Viral distribution and life strategies in the bach dang estuary, vietnam

MICROBIOLOGY OF AQUATIC SYSTEMSViral Distribution and Life Strategies in the Bach Dang Estuary, Vietnam Yvan Bettarel&Thierry Bouvier&Martin Agis&Corinne Bouvier&Thuoc Van Chu& Marine Co

MICROBIOLOGY OF AQUATIC SYSTEMS

Viral Distribution and Life Strategies in the Bach Dang

Estuary, Vietnam

Yvan Bettarel&Thierry Bouvier&Martin Agis&Corinne Bouvier&Thuoc Van Chu&

Marine Combe&Xavier Mari&Minh Ngoc Nghiem&Thuy Thanh Nguyen&

Thu The Pham&Olivier Pringault&Emma Rochelle-Newall&Jean-Pascal Torréton&

Huy Quang Tran

Received: 24 November 2010 / Accepted: 17 February 2011 / Published online: 9 March 2011

# Springer Science+Business Media, LLC 2011

Abstract Although the structure and dynamics of

plank-tonic viruses in freshwater and seawater environments are

relatively well documented, little is known about the

occurrence and activity of these viruses in estuaries,

especially in the tropics Viral abundance, life strategies,

and morphotype distribution were examined in the Bach

Dang Estuary (Vietnam) during the dry season in 2009 The

abundance of both viruses and their prokaryotic hosts

decreased significantly from upstream to downstream,

probably as the result of nutrient dilution and osmotic

stress faced by the freshwater communities The antibiotic

mitomycin-C revealed that the fraction of lysogenic cells

was substantially higher in the lower seawater part of the

estuary (max 27.1%) than in the upper freshwater area

where no inducible lysogens were observed The question

of whether there is a massive, continuous induction of

marine lysogens caused by the mixing with freshwater is considered Conversely, the production of lytic viruses declined as salinity increased, indicating a spatial succes-sion of viral life strategies in this tropical estuary Icosahedral tailless viruses with capsids smaller than

60 nm dominated the viral assemblage throughout the estuary (63.0% to 72.1% of the total viral counts), and their distribution was positively correlated with that of viral lytic production Interestingly, the gamma-proteobacteria explained a significant portion of the variance in the

<60 nm and 60 to 90 nm tailless viruses (92% and 80%, respectively), and in the Myoviridae (73%) Also, 60% of the variance of the tailless larger viruses (>90 nm) was explained by the beta-proteobacteria Overall, these results support the view that the environment, through selection mechanisms, probably shapes the structure of the prokary-otic community This might be in turn a source of selection for the virioplankton community via specific affiliation favoring particular morphotypes and life strategies

Introduction Viruses form a ubiquitous, dynamic compartment in both seawater and freshwater environments where they fulfill numerous biogeochemical and ecological functions [49] They are now considered as major players in the ecological balance of aquatic ecosystems Primarily targeting prokar-yotes, planktonic viruses interact with their hosts in two main different ways: lytic and lysogenic infection cycles [53] and, more sporadically, through chronic cycles or pseudolysogeny [57]

Although viruses are the most abundant and probably the most diverse biological entities on earth [47], relatively little is known about the patterns of viral distribution within

Y Bettarel ( *):T Bouvier:M Agis:C Bouvier:M Combe:

X Mari:O Pringault:J.-P Torréton

UMR 5119, ECOSYM, Montpellier 2 University, CNRS, IRD,

IFREMER,

Montpellier, France

e-mail: yvan.bettarel@ird.fr

T Van Chu:T T Pham

Institute of Marine Environment and Resources (IMER),

Hai Phong, Vietnam

M N Nghiem

Institute of Biotechnology (IBT),

Hanoi, Vietnam

T T Nguyen:H Q Tran

National Institute of Hygiene and Epidemiology (NIHE),

Hanoi, Vietnam

E Rochelle-Newall

UMR 7818 BIOEMCO, IRD,

Paris, France

Microb Ecol (2011) 62:143–154

DOI 10.1007/s00248-011-9835-6

and between the main types of aquatic habitat During the

last two decades, much attention has been paid to marine

coastal and oceanic waters and, to a lesser extent, lacustrine

freshwater from a virio-ecological perspective, especially in

temperate zones However, few studies have been carried

out in transition zones such as estuaries, particularly in

Asia

Estuaries are interesting areas for research because they

are among the most productive and exploited aquatic

habitats They often have strong eutrophication and salinity

gradients as a result of the dilution of nutrient-rich river

water by seawater Studying viral distribution along such

steep gradients is of great interest for microbial ecologists

because the mixing front between marine tidal water and

outflowing river water is an area of dramatic changes which

can trigger important physiological, genetic, and ecological

shifts in their microbial hosts [11,24]

Previous studies on estuarine virioplankton communities

have shown that abundance is typically correlated with the

distribution of prokaryotes for which they can be a major

correlations have been found between viral abundance and

some physical and chemical parameters such as salinity,

chlorophyll a, suspended material or nutrient

has emerged as estuaries are often characterized by a

unique, complex combination of hydrodynamic, trophic,

and thermal conditions

Recent reports have shown that viral diversity and

richness observed in aquatic biomes can change rapidly

for example, the virioplankton community structure, as

inferred from PFGE and RAPD-PCR analyses, was

reported to exhibit seasonally and spatially dynamic

case in the Charente Estuary (France) where the genetic and

morphological structure of the virioplankton community

was relatively stable [5] Unfortunately, no other reports of

viral diversity in estuaries could be found to explain why

the structure of a viral community can remain stable in

environments where biophysicochemical characteristics are

so heterogeneous Moreover, the distribution of viral life

strategies along estuarine gradients is still unknown

Clearly, further studies are needed to provide a better

understanding of the factors that govern viral abundance,

proliferation, and diversity in all estuarine systems

This study examined the distribution of viruses, the

prevalence of lytic versus lysogenic strategies, and the

morphological composition of the virioplankton

commu-nity It was conducted in the Bach Dang Estuary, one of

the main tributaries in the Red River Delta, Vietnam, a

tropical region with rapid population growth and the

main source of agricultural produce for North Vietnam

[62] It also assessed whether the variability of these viral parameters could be explained by the local environmental specificities (salinity, nutrients, tempera-ture, and dissolved organic carbon) and/or the distribu-tion and the phylogenetic composidistribu-tion of their prokaryotic hosts

Material and Methods Description of the Study Site Samples were collected on March 12, 13, and 15, 2009, between 07:00 and 10:00 during neap tides, throughout the salinity gradient of the Bach Dang Estuary, one of the main tributaries of the Red River in Vietnam (Fig.1) The Red River

is characterized by high human pressure and by the transport

of vast amounts of fine sediment during seasonal monsoons

estuary over a salinity gradient of 0 to 31 (Table1) At each station (nos 1–15, see geographical coordinates in Table 1), samples were collected in subsurface water (1.5 m depth) using a Niskin bottle Duplicate samples were analyzed for nutrient and chlorophyll a (Chl) content, as well as for bacterial and viral parameters Samples for dissolved

inorgan-ic nutrient measurements (N-NO2, N-NO3, N-NH4, P-PO4) were filtered through precombusted Whatman GF/F fiberglass filters, stored at −20°C and analyzed according to Eaton et al [18] Chl concentrations were determined by fluorometry after filtration onto Whatman GF/F filters and methanol extraction [24] Dissolved organic carbon (DOC) analyses were performed on 30 mL subsamples filtered onto precombusted GF/F filters and stored in precombusted (450°C, overnight)

40 mL glass vials with Teflon stoppers, with 35 μL 85% phosphoric acid (H3PO4) Samples were stored in the dark until analysis using a Shimadzu TOC VCPH analyzer Potassium phthalate was used as a calibration standard, and certified reference materials (Hansell Laboratory, University

of Miami) were also used to verify the instrument Salinity and temperature were measured in situ using a CTD probe (Seabird SBE 19+)

Counts of Viruses and Prokaryotes Water samples were fixed with 0.02 μm filtered buffered formaldehyde (final concentration 2% v/v) after sam-pling, immediately flash frozen in liquid nitrogen, and stored at −80°C prior to counting The number of viruses and prokaryotes in the duplicate samples of 0.3–0.8 mL was determined after retention on 0.02 μm pore size membranes (Anodisc) and staining with SYBR Gold fluorochrome (Molecular Probes, Europe, Leiden, the Netherlands) as described in detail by Patel et al [36]

Table 1 Geographical coordinates and physicochemical parameters of the sampled stations in the estuary of the Bach Dang River, Vietnam, March 2009

Station

nos.

(°C)

Chl a (μg L−1)

DOC (μM-C)

N-NO 2

(μM)

N-NO 3

(μM)

N-NH 4

(μM)

P-PO 4

(μM)

a stations where the viral lytic production and the fraction of lysogenic cells were estimated

b stations where the distribution of viral morphotypes and the phylogenetic composition of the prokaryotic communities were estimated

4 km

Figure 1 Map of the Bach

Dang River Estuary (Vietnam)

and location of the fifteen

sampling stations, March 2009

Examination of Viral Morphotypes

Planktonic viruses were observed using transmission electron

microscopy (TEM) Viruses from 500 μL aliquots of

formalin-fixed samples were harvested by repeated ultracentrifugation

of 50 μL onto grids (400 mesh Cu electron microscope grids

with carbon coated Formvar film) using an A-100/30 rotor in

an air-driven ultracentrifuge (Airfuge®, Beckman) at

105,000×g for 70 min The grids were then stained for 30 s

with uranyl acetate (2%, w/w), and viruses were examined

and measured using a JEOL 1200EX TEM operated at

80 kV and magnification from ×20,000 to ×100,000 Three

morphotypes were distinguished for shape classification of

tailed viruses (Caudovirales) Tailed viruses with isometric

heads and long noncontractile tails were considered to be

siphoviruses Tailed viruses with isometric heads and

contractile tails (presence of a neck) were considered to be

myoviruses Tailed viruses with isometric heads and short

distribution of tailless icosahedral viruses in size classes <60,

60–90, and >90 nm was also determined

Fraction of Lysogenic Cells

The fraction of lysogenic cells (FLC) was determined by

the induction of prophages using mitomycin-C [52]

Mitomycin-C was added to samples (final concentration

samples served as controls Samples were incubated at in

situ temperature; duplicate subsamples were taken with

syringes at the start of incubation (t0) and after 12 h (t12h)

and fixed with 0.02 μm filtered buffered formaldehyde (2%

final concentration) for viral and bacterial counts (see above)

The FLC was estimated from viral abundances in

mitomycin-C treated (VAm) versus control (VAc) incubations, as well as

bacterial abundance (BAt0) and burst size (BSt0): FLC¼

100 ½ðVAm VAcÞ= BSð t0 BAt0ފ [52] The burst size

chosen in this study (BS=24) was the mean value for aquatic

environments calculated by Wommack and Colwell [61] and

Parada et al [35]

Viral Lytic Production

Viral production was determined using the dilution

tech-nique described by Wilhelm et al [56] Fifty milliliters of

duplicate subsamples was filtered onto a 47-mm diameter,

0.2-μm pore size polycarbonate membrane at low pressure

(<33 kPa), using a transfer pipette to keep the bacteria in

suspension and adding ultra-filtered water (<30 kDa) to

maintain the total water volume Ultra-filtered water from

each station was added to the filter reservoir until three

volumes (150 mL total) had been flushed through the filter

By diluting the viruses and not the prokaryotes, it was

possible to determine viral production resulting from infection prior to the start of the experiment A final

50 mL volume retained in the filter housing was immediately transferred to polycarbonate bottles and incubated in the dark at in situ temperature for 12 h Two milliliters of these subsamples was collected every

3 h and fixed with 0.22 mL filtered formaldehyde (final concentration of 2%) for viral enumeration Viral production rates were inferred from the slope of the linear regression line of viral abundance versus time, for duplicate incubations

CARD-FISH Analyses of Phylogenic Diversity Samples (5 mL) were fixed with formaldehyde (2% final concentration), filtered using 0.2 μm polycarbonate filters (Whatman) and kept at −20°C until hybridization Several oligonucleotide probes (Biomers) were used: EUB 338 I+II+III, ALF968, BET42a, GAM42a, CF319a, ARCH915, targeting bacteria, alpha-, beta- and gamma-proteobacteria, the Bacteroidetes phylum and the Archaea domain, respectively The NON338 probe was used as a negative control Hybridization and mounting procedures were carried out as described by Pernthaler et

al [37] The selected groups were counted and expressed

as a percentage relative to the total abundance of DAPI-stained bacterial cells The error associated with replicate CARD-FISH counts ranged from 5% to 25% (mean 15.3%) based on a subset of three samples for which independent replicate CARD-FISH counts were con-ducted The mean error (15.3%) was taken to apply to all the CARD-FISH analyses

Statistical Analyses The data was log transformed to provide the normality and homogeneity of variance necessary for parametric analyses Simple relationships between original data sets were tested by Pearson correlation analysis The rela-tionship between salinity and nutrients was studied to determine whether the distributions followed a dilution gradient As the variables are considered to be interde-pendent, the model II regression model method [27] was used to estimate the slope of the regression lines for each nutrient The slopes of the theoretical dilution lines were calculated using the equation:

Sth¼
½ ŠNS max ½ ŠN FW

= Sð maxÞ, where Sthis the

(freshwater) Differences were considered to be nonsig-nificant if the theoretical slopes were within the 95% confidence interval

In estuarine data, variables are often internally correlated

because of their strong common link to physical variables

such as salinity [40] The relationship between bacterial

phylogenetic identity and viral morphology was analyzed

using partial correlation analysis, keeping salinity constant,

according to the Sokal and Rohlf [46] procedure This

method allows to directly estimate the degree of an

association between two variables with the effect of another

variable removed, here the salinity Considering X, Y, and Z

as the bacterial phylogenetic group, the viral morphologic

category, and the salinity, respectively, the partial correlation

of X and Y adjusted for Z (ρXY.Z) was computed as follows,

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

1 r2YZ

in which ρXY, ρXZ, and ρYZrepresent the regular correlation

coefficients of respective correlations The partial

determina-tion coefficients (the square of the partial correladetermina-tion

coefficient) were used to determine what fraction of the total

variance of the bacterial variables was explained by their

correlation with viral morphological groups [28] The

statistical analyses were performed using the statistics

program R

Results

Environmental Characteristics of the Sampling Site

concentrations decreased steadily with decreasing salinity

(Table1) Model II linear regression between nutrients and

salinity did not show significant deviations from the

straight line between river and seawater reference stations,

indicating that the concentration of the different nutrients

followed the linear mixing model and suggesting a simple

dilution process The temperature was relatively stable and

ranged from 21.2°C to 19.9°C during the sampling period

The DOC varied between 113.0 to 97.0 μM-C, with the

lowest values observed in the high salinity, seaward

stations Unlike the nutrient distribution, DOC distribution

deviated from the simple mixing model and some net

accumulation was observed in the mid-salinity ranges

Viral and Prokaryotic Abundance

Viral and prokaryotic abundances decreased sharply from

at salinity 1.4 (MVIR14.7×107vir mL−1, MPROK18.8×106

cells mL−1) and then declined to 6.5×107vir mL−1and 4.2×

closely correlated (r=0.79, p<0.05; Fig.2b)

Distribution of Lysogenic Prokaryotes and Viral Lytic Production

The FLC generally varied between 0% and 5.4% overall but rose abruptly from salinity 21 to reach 27.1% at the

there was a strong negative correlation between viral lytic

production was significantly (p<0.05) correlated with the nitrogen and phosphorus nutrient concentrations, as well as with the viral and prokaryotic concentrations (Table2) Distribution of Viral Morphotypes

Tailless icosahedral viruses were dominant within the viral community, accounting for 78.6% to 83.4% of the total viral community Only 16.6% to 21.4% of the community was comprised of tailed viruses Based on the morpholog-ical criteria defined by the International Committee on Taxonomy of Viruses [13], tailed viruses were assigned to the three families of the order Caudovirales (Siphoviridae,

proportion of these morphotypes was never more than 11%

of the total viral community The proportion of Podoviridae increased significantly with increasing salinity (r=0.65, p<

and Siphoviridae did not show any clear spatial pattern Among tailless icosahedral viruses, those with the smallest capsid (i.e., <60 nm) were clearly dominant, accounting for 63% to 72.1% of the viral counts Their proportion fell sharply along the salinity gradient (r=−0.92, p<0.005, Fig 4) and was positively correlated with the different nutrient concen-trations, as well as with viral abundance and lytic production There was also a significant negative correlation between the distribution of this group and the fraction of cells with

between 60 and 90 nm and viruses larger than 90 nm were of minor importance as they accounted for only 4% to 10% of the total viral community (Fig.4)

Relationships between Viral Morphological Groups and the Bacterial Phylogenetic Groups

To explore the relationships between morphological viral groups and bacterial composition, we used multiple regression analysis and used the resulting partial regression coefficients to assess the proportion of the variance explained by the dominant bacterial phylogenic groups once the effect of salinity is removed The gamma-proteobacteria explained a significant portion of the variance in <60 nm and 60 to 90 nm tailless viruses (92%

and 80%, respectively) and in the Myoviridae (73%) after

having removed the effect of salinity, whereas the other

phylogenetic groups explain virtually none of these

tailless larger viruses (>90 nm) was explained by the

Bacteroidetes did not explain any variability in viral

morphological groups

Discussion

Viral abundance, morphological diversity, and life strategies

all varied considerably in the Bach Dang Estuary We tried

to determine whether such spatial patterns could be

predicted by the environment and/or the distribution and

diversity of their prokaryotic hosts

As expected, the nutrient and salinity gradients were the

major forces shaping (directly or indirectly) the prokaryotic

community, and subsequently, their viral pathogens Changes in nutrients (NH4, NO2, NO4, PO4), as described

by a linear regression model, followed a simple dilution process of the Bach Dang River in coastal waters Such phenomena have been reported on several occasions by Troussellier et al [50] in the Rhone Estuary (France), in the Saint Lawrence Estuary, Canada [34], and in Chesapeake Bay, USA [41] The only exception was DOC, where some net accumulation was observed in the mid-salinity zones (not shown)

The decline in prokaryotic abundance may be the direct consequence of chemical changes in the water, including a combination of the decrease in nutrient concentration and

an increase in salinity Temperature, however, did not emerge as a strong determinant of prokaryotic distribution,

as shown by the low spatial variability recorded throughout the estuary (min–max, 19.7–21.7°C) Shiah and Ducklow [44] suggested that temperature ceases to be a limiting factor when it exceeds 20°C, which was the case in this

0 5 10 15 20

Viral abundance (107 viruses mL-1)

6 cells mL

-1 )

r = 0.79*

b

0 4 8 12 16 20

rVIR = -0.70*

rPROK = - 0.90*

Viruses Prokaryotes

6 cells mL

-1 )

7 VLPs mL

-1 )

a

Figure 2 Relationships

between prokaryotic and viral

abundance in the Bach Dang

Estuary (a) and prokaryotic and

viral abundance from upstream

to downstream stations (b)

0 5 10 15 20 25 30 35

0 10 20 30 40 50 60

8 viruses L

-1 h -1 )

a

b

Figure 3 Fraction of lysogenic

cells (a) and viral lytic

production (b) along the

estuarine salinity gradient

Table 2 Correlation relationships between environmental parameters in the Bach Dang Estuary

Salinity [Chl a] [DOC] [NO2] [NO3] [NH4] [PO4] FLC [VIR] [PROK] VP <60 nm >90 nm [Chl a] − 0.44 1

>90 nm 0.77 − 0.51 − 0.46 − 0.59 − 0.56 −0.87 − 0.67 0.58 −0.75 − 0.70 − 0.59 − 0.56 1 Tailless icosaehedral viruses smaller than 60 nm (<60) and larger than 90 nm (>90) Significant correlations (p<0.05) are indicated in bold FLC fraction of lysogenic cells, VP viral lytic production

65

70

75

0 2 4 6 8 10 12

0

2

4

6

8

10

0 2 4 6 8 10

0

2

4

6

8

10

0 2 4 6 8 10

r = -0,90*

r = -0,02

r = 0,48

r = 0,65*

r = 0,16

r = -0,04

morphotypes (% total viruses) < 60 nm tailless

60-90 nm tailless

> 90 nm tailless

Podoviridae

Myoviridae

Siphoviridae

Figure 4 Distribution of the different viral morphological groups along the estuarine salinity gradient Bar=50 nm

Table 3 Coefficient of partial determination of alpha-, beta-, and gamma-proteobacteria and Bacteroidetes (Bdetes) and viral morphological groups from all stations in the Back Dang River Estuary

Tailless viruses smaller than 60 (<60), between 60 and 90 (60–90), and larger than 90 nm (>90) Tailed viruses of the Podoviridae (Podo), Myoviridae (Myo) and Siphoviridae (Sipho)

n number of samples

*0.01≤p≤0.05

**p≤0.01

study Concentrations of DOC remained above 100 μM C

and the correlation between primary and bacterial

produc-tion was weak, suggesting that bioavailable organic carbon

was not limiting [42] Therefore, the decline in prokaryote

abundance appears to be rather the result of substantial

changes in osmolarity Indeed, the osmotic stress faced by

freshwater prokaryotes as they pass along the estuary might

explain most of the net loss of cells, as assessed

experimentally by Cissoko et al [15] Troussellier et al

[50] also showed that only a very limited number of

freshwater bacteria can maintain metabolic activity under

marine conditions

Given the high specificity of virus–prokaryotes

relation-ships [9], and that prokaryote distributions are strongly

related to salinity distributions [30], then it is also expected

that salinity will have a strong incidence on virioplankton

distribution, life cycle, and community structure [10] The

concomitant decrease in prokaryotic and viral abundance

from freshwater to seawater seems to indicate that most of

the viral estuarine communities comprised bacteriophages,

as is the case in the vast majority of aquatic habitats [53]

Similar conclusions were drawn from studies conducted in

estuaries in USA, Australia, and France where there was

significant covariance between viruses and salinity, nutrient

the other hand, the metabolic cost to freshwater prokaryotes

to resist osmotic stress may alter their susceptibility to viral

infection There is now a consensus that viral activity relies

heavily on the host being healthy in order to complete their

lytic cycle (e.g., the synthesis of the viral constituents and

salinity may alter the integrity of the capsid’s receptors and

inhibit the binding of viruses to their hosts [26], resulting in

significant reduction of viral stocks Alternatively, the

inactivation and decay of viruses themselves could be

influenced by changes in ionic strength [55] Interestingly,

the absence of environmental gradients in a temporarily

open/closed South African estuary resulted in a total

absence of temporal or spatial pattern in bacteria abundance

and viral activity [1]

Switching of Viral Life Cycle

Contrasting viral strategies were observed throughout the

estuarine gradient, characterized, from freshwater to sea

water, by a decrease of virulent lytically produced

viruses in favor of lysogenized cells Lysogeny is

typically considered as a survival strategy for viruses

that is favored during times of low resources, low host

the drop in nutrient concentrations together with that of

prokaryotic abundances might be the cause of the

prevalence of lysogenic rather than lytic viral life styles

in the more saline area Hewson et al [23] also suggested that lytic infections may be less common in the seawater than in the freshwater part of the estuary of Moreton Bay (Australia) In this study, the spectacular rise of the FLC in the salinity range of 22–30 raises a fundamental question: are the lysogens of river or marine origin? A river origin is unlikely as freshwater prokaryotes have very little

the lysogens are of marine origin, then the rapid decline in the proportion of inducible lysogens with decreasing salinity (coming from the seawater) could be partly explained by salinity-driven activation of the lysogens during the mixing between seawater and freshwater Shkilnyj and Koudelka [45] recently reported that strong shifts in salinity can trigger the switch from lysogenic to lytic cycles Under certain circumstances, therefore, mitomycin-C may only induce the fraction of residual lysogens that had not already been induced by environ-mental factors such as light conditions [8] or osmolarity Finally, the replacement of lytic by lysogenic pathways along salinity gradients suggests that estuaries are areas where there are major changes in virus–prokaryote interactions

Distribution of Viral Morphotypes Although analysis of viral morphology provides less information about viral diversity than metagenomic approaches, observing viral morphotypes was useful as

it showed that the communities of viruses varied along the whole of the estuarine gradient of the Bach Dang River One of the most striking findings was the distribution of some viral morphological groups that appeared to be closely related to both the viral life strategies and the phylogenetic affiliation of their hosts

In general, icosahedral tailless viruses were far more numerous than the tailed viruses of the Myoviridae, Siphoviridae and Podoviridae families Several other reports have also shown a dominance of tailless versus tailed phages [5,7,33,47] However, the possibility that ultracentrifugation is responsible for the tail disruption of some Caudovirales cannot be excluded [53], which might lead to the substantial underestimation of the significance

of this group In this study, the proportion of tailless viruses increased with increasing salinity while that of the Caudovirales declined (Fig.4) The decline of the tailless viruses was most marked for the smallest viruses (<60 nm) which were the dominant fraction of the viral community and typically dominate in natural freshwater or seawater

between this subgroup and the viral lytic production seems

to imply that most of these small icosahedral viruses are virulent and are the main component of the virioplankton

community controlling prokaryotes On the other hand,

from a marine perspective, the quantitative increase of this

subgroup from downstream to upstream, together with

decreasing FLC values could also be considered as the

consequence of massive salinity-initiated induction of

marine lysogens, as mentioned above Lastly, there may

be increased adsorption rates of these small viruses onto

the large amount of suspended matter in the river waters

[42] before they are diluted with seawater [23,54]

We observed significant relationships between the

proportion of certain viral morphotypes and that of some

bacterial phylogenetic groups, and partial correlation

analysis showed that these relationships are not driven by

salinity alone (Table3) The specificity of the interaction

between virus and host is strong and usually located at the

species- or genus-specific level The different main groups

within bacteria are presently identified solely on the basis

of their branching pattern in the 16S rRNA trees Some of

these groups are very large and include thousands of

species The morphological discrimination of viruses into

size classes or families also leads to large groups

comprising several genera and species One may thus not

expect any relationship between virus and host when host

are considered at such phylogenetic level Using a strictly

genealogical basis, a detailed understanding of prokaryotic

phylogeny has begun to emerge [20] Phylogenetic groups

share certain morphological and physiological

character-istics that distinguish them from each other as well as from

the organisms on the other branches For example, uptake

rates of specific fractions of DOM, membrane lipid

composition, growth rate, protein base content, or cell size

have been described for cells of certain bacterial

biochemical processes within a phylogenetic group may

well provide the base for a group level interaction with

viruses Indeed, bacteria phylogenetic identity has been

linked to their size and biovolume [48], which, in turn, have

host phylogeny to viral morphology Similarly, host

phylogeny has been related to viral life strategy [32],

which is also related to viral morphotype [12], again linking

the host phylogeny to viral morphology The lack of

relationship observed in some of the other groups

(Alpha-proteobacteria, Bacteroidetes) underlines the fact that biotic

and abiotic factors other than host identity alone can also

control viral production and abundance

Finally, our results support the view that affinities might

exist between some viral morphotypes, their replication

pathways, and the phylogenetic affiliation of their host In the

Bach Dang Estuary, these affinities seem to be driven by the

environmental conditions, and more particularly, by external

osmolarity Further studies are now required to determine

whether these affinities can be observed in other aquatic

ecosystems and more generally to assess the role of environ-mental features in shaping virus–prokaryote interactions Acknowledgements This work was financed by the EC2CO-PNEC project “HAIPHONG”, and grants from the French IRD, CNRS, Groupement De Recherche (GDR) 2476 Réseaux Trophiques Pélagi-ques, and the Vietnam Academy of Science and Technology (VAST).

We thank Jean-Yves Panché, Jean-Pierre Lefebvre, and Robert Arfi for their help with the physical oceanography measurements.

References

1 Allan EL, Froneman PW (2008) Spatial and temporal patterns

in bacterial abundance, production and viral infection in a temporarily open/closed southern African estuary Estuar Coast Shelf Sci 77:731–742

2 Almeida MA, Cunha MA, Dias JM (2007) Bacterial productivity distribution during a rainy year in an estuarine system Microb Ecol 53:208–220

3 Angly FE, Felts B, Breitbart M, Salamon P, Edwards RA, Carlson

C, Chan AM, Haynes M, Kelley S, Liu H, Mahaffy JM, Mueller

JE, Nulton J, Olson R, Parsons R, Rayhawk S, Suttle CA, Rohwer

F (2006) The marine viromes of four oceanic regions PLoS Biol 4:2121–2131

4 Auguet JC, Montanié H, Delmas D, Hartmann HJ, Huet V (2005) Dynamic of virioplankton abundance and its environmental control in the Charente estuary (France) Microb Ecol 50:337–349

5 Auguet JC, Montanie H, Lebaron P (2006) Structure of virioplankton in the Charente estuary (France): transmission electron microscopy versus pulsed field gel electrophoresis Microb Ecol 51:197–208

6 Bench SR, Hanson TE, Williamson KE, Ghosh D, Radosovich M, Wang K, Wommack KE (2007) Metagenomic characterization of Chesapeake Bay virioplankton Appl Environ Microb 73:7629–7641

7 Bettarel Y, Bouvy M, Dumont C, Sime-Ngando T (2006) Virus– bacterium interactions in water and sediment of West African inland aquatic systems Appl Environ Microbiol 72:5274–5282

8 Bettarel Y, Arfi R, Bouvier T, Bouvy M, Briand E, Colombet J, Corbin

D, Sime-Ngando T (2008) Virioplankton distribution and activity in a tropical eutrophicated Bay Estuar Coast Shelf Sci 3:425–429

9 Bettarel Y, Desnues A, Rochelle-Newall E (2010) Lytic failure in cross-inoculation assays between phages and prokaryotes from three aquatic sites of contrasting salinity FEMS Microbiol Lett 311:113–118

10 Bettarel Y, Bouvier T, Bouvier C, Carré C, Desnues A, Domaizon

I, Jacquet S, Robin A, Sime-Ngando T (2011) Ecological traits of planktonic viruses and prokaryotes along a full salinity gradient FEMS Microbiol Ecol doi: 10.1111/j.1574-6941.2011.01054.x

11 Bouvier T, del Giorgio PA (2002) Compositional changes in free-living bacterial communities along the salinity gradient in two temperate estuaries Limnol Oceanogr 47:453–470

12 Breitbart M, Felts B, Kelley S, Mahaffy JM, Nulton J, Salamon P, Rohwer F (2004) Diversity and population structure of a near-shore marine-sediment viral community Proc R Soc London Ser

B Biol Sci 271:565–574

13 Buchen-Osmond C (2003) Taxonomy and classification of viruses In: Manual of Clinical Microbiology, 8th ed, Vol 2 ASM Press, Washington DC, pp 1217–1226

14 Bühring SI, Elvert M, Witte U (2005) The microbial community structure of different permeable sandy sediments characterized by the investigation of bacterial fatty acids and fluorescence in situ hybridization Env Microbiol 7:281–293

15 Cissoko M, Desnues A, Bouvy M, Sime-Ngando T, Verling E, Bettarel Y (2008) Effects of freshwater and seawater mixing on