Content uploaded by Frederic Briand
Author content
All content in this area was uploaded by Frederic Briand on Jan 14, 2023
Content may be subject to copyright.
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
5 CIESM Workshop Monographs n°21
I - EXECUTIVE SUMMARY
This synthesis, initiated during the meeting, was consolidated thereafter by inputs received from the
participants.
1. INTRODUCTION
The workshop was held at the historical Laboratoire Arago, Observatoire Océanologique de
Banyuls-sur-mer, from 19 to 22 March 2003. Twelve scientists from nine different countries (see
list at the end of volume) attended the meeting at the invitation of CIESM.
After welcoming remarks from Gilles Boeuf, Director of the Observatory, the meeting was
opened by Frédéric Briand, Director General of CIESM, and Gerhard Herndl, Chair of CIESM
Committee on Marine Microbiology and coordinator of the workshop, who briefly presented the
context, background and objectives of this event.
1.1. Background and objectives
The initial aim of the workshop was to stimulate research on this fast-emerging subject in the Medi-
terranean Sea. The discovery that viruses are far more abundant – by three orders of magnitude (!)
– in the marine environment than previously assumed is only ten years old. By now it is also well
established that viruses are an active component of the microbial food web. As viruses may
influence microbial species composition and regulate the abundance of specific species, they
probably represent one of the driving forces of microbial successions. Therefore they directly
influence nutrient regeneration and carbon cycling through the microbial communities.
Moreover, viruses may be responsible for the occurrence of specific diseases in marine plants and
animals, particularly if specific species are present in high densities such as in mariculture.
The Mediterranean Sea, with its large range of different subsystems ranging from dystrophic
lagoons to oligotrophic deep waters, provides a unique opportunity to explore the function and
significance of viruses under contrasting environmental conditions. Through a series of presen-
tations and in-depth discussions, the workshop reviewed existing knowledge on marine viruses
and explored promising paths for future investigations, taking the specific conditions of the
Mediterranean Sea into account. After the first two days of presentations, two parallel sessions
were organized to facilitate exchanges and brainstorming. The first group focused on viral abundance
and virus-mediated mortality, the second on viral diversity. The final session included a general
discussion where both groups presented their findings. Their main conclusions are reported in this
executive summary.
2. VIRAL ABUNDANCE, PRODUCTION AND VIRUS-MEDIATED MORTALITY.
Section prepared by R. Noble, C. Brussaard, R. Danovaro, P. Lebaron, D. Mestivier,
and T.F. Thingstad
Viruses are now known to be ubiquitous, numerically dominant members of aquatic microbial
communities. Because of the nature of viruses as obligate parasites, viral infection commandeers
an important role in structuring marine microbial communities, making them drivers of nutrient
flux, elements in the global change puzzlework, and ferries of genetic information. The conse-
quences of viral infection can be studied on a wide range of scales, from global to micron-size
scales of microbial interaction, and from time scales ranging from thousands of years to a few
seconds. Viral infection affects the dynamics of “microbial loops”, and has demonstrated impacts
(sometimes dramatic) on both bacterial and phytoplankton populations, yet our understanding of
5 / 94
the consequence and complexity of these interactions is still limited. In particular, development
of an understanding of the importance of viruses to issues such as nutrient flux will require
improvement of current methods to assess viral abundance, rates of virus production, and virus-
mediated microbial mortality. All of these developments will necessitate further understanding of
more “static” virus parameters such as viral diversity, host specificity, lysogeny vs. lytic virus
types, and host range.
Because the field of viral ecology is multi-scalar, we choose to present issues in viral ecology,
and needs for future advancement of the field from a global - ecosystem - community - population
- genotype perspective. We consider the Mediterranean Sea as a potential example for an excellent
study system. With its anti-estuarine circulation, hydrographic mesoscale features with permanent
cyclonic and anti-cyclonic gyres, general west-east trophic gradient (increasingly oligotrophic
moving east), deep light penetration, seasonal and spatial shifts in pelagic food web structure,
postulated phosphorus limitation, and winter deep-water formation, the Mediterranean is a perfect
natural laboratory for the study of the interactions between microbial ecology and ocean bio-
geochemistry. Viruses are thought to influence microbial diversity, food web structure, and the
partitioning of material fluxes between dissolved and particulate forms, processes believed to be
modified by environmental factors such as UV light, temperature, and nutrient availability.
Studies of these processes in the Mediterranean thus have the potential, not only to improve our
understanding of the Mediterranean ecosystem per se, but to provide new generic knowledge on
these processes in general.
2.1. Importance of understanding large scale issues
The theme of temporal and spatial dynamics of viral communities is one that contains both a
descriptive and an analytical aspect. There is still a lack of knowledge as to how abundance,
diversity and activity of viruses vary over long-time and large-spatial scales in different aquatic
environments, the Mediterranean Sea included. Over long time scales, the aim of such studies
should be to understand the underlying mechanisms that control abundance, diversity and activ-
ity of viruses in natural ecosystems, especially in the light of the possible climate change sce-
narios. At present, no commonly accepted theory exists for this. Analytically, an understanding
of temporal and spatial dynamics will improve our ability to accurately place viruses into the con-
text of microbial food webs, and will allow to create models of their effects on nutrient cycling,
food web dynamics and biodiversity.
There are several environmental factors: UV-light, particulate organic matter (POM), dissolved
organic matter (DOM), turbulence, and temperature, for example, that have known or suspected
direct effects on viral processes such as decay, latent period, host encounter, and burst size. These
direct effects are important subjects of study, and specific mechanisms such as UV-effects on
viral decay may be particularly relevant in the high intensity solar radiation common to the
Mediterranean Sea environment.
It is generally believed that viruses interfere and interact with the rest of the ecosystem in a num-
ber of ways: increasing the diversity of host populations by preventing permanent dominance of
otherwise successful species and toxic species, diverting the flux of energy and matter from the
particulate part of the food web into the DOM-pool, and by increasing the rate of horizontal gene
transfer between hosts.
However, there is a difficult but crucial step in translating knowledge of direct effects into an
understanding of indirect effects; or how the direct effects spread via trophic interactions to other
levels in the food web. In other words, there is an actual need for field studies aimed at verifying
how isolated processes observed in laboratory experiments behave in the context of natural
ecosystems. The lack of any general theory of viral ecology makes this translation from direct to
indirect effects particularly difficult.
To understand these relationships, data and descriptions of whole systems are needed and com-
parisons of systems should be encouraged. The ultimate goal will be to explain the properties of
different systems from common generic principles. Mathematical ecosystem models will be a
useful tool in this regard.
CIESM Workshop Monographs n°21 6
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
6 / 94
2.2. Viral induced mortality and its impact on biogeochemical cycling
The main biogeochemical function of viruses is thought to be their function as catalysers for the
transformation of particulate (POM) into dissolved (DOM) organic matter as the content of the
host organisms is released during cell lysis.
Whether microorganisms are grazed upon, or die due to viral-induced cell lysis has major implica-
tions for the flow of material and energy cycling in marine pelagic food webs. DOM-release will
force the photic zone food web towards a more regenerative system. With more of the respiration
occurring at the level of small organisms, the expectation on an intermediate time scale (see
below) would be a reduction in the yield of larger organisms in the ecosystem.
To comprehend the microbial compartment of the Mediterranean pelagic ecosystem, an under-
standing of the processes retaining bioavailable forms of nitrogen (N) and phosphorus (P) within
the photic zone as the surface water is moved eastwards from Gibraltar is essential. Again, the
lack of a general food web theory makes it difficult to analyze the effects that viruses may have.
Based on the steady state argument that viral lysis increases loss rates of the organisms present,
these organisms also have to grow faster in the presence of viruses. Faster growth requires more
food, with the consequence that more of the available total nutrients would have to be distributed
into the smaller-sized organisms serving as food for larger predators. If such reasoning is correct,
viral lysis should promote food chains dominated by small-celled organisms, producing little
export and thus retaining more nutrients in the photic zone. With this expected improved nutri-
ent retention, the long term net effect of viral lysis on ecosystem yield at higher trophic levels is
not trivial. Ahigher viral abundance in eutrophic systems (typically characterised by large-celled
organisms) has been demonstrated, but adequate information on viral infections and production
in oligotrophic (long food chain) versus eutrophic (short food chain) systems is still lacking. As
a result the differences in the role of viruses in aquatic trophodynamics and biogeochemical
processes remain largely unknown.
The composition and fate of released DOM need to be examined in more detail. Viral lysis does
not only produce a selective top-down mechanism thought to affect microbial diversity, but also
potentially a bottom-up effect by alteration of the substrate spectrum for the osmotrophs (phyto-
plankton and heterotrophic bacteria). Comparison of the impact of the released carbon and nutri-
ents on the biodiversity and function of pelagic food webs with different trophic status is still
lacking.
Diversion of the flux of energy and material from particles to DOM also implies a diversion from
potentially sinking into non-sinking forms. However, bacteria attached to sinking particles are
also believed to dissolve the particles through enzymatic activities. Lysis of these bacteria may
thus also have a reducing effect on particle dissolution. The biogeochemical distribution of ele-
ments and the efficiency of the biological carbon pump depend upon the depths at which carbon
C, N, and P are transferred from sinking to non-sinking forms. In the Mediterranean circulation
pattern – with the west-flowing deeper currents of the Levantine Intermediate Water and
Mediterranean Deep Water – the biogeochemical distribution of C, N and P in the water column
is presumably sensitive to the depths at which these elements are released into non-sinking forms.
Since the depth of release is a combination of release rate and sinking velocity of the particles,
the net result of these processes on Mediterranean biogeochemistry is difficult to predict (Fig.1).
Viral lysis of coccolithophoride phytoplankton species has previously been described. One may
speculate that viral lysis, presumably releasing the slow-sinking coccoliths within the photic
zone, versus grazing by macrozooplankton that may pack the coccoliths into rapidly sinking fecal
pellets, may have very different consequences for the Ca-distribution in the sea, and thus for the
so-called alkalinity pump. Dissolution of the coccolith CaCO3increases the alkalinity. If dissolu-
tion occurs in surface waters, rapid re-sequestration into the ocean of the CO2released when the
CaCO3was formed is thus possible, and the potential large negative effect of coccolithophoride
blooms on the ocean’s ability to sequester atmospheric CO2is neutralized. The consequences for
our ability to understand the feedback between global change and ocean biogeochemistry are
large.
Several studies showed that viral-induced cell lysis of phytoplankton represents a mechanism of
DMSP / DMS release in the marine environment, suggesting that viruses may play a possible role
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
7 CIESM Workshop Monographs n°21
7 / 94
in climate regulation. The release of DMS / DMSP from lysed phytoplankton cells is of signifi-
cant importance during blooming events such as typically found in eutrophic coastal areas. In
more oligotrophic systems, the effects of anthropogenic impacts (PCBs, eutrophication, environ-
mental stressors and micropollutants) on the significance of viral-induced mortality (induction
and lysogeny, algal bloom dynamics) may be of special interest, especially for the Mediterranean
Sea with its high degree of human activities and ever increasing coastal development. For a better
understanding of the quantitative importance of viral infection on the pelagic food web, there is
strong need for methods able to specifically detect viral abundances and lysis rates of bacteria
and of phytoplankton, separately. Beside these two obvious groups of organisms, viral lysis of
other members of the microbial food web, such as pico-eukaryotes and protozoans, can also be
expected to contribute significantly to the flux of matter and energy. At this time, our under-
standing of the relative importance of viral lysis of specific groups is poor. The development of
accurate methods will allow us to also estimate the magnitude of community specific, or possibly
even species specific losses, a necessity for our research field. Virus-host systems based on key
planktonic species should be used to further optimize our knowledge of actual lysis rates.
Mortality of microbial populations due to viral lysis has further implications for substrate avail-
ability and development of host infection resistance modes. Several recent studies have demon-
strated the “usefulness” of viral infection products for successful growth by heterotrophic
CIESM Workshop Monographs n°21 8
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
Fig. 1. Potential effects of viruses on Mediterranean biogeochemistry. Background figure shows isopleths of
phosphate concentration (µM-P) in the Mediterranean and Black Seas (modified from Redfield et al., 1963). If
transfer of matter from particulate to dissolved forms via viral lysis is a quantitatively important process, viral
lysis may affect such biogeochemical distributions through different mechanisms:
1. The east-to-west oligotrophication of the surface waters is potentially influenced by viral lysis affecting
nutrient retention in the photic zone, and thus structural changes in the pelagic food web.
2. Viruses potentially affect a differential loss of N and P to non-available forms of DON and DOP, thereby
possibly playing a role in shifting the Mediterranean system towards P-limitation.
3. Viral lysis of sinking organisms or organisms attached to sinking particles will release C, N and P from
sinking to non-sinking forms.
4. Sinking to non-sinking transfer of matter may differ in the different basins due to differences in the photic
zone food web structures, producing quantitative as well as qualitative differences in the vertical fluxes of
POM.
5. Biogeochemical distributions in the Mediterranean and export of C, N, and P through Gibraltar will
depend upon whether the sinking to non-sinking transfer occurs in the Levantine Intermediate Water or in
the deep water.
6. DOC, DON, DOP release from POM in sediments may promote burial of C, N and P.
8 / 94
bacterioplankton. The availability of the lysis-derived organic material for assimilation and further
degradation has many biogeochemical implications. As in other areas, dissolved organic carbon
(DOC) accumulates in Mediterranean surface waters in the stratified season. Does this occur
because viral lysis produces recalcitrant organic molecules, because there are photochemical
processes transforming degradable into recalcitrant chemical forms, or because of accumulation
of degradable material as a consequence of P-limitation of heterotrophic bacteria? In any case,
the accumulation will increase the C-sequestration capacity of the sea. With a large fraction of
total N and P in the dissolved organic nitrogen (DON) and dissolved organic phosphorus (DOP)
pools, one could also speculate whether a small difference in the loss rates of N and P to recalci-
trant forms of DON and DOP may play a part in the so far unexplained shift towards P-limitation
in the Mediterranean Sea. The process of viral lysis in sediments is also speculated to release host
DNA into an environment with low DNA degradability. If so, viral lysis could be a process influ-
encing P cycling in the ocean’s sediment.
Pilot studies indicate that the viral lysate can affect the susceptibility of the host population for
further virus infections, something that warrants more dedicated research. The issue of sensitive
versus resistant host populations has just begun to be explored. Techniques that enable the
researcher to examine these processes in situ are vital for an accurate understanding of this
process. Another intriguing issue is the evolutionary implication of viral lysates and viral infec-
tion events, as it has been hypothesized that repackaging of cellular DNA by viruses contributes
to wide-scale spread of specific genes and that viral lysis from dominant species yields DNA
release for potential gene transfer to other microorganisms.
2.3. What are the controlling factors for microbial loop components ?
Based on traditional Lotka-Volterra formulations, one can construct simple idealised models for
the microbial part of the food web. In these, size-selective grazing will allow steady states with
coexistence of size groups of osmotrophs (phytoplankton and heterotrophic bacteria) even when
all of these compete for a common limiting substrate such as e.g. phosphate. Including size-selec-
tive predation is thus one simple way to resolve the classical Hutchinson’s paradox. Adding host
specific viruses to this description will in an analogue manner allow coexistence of host-groups
inside each size-group of osmotroph competitors. This theory has been argued to allow for more
microbial diversity than presently suggested for bacteria by methods such as denaturing gradient
gel electrophoresis (DGGE). In a sense, this reverses Hutchinson’s classical paradox to the ques-
tion of “why do we seem to observe less diversity than can be explained by simple steady state
models?”. The above principle of selective loss mechanisms “killing the winner” is, however not
readily extended to viruses since selective loss mechanisms for viruses are not presently known.
A theory has been proposed where viral lysis is the mechanism that compensates for differences
in growth rate between coexisting (dominant) bacteria, leading to the conclusion that viral abun-
dance is linked to the magnitude of physiological diversity in the bacterial community.
Interestingly, such a theory links biodiversity directly to the pattern of biogeochemical cycling of
elements. This theory seems, however, able only to explain the stable existence of one virus per
“host group”. With indications that there may be many (coexisting) viruses attacking the same
host, Hutchinson’s paradox may seem to return, but this time at the level of viruses: how can sev-
eral viruses coexist on one host population?
In the above description, diversity in the sense of number of simultaneously coexisting “species”
becomes a phenomenon that is “top-down” controlled. It does, however, not give any clues as to
the organisms that will occupy these niches. This, to a much larger extent, is presumably a ques-
tion of the relative competitive abilities among the potentially coexisting hosts. With this in mind,
predation (top-down), viral lysis (side-in), and nutrient competition (bottom-up) control different
aspects of host diversity. There is a large need for relevant experimental model systems that can
be used to elucidate these different modes of control.
An interesting aspect of the community of coexisting hosts is whether the individual species
remain at a (more or less) stable abundance, whether the individual host-virus system exhibit
Lotka-Volterra type oscillations, or whether there are spontaneously occurring oscillations that
collectively determine the stability of the system.
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
9 CIESM Workshop Monographs n°21
9 / 94
2.4. Impact of viruses at the population level
Viruses exert significant impacts on the structure of microbial populations in aquatic environ-
ments. Because of their relative specificity, viruses are capable of altering microbial community
composition over short time scales. Since viral infection is thought to be driven largely by
encounter rates, this process causes immediate and direct effects shifts in community composition
by “killing the winner” where the success of a particular microbial species can quickly translate
into death by viral infection. Direct effects of community composition shifts include the obvious
removal of specific microbial species or groups, which in the case of diverse aquatic environ-
ments can be a complex issue to resolve. Indirect effects of viral infection, such as alterations in
substrate availability, promotion of resistance strains of host cells, and shifts in competition
among microbes for available substrates, are even more complex and thus far poorly studied.
Advancements in methods to assess these effects (both direct and indirect) in natural systems are
in dire need, as well as the use of relevant model virus-host systems to generate input for deter-
ministic and probabilistic models.
There have been several important factors in virus control that have been speculated in recent
publications. The first example is the development of host resistance to viral infection. It has been
demonstrated in several studies that host cells are capable of developing resistance to viral infec-
tion over short time scales, thereby preventing decimation of the population. From experimental
model systems, it is known that resistance in bacterial hosts often has a price in the form of com-
promised competitive abilities of the host. The bacteria-phage “war” is thus not only a fascinating
subject of study from the perspective of the microbial food web dynamics, but also is an ideal
system for the study of evolutionary processes within an ecosystem context. The processes of
development of resistance and its ecosystem consequences are not well understood, and this area
deserves more attention.
Viral impacts on populations have been demonstrated under bloom conditions, where viruses have
the potential to cause immediate termination of bloom species. We know little about the factors
and interactions of this process, the importance of host range, clonal variation within host species,
morphotypes within species, and relation to other parameters such as nutrients and grazing
impacts. It has been speculated that the range of clonal variation within host populations is impor-
tant to the success of viral infection, therefore clonal variation may play an important ecological
role in biodiversity. However, clonal variation within the virus populations infecting the same
host species, or even the same strain, suggests that the effect of viruses on biodiversity is more
complex than thought so far.
Although it goes almost without saying that it is important to study virus control of populations
in natural systems, we would like to stress that model host-virus systems are vital for an optimal
understanding of the mechanisms behind virus control of host populations. Relevant model sys-
tems can be used to study decay, infection rates, latent periods, lysis rates, clonal variation in host
populations, and variables that control success of infection, microscale studies on surface charge
changes of the host cells, interactions of receptor sites and nutrients, as well as the adsorption
mechanisms. Prochlorococcus, with its synchronized life cycle and occurrence almost as a mono-
culture in the stratified season in the eastern Mediterranean Sea, is an intriguing model system
for the study of virus ecology.
There are specific issues related to population level processes that deserve mention, mostly
because of our poor general understanding. For example, because of the sheer number of virus
host encounters in marine environments, and the known processes related to gene transfer
(repackaging of cellular DNA by viruses, transduction, conversion) by viral infection, it is
thought that viruses may be responsible in large part for microbial evolution. With an under-
standing of the rates of gene transfer in natural environments, it may be possible for us to assess
relative time scales of evolution, and thereby determine if they are shorter than time scales of
global change. We suggest that by studying the process of gene transfer from a rate perspective,
and by direct linkage to a better descriptive understanding of viral diversity, it will be possible to
assess the importance of gene transfer and its role in diversity.
CIESM Workshop Monographs n°21 10
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
10 / 94
2.5. Persistence of virus populations
There have been several studies conducted to assess viral decay and/or loss of infectivity in aquatic
environments. However, these experiments have largely been conducted either on specific virus-
host systems, or by using a whole system approach over very short time- scales. Most of these
studies have been executed only in laboratory microcosms, examining single alterations in a single
environmental variables. Although these experiments have yielded some useful information, they
are far too limited to provide an understanding of viral decay and inactivation in real-world sce-
narios. Field studies have provided conflicting results related to our current concepts of viral
decay, as they provide evidence for viral turnover rates either much higher or in some cases much
lower than those seen in laboratory studies. This suggests that there are complex interactions with
several environmental parameters that can strongly influence rates of viral decay. If we focus only
on single variables, we are excluding these types of interactions from our assessments of viral decay.
Additionally, it is widely thought that mechanisms of loss of viral infectivity are completely
uncoupled from those of viral degradation and this has not been addressed in aquatic environ-
ments.
In addition to environmental variables such as UV light, enzymes (nucleases/proteases), it is
important to consider the relevance of the mechanical damage, and the process of viral adsorp-
tion and their relations with removal of viruses from the system. Still very little is known on the
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
11 CIESM Workshop Monographs n°21
Box 1 COMMON METHODS IN VIRAL ECOLOGY
For the benefit of the reader, we have compiled a list of viral ecology methods
that are commonly employed in the field of aquatic viral ecology. Each method is
listed with one or two relevant bibliographic references (among others), so that
the reader can access further details of the methods. This list is not intended to
be exhaustive, but merely a summary of some of the most routinely methods cur-
rently used.
Virus enumeration
• Transmission electron microscopy counts (Børsheim et al., 1990, AEM)
• Epifluorescence microscopy (Yo-Pro 1: Hennes and Suttle, 1995;
SYBR Green I: Noble and Fuhrman, 1998)
• Flow cytometry (Marie et al., 1999; Brussaard et al., 2000)
• Most probably number assay of infectious units (Suttle, 1993)
• Plaque assays of infectious units (Suttle, 1993)
Viral lysis rates
• Frequency of visibly infected cells (Hennes and Simon, 1995;
Weinbauer et al., 1993)
• Production of radiolabeled, virioplankton sized biomass (Steward et al.,
1992)
• Use of fluorescently labelled viruses as tracers (Noble and Fuhrman,
2001).
For detailed protocols for the above approaches, see Noble and
Steward, 2001.
• Dilution approach (Wilhelm et al., 2002)
• L&H method (Evans et al,. 2003)
Viral diversity
• Pulsed Field Gel Electrophoresis (PFGE, Steward, 2001)
• Denaturing Gradient Gel Electrophoresis (DGGE, Short and Suttle,
2002; Schroeder et al., 2003
• Nucleic acid sequencing of viral genotypes (Lu et al., 2001)
Virus-host interactions, host range, viral typing
• DNA-DNA hybridization (Wichels et al., 2002)
• Flow cytometry (Brussaard et al., 2002)
• Use of fluorescently labelled viruses as probes (Hennes et al., 1995)
11 / 94
preservation of free viruses that are attached or entrapped into an aggregate matrix. There are
several mechanisms by which this is hypothesized to occur: viruses may be associated with
exopolymeric substances, aggregates, and marine snow, and subsequently removed from the sys-
tem, or can be incorporated into the aggregate and subsequently released into the environment
during aggregate degradation. It has been also suggested, based on preliminary experiments, that
successful viral infection of Heterosigma akashiwo increases the sinking rates of the infected
cells, causing removal of the specific, potentially damaging virus strains and nutrients from the
water column. Furthermore, the importance of viruses as actual food source is largely un-
explored, but it can be hypothesized as potentially important in P-limited environments.
On larger spatial scales, it is important for us to understand rates of viral decay and inactivation
in relation to the potential of viral transport throughout the world’s oceans. Distribution of virus
types in a global context is not only dependent upon the presence of viable hosts in sufficient
abundance, transport by mixing, and encounter rates, but is also directly related to rates of viral
persistence. Understanding viral decay rates in specific systems may permit us to model the trans-
port of “globetrotter” viral types, extending the theory of supply-site ecology in aquatic systems.
2.6. Reservoirs of viruses
Deep water masses and sediments can be thought of as reservoirs of viruses, able to segregate or
exchange virus assemblages with the overlying waters. Although investigations of benthic virus-
es are still at the scientific frontier’s edge, there is increasing evidence that surface sediments,
down to abyssal depths, can host very large virus numbers. Viruses also display a sharp decrease
with depth in the sediment, but large numbers still are present in deeper sediment layers. This
would suggest that viruses can be preserved in the sediment matrix and particularly in anaerobic
sediments. Practically no information is available on decay rates of viruses in the sediments. If
slow decay rates of benthic viruses are confirmed, this would provide a potential role of sedi-
ments as reservoirs for viruses. The overall importance of studying sediment viral and microbial
communities is also related to theoretical questions. For instance, the abundance of specific
groups of bacteria in marine sediments seems to be similar to those found in the water column.
However, total abundances of bacteria are very different between sediments and water column,
with more coexisting bacteria in the sediment. Why does this difference exist, and could this dif-
ference be explained by the difference in control of microbial communities exerted by protozoan
grazing and viral lysis?
What is still unclear is whether viral assemblages in the water column and sediments are similar
or not. If one assumes that benthic bacteria are different from pelagic bacteria, then their respec-
tive viruses should also be different. Preliminary electron microscopy evidence would suggest
rather different morphological features. Is virioplankton able to infect benthic bacteria, and/or is
viriobenthos able to infect virioplankton? If so the role of sediments as reservoir of viruses would
have major implications for bacterioplankton dynamics and would suggest that the benthic-
pelagic coupling of viral infections should be taken into account for modelling pelagic processes.
This would increase the importance of physical processes in promoting viral exchange between
different marine compartments. Sediment resuspension, slope currents, and upwelling could have
a major relevance from epidemiologic perspectives, and could contribute to the exchange of viral
diversity between sediments and the water column. Finally, as sediments present completely dif-
ferent environmental (physical, chemical, trophic) conditions from the water column, does this
implies that virioplankton and viriobenthos have different viral life strategies?
2.7. Role of viral ecology in cross disciplinary studies
Viral ecology is an important, yet often unaddressed component of interdisciplinary aquatic studies.
Not only could the field of viral ecology benefit from cross-disciplinary interaction, but other
fields could benefit from understanding the wide array of processes that viruses impact in aquatic
systems. There are several specific examples of the need for intertwining studies in the field of
viral ecology with other disciplines. There is a need for incorporating specialized studies of phage
and viral function, including assessment of receptor protein structure, attachment mechanisms,
induction of lysogeny in aquatic species. On a wider scale, there is also a need for involvement
of physical oceanography expertise in the field of viral ecology, for example to help elucidate
CIESM Workshop Monographs n°21 12
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
12 / 94
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
13 CIESM Workshop Monographs n°21
Box 2. PRIORITY RESEARCH QUESTIONS
• How do abundance, diversity and activity of viruses vary over time and
space in different aquatic environments? At present, no commonly
accepted theory exists explaining the underlying mechanisms that control
abundance, diversity and activity of viruses in natural ecosystems. It is
important to understand both the direct and indirect controlling factors.
•To what extent does viral lysis contribute to and maintain observed nutrient
gradients ?
• Does viral lysis significantly influence biogeochemical distribution,
especially in comparison to large scale processes like physical mixing of
water masses ?
• What is the impact of climate change on viral infection and related
biogeochemical processes ?
• What is the fate of viruses (decay, removal rate from pelagic to sediment,
inactivation), specifically in relation to temperature, UV light, salinity, and
pollutants ?
• What are the effects of anthropogenic impacts (PCBs, eutrophication,
environmental stressors and micropollutants) on viral-induced mortality of
microorganisms ?
• How well-preserved are viruses that are trapped in aggregates or
sediments ? Are the free viral communities similar to the trapped viral
communities ?
• Is viral infection the agent of control that is responsible for maintaining
existing microbial community structures ?
• Does viral lysis promote food chains dominated by small-celled organisms,
producing little export and thus retaining more nutrients in the photic
zone ?
• What is the impact of lysogeny in aquatic systems ?
• Are viruses a significant food source, and if so, under what conditions ?
• Considering that each host species and likely each host strain can be
infected by more virus types, how does viral infection regulate host
diversity ?
• How can several viruses coexist on one host population ?
• What factors mediate and control resistance of the host cell to virus
infection ?
• What are the evolutionary implications of viral infection events and viral
lysates (repackaging of cellular DNA by viruses contributes to wide scale
spread of specific genes. Viral lysis from dominant species yields DNA
release for potential gene transfer to other microorganisms).
• How important is viral infection to the process of gene transfer? At what
rate does gene transfer take place in natural environments ?
• Can we use mathematical ecosystem models that include a thoroughly
tested virus module, and are validated against actual data, to model
interactions of viruses in natural systems ?
• What is the ecological relevance of novel viruses (e.g. picophytoplankton)?
• What controls stability in virus-host interactions, e.g. why can some viruses
coexist stably with their hosts ?
13 / 94
processes of water parcel mixing and its role in transport of viruses. There is need for involve-
ment of chemical oceanography to help elucidate the composition of viral lysis induced released
cellular organic matter and how this affects the functioning and structure of the microbial pelagic
food web.
In applied microbiology, there is a need for applying currently used methods in the field of viral
ecology to the study of eutrophication and water quality and vice versa. For example, current
studies of eutrophication and bloom formation tend to focus on nutrient inputs and “top-down”
controls such as grazing, but few of them examine the virus-related control of communities.
Viruses are capable of exerting both direct and indirect impacts in these systems and these controls
need to be further evaluated. Developments in the field of viral ecology, and more specifically
related to the impact of viral infection in controlling bloom-forming species, can be expected to
help understanding how to mitigate harmful algal bloom (HAB) species. Furthermore, there is a
need for linking the fields of phytoplankton ecology, viral ecology, nutrient cycling and physical
oceanography to better understand whole ecosystems. In a time of increased development along
the coastlines of the world, there is a recent wave of efforts to develop whole ecosystem models.
In a specific sense, there is an immediate need for understanding the linkages between the micro-
bial loop, the process of eutrophication (nutrient uptake by phytoplankton, hypoxia/anoxia, fish
kills) and the increasing occurrence of HABs, something that is currently omitted from most
coastal and estuarine modelling approaches.
3. VIRAL DIVERSITY
Section prepared by K. E. Wommack, C. A. Suttle, W. H. Wilson, M. Weinbauer,
C. Pedrós-Alió, and C. Schuett
Appreciation of biological diversity is key to many investigations of ecosystem processes.
Diversity exists on several levels, communities, populations, organisms and ultimately genes, and
can be defined with increasing levels of resolution. Because viruses are incapable of replication
outside of host infection, virus diversity is inextricably linked to host diversity. Thus, determinis-
tic studies of marine viral diversity at increasing levels of resolution (community, population, and
strain) are critical to a better appreciation of the role of viral infection and lysis in the population
biology and ecology of marine microbial host communities. In this sense, diverse communities
of marine viruses may represent a unique paradigm within ecology.
3.1. Scales of viral diversity
The past decade of research in marine viral ecology has revealed that characterization of viral
diversity is critical to constraining the ecological role of viruses within marine microbial com-
munities. Diversity of viral communities can be defined and measured on several levels which
have implications for the evolutionary and ecological impact of viruses within marine microbial
communities. Perhaps the earliest indication that viral diversity probably influences co-occurring
host diversity came through the general realization that viral strains can be very restricted in the
range of hosts they infect. The routine observation of host specificity has important implications
for understanding the interaction of viral diversity on host community diversity. Unlike other
predators of bacteria and phytoplankton, the specificity of viral infection indicates that viruses
may selectively affect the occurrence and distribution of host strains within communities of
plankton (i.e., killing the winner hypothesis) (Thingstad, 2000). Conversely, some viruses infect
a diverse range of hosts. Thus, it is easily imagined that the diversity of host ranges within viral
communities, from highly specific (strain) to broad (intergenic), is a key parameter to under-
standing the ecological impact of viral infection within marine microbial communities.
While host range is a critical, albeit poorly constrained, ecological parameter for marine viruses,
it is exceedingly difficult to determine for a large suite of viruses. Moreover, host range is only
informative to ecological studies if this parameter is known for the viral strains within a natural
community. Thus, characterizing viral diversity at a genetic level is a more tractable approach for
investigating the effect of viral infection on host community diversity. For uncultivable micro-
organisms, small sub-unit ribosomal RNA has been a critical marker for phyletic studies of
microbial communities and has revealed that the biosphere contains an amazing diversity of
CIESM Workshop Monographs n°21 14
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
14 / 94
prokaryotes and picoeukaryotes. However, because of their reduced genomes and varied life
strategies, no universal genetic marker exists with which to compare the diversity among all
viruses. Nevertheless, through molecular genetic tools, it is possible to characterize viral diversity,
from an evolutionary standpoint, using marker genes specific to broad groups of viruses.
Moreover, sequence analysis has revealed the mosaic nature of viral genomes and the complex
nature of their evolutionary history. On both these levels, single gene to whole genome, we can
begin to appreciate how the diversity of viruses affects the structure and activity of planktonic
communities.
A conceptual diagram illustrates the interconnected relationships that define how viral diversity
within marine ecosystems is characterized (Fig. 2). We define the smallest unit of viral diversity
as the genogroup which is a group of related viruses that is clearly monophyletic based on a spe-
cific marker gene. The difficulty of cultivating viruses and their hosts, as well as the modular
nature of viral evolution, makes it necessary to focus on single genes as defining the basic units
of viral populations. In this regard it is imperative to focus on viral genes with a well-defined evo-
lutionary history. At present examples of such viral genes are few, making the search for viral
gene markers essential for more sophisticated examinations of marine viral diversity. Ultimately
it is desirable to include genes in our molecular toolbox, which confer viral phenotypes of eco-
logical consequence, (e.g., host range, burst size, and latent period.)
At the level of the genogroup, evolutionary processes can be rapid, principally deriving from
point mutations in key genes that alter host range, speed of infection (latent period), life cycle
(temperate, virulent, chronic) and burst size. Each of these alterations in viral phenotype has sig-
nificant implications for host population biology. Understanding how these phenotypes (key
genes) change on environmentally relevant time scales will be critical to the accurate estimation
of the role of viral infection and lysis within marine microbial communities. At present, each of
these specific characteristics of host-viral interactions is poorly constrained for natural popula-
tions. Thus, isolation of appropriate viral-host model systems, as well as more encompassing data
sets of whole viral genome sequences are critical needs.
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
15 CIESM Workshop Monographs n°21
Fig. 2. Scales of viral diversity
15 / 94
The classic ecological definition of a population is a collection of strain (sub-species) groups that
can exchange genetic information. While the modular theory of bacteriophage evolution was pre-
dicted before the wider availability of genome sequence, it appears that the genomes of tailed
bacteriophages are comprised of genetic modules which are disseminated among apparently
dissimilar strain groups (Botstein, 1980; Hendrix et al., 1999). In effect, the genomes of tailed
phages are mosaics of connected modules that mix and match over evolutionary time scales. At
present, it is not known whether viral groups outside of the tailed bacteriophages display similar
genetic promiscuity over evolutionary time. Thus, whole genome sequencing of viruses infecting
marine eukaryotes is warranted to determine the underlying mechanisms driving evolution of
these viruses. The mosaic nature of tailed bacteriophage genomes confounds easy definition of a
viral species; nonetheless, empirical evidence demonstrates that key marker genes can define
monophyletic groups of related marine viruses. These monophyletic groups infect a defined,
albeit broad, range of hosts. Examples are the use of pol (the gene for DNApolymerase I) to char-
acterize viruses infecting eukaryotic microalgae (Chen and Suttle, 1996). It is possible that for a
given functional or genetic group of marine microorganisms there are key genes that define the
groups of viral genotypes to which they are susceptible. In the case of the Phycodnaviridae the
genetic sequence of pol has been critical for examining the distribution and characterization of
these viruses and for making inferences on their ecological impact (Short and Suttle, 2002). In
essence, more genetic markers are needed to define the population ecology of specific viral
groups. Thus far, faithful genetic markers have only been identified for viruses infecting photo-
autotrophs. In depth understanding of the impact of viral infection on the community ecology of
hydrotropic bacteria is now limited by a lack of key, monophyletic markers for this disparate
group of marine microorganisms.
At the broadest scale of diversity, collections of viral populations form a community known as the
virioplankton (Sieburth, 1979). Virioplankton composition is influenced by a number of factors,
the first of which may be the selective forces determining the assortment of co-occurring host
species. Determinants of virioplankton community composition are complex however, due to the
negative feedback between host lysis and viral production. Thus, selective viral infection may
shape host community composition on short time scales; ultimately virioplankton communities
are an assortment of individual genogroups infecting a diverse range of hosts. Examination of
viral community composition has been aided through the use of molecular fingerprinting tools
such as pulsed-field gel electrophoresis (PFGE). This molecular fingerprinting technology is
beginning to show that estuarine environments exhibit rapid (i.e. less than a month) shifts in
virioplankton composition (Wommack et al.,. 1999) while near coastal oceanic environments
have relatively stable viral communities (Steward et al.,. 2000; Riemann and Middelboe, 2002).
CIESM Workshop Monographs n°21 16
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
Fig 3. Conceptual
diagram of the
relationships between
viral and host diversity.
16 / 94
Observations of virioplankton populations over environmental gradients (e.g., the photic to
euphotic, or oxic to hypoxic) have demonstrated significant shifts in composition. Thus, even at
the relatively poor resolution of PFGE we now appreciate that viruses are dynamic members of
microbial communities. The question now is what factors and mechanisms contribute to the rel-
ative diversity and dynamism of a given virioplankton community ?
3.2. A unifying theory of viral and host diversity
A conceptual diagram of the complex ecological relationship between virus and host diversity is
shown in Fig. 3. Examinations of viral diversity provide data to infer the course of viral evolution
and can suggest possible genetic mechanisms at work (e.g. the mosaic nature of tailed phage
genomes). In turn, these molecular genetic mechanisms drive, over evolutionary time, changes in
the diversity of marine viral communities. On shorter, seasonal, timescales, changes in virio-
plankton diversity (i.e., community composition) as well as external selective pressures likely
influence the diversity of specific host populations (e.g. marine Synechococcus). However, due to
the negative relationship between viral lysis and host productivity, ultimately the assortment of
host species within a community shapes overall virioplankton diversity and composition. Finally,
because viruses are both potent killers and vectors for genetic exchange between dissimilar host
species (i.e., horizontal gene transfer), viral activity can influence co-occurring host diversity.
The challenge to marine microbial ecologists is to determine the exact nature of these relation-
ships, the conditions under which they occur and the relevant timescales for each process.
In this regard, integrative studies incorporating viral diversity measures along with estimates of
microbial activity (i.e. primary and secondary production) and environmental conditions will
provide insights on the changing role of virus infection within microbial communities. It is quite
possible that under certain conditions viral lysis plays a key role in restricting bacterial and
phytoplankton production; while under other conditions, viral lysis principally effects host com-
munity composition with only modest effects on overall production.
ECOLOGY OF MARINE VIRUSES – Banyuls-sur-mer, 19-22 March 2003
17 CIESM Workshop Monographs n°21
Box 3. RESEARCH PRIORITIES IN MARINE VIRAL DIVERSITY
Genomics
- Broaden the search for genetic markers of viral evolution and diversity.
- Identify specific genes or gene cassettes that confer ecologically
relevant phenotypes (e.g. burst size, latent period).
- Determine the extent of mosaicism and other genetic mechanisms in
viral evolution.
- Examine the evolutionary interplay between diversity of specific host
populations and their viruses.
- Work towards increasing levels of resolution and information with
regards to viral genomics by moving from comparative to functional and
ultimately to ecological studies.
- Examine the extent and impact of viral-mediated gene transfer on host
diversity and evolution.
- Better connect primary genetic data with actual viral phenotypes.
Scales
- Determine whether viral genogroups have specific bio-geographic
distributions.
- Investigate the temporal and spatial environmental gradients
constraining viral diversity.
- Constrain the extent of co-variation/co-evolution of host and viruses.
- Explore the extent to which viral genotypic diversity drives the
population genetics of host species over ecologically relevant time.
-Examine the role of viral infection in the dynamics of microbial blooms.
-Place viral activity within the larger context of annual biological cycles
in marine ecosystems.
- Identify of ecologically relevant virus/host systems.
17 / 94