232
means that obtaining viral material suffi cient for examining
the abundance, diversity and composition of aquatic viral
assemblages can be challenging (Wommack et al. 2010 ). This
was critical at the onset of aquatic viral ecology because the
need for samples containing a high-density of viruses and viral
genomes was critical to challenge the detection thresholds of
microscopic and molecular methods for the study of viruses.
14.3.1 Transmission Electron Microscopy
(TEM) and Infectivity Parameters
Following the discovery that viruses are abundant in marine
ecosystems (Torrella and Morita 1979 ; Bergh et al. 1989 ),
one of the major challenges for starting the new discipline of
environmental virology was to overcome the methodological
constraints which in the past have strongly limited the obser-
vation of both free-fl oating and intracellular (i.e. infected
cells) viruses in the environment (Bratbak and Heldal 1993 ;
Suttle 1993 ). The earlier studies that have tackled these chal-
lenges were conducted in lake ecosystems of the French
Massif Central, including Lake Pavin (Bettarel et al. 2000 ).
In a short communication, we fi rst proposed an optimized
method to ultracentrifuge and concentrate planktonic viruses
for their examination with transmission electron microscopy
(TEM) (Sime-Ngando et al. 1996 ). Because it is important,
for the accuracy of quantitative estimates of viruses and
infected cells, to avoid collecting these items in an
ultracentrifuge- pellet but on a platform using a swing-out
rotor, we propose a platform made from a commercial poly-
mer resin adaptable at the bottom of an ultracentrifugation
tube, where up to three 400-mesh copper grids (catalog no.
A03; Pelanne Instruments®) can be fi xed with a double-face
tape. This tool allows to directly harvest, viruses and cell-
hosts onto the TEM grids after a centrifugation at 4 °C either
at 120,000 x g for 2 h to keep free-occurring viruses or at
70,000 x g for 30 min to distinguish between prokaryotic
cells with and without intracellular viruses. With our method,
we provided one of the fi rst accurate descriptions of freshwa-
ter viruses dominated by particles with regular icosaedric
capsid with 6 faces and a tail, assimilated to phages that have
been previously described in marine environments (Fig.
14.2 ). These particles were about 10 times more abundant
than prokaryotes, corroborating the hypothesis of their role
in controlling the community dynamics of bacteria in fresh-
water systems (Sime-Ngando et al. 1996 ).
Our method also allowed observation of infected cells, pri-
marily of prokaryotic cells, and the determination of critical
parameters in the calculation of the infection prevalence (i.e.
the so-called FVIC, frequency of visibly infected cells ) , the
related viral production and the burst size (BS) or the average
number of viruses released when a single host cell lyses.
These parameters are generally derived from TEM observa-
tion of visibly mature phages within intact or thin- sectioned
host cells (Fig. 14.3 ), a method that has the advantage of
accounting for the effect of the environment, compared to the
application of BS derived from laboratory cultures. Similar to
marine and other freshwater ecosystems , FVIC in Lake Pavin
is generally less that 5 % but, at a regional local scale, is
higher (Mean 2 %, range 0–4 %) compared to the eutrophic
Lake Aydat (Mean 1 %, range 0–3 %). The seasonal variabil-
ity in both types of lakes is however roughly the same, with
maximum in late spring (May to June) and early autumn
(September to October). It is thus possible that in oligotrophic
lakes, where substrates are in short supply for bacterial pro-
duction, bacterial lysates might represent an important source
of dissolved organic matter or inorganic nutrients, the quan-
tity of which would depend on the frequency of viral infec-
tion (Bettarel et al. 2004 ). This differs from studies conducted
in marine systems, which have reported higher FVIC values
with increasing productivity (Steward et al. 1992 ), or no evi-
dence for a relationship between FVIC and productivity
(Noble and Fuhrman 1997 ). In Fig. 14.4 , we plotted mean
FVIC values for freshwaters in relation to the trophic states of
the lakes. Although the gaps between each trophic level on
the x axis cannot be considered equivalent, there is no appar-
ent trend similar to that which has been found in the marine
environment, suggesting that for lakes, productivity alone is a
poor predictor of the level of viral infection.
In contrast to FVIC in freshwaters, the number of viruses
produced per infected cells (e.g. burst size) , as measured in
TEM, increased with increasing productivity, ranging from
about 5–60 viruses prokaryote −1 in oligotrophic marine
waters to 5–100 viruses prokaryote −1 in productive freshwa-
ters, suggesting that the burst size of infected bacteria would
be higher in more productive environments (Sime-Ngando
et al. 2003 ). In Lake Pavin, BS averaged about 25 viruses
prokaryote −1 (range: 15–50 viruses prokaryote −1 ), with rela-
tively low fl uctuations with depth and over the year when
excluding the peaks that generally occurred in spring time
(Bettarel et al. 2004 ). The trophy-trend in BS is not unrea-
sonable as both cell size and growth rate are generally greater
in eutrophic than in oligotrophic environments, but can be
strongly constrained by the variability in capsid size for dif-
ferent occurring viral populations. Based on reports from a
variety of different aquatic environments, Parada et al. ( 2006 )
calculated a mean BS of 24 and 34 for marine and freshwater
environments, respectively.14.3.2 Epifl uorescence Microscopy (EM)The interest of biologists for the role of viruses in ecological
processes has rapidly raised the requirement of reliable
methods for estimating the level of these biological entities
in aquatic systems. The earlier estimates of viral levels were
obtained by concentrating the viruses by ultracentrifuga-
tion (Børsheim et al. 1990 ) or by ultrafi ltration (Suttle et al.T. Sime-Ngando et al.