The balance between lytic and lysogenic viral

Download PDF
10 downloads 0 Views 2MB Size Report
Comment
To kill or not to kill: The balance between lytic and lysogenic viral infection is ... water samples collected on the Canadian Arctic Shelf, southern Beaufort Sea.
Limnol. Oceanogr., 58(2), 2013, 465–474 2013, by the Association for the Sciences of Limnology and Oceanography, Inc. doi:10.4319/lo.2013.58.2.0465

E

To kill or not to kill: The balance between lytic and lysogenic viral infection is driven by trophic status Je´roˆme P. Payeta,1 and Curtis A. Suttle a,b,* a Department

of Earth and Ocean Sciences, University of British Columbia, Vancouver, British Columbia, Canada of Microbiology and Immunology and Botany, University of British Columbia, Vancouver, British Columbia, Canada

b Departments

Abstract Experiments were conducted to investigate spatiotemporal patterns in lytic and lysogenic viral infection using water samples collected on the Canadian Arctic Shelf, southern Beaufort Sea. Viral production (VP) and viralinduced mortality of bacteria (VMB) were determined using a viral reduction approach during a full seasonal cycle, while the percentage of lysogenic bacteria (PLB) in spring and summer was determined in virus-reduced samples by induction with mitomycin C. Overall, VP (range: 0.3 3 108–77 3 108 viruses L21 d21), VMB (range: 0.2 3 107–43 3 107 bacteria L21 d21), and PLB (range: 4–38%) displayed marked spatiotemporal variations concomitant with changes in chlorophyll a, bacterial abundance, and production. Highest VP and VMB occurred in summer when the water was warmest, stratified, and most productive, and when viruses removed up to 29% of bacterial standing stock d21 and released up to 4.3 mg of organic carbon L21 d21. In contrast, the highest PLB occurred in spring when the water was colder, well mixed, and oligotrophic. Correlative and regression analyses indicated viral lytic and lysogenic variables were significantly coupled with chlorophyll a and the abundance, production, and growth rate of bacteria, implying that viral lytic and lysogenic lifestyles were dependent on system productivity. Furthermore, lytic VP and the proportion of lysogenized bacteria were inversely related, suggesting a dynamic interplay between viral infection pathways. Lytic infection was more pronounced when system productivity was high, while lysogeny prevailed when system productivity was low. These data demonstrate the important role of viruses in bacterial mortality and carbon cycling in the Arctic Ocean, and show how their effect is influenced by trophic status.

(Paul 2008). Several studies have examined lysogeny in marine microbial communities using induction assays of cultivable and bacterial communities in a wide variety of environments, and have observed that the percentage of prokaryotes harboring inducible prophages ranges from undetectable to . 80% (Weinbauer 2004; Paul 2008); however, few studies have compared both lytic and lysogenic infection, and little is known about the influence of environmental factors on the relative importance of these two modes of infection, particularly in arctic waters. The few studies that have examined both viral lifestyles indicate that the relative importance of lytic and lysogenic infection is related to changes in environmental conditions as well as host abundance and growth rates (Weinbauer et al. 2003; Bongiorni et al. 2005; Long et al. 2008). Relative to open-ocean systems, coastal waters are dynamic heterogeneous sites that play a critical role in the cycling of organic matter due to high primary production and allochthonous inputs from rivers. The Canadian Arctic Shelf (CAS), southeastern Beaufort Sea in the Arctic Ocean, is a large estuarine system that receives continuous inputs of freshwater, sediment, nutrients, and organic matter from adjacent rivers (Carmack and Macdonald 2002; Stein and Macdonald 2004). The CAS supports high biological productivity during the spring– summer period, concomitant with higher solar irradiance, lower ice cover, and higher river discharges (Carmack et al. 2004; Garneau et al. 2008; Brugel et al. 2009). In this region, microbial assemblages of viruses and cells rapidly respond to shifts in environmental conditions (Alonso-Sa´ez et al. 2008; Garneau et al. 2008; Payet and Suttle 2008).

Viruses that infect bacteria (phages) are the most abundant, diverse, and pervasive biological entities in the world’s oceans (Fuhrman 1999; Suttle 2005). They are substantial mortality agents of bacteria, and therefore affect global biogeochemical processes and energy fluxes (Fuhrman 1999; Wilhelm and Suttle 1999; Suttle 2005). The effect of phages on ecological and biogeochemical processes is affected by their life cycle. In the lytic cycle, viral replication begins immediately after infection, leading to phage production and lysis of the host cell. In marine environments, it is estimated that 20–50% of bacterial biomass is lost daily due to viral lytic infection (Suttle 2005). This viral lysis fuels primary and secondary productivity through the release of organic matter and nutrients, which shunts substrates to uninfected microbes (Middelboe et al. 1996; Weinbauer et al. 2011; Shelford et al. 2012). In lysogeny, the genetic material of temperate phages is integrated into the host genome as a prophage and subsequently transmitted vertically during cell division. The lysogenic conversion can alter the host phenotype and confer resistance against other viral infections (Bru¨ssow et al. 2004; Chibani-Chennoufi et al. 2004; Weinbauer 2004). Later, the prophage can be induced into the lytic pathway either spontaneously or by physical or chemical factors

* Corresponding author: [email protected] 1 Present address: Department of Microbiology, Oregon State University, Corvallis, Oregon

465

466

Payet and Suttle

Fig. 1. Map of the study area on the CAS, showing sampling stations on the Mackenzie Shelf–Amundsen Gulf system. Sta. 200 in Franklin Bay was sampled seasonally from 04 November 2003 to 10 August 2004. Stas. 912, 803, 718, 650, 415, 200, and 106 were sampled from 04 July to 10 August 2004 along the cruise track (dashed arrow) from the Mackenzie River plume to the Amundsen Gulf. Stations were grouped into three regions according to their locations on the shelf: black circles, RP; white circles, MS; and gray circles, G.

Given its high variability in biological productivity and hydrological conditions, the CAS is an ideal site to investigate environmental factors influencing viral life strategies. In this paper, we report the results of experiments that examine the prevalence of inducible lysogens and lytic virus production in waters overlying the CAS. Specifically, this work investigates seasonal and spatial patterns of lysogenic and lytic viral infection, and explores whether these patterns are related to changes in trophic and environmental conditions. In this comparative study, our results suggest that there is a dynamic interplay between viral lifestyle and environmental conditions, and that both lytic and lysogenic viral infection have significant implications for host cells and carbon cycling in marine coastal Arctic systems.

Water samples were obtained with a rosette sampler equipped with 12 liter Niskin or Go-Flo bottles, a conductivity–temperature–depth system (CTD, Seabird Electronics), and a fluorometer (SeaPoint) to determine in situ chlorophyll a (Chl a) fluorescence. Samples were collected either at the Chl a–maximum layer (from 2 to 56 m), or, if no Chl a maximum was present, from the surface (from 2 to 5 m). Approximately 40 liters of seawater was filtered through a 120 mm-pore-size Nitex mesh to remove large particles, transferred into 20 liter polycarbonate carboys, and subsequently used for the experiments described below. Subsamples for determination of Chl a, heterotrophic bacterial abundance (BA), bacterial production (BP), and viral abundance (VA) were taken immediately following collection of the seawater.

Methods

Chl a concentrations—Depending on the phytoplankton biomass, 0.05–3 liters of seawater was filtered through Whatman GF/F filters for fluorometric analysis of Chl a concentration as outlined by Garneau et al. (2008).

Sample collection—Water samples for experiments (see below) were collected from the CAS between 04 November 2003 and 10 August 2004 during the Canadian Arctic Shelf Exchange Study (CASES; Fortier and Cochran 2008; Vincent and Pedro´ s-Alio´ 2008). Water samples were collected on seven dates at Sta. 200 in Franklin Bay during fall (04 and 19 November 2003), winter (28 January and 14 February 2004), spring (15 and 21 May 2004), and summer (06 August 2004), and at seven stations (912, 803, 718, 650, 415, 200, and 106) across the CAS from 04 July to 10 August 2004 (Fig. 1). The stations stretched from the mouth of the Mackenzie River to the Amundsen Gulf and were grouped by location into river plume (RP), mid-shelf (MS), and gulf (G) (Fig. 1).

Enumeration of viruses and bacteria—Viruses and bacteria were counted by epifluorescence microscopy (EFM) or flow cytometry (FC) in glutaraldehyde-fixed (0.5% final concentration, electron microscopy grade) samples using established methods (Noble and Fuhrman 1998; Brussaard et al. 2010; Suttle and Fuhrman 2010). Both methods were shown to yield equivalent results (Payet and Suttle 2008). Briefly, EFM samples were fixed with glutaraldehyde for , 30 min, filtered through 0.02 mm-pore-size filters (Anodisc, Whatman), stained with SYBR Green I (10,0003 in dimethyl sulfide, Molecular Probes; concentration of staining solution,

Lytic and lysogenic viral infection 4 3 1022 of stock in sterile 0.02 mm filtered MilliQ) for 15 min in the dark and mounted on a glass slide with 0.1% pphenylenediamine (made freshly from a frozen 10% aqueous stock, Sigma Chemical) in 50% glycerol–50% phosphatebuffered saline (0.05 mol L21 Na2HPO4, 0.85% NaCl; pH 7.5). For each sample, duplicate filters were prepared and slides were analyzed using an Olympus AX70 EFM with a wide-blue filter set (excitation at 450–480 nm, 515 nm cutoff). Viruses and bacteria were enumerated on the same slide and a minimum of 200 particles was counted in at least 20 random fields. FC samples were analyzed using a Becton Dickinson FACSCaliburTM platform equipped with an air-cooled argon laser (15 mW, 488 nm) as outlined in Payet and Suttle (2008). Viruses and bacteria were preserved for , 30 min in glutaraldehyde, flash frozen in liquid nitrogen, and then stored at 280uC until analysis. Prior to analysis, samples were quickly thawed, diluted (1–10-fold for bacteria and 20–50-fold for viruses) in sterile buffer (10 mmol L21 Tris, 1 mmol L21 ethylene-diamine-tetraacetic acid, pH 8.0), then stained with SYBR Green I for 15 min at 80uC for viruses and at room temperature for bacteria in the dark. Fluorescent beads (Fluoresbrite Microparticles Polysciences) of 0.92 mm diameter were also added as an internal standard. The samples were analyzed for 1 min or until 10,000 events were acquired for viruses and 20,000 events for bacteria at event rates between 200 and 800 particles s21 with the discriminator set on green fluorescence (FL1). Discrimination of bacteria and viruses was based on their signature in scatter plot of side scatter vs. FL1 as shown elsewhere (Payet and Suttle 2008). BP and bacterial growth rates—In situ heterotrophic bacterial carbon production (BP, mg C L21 d21) was measured by the rates of incorporation of [3H]-thymidine and/or [14C]-L-leucine using a bacterial carbon quota of 10 fg C cell21, and was obtained from Garneau et al. (2008, 2009). Bacterial growth rate (BGR, d21) was calculated as the ratio between BP and bacterial biomass (mg C L21). Biomass was calculated by multiplying BA by the bacterial carbon quota of 10 fg C cell21. Viral production and viral lytic activities—Experiments to assess rates of lytic viral production (VP) were based on the virus reduction approach (Wilhelm et al. 2002). Briefly, 500 mL of seawater was filtered through a 0.2 mm-pore-size polycarbonate filter (47 mm diameter, Millipore) and mounted onto a Sterifil unit (Millipore) under low vacuum (, 200 mm Hg). Throughout this process, particles , 0.2 mm in diameter (i.e., size range of most viruses infecting bacteria) were reduced while particles in the size range of 0.2–120 mm were retained in suspension above the filter. The retentate was washed by adding 1 liter of ultrafiltered seawater (, 30 kDa cutoff, Amicon M12 System equipped with S10Y30 cartridge) produced from the same seawater sample. When 50 mL of retentate remained, the volume was brought back to 500 mL by addition of ultrafiltered seawater, reducing virus abundances, on average, by 98% 6 2% (range: 94.4–99.8%, n 5 39). For each experiment, the filtration approach was carried out in triplicate and the retentates pooled (final volume, 1.5 liters)

467

Table 1. Equations used to estimate lytic viral activity variables. VAtf 5 VA at final time point; VAt0 5 VA at initial time 0; BAa 5 ambient BA; VMBc 5 VMB expressed in mg C L21 d21 with a bacterial carbon quota of 10 fg C cell21. VP (viruses L21 d21)5[(VAtf2VAt0)/tf]3(BAa/BAt0) VT (d21)5VP/VA VMB (bacteria L21 d21)5VP/BS OCR (mg C L21 d21)5VMB310 fg C cell21 %BSSR (d21)51003(VMB/BA) %BPR (d21)51003(VMBc/BP)

and then aliquoted into three sterile 1 liter polypropylene Whirlpack bags. Sample filtration was carried out in the ship’s cold room, at near in situ temperatures (21.5uC to 4uC). The bags were incubated in a Sanyo versatile environmental test chamber in darkness at in situ temperatures (21.6uC to 9uC) for 48–72 h. Subsamples were taken at ca. 6–12 h intervals for the first 24 h of incubation and then at ca. 12–24 h intervals during the following days. For each incubation VP estimates were determined from the slope of a least-square linear regression fitted to changes in VA vs. time, following Wilhelm et al. (2002). The slopes of the three replicates were averaged and corrected for loss of bacterial hosts relative to in situ bacterial standing stocks during the filtration (e.g., Table 1). On average, 52% 6 14% (range: 28–74%, n 5 39) of in situ bacterial standing stocks were lost during the filtration procedure (data not shown). The viral turnover time (VT, d21) was calculated as the ratio between VP and VA (e.g., Table 1). The viral-induced mortality of bacteria (VMB, bacteria L21 d21) and the percentage of bacterial standing stock removed (%BSSR, d21) due to viral lysis were calculated as shown in Table 1, using an average burst size (BS, i.e., number of viruses released per lytic event) estimate of 18. This BS estimate has been reported to be the average value in Arctic marine waters (Middelboe et al. 2002). The amount of organic carbon released (OCR, mg C L21 d21) upon viral lysis of bacterial cells and the percentage of BP removed due to viral lysis (%BPR, d21) were estimated assuming a bacterial carbon content of 10 fg C cell21 (Table 1). Prophage induction experiments—The percentage of lysogens that could be induced was assessed by using the virus reduction approach outlined above, with the modification that triplicate samples were treated with mitomycin C (1 mg mL21 final concentration, Sigma M-0503) prior to incubation. Mitomycin C is an effective prophage inducer and it has been widely used (reviewed in Paul 2008; Paul and Weinbauer 2010). Induction experiments were run in parallel to VP experiments, so that untreated VP incubations served as controls for the induction incubations. Subsamples for determination of VA and BA were taken at the same time points as in the VP experiments during the 48–72 h incubation period. The percentage of lysogenic bacteria (PLB) was calculated as PLB(%)~½(VAMitC {VAC =(BS BAt0 )"|100

ð1Þ

468

Payet and Suttle

Fig. 2. Spatial variations in (a) T (black circle), S (gray circle), and Chl a (triangle), (b) BA (black circle), BP (gray circle), and BGR (triangle), (c) VA (black circle), lytic VP (white circle), VMB (gray circle), and %BSSR (triangle up), (d) %BPR (diamond), OCR (square), and VT (triangle down), and (e) PLB across three regions in summer: RP, MS, and G. Error bars represent standard deviations (n 5 3).

where VAMitC and VAC are the virus counts in mitomycin C–treated and control incubations, respectively; BS is the burst size (18); and BAt0 is the bacterial count in the control at the start of the incubation. Statistical analysis—Due to violation of the assumptions of homoscedasticity and homogeneity of variances and unequal sample sizes among groups, the nonparametric Kruskal-Wallis (KW) analysis of variance (ANOVA) on ranks test followed by Dunn’s post hoc multiple comparison was used to assess for differences among variables. Spearman rank correlation coefficients (rs) were used to determine the correlation among parameters. Univariate linear regressions were used for calculations of VP estimates within incubations (see above). Relationships between selected variables were examined by nonlinear regressions. Statistical analyses were performed using Systat 11H (Systat Software) and Origin 8.1 (OriginLab). When necessary, the results were expressed as the means 6 standard deviations.

Results Environmental conditions and Chl a concentration—There were distinct spatial patterns in temperature (T) and salinity (S) across the shelf (Fig. 2a). Relatively warm freshwater inflow from the Mackenzie River steeply increased surface stratification and turbidity in the RP during summer, resulting in warmer T (7.9–8.4uC, n 5 2) and lower S (14.8–23.8, n 5 2) (Fig. 2a). In contrast, surface waters in the MS and G regions were typical of marine arctic surface waters, with low T ranging from 20.8uC to 5.8uC (n 5 5) and high mean S ranging from 25.8 to 29.8 (n 5 5) (Fig. 2a). Chl a exhibited strong spatial patterns on the shelf with concentrations ranging from 0.1 to 1.4 mg L21 with a 2–3fold increase in the RP (range: 1.4–1.2 mg L21, n 5 2) relative to the MS and G regions (range: 0.1–0.5 mg L21, n 5 5) (Fig. 2a). At Sta. 200, sea ice formation caused mixing of the surface layers from mid-November to late April, with T

Lytic and lysogenic viral infection

469

Fig. 3. Seasonal variations in (a) T (black circle), S (gray circle), and Chl a (triangle), (b) BA (black circle), BP (gray circle), and BGR (triangle), (c) VA (black circle), lytic VP (white circle), VMB (gray circle), and %BSSR (triangle up), (d) %BPR (diamond), OCR (square), and VT (triangle down), and (e) PLB at Sta. 200 during fall, winter, spring, and summer. Error bars represent standard deviations (n 5 3). nd 5 non-determined.

and S ranging from 21.7uC to 21.4uC and from 26.3 to 31.9 (Fig. 3a). In mid-May, T and S slightly changed coinciding with the onset of sea ice melt. In summer, there was a marked increase in surface stratification as the sea ice receded at Sta. 200, with warmer surface T (2.8uC) and lower S (26.1) relative to other seasons (Fig. 3a). Chl a exhibited strong seasonal patterns at Sta. 200, with concentrations decreasing , 3-fold from fall (range: 0.06– 0.21 mg L21, n 5 2) to winter (range: 0.03–0.04 mg L21, n 5 2) and peaking , 7-fold higher in late spring (range: 0.25– 0.34 mg L21, n 5 2) and summer (0.22 mg L21) (Fig. 3a). Over all samples, Chl a was significantly positively correlated to T (rs 5 0.63, p , 0.05, n 5 13) and inversely correlated to S (rs 5 20.74, p , 0.01, n 5 13). BA, BP, and BGR—BA, BP, and BGR showed marked seasonal and spatial variations (Figs. 2b, 3b). BA, BP, and BGR ranged from 1.3 3 108 to 14 3 108 cells L21, from 0.02 to 3.3 mg C L21 d21, and from 0.01 to 0.26 d21, respectively (Figs. 2b, 3b). Across the shelf, BA, BP, and BGR were , 1.5–3-fold higher in the RP region relative to

the MS and G regions (Fig. 2b). No significant differences in BA, BP, and BGR were found between the MS and G regions (KW ANOVA on ranks with Dunn’s test, p . 0.05) (Fig. 2b). At Sta. 200, BA, BP, and BGR followed similar seasonal patterns as for Chl a, with significantly lowest and highest values measured in winter and summer, respectively (KW ANOVA on ranks with Dunn’s test, p , 0.05) (Fig. 3b). Overall, BA, BP, and BGR were significantly correlated to T, S, and Chl a (rs . 0.70, p , 0.001). VA, VP, VMB, and VT—VA and estimates of VP and VT displayed strong spatiotemporal variations (Figs. 2c,d, 3c,d). In general, VA, VP, and VT showed similar variations to those of Chl a and bacterial variables (BA, BP, and BGR) (e.g., Figs. 2b, 3b). VA, VP, and VT ranged from 2.7 3 109 to 27 3 109 viruses L21, from 0.3 3 108 to 77 3 108 viruses L21 d21, and from 0.01 to 0.52 d21, respectively. Highest VA, VP, and VT occurred in the RP region, where VA and VP significantly increased relative to the MS and G regions (KW ANOVA on ranks with Dunn’s

470

Payet and Suttle

Fig. 4. Relationships between (a) VP and bacterial variables (BA, black circle; BP, white circle; BGR, triangle) and (b) VP and Chl a (gray circle). Allometric power-law relationships are VP 5 (0.3 6 0.1) 3 BA exp (2.1 6 0.1), r2 5 0.95, p , 0.001, n 5 12 (black curve, a); VP 5 (0.3 6 0.1) 3 BP exp (1.4 6 0.1), r2 5 0.92, p , 0.001, n 5 12 (dotted curve, a); VP 5 (2.4 6 1.5) 3 BGR exp (1.9 6 0.2), r2 5 0.98, p , 0.001, n 5 12 (gray curve, a); VP 5 (3.7 6 2.2) 3 Chl a exp (1.7 6 0.3), r2 5 0.96, p , 0.001, n 5 12 (black curve, b). The error bars representing the standard deviations (n 5 3) are not shown for clarity.

test, p , 0.05) (Fig. 2c,d). No significant differences were found in mean VA, VP, and VT between the MS and the G regions (KW ANOVA on ranks with Dunn’s test, p . 0.05) (Fig. 2c,d). At Sta. 200, VA, VP, and VT showed similar seasonal trends (Fig. 3c,d) and were significantly lower in winter compared to fall, spring, and summer (KW ANOVA on ranks with Dunn’s test, p , 0.05). Assuming a BS of 18 (see Methods), estimates of VMB ranged between 0.2 3 107 and 43.2 3 107 bacteria L21 d21 (Figs. 2c, 3c). This implied that on average from 1.4% to 29% of the bacterial standing stock was removed daily (%BSSR) due to viral lysis (Figs. 2c, 3c). Similar to VP, highest mortality rates (%BSSR) occurred in summer and in the RP region compared to other seasons and regions (KW ANOVA on ranks with Dunn’s test, p , 0.05) (Figs. 2c, 3c). Based on a bacterial carbon quota of 10 fg C cell 21 (see Methods), between 0.02 and 4.3 mg C L21 d21 were released (OCR) through viral lysis of bacteria (Figs. 2d, 3d). Furthermore, the mean %BPR ranged between 31% and 156% daily (Figs. 2d, 3d). VA, VP, VMB, VT, %BSSR, and OCR were significantly positively correlated with BA, BP, BGR, T, and Chl a (0.53 , rs , 0.92, p , 0.05, n 5 13) and significantly negatively correlated with S (20.91 , rs , 20.67, p , 0.05, n 5 13). Since VMB is calculated from VP (e.g., see Table 1), these variables are autocorrelated. Univariate nonlinear models predicted positive allometric power-law increases of VP against BA, BP, BGR, and

Chl a (Fig. 4a,b), and provided better fits and distributions of residuals than did linear models (not shown). Induction of lysogenic viruses—In mitomycin C–treated incubations, VA increased on average , 2–4-fold at , 6– 24 h postincubation compared to controls, while BA decreased on average , 1–2-fold compared to controls (not shown). Estimates of the PLB ranged between 4% and 38% (Figs. 2e, 3e) and displayed significant seasonal and spatial patterns (KW ANOVA on ranks with Dunn’s test, p , 0.05). Across the shelf, the mean PLB estimates were , 4– 5-fold lower in the RP (4.4% 6 0.8%, n 5 6) relative to the MS (21.6% 6 8.2%, n 5 9) and G (12.6% 6 1.1%, n 5 6), respectively (Fig. 2e). At Sta. 200, PLB estimates were on average , 3-fold higher in late spring (35.2% 6 6.2%, n 5 6) than in summer (11.8% 6 2.0%, n 5 3) (Fig. 3e). Spearman’s rank correlation analysis revealed that PLB was negatively correlated with VA (rs 5 20.67, p , 0.01, n 5 9), BA (rs 5 20.85, p , 0.001, n 5 9), BP (rs 5 20.92, p , 0.001, n 5 9), BGR (rs 5 20.72, p , 0.001, n 5 9), Chl a (rs 5 20.65, p , 0.05, n 5 9), and T (rs 5 20.85, p , 0.001, n 5 9), and positively correlated with S (rs 5 0.90, p , 0.001, n 5 9). Nonlinear regression analyses revealed significant negative power-law relationships between PLB and BA, BP, BGR, and Chl a (Fig. 5a,b). Viral lysis–lysogeny relationships—Interestingly, there were significant relationships between the PLB and VP

Lytic and lysogenic viral infection

471

Fig. 5. Relationships between (a) PLB and bacterial variables (BA, black circle; BP, white circle; BGR, triangle) and (b) PLB and Chl a (gray circle). Allometric power-law relationships are PLB 5 (252 6 119) 3 BA exp (21.5 6 0.3), r2 5 0.79, p , 0.001, n 5 9 (black curve, a); PLB 5 (13.8 6 1.5) 3 BP exp (20.6 6 0.1), r2 5 0.88, p , 0.001, n 5 9 (dotted curve, a); PLB 5 (1.1 6 1.1) 3 BGR exp (21.2 6 0.2), r2 5 0.78, p , 0.01, n 5 9 (gray curve, a); PLB 5 (6.9 6 4.1) 3 Chl a exp (20.7 6 0.2), r2 5 0.45, p , 0.05, n 5 9 (black curve, b). The error bars representing the standard deviations (n 5 3) are not shown for clarity.

(Fig. 6a) and between the PLB and %BSSR (Fig. 6b) described by a negative power-law asymptotic function, indicating that lytic VP was inversely related to the proportion of lysogenic cells that could be induced.

Discussion The effect of phages on bacterial communities depends on whether infection is lysogenic or lytic, and although both replication pathways are important in the sea (Weinbauer 2004; Paul 2008), they have rarely been examined in the same study. Since viral infection is

dependent upon density and metabolic activity of hosts, any changes in the abiotic environment may affect host communities, with selective consequences for viruses that have a lytic vs. a lysogenic lifestyle. Change in the balance between lytic and lysogenic infection is a dynamic phenomenon, and rapidly responds to changes in environmental conditions (Bongiorni et al. 2005; Long et al. 2008; Weinbauer et al. 2009). Results of this investigation reveal a dynamic interplay between lytic and lysogenic viral infection that is driven by changes in system productivity. A series of experiments on the CAS show that the balance between lytic and lysogenic infection is affected by

Fig. 6. Relationships between (a) PLB and VP and (b) PLB and %BSSR. Allometric powerlaw relationships are PLB 5 (82 6 17) 3 VP exp (20.7 6 0.1), r2 5 0.89, p , 0.001, n 5 9 (a); PLB 5 (339 6 96) 3 %BSSR exp (21.4 6 0.1), r2 5 0.92, p , 0.001, n 5 9 (b). The error bars represent the standard deviations (n 5 3). Triangle up, RP; gray circle, MS; black diamond, G in the spring; gray diamond, G in the summer.

472

Payet and Suttle

hydrodynamic and ecological variables. These results and their implications are discussed in detail below. Viral lytic production and lysis of heterotrophic bacteria— VP was measured using the virus reduction approach (Wilhelm et al. 2002), and despite its methodological limitations (Winget et al. 2005), it has become the preferred method for inferring viral-mediated cell lysis (Weinbauer et al. 2010). Seasonal and spatial patterns in VP were observed in conjunction with changes in phytoplankton and bacterial biomass and productivity. The average VP of 17 3 108 viruses L21 d21 was at the lower end of the ranges reported for other marine systems (Wilhelm et al. 2002; Rowe et al. 2008; Winget et al. 2011), but was comparable to levels reported for marine arctic waters (Steward et al. 1996; Middelboe et al. 2002; Wells and Deming 2006). VMB averaged 9.4 3 107 6 1.3 3 107 (n 5 39) bacteria L21 d21 and removed 12% 6 8% (n 5 39) of bacterial standing stock daily, thus shunting , 0.9 6 1.3 (n 5 39) mg C L21 d21 from bacteria to pools of dissolved and particulate organic carbon, thereby fueling the growth of uninfected bacteria, with consequent effects on carbon cycling and energy fluxes. In winter on the CAS, low solar irradiance and temperature as well as a thick ice cover constrain productivity, resulting in low abundances of phytoplankton, bacteria, and viruses. Nevertheless, heterotrophic bacterial activity persists in the surface layers at Sta. 200 (Alonso-Sa´ez et al. 2008; Garneau et al. 2008), allowing for the low measured rates of VP, and hence cell lysis. In turn, lysis provides a continuous nutrient supply that contributes to the persistence of BP in the winter. Wells and Deming (2006) also found low rates of VP in deep waters at Sta. 200, showing that viral lysis of bacteria occurs throughout the water column in winter. During summer, warmer organic-rich inflow from the Mackenzie River led to strong surface stratification in the RP region and elevated autotrophic and heterotrophic microbial abundance and productivity (Garneau et al. 2006; Vallie´res et al. 2008; Garneau et al. 2009), leading to higher VP, VT, and VMB. In turn, increased abundances of viruses and hosts resulted in higher virus-host contact rates, increased mortality, and higher rates of VP. Regression showed strong positive associations between viral lytic activity and microbial abundance and production (Fig. 5), with significant power-law trends of increased rates of viral lysis and bacterial mortality in response to higher levels of autotrophic and heterotrophic microbial biomass and productivity. This is consistent with lytic viral infection being driven by host density and growth rates, as has been noted for other marine systems (Weinbauer et al. 2003; Bongiorni et al. 2005; Rowe et al. 2008). Prevalence of lysogeny—While several agents may be used to induce prophages (e.g., ultraviolet radiation, antibiotics, and hydrogen peroxide), mitomycin C has been widely used to estimate the proportion of lysogenic bacteria in natural communities (Jiang and Paul 1996; Weinbauer and Suttle 1999; Paul and Weinbauer 2010). We chose to use mitomycin C, so our work was comparable to that of

others and because it has been shown to be a more effective inducing agent than ultraviolet radiation or hydrogen peroxide (Weinbauer and Suttle 1996, 1999); hence, it is reasonable to assume that the spatial and seasonal changes observed in the percentage of inducible lysogens (Figs. 2e, 3e) reflect changes in the relative abundance of lysogens overall. Nonetheless, not all prophages are induced by mitomycin C (Paul 2008); consequently, the estimated PLBs represent minima. The PLB averaged 17% 6 12% (n 5 27), indicating a relatively high proportion of lysogenized bacteria in Arctic waters, as has been observed in other high-latitude marine systems (Weinbauer et al. 2003, 2009). Likewise, environmental metagenomic data indicate that viruses in the Arctic Ocean are distinct from those in other oceanic biomes, and contain a higher proportion of homologs to prophage sequences (Angly et al. 2006). Particularly striking was the much greater proportion of lysogenized cells that were present during spring when BA, BP, and BGR were low. In contrast, the proportion of lysogens was much lower in summer, when BA and BP were higher. This was mirrored in the spatial data, with the lowest proportion of lysogens found in the productive plume waters and the highest in the less productive waters of the Amundson Gulf. Similar seasonal trends in the proportion of lysogens have been documented in lower latitude systems (Cochran and Paul 1998; Williamson et al. 2002; Long et al. 2008), where lysogeny prevails under unfavorable growth conditions. Overall, the data indicate that there is a strong negative association between the proportion of bacteria that were induced and the productivity of the system, with significant power-law trends of decreased rates of lysogenic infection in response to higher levels of autotrophic and heterotrophic microbial biomass and productivity (Fig. 5). Ultimately, environmental conditions resulting in lower BAs and BGRs appear to drive the decision towards lysogenic infection. Ecological consideration of the viral lifestyle strategy— The data show a strong inverse relationship between rates of lytic virus production and the proportion of bacteria that contain inducible lysogens (Fig. 6), with lytic VP being much higher when BA and BP are highest. Such results imply that the physiological state of the bacterial community and, hence, environmental conditions play a strong role in determining whether viruses lysogenize or lyse their hosts. These data are consistent with the idea that lysogeny allows viruses to survive within host cells under oligotrophic conditions, while lytic infection allows viruses to rapidly propagate when host abundance and productivity are high. These findings illustrate that viral lysis and lysogeny are dynamic processes that respond on relatively short temporal and spatial scales to differences in environmental conditions, and particularly to changes in system productivity. In particular, the results suggest that lysogeny may be an important factor that allows viruses to survive the long period of low productivity associated with the arctic winter. Weinbauer et al. (2003) also found negative power-law relationships between lytic VP and the

Lytic and lysogenic viral infection proportion of inducible lysogens in other environments, demonstrating similarities in viral infection pathways between the Arctic Ocean and lower latitude marine systems. Our results indicate that the balance between lytic and lysogenic viral infection is highly responsive to environmental changes in both temperate and polar marine coastal systems. In turn, this balance will influence the population dynamics of host communities, as well as pathways and rates of nutrient cycling. Acknowledgments We thank the captains, crew, and scientists aboard the R/V Amundsen for their support, A. I. Culley and A. M. Comeau for field sampling, and A. C. Ortmann and A. M. Chan for logistic support and facilitation during the expedition. We thank W. F. Vincent and M.-E´. Garneau for providing published chlorophyll a and bacterial production data. We also thank C. J. Charlesworth for editing and reviewing drafts of the manuscript. We thank the two anonymous reviewers for their detailed comments and insightful suggestions. This study was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC) through a network grant for the Canadian Arctic Shelf Exchange Study (CASES) project and a Discovery Grant to C.A.S. This is a contribution to the CASES project under the overall direction of L. Fortier.

References ALONSO-SA´EZ, L., O. SANCHEZ, J. GASOL, V. BALAGUE, AND C. PEDRO´S-ALIO´. 2008. Winter-to-summer changes in the composition and single-cell activity of near-surface Arctic prokaryotes. Environ. Microbiol. 10: 2444–2454, doi:10.1111/j.1462-2920.2008.01674.x ANGLY, F. E., AND OTHERS. 2006. The marine viromes of four oceanic regions. PLoS Biol. 4: 2121–2131, doi:10.1371/ journal.pbio.0040368 BONGIORNI, L., M. MAGAGNINI, M. ARMENI, R. NOBLE, AND R. DANOVARO. 2005. Viral production, decay rates, and life strategies along a trophic gradient in the North Adriatic Sea. Appl. Environ. Microbiol. 71: 6644–6650, doi:10.1128/ AEM.71.11.6644-6650.2005 BRUGEL, S., AND OTHERS. 2009. Phytoplankton biomass and production in the southeastern Beaufort Sea in autumn 2002 and 2003. Mar. Ecol. Prog. Ser. 377: 63–77, doi:10.3354/ meps07808 BRUSSAARD, C. P. D., J. P. PAYET, C. WINTER, AND M. G. WEINBAUER. 2010. Quantification of aquatic viruses by flow cytometry, p. 102–109. In S. W. Wilhelm, M. G. Weinbauer, and C. A. Suttle [eds.], Manual of aquatic viral ecology. ASLO, doi:10.4319/mave.2010.978-0-9845591-0-7 BRU¨SSOW, H., C. CANCHAYA, AND W.-D. HARDT. 2004. Phages and the evolution of bacterial pathogens: From genomics rearrangements to lysogenic conversion. Microbiol. Mol. Biol. Rev. 68: 560–602, doi:10.1128/MMBR.68.3.560-602.2004 CARMACK, E. C., AND R. W. MACDONALD. 2002. Oceanography of the Canadian shelf of the Beaufort Sea: A setting for marine life. Arctic 55: 29–45. ———, ———, AND S. JASPER. 2004. Phytoplankton productivity on the Canadian Shelf of the Beaufort Sea. Mar. Ecol. Prog. Ser. 227: 37–50, doi:10.3354/meps277037 CHIBANI-CHENNOUFI, S., A. BRUTTIN, M.-L. DILLMANN, AND H. BRU¨SSOW. 2004. Phage-host interaction: An ecological perspective. J. Bacteriol. 186: 3677–3686, doi:10.1128/ JB.186.12.3677-3686.2004

473

COCHRAN, P. K., AND J. H. PAUL. 1998. Seasonal abundance of lysogenic bacteria in a subtropical estuary. Appl. Environ. Microbiol. 64: 2308–2312. FORTIER, L., AND J. K. COCHRAN. 2008. Introduction to special section on annual cycles on the Arctic Ocean Shelf. J. Geophys. Res. 113: C03S00, doi:10.1029/2007JC004457 FUHRMAN, J. A. 1999. Marine viruses and their biogeochemical and ecological effects. Nature 299: 541–548, doi:10.1038/ 21119 GARNEAU, M.-E`., S. ROY, C. LOVEJOY, Y. GRATTON, AND W. F. VINCENT. 2008. Seasonal dynamics of bacterial biomass and production in a coastal arctic ecosystem: Franklin Bay, western Canadian Arctic. J. Geophys. Res. 113: C07S91, doi:10.1029/2007JC004281 ———, W. F. VINCENT, L. ALONSO-SA´EZ, AND Y. GRATTON. 2006. Prokaryotic community structure and heterotrophic production in a river-influenced coastal arctic ecosystem. Aquat. Microb. Ecol. 42: 27–40, doi:10.3354/ame042027 ———, ———, R. TERRADO, AND C. LOVEJOY. 2009. Importance of particle-associated bacterial heterotrophy in a coastal Arctic ecosystem. J. Mar. Syst. 75: 185–197, doi:10.1016/ j.jmarsys.2008.09.002 JIANG, S. C., AND J. H. PAUL. 1996. Occurrence of lysogenic bacteria in marine microbial communities as determined by prophage induction. Mar. Ecol. Prog. Ser. 142: 27–38, doi:10.3354/meps142027 LONG, A., L. D. MCDANIEL, J. MOBBERLEY, AND J. H. PAUL. 2008. Comparison of lysogeny (prophage induction) in heterotrophic bacterial and Synechococcus populations in the Gulf of Mexico and Mississippi river plume. ISME J. 2: 132–144, doi:10.1038/ismej.2007.102 MIDDELBOE, M., N. JØRGENSEN, AND N. KROER. 1996. Effects of viruses on nutrient turnover and growth efficiency of noninfected marine bacterioplankton. Appl. Environ. Microbiol. 62: 1991–1997. ———, T. NIELSEN, AND P. BJØRNSEN. 2002. Viral and bacterial production in the North Water: In situ measurements, batchculture experiments and characterization and distribution of a virus-host system. Deep-Sea Res. II 49: 5063–5079, doi:10.1016/S0967-0645(02)00178-9 NOBLE, R. T., AND J. A. FUHRMAN. 1998. Use of SYBR Green I for rapid epifluorescence counts of marine viruses and bacteria. Appl. Environ. Microbiol. 59: 905–911. PAUL, J. H. 2008. Prophages in marine bacteria: Dangerous molecular time bombs or the key to survival in the seas? ISME J. 2: 579–589, doi:10.1038/ismej.2008.35 ———, AND M. G. WEINBAUER. 2010. Detection of lysogeny in marine environments, p. 30–33. In S. W. Wilhelm, M. G. Weinbauer, and C. A. Suttle [eds.], Manual of aquatic viral ecology. ASLO. PAYET, J. P., AND C. A. SUTTLE. 2008. Physical and biological correlates of virus dynamics in the southern Beaufort Sea and Amundsen Gulf. J. Mar. Syst. 74: 933–945, doi:10.1016/ j.jmarsys.2007.11.002 ROWE, J. M., AND OTHERS. 2008. Constraints on virus production in the Sargasso Sea and North Atlantic. Aquat. Microb. Ecol. 52: 233–244, doi:10.3354/ame01231 SHELFORD, E. J., M. MIDDELBOE, E. F. MØLLER, AND C. A. SUTTLE. 2012. Virus-driven nitrogen cycling enhances phytoplankton growth. Aquat. Microb. Ecol. 66: 41–46, doi:10.3354/ame01553 STEIN, R., AND R. W. MACDONALD [EDS.], The organic carbon cycle in the Arctic Ocean. Springer. STEWARD, G. F., D. SMITH, AND F. AZAM. 1996. Abundance and production of bacteria and viruses in the Bering and Chukchi Seas. Mar. Ecol. Prog. Ser. 131: 287–300, doi:10.3354/ meps131287

474

Payet and Suttle

SUTTLE, C. A. 2005. Viruses in the sea. Nature 437: 356–361, doi:10.1038/nature04160 ———, AND J. A. FUHRMAN. 2010. Enumeration of virus particles in aquatic or sediment samples by epifluorescence microscopy, p. 145–153. In S. W. Wilhelm, M. G. Weinbauer, and C. A. Suttle [eds.], Manual of aquatic viral ecology. ASLO, doi:10.4319/mave.2010.978-0-9845591-0-7 VALLIE´RES, C., L. RETAMAL, C. OSBURN, AND W. F. VINCENT. 2008. Bacterial production and microbial food web structure in a large arctic river and the coastal Arctic Ocean. J. Mar. Syst. 74: 756–773, doi:10.1016/j.jmarsys.2007.12.002 VINCENT, W. F., AND C. PEDRO´S-ALIO´. 2008. Sea ice and life in a river-influenced arctic shelf ecosystem. J. Mar. Syst. 74: 739–1024, doi:10.1016/j.jmarsys.2008.03.003 WEINBAUER, M. G. 2004. Ecology of prokaryotic viruses. FEMS Microbiol. Rev. 28: 127–181, doi:10.1016/j.femsre.2003.08.001 ———, J.-M. ARRIETA, C. GRIEBLER, AND G. J. HERNDL. 2009. Enhanced viral production and infection of bacterioplankton during an iron induced phytoplankton bloom in the Southern Ocean. Limnol. Oceanogr. 54: 774–784, doi:10.4319/ lo.2009.54.3.0774 ———, I. BRETTAR, AND M. HO¨FLE. 2003. Lysogeny and virusinduced mortality of bacterioplankton in surface, deep, and anoxic marine waters. Limnol. Oceanogr. 48: 1457–1465, doi:10.4319/lo.2003.48.4.1457 ———, J. M. ROWE, AND S. W. WILHELM. 2010. Determining rates of virus production in aquatic systems by the virus reduction approach, p. 1–8. In S. W. Wilhelm, M. G. Weinbauer, and C. A. Suttle [eds.], Manual of aquatic viral ecology. ASLO, doi:10.4319/mave.2010.978-0-9845591-0-7 ———, AND C. A. SUTTLE. 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. ———, AND ———. 1999. Lysogeny and prophage induction in coastal and offshore bacterial communities. Aquat. Microb. Ecol. 18: 217–225, doi:10.3354/ame018217

———, AND OTHERS. 2011. Synechococcus growth in the ocean may depend on the lysis of heterotrophic bacteria. J. Plankton Res. 33: 1465–1476, doi:10.1093/plankt/fbr041 WELLS , L. E., AND J. W. DEMING. 2006. Significance of bacteriovory and viral lysis in bottom waters of Franklin Bay, Canadian Arctic, during winter. Aquat. Microb. Ecol. 43: 209–221, doi:10.3354/ame043209 WILHELM, S. W., S. M. BRIGDEN, AND C. A. SUTTLE. 2002. A dilution technique for the direct measurement of viral production: A comparison in stratified and tidally mixed coastal waters. Microbial. Ecol. 43: 168–173, doi:10.1007/ s00248-001-1021-9 ———, AND C. A. SUTTLE. 1999. Viruses and nutrient cycles in the sea. Bioscience 49: 781–788, doi:10.2307/1313569 WILLIAMSON, S. J., L. A. HOUCHIN, L. D. MCDANIEL, AND J. H. PAUL. 2002. Seasonal variation in lysogeny as depicted by prophage induction in Tampa Bay, Florida. Appl. Environ. Microbiol. 68: 4307–4314, doi:10.1128/AEM.68.9.43074314.2002 WINGET, D. M., R. R. HELTON, K. E. WILLIAMSON, S. R. BENCH, S. J. WILLIAMSON, AND K. E. WOMMACK. 2011. Repeating patterns of virioplankton production within an estuaring ecosystem. Proc. Natl. Acad. Sci. USA 106: 11506–11511, doi:10.1073/pnas.1101907108 ———, K. E. WILLIAMSON, R. R. HELTON, AND K. E. WOMMACK. 2005. Tangential flow diafiltration: an improved technique for estimation of virioplankton production. Aquat. Microb. Ecol. 41: 221–232, doi:10.3354/ame041221

Associate editor: Wade H. Jeffrey Received: 21 April 2012 Accepted: 05 November 2012 Amended: 07 November 2012