Modulation of host ROS metabolism is essential for viral infection of a bloom-forming coccolithophore in the ocean

Abstract

The cosmopolitan coccolithophore Emiliania huxleyi is a unicellular eukaryotic alga responsible for vast blooms in the ocean. These blooms have immense impact on large biogeochemical cycles and are terminated by a specific large double-stranded DNA E. huxleyi virus (EhV, Phycodnaviridae). EhV infection is accompanied by induction of hallmarks of programmed cell death and production of reactive oxygen species (ROS). Here we characterized alterations in ROS metabolism and explored its role during infection. Transcriptomic analysis of ROS-related genes predicted an increase in glutathione (GSH) and H₂O₂ production during infection. In accordance, using biochemical assays and specific fluorescent probes we demonstrated the overproduction of GSH during lytic infection. We also showed that H₂O₂ production, rather than superoxide, is the predominant ROS during the onset of the lytic phase of infection. Using flow cytometry, confocal microscopy and multispectral imaging flow cytometry, we showed that the profound co-production of H₂O₂ and GSH occurred in the same subpopulation of cells but at different subcellular localization. Positively stained cells for GSH and H₂O₂ were highly infected compared with negatively stained cells. Inhibition of ROS production by application of a peroxidase inhibitor or an H₂O₂ scavenger inhibited host cell death and reduced viral production. We conclude that viral infection induced remodeling of the host antioxidant network that is essential for a successful viral replication cycle. This study provides insight into viral replication strategy and suggests the use of specific cellular markers to identify and quantify the extent of active viral infection during E. huxleyi blooms in the ocean.

Introduction

Phytoplankton blooms are of immense ecological importance, as they serve as the foundation for marine food webs, are major sinks for atmospheric CO₂ and are responsible for approximately half of the primary productivity on Earth (Field et al., 1998; Behrenfeld et al., 2006). Opposing the classical view of phytoplankton cells as only passive drifters, recent research conducted on diverse phytoplankton have shown their ability to sense and respond to changing environmental conditions by employing specific acclimation mechanisms (Falciatore et al., 2000; Vardi et al., 2002, 2006; Montsant et al., 2007; Taylor, 2009; McLachlan et al., 2012; Graff van Creveld et al., 2014).

Viruses have a key role in regulating algal bloom dynamics in addition to other common ecosystem removal pathways such as grazing, nutrient availability and vertical sinking (Bidle, 2015). Viral infection that leads to cell lysis is estimated to turn over at least 20% of the photosynthetic biomass (Brussaard et al., 1995; Agusti et al., 1998; Suttle, 2005; Lehahn et al., 2014). Thus, viruses have a huge impact on ocean biogeochemical cycles by short-circuiting the flux of carbon and nutrients from phytoplankton and bacteria to higher trophic levels, shunting it to the pool of dissolved and particulate organic matter, and making it more available for recycling by microbial respiration (Fuhrman, 1999; Wilhelm and Suttle, 1999; Suttle, 2007). Marine viruses are known to regulate phytoplankton population size, distribution and composition by eliminating more dominant phytoplankton species and are considered major vehicles for horizontal gene transfer between distantly related species (Lindell et al., 2005; Suttle, 2005; Rohwer and Thurber, 2009; Sharon et al., 2009; Winter et al., 2010; Breitbart, 2012).

E. huxleyi is the most abundant and widespread coccolithophore species in the modern oceans (Holligan et al., 1983; Balch et al., 1992; Brown and Yoder, 1994). It forms massive annual spring blooms that are detectable from space by satellites and affects global climate and biogeochemical cycles of carbon and sulfur (Hatton et al., 2004; Rost and Riebesell, 2004; Tyrrell and Merico, 2004). E. huxleyi blooms are routinely infected and terminated by specific giant double-stranded DNA coccolithoviruses (Phycodnaviridae) (Bratbak et al., 1993; Brussaard et al., 1996; Schroeder et al., 2002; Wilson et al., 2002; Lehahn et al., 2014), the E. huxleyi virus (EhV), which is part of the nucleo-cytoplasmic large DNA virus (Asfarviridae, Ascoviridae, Iridoviridae, Marseilleviridae, Megaviridae, Mimiviridae, Pandoraviridae, Phycodnaviridae, Pithoviridae and Poxviridae) clade. EhV encodes for an almost complete de-novo biosynthetic pathway for ceramide (Wilson et al., 2005), a sphingolipid known to induce apoptosis in animals and plants (Pettus et al., 2002; Liang et al., 2003). This pathway was recently shown to have a critical role in EhV replication and in induction of E. huxleyi programmed cell death in cultures and during natural bloom demise (Pagarete et al., 2009; Vardi et al., 2009, 2012). The viral-induced programmed cell death includes membrane blebbing, induction of caspase-like activity, expression of metacaspases and accumulation of reactive oxygen species (ROS) and specific viral-derived glycosphingolipids (Evans et al., 2006; Bidle et al., 2007; Vardi et al., 2009, 2012). Nonetheless, the cellular signaling and infection mechanisms during the interaction of E. huxleyi with its specific virus are poorly understood.

ROS are highly reactive forms of oxygen (radicals and non-radicals), usually regular by-products of oxygen metabolism from photosynthesis, oxidative phosphorylation or peroxisomal activity (Halliwell and Gutteridge, 1986; Asada, 2006; Halliwell, 2006). In high doses, ROS can be very harmful to living cells, ultimately leading to induction of cell death (Van Breusegem and Dat, 2006). In order to avoid the damage caused by ROS, aerobic organisms developed an antioxidant machinery to control the levels of different reactive species. This antioxidant system is composed of ROS-scavenging enzymes and small molecules. Among the ROS-scavenging enzymes are superoxide dismutase, ascorbate peroxidase (APX), catalase (CAT), glutathione (GSH) peroxidase and peroxiredoxin. These enzymes are encoded by the antioxidant gene network and are present in almost all organelles (Apel and Hirt, 2004; Mittler et al., 2004; Halliwell, 2006). Among the small-molecule antioxidants are GSH, ascorbic acid and nicotinamide adenine dinucleotide phosphate (NADPH). These molecules comprise the reductive pool of the cell and react directly with oxidizing agents or through enzymatic reactions mediated by the different antioxidant enzymes listed above (Mittler, 2002). Controlled by the cellular antioxidant system, ROS is a key signaling factor that mediates acclimation responses to environmental stress conditions (Foyer and Noctor, 2003; Mittler et al., 2004; D’Autréaux and Toledano, 2007; Dangoor et al., 2012). Recent reports proposed an important role for ROS metabolism in algal response to environmental stress, such as iron limitation, allelopathy, UV stress, photo-acclimation and oxidative stress (Rijstenbil, 2002; Sukenik et al., 2002; Vardi et al., 2002, 2007; Janknegt et al., 2008; Li et al., 2009; Murik and Kaplan, 2009; Fischer et al., 2012; Thamatrakoln et al., 2012; Graff van Creveld et al., 2014; Murik et al., 2014). As environmental stress conditions can eventually determine the fate of large-scale oceanic blooms, ROS-mediated responses may have an impact on large biogeochemical cycles.

In this study, we investigate the role of ROS metabolism in E. huxleyi cells during viral infection. By mapping gene expression data onto antioxidant metabolic pathways, we predicted an increase in GSH biosynthesis, dehydroascorbic acid reduction and H₂O₂ metabolism during lytic infection. Flow cytometry analysis, confocal microscopy and multispectral imaging flow cytometry coupled to biochemical assays showed that H₂O₂ and GSH accumulate in different subcellular localizations only within a subpopulation of infected cells. When the induction of ROS production was inhibited, viral production and cell death were strongly reduced. These findings reveal the importance of ROS and redox metabolism in the viral infection mechanism of E. huxleyi.

Materials and methods

Culture growth and viral infection

The non-calcifying E. huxleyi strain CCMP2090 was used for this study. Cells were cultured in K/2 medium (Keller et al., 1987) and incubated at 18 °C with a 16:8 h. light–dark illumination cycle. A light intensity of 100 μmol photons m⁻² s⁻¹ was provided by cool white light-emitting diode lights. All experiments were performed with exponential phase cultures (5 × 10⁵–1 × 10⁶ cells ml⁻¹). The virus used for this study is EhV86 (Schroeder et al., 2002). In all infection experiments, E. huxleyi was infected with 1:50 volumetric ratio of viral lysate to culture (multiplicity of infection of ~1:1 viral particles per cell).

Bioinformatic analysis

A list of candidate genes involved in ROS and antioxidant metabolism was used according to (Mittler et al., 2004). Protein sequences of the target gene from human, Arabidopsis thaliana, Saccharomyces cerevisiae and, if necessary, additional species were compared with the E. huxleyi genome (Read et al., 2013) on the JGI genome website using TBlastN (Altschul et al., 1997). TBlastN hits on the E. huxleyi genome were manually defined, aligned and compared with different closely related sequences as previously described by Feldmesser et al. (2014). Expression data for putative genes participating in ROS and antioxidant metabolism was extracted from the database produced by Rosenwasser et al. (2014) and presented as the calculated log2 of the expression relative to uninfected control in the same time point.

Peroxidase activity

Peroxidase activity was measured using Amplex Red (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions.

Flow cytometry

Flow cytometry analysis was done on either Eclipse (Sony Biotechnology Inc., Champaign, IL, USA) or LSRII (BD, Franklin Lakes, NJ, USA) flow cytometer, both equipped with 405 and 488 nm solid-state air-cooled lasers, and with standard optic filter set-up. For each sample a minimum of 5000 events were acquired.

Enumeration of cell abundance

Cells were monitored and quantified using an Eclipse (iCyt) flow cytometer. Cells were identified by plotting the chlorophyll fluorescence (663–737 nm) versus green fluorescence (500–550 nm) or side scatter.

Enumeration of viral abundance

For extracellular viral counts, samples were fixed with a final concentration of 0.5% glutaraldehyde for 30 min at 4 °C, then plunged into liquid nitrogen and stored at −80 °C until analysis. After thawing, samples were stained with SYBR gold (Invitrogen) that was diluted 1:10 000 in Tris–EDTA buffer, incubated for 20 min at 80 °C and cooled to room temperature (RT). Samples were analyzed by an Eclipse flow cytometer (excitation (ex): 488 nm and emission (em): 500–550 nm), minimum of 50 000 events were collected.

Cell death analysis

For cell death analysis, samples were stained with a final concentration of 1 μm Sytox Green (Invitrogen), incubated in the dark for 1 h at RT and analyzed by an Eclipse flow cytometer (ex: 488 nm and em: 500–550 nm). An unstained sample was used as control to eliminate the background signal.

Detection of intracellular ROS

For general ROS analysis, samples were stained with 5 μm CM-H2DCFDA (Invitrogen). Samples were incubated at RT in the dark for 1 h before analysis by an Eclipse flow cytometer (ex: 488 nm and em: 500–550 nm) and compared with an unstained sample. Bes-H₂O₂-AC (Waco, Osaka, Japan) at a final concentration of 50 μm and Bes-SO-AM (Waco) at a final concentration of 30 μm were used for specific detection of H₂O₂ and superoxide, respectively. Samples were incubated with the different stains for 30 min at RT in the dark before being analyzed by an Eclipse flow cytometer (ex: 488 nm and em: 500-550 nm). An unstained sample was used as control to eliminate the background signal.

Total GSH, GSH, GSSG and GSH/GSSG ratio

GSH was measured by the recycling assay, GSH reductase-dependent reduction of 5,5′-dithiobis(2-nitro-benzoic acid) (Ellman’s reagent; Tietze, 1969) adjusted for microplate assay (Queval and Noctor, 2007). Briefly, 400 ml cultures were pelleted by mean of centrifugation (17 000 g, 10 min, 4 °C) plunged in to liquid nitrogen and were kept in −80 °C. Extractions were performed in 1 m HClO₄ to minimize GSH oxidation. On adjustment of pH to 6 by phosphate buffer and KCO₃, part of the extract was reacted with 2-vinylpyridine for quantification of oxidized GSH (GSSG). Both (total and GSSG) were quantified by a standard curve of known concentrations of GSH and GSSG for the reaction kinetics. Quantities of total GSH, GSSG and GSH were calculated from concentration and volumes, and normalized to cell abundance at time of collection.

Detection of intracellular GSH

Monochlorobimane (mBCl, Sigma-Aldrich, Rehovot, Israel) was added to the samples to a final concentration of 50 μm and incubated for 30 min at RT in the dark, then detected using an Eclipse flow cytometer (ex: 405 nm and em: 430–480 nm). An unstained sample was used as control to eliminate the background signal.

Co-staining for GSH and H₂O₂

mBCl and Bes-H₂O₂ (50 and 30 μm, respectively) were applied to the cells simultaneously. For each treatment, an unstained sample and a single stain for each dye was assessed by the same method. All samples were analyzed by an LSRII flow cytometer (BD) using the blue (ex: 405 nm and em: 430–480 nm) and green (ex: 488 nm and em: 500–550 nm) channels. Stained samples were also observed by confocal microscopy using a Nikon Eclipse Ti-E microscope (Tokyo, Japan) with Nikon A1+ confocal laser system equipped with CFI PLAN APO VC × 60A WI and CFI Plan Apochromatic × 100 Lambda objectives. For mBCl excitation a 405 nm 100mW laser was used with 470nm emission collection filter, for Bes-H2O2 excitation a Sapphire 488 nm 50 mW laser was used with 525 nm emission collection filter and for chlorophyll excitation a 640nm 100mW laser was used with 700nm emission collection filter. Acquired images were analyzed using NIS-Elements AR imaging software. For multispectral imaging flow cytometry, cells were stained as described and 5000 cells were collected from each sample using ImageStreamX flow cytometer (Amnis Corp., Seattle, WA, USA; Millipore, Darmstadt, Germany). Data were analyzed using image analysis software (IDEAS 6.0; Amnis Corp., Millipore). Images were compensated for fluorescent dye overlap by using single-stain yeast cells as non-chlorophyll-containing controls. Cells were gated for single cells using the area and aspect ratio features, and for focused cells using the Gradient RMS feature, as previously described (George et al., 2006). Cells were gated for positive signal of chlorophyll, mBCl and Bes-H₂O₂ (for GSH and H₂O₂, respectively). The co-localization for the two dyes was quantified using the bright detail similarity feature (the log base 2 transformed Pearson’s correlation coefficient of the localized bright spots in the two input images; George et al., 2006).

Cell sorting

Cells were sorted according to their mBCl and Bes-H₂O₂ signals using SORP FACSAriaII cell sorter (BD) equipped with 405 and 488 nm solid-state air-cooled lasers, and with standard optic filter set-up. About one million cells were sorted in purity mode from each stained population. Cells were collected on GF/F filters, plunged in liquid nitrogen and stored in −80 °C before quantitative PCR analysis for viral abundance.

Enumeration of intracellular viral abundance

For intracellular viral DNA quantification, DNA was extracted from GF/F filters as previously described (Schroeder et al., 2002) and then used for quantitative PCR analysis with the major capsid protein primers (Pagarete et al., 2009). All reactions were performed in four biological replicates and technical duplicates. For all reactions, Platinum SYBER Green qPCR SuperMix-UDG with ROX (Invitrogen) was used as described by the manufacturer. Reactions were performed on StepOnePlus real-time PCR systems (Applied Biosystems, Foster City, CA, USA) as follows: 50 °C for 2 min, 95 °C for 2 min, 40 cycles of 95 °C for 15 s and 60 °C for 30 s. Results were calibrated against serial dilutions of EhV201 DNA at known concentrations, enabling exact enumeration of viral abundance, and normalized to cell numbers collected on filters.

Detection of carbonylated proteins

OxyBlot Protein Oxidation Detection Kit (Millipore, Darmstadt, Germany) was used according to the manufacturer’s instructions for the assessment of cellular oxidative damage.

Application of peroxidase inhibitor and ROS scavenger

Esculetin, an inhibitor of peroxidase activity (Dunand et al., 2007; Zhang et al., 2008), was dissolved in dimethyl sulfoxide and applied to cultures to a final concentration of 50 μM, 1 h before infection. Potassium iodide, an H₂O₂ scavenger (Dunand et al., 2007), was dissolved in distilled water and added to cultures to a final concentration of 1 mM, 1 h before infection and on every consecutive day.

Statistical analysis

All reported P-values were determined using a two-tailed unpaired Student’s t-test and presented in figures were *P<0.1, **P<0.05 and ***P<0.01. In all figures, error bars represent s.e. and ‘n’ defined as the number of unrelated replicas of each treatment.

Results

Expression profile of ROS metabolism genes during EhV infection

To explore the role of ROS during viral infection, we examined the expression profile of selected genes that participate in cellular ROS metabolism (Mittler et al., 2004). Complete DNA sequences of the target genes were manually defined based on genomic sequences, expressed sequence tags and RNA sequencing data as described in the Materials and Methods section. Expression data for each of the defined genes was generated based on RNA sequencing transcriptome analysis of RNA derived from E. huxleyi cells at 1 and 24 h post infection (hpi) with the lytic virus EhV201 or the non-lytic virus EhV163, and normalized to expression levels in uninfected cells at the same time point (Feldmesser et al., 2014; Rosenwasser et al., 2014). Changes were observed in expression levels of several transcripts involved in the biosynthesis and recycling of GSH and ascorbate, important low-molecular-weight antioxidants (Foyer and Noctor, 2005). Genes that encode for glutamate cysteine ligase and GSH synthase, key enzymes in the GSH biosynthesis pathway, were increased 2.4-fold at 24 hpi (Figure 1). The expression of GSH reductase transcript increased 1.4-fold at 24 hpi (Figure 1). The ascorbate recycling enzyme dehydroascorbate reductase had no homolog in the E. huxleyi genome. GSH S-transferase omega (GSTO) gene, which exhibited dehydroascorbate reductase activity in human and other animal systems (Board et al., 2000), increased its expression levels by 1.7-fold at 24 hpi. Although genes involved in production or recycling of ascorbate and GSH were generally upregulated, the expression level of APX, an enzyme that depends on ascorbate as an electron donor, increased up to 1.4-fold at 1 hpi but was three times lower than uninfected control at 24 hpi (Figure 1). On the other hand, transcript levels of the GSH peroxidase enzyme that uses GSH as a reducing agent increased by 1.7-fold at 1 hpi and 1.3-fold at 24 hpi (Figure 1).

Figure 1

Gene expression profiles of ROS and redox-related transcripts during E. huxleyi viral infection. E. huxleyi infected with the lytic (L) virus EhV201 or with the non-lytic (NL) virus EhV163. Red, upregulation; blue, downregulation; scale values refer to log base 2 of reads per milion (RPM) normalized to uninfected control at the same time point. AA, ascorbic acid; APX, ascorbate peroxidase; CAT, catalase; DHA, dihydroascorbate; GCL, glutamate cysteine ligase; GPX, GSH peroxidase; GR, GSH reductase; GSH, glutathione; GSS, GSH synthase; GSSG, oxidized GSH; GSTO, GSH S-transferase omega; MDHA, monodihydroascorbate; NADPH-nicotinamide adenine dinucleotide phosphate; NOX, NADPH oxidase; Px, peroxidase; SOD, superoxide dismutase.

Expression levels of NADPH oxidase, a producer of superoxide, increased by 4.1-fold at 24 hpi, whereas superoxide dismutase and CAT expression, both antioxidant enzymes that scavenge superoxide and H₂O₂, respectively, were 3.2- and 2.5-fold lower then uninfected control, respectively, at 24 hpi (Figure 1). The most dominant elevation in expression, up to 7-fold at 24 hpi, was found for transcripts encoded by gene homologs of peroxidases (Figure 1). Some of these genes are known to produce ROS as H₂O₂ or superoxide radical, whereas other functions attributed to this gene family are antioxidant activity and specifically the catalysis of H₂O₂ degradation. Only minor changes in expression of these genes were found in E. huxleyi cultures treated with the non-lytic virus EhV163, with the exception of NADPH oxidase and peroxidase transcript levels that decreased in expression. Taken together, our data reveal a major remodeling in expression of transcripts related to ROS metabolism that occurs in E. huxleyi cells as a result of viral infection.

Viral infection is associated with host co-generation of H₂O₂ and GSH

The simultaneous rise in ROS and GSH levels during infection was associated with induction of host cell death and viral production. A decline in host cell abundance was observed from 72 hpi onwards (Figure 2a), together with accumulation of EhV in the culture media (Figure 2b), and rapid increase of cell death detected using Sytox Green as a proxy for compromised membrane integrity (Figure 2c). The onset of the lytic phase of infection (72 hpi) was characterized by significant induction in ROS accumulation as detected by CM-H₂DCF DA (Figure 2d, Student’s t-test, P-value=0.01). In order to assess whether specificity in ROS chemical species have a role in onset of lytic infection, we used specific ROS stains to either superoxide (detected by Bes-SO) or H₂O₂ (detected by Bes-H₂O₂). Where superoxide levels varied only slightly (Figure 2e), levels of H₂O₂ exhibited a profound increase during lytic phase of viral infection (Figure 2f). In order to examine the impact of oxidative stress on protein oxidation, we analyzed carbonylated proteins detected by OxyBlot during infection. This analysis showed accumulation of oxidized proteins in infected cells already at early stages (>4 hpi), whereas uninfected cells started accumulating damaged proteins only at later stages of growth (Supplementary Figure S1).

Figure 2

H₂O₂ is the predominant ROS produced during viral infection of E. huxleyi. Cell abundance (a), extracellular viral abundance (b), cell death assessed by Sytox staining (c), ROS assesed by CM-H2DCFDA staining (d), superoxide measured by Bes-SO staining (e) and H₂O₂ measured by Bes-H₂O₂ staining (f) of E. huxleyi cells infected with lytic virus (empty circles) and uninfected control (full circles). Each data point was calculated based on average flow cytometry fluorescence or count measurements of at least 5000 events per sample point and presented as mean ±s.e., n=3. P-values were determined using two-tailed unpaired Student’s t- test (*P<0.1, **P<0.05 and ***P<0.01).

The fluorescent stain mBCl was further used to complement ROS measurement with quantification of intracellular GSH levels. Although GSH levels in uninfected control cells decreased slightly as the culture aged, GSH accumulated in infected cells beginning at 48 hpi (Figure 3a). The induction in GSH and H₂O₂ levels accompanied the onset of the lytic phase of infection. In order to compliment the in vivo fluorescent-based assay, we employed a biochemical approach to quantify the various GSH pools in the cells during viral infection. Using this biochemical assay, which is based on the GSH reductase-dependent reduction of 5,5′-dithiobis(2-nitro-benzoic acid) (Ellman’s reagent; Tietze, 1969; Queval and Noctor, 2007) (Figures 3b–e), we detected an increase in total GSH (GSH+GSSG, Figure 3b) in cell extracts derived from lytic phase of infection. This result was supported by previous research detecting total GSH increase during infection by means of high-performance liquid chromatography (Gledhill et al., 2012). In agreement with the in vivo GSH assessment using mBCl fluorescence stain (Figure 3a), at 72 hpi levels of GSH were significantly higher in infected cells in comparison with uninfected cells (Figure 3c, Student’s t-test, P-value=0.02). Furthermore, large increase in GSSG was detected during viral infection (Figure 3d). These changes in the relative pools of GSH were reflected in a marked decline in the GSH/GSSG ratio (Figure 3e), a ratio that is considered as proxy for the oxidative stress (Noctor and Foyer, 1998; Kranner et al., 2006). Alterations in GSH/GSSG as a result of ROS production can regulate the activation of many biological processes, such as transcription, posttranslational modification and protein–protein interactions, by affecting the oxidation state of thiol groups. Taken together, the profound ROS generation during viral infection had a major consequence on cellular redox metabolism and antioxidant networks as determined by the GSH pools. The gene expression analysis reflected this remodeling where GSH biosynthetic pathways were found to divert towards production and recycling, while genes responsible for H₂O₂ degradation (CAT and APX) decreased in expression level (Figure 1).

Figure 3

Quantification of total, reduced (GSH) and oxidized (GSSG) pools of GSH, during E. huxleyi viral infection. E. huxleyi cells infected with lytic virus (empty circles or bars) or uninfected control (full circle or bars) stained by mBCl for quantification of level of reduced GSH (a). Each data point was calculated based on average flow cytometry fluorescence measurements of at least 5000 events per sample point and presented as mean ±s.e., n=3. Biochemical-based quantification of GSH total (GSH+GSSG), reduced (GSH) and oxidized (GSSG) GSH per cell and GSH/GSSG ratio (b–e) in cell extract from control or infected cells, presented as mean ±s.e., n=4. P-values were determined using two-tailed unpaired Student’s t-test (*P<0.1, **P<0.05 and ***P<0.01).

H₂O₂ and GSH accumulate simultaneously in different subcellular localizations of infected cells

The accumulation of GSH (Figure 3a) simultaneously with H₂O₂ production (Figure 2f) raise questions regarding its origin and cellular localization of pro- and antioxidants in infected cells. Flow cytometry analysis of cells co-stained for both H₂O₂ and GSH revealed that the elevation in H₂O₂ occurred in the same subpopulation with increased GSH content (Figure 4). Although in the first day of infection no differences were found between infected cells and control, at 72 hpi 38% of the infected cells were found positive for both H₂O₂ and GSH, as compared with 7% of control cells (Figure 4a). In order to get direct evidence that the dual-stained cells were indeed infected, we sorted GSH- and H₂O₂-positive cells from culture infected for 72 hpi. These cells were subjected to quantitative PCR analysis of viral DNA using major capsid protein primers. This dual-stained subpopulation had fourfold higher viral DNA copy number per cell as compared with cells sorted from GSH- and H₂O₂-negative subpopulation (Figure 4b).

Figure 4

H₂O₂ and GSH co-generated in viral-infected cells. Flow cytometry dot plots showing population dynamics during the course of viral infection. Cells were stained for GSH (mBCl) and H₂O₂ (Bes-H₂O₂). Representative data from at least five experiments are shown (a). AFU, arbitrary fluorescent units. Bars showing major capsid protein (MCP) copy number per cell obtained by quantitative PCR (qPCR) analysis for the different GSH and H₂O₂ population sorted (indicated by arrows) out of the two gated population (GSH− H₂O₂− and GSH+ H₂O₂+) at 72 hpi with EhV 86 (b), presented as mean ±s.e., n=4. P-values were determined using two-tailed unpaired Student’s t-test (**P<0.05).

Using confocal microscopy, we subsequently examined the subcellular localization of GSH and H₂O₂ in infected cells. The GSH signal was punctuated in both control and infected cells, while a strong H₂O₂ signal extended only to a vast fraction of infected cells (Figures 5a and b). Interestingly, the two fluorescent signals had different localization within the infected cells and both were excluded from the chloroplast (Figures 5a and b). We further used a multispectral imaging flow cytometry, in order to quantify this observation in a high-throughput manner (>5000 infected E. huxleyi cells). In order to estimate the co-localization of the blue (GSH) and green (H₂O₂) signals, the bright detail similarity feature was used as described in the Materials and Methods section. Representative cells with high signal for both H₂O₂ and GSH were chosen for their median similarity score (Figure 5c). In order to calibrate the bright detail similarity value as a co-localization score between GSH and H₂O₂, the same procedure was applied to cells stained with two different DNA stains (Hoechst and Syto13). In comparison with the high-value DNA-stained cells score for their co-localization in the nucleus (population median of 2.0; Figure 5d, blue histogram), infected E. huxleyi cells with elevated GSH and H₂O₂ signals showed very low, bright detail similarity (population median of 0.8; Figure 5d, red histogram). The low score measured for the signals is indicating low co-localization between the two molecules in the cell.

Figure 5

Subcellular localization of GSH and H₂O₂ in infected E. huxleyi cells. Confocal fluorescence micrographs detecting subcellular H₂O₂ (Bes-H₂O₂, green fluorescence), GSH (mBCl, blue fluorescence), chlorophyll (red fluorescence) and DIC images of the same fields, in uninfected control (upper panels) or EhV-infected E. huxleyi cells (lower panels). Scale bars represent 1 μm (a). (b) The same samples as in a shown in higher magnification; scale bars represent 1 μm. Imaging flow cytometry (Image Stream X) micrographs of representative cells selected for their median bright detail similarity score (0.8), channels of bright field (BF), green (H₂O₂), red (chlorophyll, Chl) and blue (GSH) are presented; scale bar represent 7 μm (c). Population distribution of 5000 cells selected for their bright detail similarity between the green and blue fluorescence channels representing spatial co-localization between H₂O₂ and GSH in the cells (red histogram). As positive control for co-localization bright detail similarity score of cells stained with both Syto13 (green DNA stain) and Hoechst (blue DNA stain) are presented (blue histogram) (d).

Elevated ROS levels are essential for a successful viral replication cycle

To corroborate the high expression level of peroxidase genes with enzyme activity, we measured peroxidase activity during infection and found it to increase in the infected culture 2.4-fold above the levels measured in uninfected control (Supplementary Figure S2). Consequently, we used esculetin, a known inhibitor of peroxidase activity, in order to examine the functional role of ROS in the E. huxleyi–EhV (Sirois and Miller, 1972; Dunand et al., 2007; Potapovich et al., 2012). Esculetin treatment completely abolished both viral-induced cell death (Figure 6a) and viral production (Figure 6b). Interestingly, esculetin was highly efficient in suppressing H₂O₂ generation and in preventing GSH accumulation (Figures 6c and d). Importantly, suppression of viral release was not due to cytotoxicity of esculetin, as treatment of an uninfected culture with the substance had only a minor effect on cell growth, while esculetin treatment in infected cells showed the same growth as uninfected control (Figure 6a). Furthermore, a similar trend was found in treatment of infected cells with potassium iodide, an H₂O₂ scavenger, significantly decreasing cell death and viral productivity (Supplementary Figures S3a and b, Student’s t-test, see figure for P-values). Taken together, our data strongly support the pivotal role of ROS metabolism orchestrated by a complex stoichiometry between ROS generation and antioxidant regulation during the lytic phase of EhV infection.

Figure 6

Inhibition of peroxidase activity reduced cell death, viral production, H₂O₂ and GSH accumulation. E. huxleyi cell abundance (a), EhV viral abundance (b), cellular H₂O₂ (c) and GSH (d) of cultures infected with the lytic virus (empty symbols) and uninfected control (full symbols) treated with 50 μm esculetin (tringles) or untreated cells (circles). Each data point was calculated based on average flow cytometry fluorescence or count measurements of at least 5000 events per sample point and presented as mean ±s.e., n=3. P-values were determined using two-tailed unpaired Student’s t-test (*P<0.1, **P<0.05 and ***P<0.01) calculated in each time point between esculetin-treated and -untreated infected E. huxleyi cultures.

Discussion

In this study, we explored the role of ROS metabolism during host–virus interaction of the bloom-forming coccolithophore E. huxleyi. ROS was previously proposed to be associated with viral infection and specifically with induction of hallmarks of programmed cell death during the lytic phase of infection in E. huxleyi cultures (Evans et al., 2006) and mesocosm experiments (Vardi et al., 2012). Recent reports have shown a reduction in the photosystem II efficiency, chlorophyll degradation and carotenoid production during E. huxleyi infection, supporting a possible source for ROS generation (Evans et al., 2006; Bidle et al., 2007; Llewellyn et al., 2007; Bale et al., 2013; Kimmance et al., 2014). Our transcriptome analysis suggested an additional enzymatic source for ROS production by induction of NADPH oxidase (Figure 1). Reduction in APX and CAT expression implied a compromised antioxidant capacity and H₂O₂ accumulation that may mediate viral-induced cell death during lytic phase of viral infection (Figure 1).

As diverse environmental stress conditions trigger ROS, one of the key questions is how specificity in ROS signaling is being achieved in order to regulate cell fate in marine phytoplankton. Our data demonstrated that H₂O₂, and not superoxide, is the dominant intracellular ROS that accumulates during the onset of the lytic phase of viral infection (Figures 2e and f). The interplay between H₂O₂ and superoxide was found to be an essential factor in determining cell fate in different model systems (Bloomfield and Pears, 2003; Lardy et al., 2005; Owusu-Ansah and Banerjee, 2009; Tsukagoshi et al., 2010). The transcriptomic analysis presented here suggested that viral infection triggered H₂O₂ production, presumably by activation of a combination of NADPH oxidase and/or specific peroxidases, while decreasing the degradation of H₂O₂ by downregulating CAT and APX (Figure 1). Such changes in expression patterns can enable transduction of specific ROS signaling cascades that might orchestrate the final stage of viral infection and cell lysis (Veal et al., 2007; Go and Jones, 2008; Shao et al., 2008; Mittler et al., 2011; Siddique et al., 2014).

ROS has an important role in facilitating different stages of viral infection cycle in diverse virus host systems such as Kaposi’s sarcoma-associated herpes virus, hepatitis C virus and HIV (Perl and Banki, 2000; Ye et al., 2011; Bottero et al., 2013; Ma et al., 2013). On the other hand, in plants, during tobacco mosaic virus infection, ROS signaling was found to initiate plant defense hypersensitive response (Allan et al., 2001; Yoda et al., 2003; Love et al., 2005). Interestingly, chlorella virus, a close relative of EhV within the nucleo-cytoplasmic large DNA virus family, was found to encode a superoxide dismutase protein (Kang et al., 2014). Our data show that ROS, and specifically H₂O₂, is a crucial factor during EhV infection, whereby inhibition of peroxidase activity using esculetin or scavenging of H₂O₂ using potassium iodide resulted in a reduction in host cell death and a decline in viral production (Figures 6a and b, and Supplementary Figures S3a and b). These results add a new angle to the role of ROS as a key factor in the EhV viral replication cycle.

Typically, a rise in ROS production during stress occurs concomitant with a decrease in the antioxidant capacity (Foyer and Noctor, 2011). In contrast, we report that the rise in ROS concentration was accompanied by maintaining high GSH quota and a rise in GSSG and total GSH during the late stage of viral infection under a highly oxidizing cellular environment (Figures 3–5). These data may reflect the observed pattern of gene expression in the infected host’s transcriptome driving the GSH-ascorbate cycles. The essential pathways that maintain the reducing pools of GSH and ascorbate, such as GSH biosynthesis genes (GCL and GSS), as well as GSH and ascorbate recycling genes (GR and GSTO), were upregulated during infection (Figure 1). The observed overproduction of total GSH in parallel to decline in the GSH/GSSG ratio may point to increase in the flux through the GSH biosynthesis and GSH/ascorbate pathways. These changes might serve the purpose of maintaining a high ratio of reduced to oxidized ascorbate, which was reported to precondition the cell sensitivity to oxidative stress, regulating cell death (Murik and Kaplan, 2009; Murik et al., 2014). Similarly, keeping a high cellular GSH concentration can serve an important function in maintaining host metabolism and newly synthesized viral proteins under the highly oxidative microenvironment of the infected cell. It was recently shown that a decrease in cellular GSH concentrations can serve as an early event in the apoptotic cascade, and that low antioxidant capacity can lead to initiation of a programmed death cascade (Circu and Aw, 2008; Franco and Cidlowski, 2009). Thus, viral replication might benefit from high intracellular GSH content (Gledhill et al., 2012), prolonging cell viability and circumventing premature cell death that might be induced as a defense strategy of the host.

Furthermore, EhV replication has a high demand for nucleotides to support the synthesis and replication of its large genome (Wilson et al., 2005) and high burst size (Bratbak et al., 1993). This metabolic demand can be fulfilled by converting ribonucleotides to deoxy ribonucleotides via ribonucleotide-diphosphate reductase activity, which in turn has to be reduced by thioredoxin or by glutaredoxin, in the second drawing the reducing power out of the GSH pool (Gon et al., 2006). This possibility underlines the importance of the increase in GSH metabolism during infection, while also suggesting why this virus harbors an RNR gene in its genome (Wilson et al., 2005).

ROS metabolism was recently shown to be an important feature of algal response to environmental stress by mediating important cellular responses, which eventually determine the fate of cells. ROS take part in the response of diatoms to iron limitation (Thamatrakoln et al., 2012), in dinoflagellate response to CO₂ limitation and to mediate allelopathic interactions with toxic cyanobacteria (Vardi et al., 1999; Sukenik et al., 2002; Vardi et al., 2002), in response to UV radiation and photoacclimation (Rijstenbil, 2002; Janknegt et al., 2008; Li et al., 2009; Fischer et al., 2012), and as a secondary burst during oxidative stress and senescence (Vardi et al., 2007; Murik and Kaplan, 2009; Murik et al., 2014). We suggest that during the host–virus interactions, EhV subverts the cellular antioxidant and ROS metabolism for its benefit, to ensure an optimal replication cycle. Taken together, the results presented here clearly imply a pivotal role for ROS metabolism during this host–virus interaction. ROS can mediate the cross-talk between the main cellular pathways that are triggered during EhV infection: autophagy (Schatz et al., 2014), sphingolipid metabolism (Vardi et al., 2009) and programmed cell death (Bidle et al., 2007), and therefore will determine the fate of the viral-infected blooms and influence the cycling of nutrients and carbon within microbial food webs in the marine ecosystem. We propose using this unique set of cellular markers as a novel tool to assess viral infection at different phases of algal blooms in the ocean.

Acknowledgements

We thank Daniella Schatz for fruitful discussions and commenting on the manuscript. We thank Ester Feldmesser for assistance with data analysis. This research was supported by the European Research Council (ERC) StG (INFOTROPHIC grant number 280991), the Israeli Science Foundation (ISF) Legacy Heritage fund (grant number 1716/09) and the generous support of Edith and Nathan Goldenberg Career Development Chair both awarded to AV.

Competing interests

The authors declare no conflict of interest.

Supplementary Information accompanies this paper on The ISME Journal website

Supplementary information The online version of this article (doi:10.1038/ismej.2015.228) contains supplementary material, which is available to authorized users.