1. Introduction
Marine phycovirology, i.e., the study of viruses infecting marine eukaryotic algae, started with the lytic viruses infectious to the picophytoplankter
Micromonas pusilla [
1,
2,
3,
4,
5]. The genus
Micromonas (class Mamiellophyceae) is ubiquitous, occurring from tropical to polar regions, and is readily infected by viruses [
3,
6,
7,
8,
9]. The majority of
Micromonas virus isolates belong to the double-stranded DNA (dsDNA) prasinoviruses [
3,
4,
5,
9], although a dsRNA
Micromonas virus has also been reported [
10,
11]. The prasinoviruses are considered the most abundant group of marine phycodnaviruses [
12] and virus abundances show synchrony with their hosts’ temporal dynamics consistent with infection [
13,
14].
Micromonas is a globally important prasinophyte, which typically dominates the picophytoplankton fraction in marine Arctic waters [
15,
16,
17,
18,
19,
20,
21,
22]. Previous studies have shown that Arctic
Micromonas forms a separate ecotype from lower latitude strains [
16,
21] adapted to grow at temperatures between 0 and 12 °C (with an optimum around 6–8 °C [
16]). Considering Arctic sea surface temperature over the year to be in the range of −1 to a maximum 7 °C [
23,
24,
25] and steadily increasing as a result of global warming (0.03–0.05 °C per year over the 21st century [
24]), the
Micromonas polar ecotype species (tentatively named
M. polaris; [
26]) can be expected to belong to the picophytoplankton predicted to benefit from a warming Arctic region [
24,
27,
28,
29]. Despite this predicted increase in abundance and relative share of picophytoplankton in the changing Arctic Ocean, it is still unclear how the viruses infecting the picophytoplankton are affected by changes in temperature. Little is known about Arctic phycoviruses in general, and to our knowledge, no viruses infectious to Arctic
Micromonas species have yet been brought into culture [
30,
31,
32].
Changes in an environmental variable, such as temperature, may directly affect virus infectivity and/or more indirectly impact virus proliferation due to alterations in the metabolic activity of the host [
33]. Thus far the thermal stability of psychrophilic marine virus-host interactions has only been assessed for several phage-bacterium systems [
34,
35], despite the potential for special physiological adaptations by cold-adapted hosts and viruses [
36,
37]. It is likely that different viruses infecting the same host strain have distinct responses to shifting environmental factors and therefore environmental change may drive virus selection and host population dynamics. Nagasaki and Yamaguchi [
38] found that the temperature ranges for successful infection were different for two virus strains infecting the raphidophyte
Heterosigma akashiwo and that the host strain sensitivity to infection varied according to the temperature. Furthermore, temperature regulates growth by controlling cellular metabolic activity [
39], which has been proportionally related to latent period length and burst sizes for
Vibrio natriegens phages [
40]. Recently, Demeroy and colleagues [
41] demonstrated that temperature-regulated growth rates of
Micromonas strains that originated from the English Channel were responsible for shortened latent periods and increased viral burst sizes upon infection. Ongoing change in the Arctic necessitates a better understanding of how Arctic phycoviruses are affected by temperature.
Here we report on the isolation of four
Micromonas viruses from the Arctic. In addition to determining their viral characteristics (capsid morphology and size, genome type and size, latent period, phylogeny, host range, burst size, virion inactivation upon chloroform and freezing treatment), we investigated the impact of temperature change on virus infectivity and production. We hypothesize that (i) viral infectivity will increase with temperature, and (ii) increasing temperatures will stimulate virus production (shorter latent periods and higher burst sizes). For testing the latter hypothesis, we performed one-step virus growth experiments at a range of temperatures representative of the extremes over the polar growth season (0.5–7 °C) [
23].
4. Discussion
To our knowledge this is the first report of the isolation and characterization of phycoviruses from polar marine waters. Similar to other reports of
Micromonas virus isolates, and consistent with the
Phycodnaviridae, the virus-like particles accumulated in the cytoplasm of the host cells [
9]. The particles of the four MpoV were morphologically similar (no significant variance in virus particle size, i.e., on average 120 nm), and all contained a lipid membrane (sensitive to chloroform). Lipid-containing MpoVs were first reported by Martínez Martínez and colleagues [
9]. These authors were able to clearly and convincingly divide nineteen newly isolated
Micromonas viruses, across an area spanning the North Sea to the Mediterranean Sea, into two groups based on (i) their sensitivity to infection of LAC38 or CCMP1545, (ii) genome size (206 ± 6 Kb (
n = 12) vs. 191 ± 4 Kb (
n = 8)), and (iii) presence of a lipid membrane. Strikingly, all LAC38-infecting viruses with larger genomes contained a lipid membrane, whereas the smaller genome sized CCMP1545-infecting ones did not. Our data show a similar larger genome size for MpoV-44T which infects LAC38 (205 ± 2 Kb, in contrast to the 191 ± 3 Kb genomes of the other MpoVs), but in our case all MpoVs contained a lipid membrane.
Molecular phylogeny inferred from the amino acid sequences of the DNA polymerase gene B fragments established that the four Arctic MpoV isolates grouped distinctly with the other dsDNA
Micromonas viruses belonging to the genus Prasinovirus. The genus Prasinovirus infects
Ostreococccus and
Micromonas species and is one of the six virus genera belonging to the
Phycodnaviridae family; eukaryotic algal viruses with large dsDNA genomes (100–560 Kbp) [
65,
66]. A recent metagenomic survey (Tara Ocean Expedition) revealed that prasinoviruses are the most abundant group of phycodnaviruses in the oceans [
12]. The newly isolated MpoVs did not group together; instead MpoV-44T and 46T were phylogenetically distinct, both from each other and from MpoV-45T and 47T. Furthermore, it is clear from the phylogenic analysis based on
polB that the Arctic
Micromonas viruses do not form a separate cluster. Screening the Tara Oceans’ contigs and the KEGG Environmental database for homologs of the amplified nucleotide region of our MpoV isolates revealed a worldwide distribution and high diversity on thermal stability (i.e., from the Greenland Sea to the temperate regions to Antarctica and in waters from −1.6 to 17.3 °C,
Table S3). These results confirm that the
Micromonas viruses and their relatives are globally dispersed, show a high degree of genotypic diversity, and are ecologically relevant.
There was no general relationship between the phylogenies of the virus and host strains as revealed by Martínez Martínez et al. [
9] for viruses infecting temperate
Micromonas strains, but MpoV-44T could be distinguished from the other Arctic MpoV isolates based on its capability to virally infect
M. commoda LAC38. Although LAC38 was being cultured at low temperature (3 °C) at the time of virus isolation, it is originally a temperate
Micromonoas strain (Baltic Sea) [
67]. We cannot exclude that the isolation of MpoV-44T on a different host is underlying its intrinsic differences to the other MpoVs isolated 8–9 years later using a local Arctic
Micromonas host strain. These differences may also be due to MpoV-44T having been isolated during midwinter and years before the other MpoVs (isolated in summertime during two consecutive years). Successional patterns for marine virus communities with associations to temperature and host dynamics have been demonstrated (e.g., [
68]). Even though Arctic
Micromonas still grows well at low temperatures (0.4 d
−1 at 0.5 °C, this study) and low light (0.2 d
−1, [
16]), photosynthesis may not be possible during part of the Arctic winter. Several
Micromonas species however exhibit phagotrophy (e.g.,
M. polaris CCMP2099; [
69]) that could serve as an alternative energy source to maintain growth and/or virus production during the winter (see also [
22]). Yet, our study shows that MpoV-44T is well adapted to relatively fast and high production of infective progeny at relatively higher temperatures, which makes it more likely that advection of relatively warm Atlantic water from the West Spitsbergen Current (WSC) [
70] was responsible for being able to isolate MpoV-44T in winter. Water temperature in autumn of the year of isolation was in fact higher than the average of the preceding years [
71]. Additionally, the ability of MpoV-44T to successfully infect
Micromonas strains growing at higher temperatures up to 20 °C seems indicative that this specific virus has a high temperature tolerance. Still, a relatively high temperature optimum for a virus occurring in a cold environment could theoretically be an adaptation to be less virulent in order to avoid extinction of the host [
14,
72]. The relatively long latent periods and reduced infectivity of MpoV-44T at low temperatures would effectuate such low virulence for the slow growing hosts during the winter months. Then in the following more productive season, when the host growth rates increase, the latent period of MpoV-44T shortens and burst sizes increase to be able to keep in sync with host growth. Intriguingly, MpoV-47T also displayed thermostability, with an infectivity optimum at 7 °C. MpoV-47T was isolated only 2 months later than the temperature sensitive MpoV-45T (infectivity optimum at 0 °C), indicative of a high degree of diversity of virus thermostability in Arctic waters.
At temperatures above zero, the infectivity data do not confirm our first hypothesis that MpoV infectivity increases with temperature. The viral response to higher temperatures (0–7 °C) was highly variable and strain-specific. Two of the four virus isolates (MpoV-45 and 46T) even showed a loss of infectivity above 0 °C. All of the virus isolates, except for MpoV-44T, were able to maintain over 25% of their infectivity after being frozen at –20 °C. This suggests that these viruses are well able to withstand the freezing process during ice formation, a property which would maintain high titers during winter. Similar findings have been reported for Arctic marine bacteriophages [
73] and dsDNA algal viruses during winter in a seasonally frozen pond [
74]. Cottrell and Suttle [
5] reported relatively high decay rates for MpV-SP1, however, this virus strain originated from subtropical waters and the decay rates were largely determined by sunlight (UV intensity). There is limited knowledge of dsDNA algal virus thermal stability and the mechanisms underlying the loss of infectivity have not been elucidated [
9,
41,
53,
75,
76,
77]. Only a few cold-active viruses (i.e., viruses that successfully infect hosts at 4 °C or below) have been brought into culture and all are bacteriophages [
35,
78,
79]. Variability in MpoVs’ temperature sensitivity demonstrates a specificity of infection efficiency related to temperature. Hypothesizing that our data are generally applicable, seasonal temperature shifts could regulate
Micromonas host and virus succession. The infectivity loss at higher temperature (7 °C) for MpoV-45T and MpoV-46T shortens their window of optimal activity during the warmer summer months. Considering that the range of temperatures we tested are ecologically relevant for the Artic seas (water temperatures between –1 and 7 °C; [
23]), our results indicate the need to determine the causal processes such as the means of virus entry and conformational changes in the virus particle (e.g., viral capsid proteins and lipid membrane properties).
When propagated on the putative
M. polaris strain TX-01, MpoV-44T displayed a much longer median latent period than the other Arctic MpoVs (39 vs. 18 h, respectively). A comparably long latent period had, until recently, only been described for the dsRNA virus MpRV infecting
M. commoda LAC38 (36 h; [
10]). However, Baudoux and colleagues [
14] reported a latent period of 27–31 h for a dsDNA virus infecting
Micromonas isolates from the English Channel growing at 20 °C. The latent period of MpoV-44T displayed a strong temperature-dependence, i.e., with a temperature increase of 4 °C, the time of viral release decreased by >15 h (latent period 12–18 h; approximately 50% of the latent period at 3 °C). Although Baudoux et al. [
14] did not find a correlation between differences in latent period (or burst size) with the host growth rates for the isolated MicVs, Demory et al. [
41] reported for the virus-host model system Mic-B/MicV-B (virus infecting largely Clade B strains) an inverse relationship of the viral latent period with the host growth rate (whereby the growth rates were affected by the host culture temperature). We did not find such a relationship with growth rate for the latent periods of MpoV-44T growing on TX-01, but did find a similar significant linear relationship when infecting host RCC2258 (r
2 = 0.952,
p = 0.018). Virus MpoV-45T did not show a dependency on host growth rate (TX-01, RCC2258, and RCC2257), but instead showed a shortened latent period on host RCC2257 at the highest temperature (7 °C compared to 3 °C). Furthermore, the latent period of MpoV-44T was strongly affected by temperature whereas increasing temperature only shortened the MpoV-45T latent period on RCC2257. On host TX-01, the latent period was unaffected over the whole range of 0.5 to 7 °C, however, we found a steeper increase of viruses with increasing temperature, i.e., virus production rates of 0.18, 0.37, 1.4, and 1.6 × 10
5 viruses h
−1 between 16 and 30 h for 0.5, 2.5, 3.5, and 7.0 °C, respectively. While the data confirm our second hypothesis that the temperature increase stimulates MpoV production (through shortened latent periods, enhanced production rate and/or higher burst sizes), there is nonetheless a virus-specific response for the range of temperatures tested. Instead we found a high variability in response for the different virus isolates and host strains. When looking at a temperature increase from 3 to 7 °C, most virus-host combinations in our study showed enhanced viral burst sizes with a temperature-regulated increase in the host growth rates (0.53–0.59 d
−1 at 3 °C and approximately 1.2-fold higher at 7 °C). However, MpoV-45T infecting TX-01 did not show this increase from 3 to 7 °C, but did exhibit an increasing burst size with temperatures from 0.5 up to 3.5 °C (growth rate TX-01 increased from 0.40 to 0.66 d
−1, respectively). Hence, the optimum temperature for virus production was not the same as the optimum host growth temperature. Wells and Deming [
78] showed a similar situation in which phage 9A, infecting the psychrophile
Colwellia psychrerythraea strain 34H, had a burst size optimum at −1 °C while the host’s growth rate optimum was at 8 °C. These authors suggested that specific virally encoded enzymes have their own optimum temperatures.
Temperature strongly regulates Arctic
Micromonas growth rates with increasing growth rates up to 7 °C (this study; [
16,
29]). These temperatures are at or above the present summer sea surface temperatures in the lower latitude regions of the Arctic [
23,
25]. At the time that TX-01 was isolated (half April), in situ picophotoeukaryotic gross growth rates were 0.58 d
−1 at temperatures between 1 and 2 °C (Maat and Brussaard, unpublished data). By the end of May, the temperatures and consequently the growth rates had increased to 2–3 °C and 1.1 d
−1, respectively. This 20–50% growth rate increase is similar to TX-01 in the present study over the same temperature range. Our results indicate that over an Arctic growing season with increasing temperatures and host growth rates, viral activity can be expected to increase as a result of the decreasing latent periods and increasing burst sizes. Our data imply that temperature could affect host and virus diversity (strain dynamics), as for MpoV-45T with host TX-01 the latent period and viral burst size did not further change above 3 °C, but with other hosts (RCC2257 and 2258) and other viruses (MpoV-44T) the latent periods shortened and/or burst sizes increased to several extents. The different susceptibilities of viral infectivity to temperature, with a tolerance for the highest temperatures for MpoV-44T, strengthened this idea even more. Tarutani et al. [
80] showed how within the observed abundances of
Heterosigma akashiwo and the lytic virus HaV, a successional shift in clonal composition occurred due to differences in susceptibility/resistance of the host to the viruses. Based on our data a similar shift in strain and clonal composition of
Micromonas and MpoV may occur, not only due to differences in susceptibility to the viruses but also because of differences in virus proliferation success at different temperatures [
33].
In summary, the present study describes the first isolation and characterization of viruses infecting a cold-adapted polar phytoplankter. The relevance seems high, as it concerns the ubiquitous genus
Micromonas which belongs to the picophytoplankton fraction and is expected to be favored under future Arctic conditions (due to warming and freshening induced vertical stratification; [
24,
27,
28,
29]). The Arctic region is warming to a greater extent than lower latitudinal marine waters [
23,
24] and current summer sea surface temperatures (August 2016) as high as 5 °C above the 1982–2010 mean [
25] have been observed.
Micromonas growth rates will enhance faster and earlier in the season and our study indicates that viral production will likely do the same. We show variable infection dynamics in response to temperature for the different virus-host strain systems examined, which complicates the assessment of the environmental relevance of each isolate. However, we do like to advocate that virus (and host) isolation, characterization of virus-host dynamics, and responses to changing ecologically relevant environmental factors are fundamentally essential to understanding the role of algal viruses in (Arctic) marine waters. The newly isolated viruses make it possible to comprehensively investigate the interactions of these unique virus-host combinations under climate change relevant environmental variables. Joli et al. [
22] showed the importance of Arctic
Micromonas viruses by metagenome sequencing. It would be interesting to investigate the ecological relevance of the strains tested in our study using molecular approaches. In the natural environment, selective effects of temperature may drive (intra)species diversity, potentially affecting the ability of
Micromonas to respond to the changing conditions of the vulnerable Arctic. Modeling studies could help to comprehend (and predict) the extent to which the Arctic phytoplankton community would be influenced by changes in infection dynamics associated with temperature changes.