marine picoplankton by flow cytometry using the nucleic acid stain SYBR Green I. Appl Environ Microbiol 63: 186–193. McPhaden MJ, Zebiak SE & Glantz MH (2006) ENSO as an integrating concept in earth science. Science 314: 1740– 1745. Middelboe M & Jo¨rgensen NOG (2006) Viral lysis of bacteria: an important source of dissolved amino acids and cell wall compounds. J Mar Biol Assoc UK 86: 605–612. Middelboe M, Jo¨rgensen NOG & Kroer N (1996) Effects of viruses on nutrient turnover and growth efficiency of noninfected marine bacterioplankton. Appl Environ Micorbiol 62: 1991–1997. Murray AG & Jackson GA (1992) Viral dynamics: a model of the effects of size, shape, motion and abundance of singlecelled planktonic organisms and other particles. Mar Ecol Prog Ser 89: 103–116. Noble RT & Fuhrman JA (1998) Use of SYBR Green I for rapid epifluorescence counts of marine viruses and bacteria. Aquat Microb Ecol 14: 113–118. Noble RT & Fuhrman JA (2000) Rapid virus production and removal as measured with fluorescently labeled viruses as tracers. Appl Environ Microbiol 66: 3790–3797. Pace ML, del Giorgio P, Fisher D, Condon R & Malcom H (2004) Estimates of bacterial production using the leucine incorporation method are influenced by differences in protein retention of microcentrifuge tubes. Limnol Oceanogr Methods 2: 55–61. Poorvin L, Rinta-Kanto JM, Hutchins DA & Wilhelm SW (2004) Viral release of Fe and its bioavailability to marine plankton. Limnol Oceanogr 49: 1734–1741. Rinta-Kanto JM, Ouellette AJA, Boyer GL, Twiss MR, Bridgeman TB & Wilheim SW (2005) Quantification of toxic Microcystis spp. during the 2003 and 2004 blooms in western Lake Erie using quantitative real-time PCR. Environ Sci Technol 39: 4198–4205. Rowe JM, Saxton MA, Cottrell MT et al. (2008) Constraints on viral production in the Sargasso Sea and North Atlantic. Aquat Microb Ecol 52: 233–244. Steinberg DK, Carlson CA, Bates NR, Johnson RJ, Michaels AF & Knap AH (2001) Overview of the US JGOFS Bermuda Atlantic Time-series Study (BATS): a decade-scale look at ocean biology and biogeochemistry. Deep Sea Res Part 2 Top Stud Oceanogr 48: 1405–1447. Suttle CA & Chan AM (1993) Marine cyanophages infecting oceanic and coastal strains of Synechococcus – abundance, morphology, cross-infectivity and growth characteristics. Mar Ecol Prog Ser 92: 99–109.


369

Tamura K, Dudley J, Nei M & Kumar S (2007) MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol 24: 1596–1599. Thingstad TF & Lignell R (1997) Theoretical models for the control of bacterial growth rate, abundance, diversity and carbon demand. Aquat Microb Ecol 13: 19–27. Vaulot D, Courties C & Partensky F (1989) A simple method to preserve oceanic phytoplankton for flow cytometric analyses. Cytometry 10: 629–635. Wang CL & Enfield DB (2001) The tropical western hemisphere warm pool. Geophys Res Lett 28: 1635–1638. Weinbauer MG & Suttle CA (1996) Potential significance of lysogeny to bacteriophage production and bacterial mortality in coastal waters of the Gulf of Mexico. Appl Environ Microbiol 62: 4374–4380. Weinbauer MG, Winter C & Ho¨fle MG (2002) Reconsidering transmission electron microscopy based estimates of viral infection of bacterioplankton using conversion factors derived from natural communities. Aquat Microb Ecol 27: 103–110. Weinbauer MG, Rowe JM & Wilhelm SW (2010) Determining rates of virus production in aquatic systems by the virus reduction approach. Manual of Aquatic Viral Ecology (Wilhelm SW, Weinbauer MG & Suttle CA, eds), pp. 1–8. ASLO, Waco, TX. Welschmeyer NA (1994) Fluorometric analysis of chlorophyll a in the presence of chlorophyll b and pheopigments. Limnol Oceanogr 39: 1985–1992. Wen K, Ortmann AC & Suttle CA (2004) Accurate estimation of viral abundance by epifluorescence microscopy. Appl Environ Microbiol 70: 3862–3867. Wilhelm SW & Matteson AR (2008) Freshwater and marine virioplankton: a brief overview of commonalities and differences. Freshw Biol 53: 1076–1089. Wilhelm SW & Poorvin L (2001) Quantification of algal viruses in marine samples. Methods Microbiol vol. 30, Marine Microbiology (Paul JH, ed), pp. 53–66. Academic Press, London, UK. Wilhelm SW & Suttle CA (1999) Viruses and nutrient cycles in the sea. Bioscience 49: 781–788. Wilhelm SW, Weinbauer MG, Suttle CA & Jeffrey WH (1998) The role of sunlight in the removal and repair of viruses in the sea. Limnol Oceanogr 43: 586–592. Wilhelm SW, Brigden SM & Suttle CA (2002) A dilution technique for the direct measurement of viral production: a comparison in stratified and tidally mixed coastal waters. Microb Ecol 43: 168–173. Wilhelm SW, Jeffrey WH, Dean AL, Meador J, Pakulski JD & Mitchell DL (2003) UV radiation induced DNA damage in marine viruses along a latitudinal gradient in the southeastern Pacific Ocean. Aquat Microb Ecol 31: 1–8. Wilson WH, Francis I, Ryan K & Davy SK (2001) Temperature induction of viruses in symbiotic dinoflagellates. Aquat Microb Ecol 25: 99–102. Winter C, Herndl GJ & Weinbauer MG (2004) Diel cycles in viral infection of bacterioplankton in the North Sea. Aquat Microb Ecol 35: 207–216.


370

Winter C, Moeseneder MM, Herndl GJ & Weinbauer MG (2008) Relationship of geographic distance, depth, temperature, and viruses with prokaryotic communities in the eastern Tropical Atlantic ocean. Microb Ecol 56: 383– 389. Wommack KE & Colwell RR (2000) Virioplankton: viruses in aquatic ecosystems. Microbiol Mol Biol Rev 64: 69–114. Wommack KE, Hill RT, Kessel M, Russek-Cohen E & Colwell RR (1992) Distribution of viruses in the Chesapeake Bay. Appl Environ Microbiol 58: 2965–2970.

Supporting Information Additional Supporting Information may be found in the online version of this article: Fig. S1. Phylogenetic classification of pyrosequences from the Ribsomal Database Project classifier analyses. Fig. S2. Principle component analyses of 16S rRNA gene sequence data obtained from the UniFrac service. Table S1. PCR primers and bar codes.


J.M. Rowe et al.

Table S2. Metrics for pyrosequenced 16S rRNA gene samples. Table S3. Spearman rank correlation analysis of parameters measured in the Sargasso Sea (a selection of these data was previously published in Rowe et al., 2008). Table S4. Spearman rank correlation analysis of parameters measured in the North Atlantic (a selection of these data was previously published in Rowe et al., 2008). Table S5. Spearman rank correlation analysis of parameters measured in the Western Pacific. Table S6. Spearman rank correlation analysis of parameters measured in the Sargasso Sea, North Atlantic, and Western Pacific (a selection of these data was previously published in Rowe et al., 2008). Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
