ArticlePDF Available

Abstract and Figures

Microbial viruses can control host abundances via density-dependent lytic predator-prey dynamics. Less clear is how temperate viruses, which coexist and replicate with their host, influence microbial communities. Here we show that virus-like particles are relatively less abundant at high host densities. This suggests suppressed lysis where established models predict lytic dynamics are favoured. Meta-analysis of published viral and microbial densities showed that this trend was widespread in diverse ecosystems ranging from soil to freshwater to human lungs. Experimental manipulations showed viral densities more consistent with temperate than lytic life cycles at increasing microbial abundance. An analysis of 24 coral reef viromes showed a relative increase in the abundance of hallmark genes encoded by temperate viruses with increased microbial abundance. Based on these four lines of evidence, we propose the Piggyback-the-Winner model wherein temperate dynamics become increasingly important in ecosystems with high microbial densities; thus 'more microbes, fewer viruses'.
Content may be subject to copyright.
00 MONTH 2016 | VOL 000 | NATURE | 1
ARTICLE doi:10.1038/nature17193
Lytic to temperate switching of viral
communities
B. Knowles1*, C. B. Silveira1,2*, B. A. Bailey3, K. Barott4, V. A. Cantu5, A. G. Cobián-Güemes1, F. H. Coutinho2,6, E. A. Dinsdale1,7,
B. Felts3, K. A. Furby8, E. E. George1, K. T. Green1, G. B. Gregoracci9, A. F. Haas1, J. M. Haggerty1, E. R. Hester1, N. Hisakawa1,
L. W. Kelly1, Y. W. Lim1, M. Little1, A. Luque3,5,7, T. McDole-Somera8, K. McNair5, L. S. de Oliveira2, S. D. Quistad1, N. L. Robinett1,
E. Sala10, P. Salamon3,7, S. E. Sanchez1, S. Sandin8, G. G. Z. Silva5, J. Smith8, C. Sullivan11, C. Thompson2, M. J. A. Vermeij12,13,
M. Youle14, C. Young15, B. Zgliczynski8, R. Brainard15, R. A. Edwards5,7,16, J. Nulton3, F. Thompson2 & F. Rohwer1,7
Microbial viruses infect about 1023 cells per second in the world’s
oceans and the majority of microbial cells are infected at any given
time
1,2
. What determines the proportion of lytic versus lysogenic infec-
tions is not well understood, despite the known importance of lytic/
lysogenic fate in driving ecological and biogeochemical outcomes1,3–6.
Kill-the-Winner (KtW) models of lytic infection predict that den-
sity- and frequency-dependent viral predation suppresses blooms
of rapidly growing hosts, increasing host diversity6–9. A number
of studies provide empirical support for these predictions7,10–13. In
contrast, temperate viral dynamics in the environment are much less
studied, and the relationship between lysogeny and host density is
unclear. Provirus induction studies indicate that lysogeny is more
frequent with low host density14–16. As such, it was established that
viral communities transition from lysogeny to lytic dominance as host
densities rise4,10,12,14,15,17.
Coral reefs offer a unique opportunity to probe the relationship
between microbial host densities and the relative frequency of lytic
versus temperate viral life cycles. Anthropogenic stressors can shunt
these ecosystems into degradative regimes that result in changes in
viral and microbial community composition, and rising microbial
energy demand and densities, a state described as microbialized18–23.
On heavily microbialized reefs, microbial abundances increase five-
to tenfold, which increases predicted virus–host encounter rates18,24.
Density-dependent lytic KtW models predict that reef microbializa-
tion should therefore correlate with increased lytic viral predation,
resulting in a predicted increased virus to microbe ratio. Here we use
four independent analyses—direct counts, literature meta-analyses,
experiments, and viral community metagenomics—to show that
increased host density is instead accompanied by a transition from
lytic to temperate dynamics. On this basis we propose an extension
of the KtW models, the Piggyback-the-Winner (PtW) model, which
reflects the increased contribution of temperate viruses in ecosystems
with high host abundance, yielding ‘more microbes, fewer viruses’.
Viral and microbial abundance
Microbial and viral abundances were measured in 223 Pacific and
Atlantic coral reef samples (Fig. 1a). The density of virus-like par-
ticles (VLPs) was significantly higher than that of the microbes
(t = 19.61, degrees of freedom (d.f.) = 236.96, P < 2.20 × 1016;
Welch two sample t-test) and ranged from 9.03 × 105 to 3.86 × 107
(7.08 × 106 ± 3.01 × 105, mean ± standard error of the mean, s.e.m.)
VLPs ml–1 versus 8.08 × 104 to 6.75 × 106 (1.09 × 106 ± 5.53 × 104,
mean ± s.e.m.) microbes ml–1. The log–log plot of these VLP and
microbe abundances had a slope <1 (m = 0.59, t = 14.82, d.f. = 221, P
(t-test; m 1) = 4.08 × 10
21
; R
2
= 0.50; slope significantly different
from m = 1 by linear regression with t-test; Fig. 1a), indicating a down-
ward concave relationship between these variables. As a result, the
virus to microbe ratio (VMR) decreased significantly (analysed against
host density, both log-transformed; m = 0.37; t = 9.52, d.f. = 221,
P < 2.00 × 1016; R2 = 0.29; linear regression) from a ratio of 25 to 2
VLPs per microbe (7.44 ± 0.24, mean ± s.e.m.) as microbial abundance
increased from ~1 × 105 to greater than 6 × 106.
Microbial viruses can control host abundances via density-dependent lytic predator–prey dynamics. Less clear is how
temperate viruses, which coexist and replicate with their host, influence microbial communities. Here we show that
virus-like particles are relatively less abundant at high host densities. This suggests suppressed lysis where established
models predict lytic dynamics are favoured. Meta-analysis of published viral and microbial densities showed that this
trend was widespread in diverse ecosystems ranging from soil to freshwater to human lungs. Experimental manipulations
showed viral densities more consistent with temperate than lytic life cycles at increasing microbial abundance. An analysis
of 24 coral reef viromes showed a relative increase in the abundance of hallmark genes encoded by temperate viruses with
increased microbial abundance. Based on these four lines of evidence, we propose the Piggyback-the-Winner model
wherein temperate dynamics become increasingly important in ecosystems with high microbial densities; thus ‘more
microbes, fewer viruses’.
1Department of Biology, San Diego State University, 5500 Campanile Drive, San Diego, California 92182, USA. 2Biology Institute, Rio de Janeiro Federal University, Av. Carlos Chagas Filho 373,
Rio de Janeiro, Rio de Janeiro 21941-599, Brazil. 3Department of Mathematics and Statistics, San Diego State University, 5500 Campanile Drive, San Diego, California 92182, USA. 4Hawaii Institute
of Marine Biology, University of Hawaii at Manoa, 46-007 Lilipuna Road, Kaneohe, Hawaii 96744, USA. 5Computational Science Research Center, San Diego State University, 5500 Campanile Drive,
San Diego, California 92182, USA. 6Radboud University Medical Centre, Radboud Institute for Molecular Life Sciences, Centre for Molecular and Biomolecular Informatics, 6525HP Nijmegen,
The Netherlands. 7Viral Information Institute, San Diego State University, 5500 Campanile Drive, San Diego, California 92182, USA. 8Scripps Institution of Oceanography, 8622 Kennel Way,
La Jolla, California 92037, USA. 9Marine Sciences Department, Sao Paulo Federal University - Baixada Santista, Av. Alm. Saldanha da Gama, 89, Santos, São Paulo 11030-400, Brazil. 10National
Geographic Society, 1145 17th St NW, Washington D.C. 20036, USA. 11Department of Biology, University of California San Diego, 9500 Gilman Drive, La Jolla, California 92093, USA. 12CARMABI
Foundation, Piscaderabaai z/n, Willemstad, Curacao, Netherlands Antilles. 13Aquatic Microbiology, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, 1098XH
Amsterdam, The Netherlands. 14Rainbow Rock, Ocean View, Hawaii 96737, USA. 15Coral Reef Ecosystem Division-PIFSC-NOAA, 1845 Wasp Blvd, Honolulu, Hawaii 96818, USA. 16Department of
Computer Science, San Diego State University, 5500 Campanile Drive, San Diego, California 92182, USA.
*These authors contributed equally to this work.
© 2016 Macmillan Publishers Limited. All rights reserved
2 | NATURE | VOL 000 | 00 MONTH 2016
ARTICLE
RESEARCH
Recent models were used to contrast our counts with predicted
viral–host relationships8,9. Weitz and Dushoff (2008)8, in which burst
size is proportional to density-dependent microbial growth rate, pre-
dicts a negative relationship between viral and host density as viral
predation causes declining host density with rising density-dependent
host growth rate (Fig. 1b; details of steady state solution in Materials
and Methods). The KtW-like model by Thingstad et al. (2014)9, that
incorporates terms for nested resistance to viral infection amongst
multiple host strains
9,13
, predicts an approximately downward concave
relationship between viral and host abundances with the increasing
dominance of slow growing, resistant hosts suppressing lytic dynam-
ics as host density rises (Fig. 1b). The Piggyback-the-Winner (PtW)
model introduced here predicts a relationship between VLP and host
densities similar to Thingstad et al. (2014)
9
, but lytic dynamics are sup-
pressed at high host density and density-dependent growth rate owing
to the increased prevalence of lysogeny (modelled as lower specific
viral production rates per infection) and super-infection exclusion
rather than resistance.
Diversity and functional composition of microbial
communities
Viral predation is thought to stimulate species-level host diversity
through lineage-specific predation targeting dominant lineages, pro-
moting community evenness5,25. However, when microbial diversity
in 66 microbiomes from across the Pacific was probed, a weak and
significantly negative relationship between host density and taxonomic
diversity was observed (Fig. 1c; microbial abundance log-transformed;
m = 0.29, t = 2.60, d.f. = 64, P = 0.01; R2 = 0.09; linear regression).
This also indicates that lytic dynamics are suppressed when both
density-dependent (that is, total encounter rates) and frequency-
dependent (that is, the relative density of a given host) would both
favour lytic activity.
A recent model suggests that elevated host densities lead to an
increase in host resistance to viral infection
9
. However, investigation
of 66 Pacific microbiomes yielded weak relationships and no sup-
port for increased host resistance via CRISPRs or potential horizon-
tal transfer of resistance (per cent competence genes) in the mixed
microbial community metagenomes (CRISPRs: Fig. 1d; m = 26.17,
t = 1.44, d.f. = 64, P = 0.15; R
2
= 0.03; per cent competence genes:
Extended Data Fig. 1a; m = 0.25, t = 2.40, d.f. = 64, P = 0.02;
R
2
= 0.08; microbial abundance log-transformed and linear regres-
sions in both analyses). These data indicate that immunity to viral
infection does not change with host density as predicted by Thingstad
et al. (2014)9, and is not promoted by horizontal transfer of resist-
ance genes as predicted in King-of-the-Mountain dynamics9,26.
Rather than host-mediated resistance to viral infection, the observed
decrease in VMR with increasing microbial abundance may be
driven by an alternative strain-level diversification mechanism,
such as increasing resistance via lysogeny. Lysogeny, with its implicit
super-infection immunity dynamic, would yield similar predictions
to Thingstad et al. (2014)9, albeit through a different mechanism, and
could complement the nested infection design of the Thingstad et al.
(2014)9 model in future studies of resistance/growth trade-offs.
Viral and host abundances in other ecosystems
Data from 22 independent studies were compiled for a meta- analysis
to determine the generality of the ‘more microbes, fewer viruses’
obser vation. These studies spanned five orders of magnitude of microbial
and VLP densities (Fig. 2; summary statistics in Extended Data Table 1;
references in References for Methods). Analysis of log-transformed
microbial and VLP abundances yielded slopes of significantly <1 in eight
of the eleven environments. VMR therefore declined with increasing
microbial density in disparate coastal and estuarine, coral reef, deep
ocean, open ocean, temperate lake, animal-associated, sediment, and soil
systems, consistent with our coral reef observations. This trend was also
observed in the cystic fibrosis lung
27
. Together, these results show that
‘more microbes, fewer viruses’ is a common phenomenon. When viewed
across the full range of host densities, peak VMR values were observed
at ~10
6
microbes ml
–1
or g
–1
of sample. VMR declines as host density
decreases or increases from this value (Fig. 2, final panel).
The relationship between microbial and viral densities was further
examined through an analysis of published values of the fraction of
lysogenic cells determined by mitomycin C induction
4,28–30
. Although
a sometimes-significant negative relationship exists at a within-study
level, examination across the full range of host abundances studied
revealed no significant slope (Extended Data Fig. 2). The model that
low host density favours lysogeny is not well supported by induction
data when viewed globally; there is reason to re-examine the drivers of
lysogeny with lines of evidence independent of established methods.
Experimental manipulation of host growth rate
Our counts data contrast with predicted density-dependent lytic
predation. Further, the models examined in Fig. 1b predict differ-
ent relationships between microbial density, density-dependent host
growth rate and viral lytic activity (measured as VMR). The actual
relationship between these variables was probed with incubation exper-
iments using seawater sampled from a pristine coral reef (Palmyra;
120-h time series) and a degraded embayment (Mission Bay; 72-h time
series). Data were pooled within sites as high variability led to a lack of
significant impact of dissolved organic carbon addition on host density
(t = 0.82, d.f. = 32.18, P = 0.42; Welch two sample t-test with microbial
0.10
0.25
0.50
0.75
1.00
2.50
5.00
VLPs per ml (× 107)
a
0.25
0.50
0.75
1.00
2.50
5.00
7.50
0.10
Microbes per ml (× 106)
Microbes per ml (× 10
6
) Microbes per ml (× 10
6
)
b
Viruses (log10; arbitrary units)
Microbes (log10; arbitrary units)
m = 0.59
Pm ≠ 1 < 0.01
R2 = 0.50
Observed Predicted
m = –0.29
R2 = 0.09
Pm ≠ 0 = 0.01
Host growth rate
Piggyback-the-Winner
Thingstad et al.
Weitz & Dushoff
m = –26.17
Pm ≠ 0 = 0.15
R2 = 0.03
d
150
100
50
0
CRISPR elements (p.p.m.)
0.50
0.75
1.00
2.50
5.00
0.50
0.75
1.00
2.50
5.00
c
Microbial species diversity (H)
3.25
3.50
3.75
Figure 1 | Virus-like particle (VLP) relative abundance declines with
increasing host density despite lower microbial diversity and similar
host sensitivity to infection, contrary to predictions of lytic models.
a, Log-transformed VLP versus microbial densities have an m < 1
relationship (n = 223 independent measures); the dashed reference line
depicts a 10:1 relationship. b, Steady-state microbial and viral abundances
and schematic microbial growth rate predicted by three modified
Lotka–Volterra models: Piggyback-the-Winner (red), Thingstad et al.
(2014; black9), and Weitz and Dushoff (2008; blue8). c, Shannon microbial
species diversity versus host density (H; n = 66 independent measures).
d, Abundance of CRISPR elements in the microbial metagenomes (n = 66
independent measures). All slopes (m), R2, and P values describe linear
regressions testing against a slope of 0, except a which shows the P value
from a two-sided t-test against a slope 1. Black best-fit lines with grey
99% prediction intervals from linear regressions are shown (a, c, and d).
© 2016 Macmillan Publishers Limited. All rights reserved
00 MONTH 2016 | VOL 000 | NATURE | 3
ARTICLE RESEARCH
abundance log-transformed; Extended Data Fig. 3a) or VMR (t = 0.17,
d.f. = 27.70, P = 0.87; Welch two sample t-test).
The experimental data matched our field observations: slopes
significantly <1 were observed between log-transformed VLP and
microbial densities in both incubations (Fig. 3a; Mission Bay: m = 0.56,
t = 6.65, d.f. = 10, P (t-test; m 1) = 3.59 × 104; R2 = 0.82; Palmyra:
m = 0.63, t = 4.20, d.f. = 23, P (t-test; m 1) = 2.25 × 10
2
; R
2
= 0.43;
linear regression). These incubations are therefore more similar to the
data set characterized by Wilcox and Fuhrman (1994)
3
as non-lytic
(Fig. 3c; m = 0.13, t = 0.64, d.f. = 26, P (t-test; m 1) = 2.65 × 10
4
;
R2 = 0.02; linear regression) than the Hennes et al. (1995)31
putatively lytic data set (Fig. 3c; m = 1.19, t = 1.59, d.f. = 4, P (t-test;
m 1) = 0.81; R2 = 0.39; linear regression). Hennes et al. (1995)31
and Wilcox and Fuhrman (1994)
3
attribute their lytic (where VMR
rose by ~40) and non-lytic dynamics to elevated and lowered micro-
bial densities, respectively. In contrast, we did not observe a similar
rise in VMR despite exceeding an order of magnitude higher host
densities (Fig. 3b and 3d) and five times faster net growth rates than
Hennes et al. (1995)
31
(9.71 × 10
6
and 1.77 × 10
6
cells h
1
in Palmyra
and Mission Bay incubations, respectively; Hennes et al. (1995)31:
1.74 × 10
6
cells h
1
). These protist predator-free incubations (that is,
0.8 μm filtered) showed no significant increase in VMR with increas-
ing host density, indicating that viral–host interactions alone are
sufficient to drive the approximately downward concave relationship
between VLP and host densities (Fig. 3a, b).
Temperate genes, diversity, and virulence
Metagenomes of viral communities (viromes) from 24 Pacific and Atlantic
coral reefs were sequenced (Extended Data Table 2). High variability and
high leverage points were observed in the relationship of all viral bioin-
formatic metrics and host density, requiring the use of robust regression
followed by bootstrap confidence interval estimation (RR-B) due to its
insensitivity to high leverage, peripheral values. The per cent abundance
of viral integrase, excisionase, and prophage reads increased significantly
with microbial density (Fig. 4a, b, and Extended Data Fig. 4a) at the 90%
confidence level (Fig. 4a, per cent integrase, m = 1.23, 90th percentile
confidence interval (CI; 0.01, 2.69), Fig. 4b, per cent excisionase,
m = 0.04, 90th CI (0.02, 0.10); Extended Data Fig. 4a, % prophage,
m = 0.13, 90th CI (0.03, 0.44); RR-B against log- transformed host
density; R2 are not appropriate for robust regressions and are omitted).
Increased cell density was associated with a significant decline in
functional diversity of the viral communities, an indicator of temperate
viral communities
32
, as measured by the Shannon (H) index of puta-
tive coding genes in the viromes (Fig. 4c, m = 3.54, 90th CI (6.14,
1.73); RR-B against log-transformed host density). Furthermore,
the lower diversity, more temperate viral communities carry more
virulence genes than the more diverse and lytic viral communities
found at lower host densities (Fig. 4d, m = 1.09, 90th CI (0.46, 3.01);
RR-B against log-transformed host density). Further, while we have
conservatively used linear regression to analyse these relationships, the
data suggests an exponential relationship between host density and %
integrase, % excisionase, and % virulence genes, and a decay function
between host density and viral functional diversity. These observed
trends were not a result of overall viral community genome size reduc-
tion, as the average viral genome size determined by the Genome
relative Abundance and Average Size tool was unaffected by micro-
bial abundance (m = 9190, t = 1.02, d.f. = 22, P = 0.32; R
2
= 0.04;
linear regression of genome length (bp) against log-transformed
host density; mean estimated viral genome length = 42.07 ± 2.45 kb,
mean ± s.e.m.). Rather, viral communities changed with increasing
temperateness; viral communities clustered geographically as low-
cell-density Atlantic viromes grouped away from Pacific viromes
(Extended Data Fig. 4b).
Discussion
The observed decline in the virus to microbe ratio (VMR) with ele-
vated host abundances on microbialized coral reefs (Fig. 1) is consistent
with lowered lytic activity at high host density (Fig. 1b). This trend
was observed in eight of eleven (>70%) other disparate environments
(Fig. 2). No support was found for competitive exclusion of viral pred-
ators by heterotrophic protists6,33,34 (Fig. 3a), the rise or transfer of
resistance to viral infection9,26 (Fig. 1d and Extended Data Fig. 1b),
Microbes per ml or g of sampl
e
VLPs per ml or
g
of sampl
e
10
10
10
10
6
10
10
10
10
6
10
10
10
10
6
10
4
10
6
10
8
10
10
10
4
1
0
6
10
8
10
10
1
0
4
10
6
1
0
8
10
10
10
4
10
6
10
8
10
10
Soil
m
= –0.0
7
P
m
P
1
<
0
.
01
R
2
=
0
.
01
= 0.31
< 0.0
1
R =
0
.1
3
Temperate lake/river
= 0.9
3
= 0.15
R
=
0
.
93
Soil pore water
Animal
m
=
0
.
39
P
m
P
1
<
0
.
01
R
2
=
0
.
44
m
m
=
0
.
69
P
m
P
1
=
0
.
02
R
2
=
0
.
18
Coastal/estuarine Deep ocean
m
=
0
.
81
P
m
P
1
=
0
.
03
R
2
=
0
.
8
7
Coral reef
0
50
1
00
150
VMR
m = 0.2
0
P
m
P
1
<
0
.
01
R
2
=
0
.
11
Sediment
Drinking water
= 0.8
5
=
0
.2
6
= 0.70
Open ocean
= 0.6
9
< 0.0
1
= 0.57
= 1.62
<
0
.
01
= 0.5
9
Polar lakes
All ecosystems
Figure 2 | The relative decline in virus-like particles (VLPs) with
increasing host density is common in disparate environmental systems.
Published microbial and VLP densities, and calculated virus to microbe
ratio (with all environments pooled; final panel) are plotted by ecosystem.
n = 23, 139, 27, 18, 22, 1397, 85, 71, 18, 35, and 46 independent measures
for Animal-associated, Coastal/estuarine, Coral reef, Deep ocean,
Drinking water, Open ocean, Polar lakes, Sediment, Soil, Soil pore water,
and Temperate lake/river environments, respectively; pooled n = 1,881.
Dashed lines depict 10:1 linear relationships; blue lines of best fit and pink
99% prediction intervals from linear regression are shown. All slopes (m)
and R2 values describe linear regressions, and P values are derived from a
two-sided t-test against a slope 1; details, including false-detection rate
corrected values in Extended Data Table 1.
© 2016 Macmillan Publishers Limited. All rights reserved
4 | NATURE | VOL 000 | 00 MONTH 2016
ARTICLE
RESEARCH
greater species-level host diversity
5
(Fig. 1c), or increasing viral decay
5
(Extended Data Fig. 3b). Rather, multiple independent bioinformatic
analyses of our viromes from this study, reinforced by viromes from
other ecosystems35–39, indicated an increased relative abundance of
temperate viruses in communities with high microbial densities (Fig. 4).
Empirical tests of alternative models to Piggyback-the-Winner (PtW)
showed weak or ambiguous relationships while correlations sup-
porting PtW were significant (for example, R2 of <0.09 in Fig. 1c, d
and Extended Data Fig. 2, compared with R2 > 0.56 in Figs 1a and
3a). All four independent lines of evidence examined here—direct
counts, literature meta-analyses, experiments, and viral community
metagenomics—provide significant support for PtW.
The established model in viral ecology predicts that lytic dynamics
dominate at high host density, whereas lysogeny is favoured at low
densities4,14,17. We propose an extension of these Kill-the-Winner
(KtW) models, Piggyback-the-Winner (PtW), wherein temperateness
is favoured at high host densities as viruses exploit their hosts through
lysogeny rather than killing them. As viral and host densities increase,
lysogen resistance to superinfection by related viruses becomes
increasingly important
9
. In this scenario, the energetic costs of gen-
erating resistance to infection through carrying proviruses should be
less than through mutation40,41. Further, lysogeny can decouple micro-
bial taxonomic and functional composition through horizontal gene
transfer
42
. Virulence genes encode functions that harm eukaryotes;
the increasing virulence content of viral communities under PtW
dynamics (Fig. 4d) suggests that lysogens could evade protistan
predation in addition to viral lysis. Suppressed top-down viral
and protistan predation under PtW dynamics is likely to facilitate
microbialization and ecosystem decline22,23,43.
The ‘narwhal-shaped’ distribution that results when VMR is plotted
against microbe density in multiple environments (Fig. 2, final panel)
suggests that host densities observed in the ocean (~5 × 10
5
to 1 × 10
6
cells ml1) favour lytic KtW dynamics. Lower and higher host densities
show a suppressed VMR. Thus, we predict that a Piggyback-the-Losers
(PtL) dynamic extends the lytic-to-temperate shift to communities
with low host densities. The diversity of environments across which
the PtL–KtW–PtW dynamic is observed suggests that whichever
viral–host dynamic prevails within a system, PtL, KtW or PtW, has
major effects on processes as diverse as ecosystem function and disease
progression34,43–46.
Online Content Methods, along with any additional Extended Data display items and
Source Data, are available in the online version of the paper; references unique to
these sections appear only in the online paper.
Received 19 May 2015; accepted 3 February 2016.
Published online 16 March 2016.
1. Suttle, C. A. Marine viruses — major players in the global ecosystem.
Nature Rev. Microbiol. 5, 801–812 (2007).
2. Proctor, L. M. & Fuhrman, J. A. Viral mortality of marine bacteria and
cyanobacteria. Nature 343, 60–62 (1990).
3. Wilcox, R. M. & Fuhrman, J. A. Bacterial viruses in coastal seawater: lytic rather
than lysogenic production. Mar. Ecol. Ser. 114, 35–45 (1994).
4. Payet, J. & Suttle, C. A. To kill or not to kill: the balance between lytic and
lysogenic viral infection is driven by trophic status. Limnol. Oceanogr. 58,
465–474 (2013).
5. Wommack, K. E. & Colwell, R. R. Virioplankton: viruses in aquatic ecosystems.
Microbiol. Mol. Biol. Rev. 64, 69–114 (2000).
6. Thingstad, T. F. Elements of a theory for the mechanisms controlling
abundance, diversity, and biogeochemical role of lytic bacterial viruses
in aquatic systems. Limnol. Oceanogr. 45, 1320–1328 (2000).
a
VLPs per mlVLPs per ml
1010
108
108
107
106
105
108
107
106
105
106
104
1010
108
106
104
105106107108
Microbes per ml
c
b
VMR
50
40
30
20
10
0
60
Microbes per ml
050 100 150
Incubation time (h)
Microbes per ml
d
Palmyra
m = 0.63
Pm ≠ 1 = 0.02
R2 = 0.43
m = 0.56
Pm ≠ 1 < 0.01
R2 = 0.82
Mission Bay
Individual incubations
m = 0.13
Pm ≠ 1 < 0.01
R2 = 0.02
Wilcox & Fuhrman
Published mean values
m = 1.19
Pm ≠ 1 = 0.81
R2 = 0.39
Hennes et al.
Figure 3 | Density dependence does not drive viral predation. a, c, Viral
and host densities (individual counts shown) follow an m < 1 distribution
despite high host densities (a; Mission Bay (stars) and Palmyra (circles)
incubations; n = 12 and 25, respectively, from repeated measures),
compared to putatively ‘lytic’ (slope = 1) and ‘non-lytic’ (m < 1) published
data (c; Hennes et al. (1995)31 (triangles) and Wilcox and Fuhrman (1994)3
(squares), respectively; mean values shown; n = 6 and 28, respectively,
from repeated measures). b, d, Microbe density and VMR over time in
Mission Bay and Palmyra (b; individual values; n = 3 and 5 per time-point,
respectively, except for time zero, when n = 1), and published putative lytic
and non-lytic incubations (d; mean values; n = 1 and 4 per time-point,
respectively) plotted over a thin plate spline. a, c, Dashed 10:1 lines, solid
lines of best fit, with 99% prediction intervals in grey; all slopes (m) and
R2 values describe linear regressions, and P values are derived from a
two-sided t-test against a slope 1. Individual incubation data are
shown in Extended Data Fig. 3a. Mission Bay and Palmyra incubation
experiments were each conducted once.
a
cd
Microbes per ml (× 106) Microbes per ml (× 106)
Microbes per ml (
×
106) Microbes per ml (
×
106)
0.50
1.00
0.25
2.50
5.00
0.75
2
0
3
1
Viral per cent integrase reads
y = 1.23x – 6.59
95% CI (–0.07, 3.07)
90% CI (0.01, 2.69)
0.50
1.00
0.25
2.50
5.00
0.75
0.05
0.10
0
0.15
Viral per cent excisionase reads
y = 0.04x – 0.22
95% CI (0.01, 0.11)
90% CI (0.02, 0.10)
0.50
1.00
0.25
2.50
5.00
0.75
8
10
6
12
Viral functional diversity (H)
y = –3.54x + 29.65
95% CI (–6.77, –0.49)
90% CI (–6.14, –1.73)
0.50
1.00
0.25
2.50
5.00
0.75
2
0
3
1
Viral per cent virulence reads
y = 1.09x – 5.83
95% CI (0.37, 3.46)
90% CI (0.46, 3.01)
b
Figure 4 | Temperate features in viromes increase with host density.
ad, The relationship between log-transformed microbial density and the
percent abundance of integrase (a), excisionase (b), and virulence reads in
viromes (d), normalized by total sequences in each sample, and Shannon
(H) viral functional diversity (c) (n = 24 independent measures for all
analyses). The linear equations and lines of best fit from robust regression
and bootstrapped 95% and 90% confidence intervals (CIs) for the slopes
are shown. Goodness of fit metrics are inappropriate for robust regression
and are omitted.
© 2016 Macmillan Publishers Limited. All rights reserved
00 MONTH 2016 | VOL 000 | NATURE | 5
ARTICLE RESEARCH
7. Rodriguez-Brito, B. et al. Viral and microbial community dynamics in four
aquatic environments. ISME J. 4, 739–751 (2010).
8. Weitz, J. S. & Dusho, J. Alternative stable states in host–phage dynamics.
Theor. Ecol. 1, 13–19 (2008).
9. Thingstad, T. F., Våge, S., Storesund, J. E., Sandaa, R.-A. & Giske, J. A theoretical
analysis of how strain-specic viruses can control microbial species diversity.
Proc. Natl Acad. Sci. USA 111, 7813–7818 (2014).
10. Weinbauer, M. G. & Höe, M. G. Signicance of viral lysis and agellate grazing
as factors controlling bacterioplankton production in a eutrophic lake.
Appl. Environ. Microbiol. 64, 431–438 (1998).
11. Bratbak, G., Egge, J. K. & Heldal, M. Viral mortality of the marine alga Emiliania
huxleyi (Haptophyceae) and termination of algal blooms. Mar. Ecol. Prog. Ser.
93, 39–48 (1993).
12. Evans, C. & Brussaard, C. P. D. Viral lysis and microzooplankton grazing of
phytoplankton throughout the Southern Ocean. Limnol. Oceanogr. 57,
1826–1837 (2012).
13. Needham, D. M. et al. Short-term observations of marine bacterial and viral
communities: patterns, connections and resilience. ISME J. 7, 1274–1285
(2013).
14. Jiang, S. C. & Paul, J. H. Signicance of lysogeny in the marine environment:
studies with isolates and a model of lysogenic phage production. Microb. Ecol.
35, 235–243 (1998).
15. Paul, J. H. Prophages in marine bacteria: dangerous molecular time bombs or
the key to survival in the seas? ISME J. 2, 579–589 (2008).
16. Paul, J. H. & Weinbauer, M. Detection of lysogeny in marine environments.
Man. Aquat. Viral Ecol. 4, 30–33 (2010).
17. Maurice, C. F., Bouvier, C., de Wit, R. & Bouvier, T. Linking the lytic and lysogenic
bacteriophage cycles to environmental conditions, host physiology and their
variability in coastal lagoons. Environ. Microbiol. 15, 2463–2475 (2013).
18. Dinsdale, E. A. et al. Microbial ecology of four coral atolls in the Northern
Line Islands. PLoS One 3, e1584 (2008).
19. Smith, J. E. et al. Indirect eects of algae on coral: algae-mediated, microbe-
induced coral mortality. Ecol. Lett. 9, 835–845 (2006).
20. Thurber, R. V. et al. Metagenomic analysis of stressed coral holobionts.
Environ. Microbiol. 11, 2148–2163 (2009).
21. Kelly, L. W. et al. Black reefs: iron-induced phase shifts on coral reefs. ISME J. 6,
638–649 (2012).
22. McDole, T. et al. Assessing coral reefs on a Pacic-wide scale using the
microbialization score. PLoS One 7, e43233 (2012).
23. Barott, K. L. & Rohwer, F. L. Unseen players shape benthic competition on
coral reefs. Trends Microbiol. 20, 621–628 (2012).
24. Alongi, D. M. et al. Phytoplankton, bacterioplankton and virioplankton structure
and function across the southern Great Barrier Reef shelf. J. Mar. Syst. 142,
25–39 (2015).
25. Thingstad, T. F. & Lignell, R. Theoretical models for the control of bacterial
growth rate, abundance, diversity and carbon demand. Aquat. Microb. Ecol. 13,
19–27 (1997).
26. Giovannoni, S., Temperton, B. & Zhao, Y. Giovannoni et al. reply. Nature 499,
E4–E5 (2013).
27. James, C. E. et al. Lytic activity by temperate phages of Pseudomonas
aeruginosa in long-term cystic brosis chronic lung infections. ISME J. 9,
1391–1398 (2015).
28. Muck, S. et al. Fracture zones in the Mid Atlantic Ridge lead to alterations in
prokaryotic and viral parameters in deep-water masses. Front. Microbiol. 5,
264 (2014).
29. Bongiorni, L., Magagnini, M., Armeni, M., Noble, R. & Danovaro, R. Viral
production, decay rates, and life strategies along a trophic gradient in the
North Adriatic Sea. Appl. Environ. Microbiol. 71, 6644–6650 (2005).
30. Williamson, S. J., Houchin, L. A., McDaniel, L. & Paul, J. H. Seasonal variation
in lysogeny as depicted by prophage induction in Tampa Bay, Florida.
Appl. Environ. Microbiol. 68, 4307–4314 (2002).
31. Hennes, K. P., Suttle, C. A. & Chan, A. M. Fluorescently labeled virus probes
show that natural virus populations can control the structure of marine
microbial communities. Appl. Environ. Microbiol. 61, 3623–3627 (1995).
32. McDaniel, L. D., Rosario, K., Breitbart, M. & Paul, J. H. Comparative
metagenomics: natural populations of induced prophages demonstrate highly
unique, lower diversity viral sequences. Environ. Microbiol. 16, 570–585 (2014).
33. Fuhrman, J. A. Microbial community structure and its functional implications.
Nature 459, 193–199 (2009).
34. Weitz, J. S. et al. A multitrophic model to quantify the eects of marine viruses on
microbial food webs and ecosystem processes. ISME J. 9, 1352–1364 (2015).
35. Breitbart, M. et al. Metagenomic analyses of an uncultured viral community
from human feces. J. Bacteriol. 185, 6220–6223 (2003).
36. Reyes, A. et al. Viruses in the faecal microbiota of monozygotic twins and
their mothers. Nature 466, 334–338 (2010).
37. Angly, F. E. et al. The marine viromes of four oceanic regions. PLoS Biol. 4, e368
(2006).
38. Breitbart, M. et al. Diversity and population structure of a near-shore
marine-sediment viral community. Proc. R. Soc. Lond. B 271, 565–574
(2004).
39. Brum, J. R., Hurwitz, B. L., Schoeld, O., Ducklow, H. W. & Sullivan, M. B.
Seasonal time bombs: dominant temperate viruses aect Southern Ocean
microbial dynamics. ISME J. 10, 437–449 (2015).
40. Våge, S., Storesund, J. E. & Thingstad, T. F. Adding a cost of resistance
description extends the ability of virus-host model to explain observed
patterns in structure and function of pelagic microbial communities.
Environ. Microbiol. 15, 1842–1852 (2013).
41. Avrani, S. & Lindell, D. Convergent evolution toward an improved growth rate
and a reduced resistance range in Prochlorococcus strains resistant to phage.
Proc. Natl Acad. Sci. USA 112, E2191–E2200 (2015).
42. Kelly, L. W. et al. Local genomic adaptation of coral reef-associated
microbiomes to gradients of natural variability and anthropogenic stressors.
Proc. Natl Acad. Sci. USA 111, 10227–10232 (2014).
43. Silveira, C. B. et al. Microbial and sponge loops modify sh production in
phase-shifting coral reefs. Environ. Microbiol. 17, 3832–3846 (2015).
44. Barr, J. J. et al. Bacteriophage adhering to mucus provide a non-host-derived
immunity. Proc. Natl Acad. Sci. USA 110, 10771–10776 (2013).
45. Whiteson, K. L. et al. The upper respiratory tract as a microbial source for
pulmonary infections in cystic brosis. Parallels from island biogeography.
Am. J. Respir. Crit. Care Med. 189, 1309–1315 (2014).
46. Willner, D. et al. Metagenomic analysis of respiratory tract DNA viral
communities in cystic brosis and non-cystic brosis individuals. PLoS One 4,
e7370 (2009).
Acknowledgements This paper is dedicated to the memory of Mike Furlan,
mentor, friend, and colleague. We are grateful to the National Oceanographic
and Atmospheric Administration Coral Reef Ecosystem Division for supporting
this research, and to the captains and crews of the NOAA ship Hi’ialakai and the
Hanse Explorer. Thanks to J. Payet for providing viral and microbial abundance
data. Sampling was carried out under research permits from the US Fish and
Wildlife Service, Palmyra Atoll National Wildlife Refuge, the Environment and
Conservation Division of the Republic of Kiribati (n. 021/13) and ICMBio, Brazil
(n. 27147-2). This work was funded by the Canadian Institute for Advanced
Research Integrated Microbial Biodiversity Program Fellowship Award 141679
(to F.R.) and National Science Foundation grants OISE-1243541 and
DEB-1046413 (to F.R.), CNS-1305112 and MCB-1330800 (to R.A.E.), DUE-
1323809 (to E.A.D.), Gordon and Betty Moore Foundation Investigator Award
GBMF-3781 (to F.R.), and the Brazilian National Research Council (CNPq; to F.T.)
and Brazilian National Research Council Science Without Borders Program
(CNPq/CAPES; to C.B.S.).
Author Contributions F.R., B.K., C.B.S., and F.T. conceptualized the project; B.K.,
F.R., C.B.S., and M.Y. wrote the manuscript; B.K., C.B.S., V.A.C., A.G.C.-G., K.T.G,
K.M., G.G.Z.S., S.D.Q., Y.W.L., S.E.S., F.H.C., E.R.H. , N.L.R., B.A.B., B.F., A.L., P.S., J.N.,
C.Y., E.E.G., M.L., K.A.F., L.S.O., T.M.-S., J.M.H., B.Z., A.F.H., M.J.A.V., K.B., C.S., R.A.E.,
and F.R. performed sample collection, processing, experiments, and analysis;
N.H. provided graphics and GIS analysis; E.A.D., L.W.K., S.S., J.S., R.B., C.T., G.B.G.,
J.N., E.S., R.A.E., F.T., and F.R. provided intellectual guidance and funding during
the development of the research.
Author Information The viromes and microbiomes used in this paper are
accessible at MG-RAST (http://metagenomics.anl.gov/) under the Piggyback-
the-Winner project. Virome accession numbers: 4683670.3, 4683674.3,
4683677.3, 4683680.3, 4683683.3, 4683684.3, 4683686.3, 4683690.3,
4683702.3, 4683703.3, 4683704.3, 4683706.3, 4683712.3, 4683720.3,
4683739.3, 4683744.3, 4683745.3, 4683746.3, 4683747.3, 4683731.3,
4683733.3, 4683734.3, 4683718.3, 4684617.3. Microbiome accession
numbers: 4683666.3, 4683667.3, 4683668.3, 4683669.3, 4683671.3,
4683672.3, 4683673.3, 4683675.3, 4683676.3, 4683678.3, 4683679.3,
4683681.3, 4683682.3, 4683685.3, 4683687.3, 4683688.3, 4683689.3,
4683691.3, 4683692.3, 4683693.3, 4683694.3, 4683695.3, 4683696.3,
4683697.3, 4683698.3, 4683699.3, 4683700.3, 4683701.3, 4683705.3,
4683707.3, 4683708.3, 4683709.3, 4683710.3, 4683711.3, 4683713.3,
4683714.3, 4683715.3, 4683716.3, 4683717.3, 4683719.3, 4683721.3,
4683722.3, 4683723.3, 4683724.3, 4683725.3, 4683726.3, 4683727.3,
4683728.3, 4683729.3, 4683732.3, 4683735.3, 4683736.3, 4683737.3,
4683738.3, 4683740.3, 4683741.3, 4683742.3, 4683743.3, 4683748.3,
4683749.3, 4683750.3, 4683751.3, 4683752.3, 4683753.3, 4683754.3,
4684616.3. Reprints and permissions information is available at www.nature.
com/reprints. The authors declare no competing financial interests. Readers are
welcome to comment on the online version of the paper. Correspondence and
requests for materials should be addressed to F.R. (frohwer@gmail.com) or
B.K. (benjaminwilliamknowles@gmail.com).
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE
RESEARCH
METHODS
Viral and microbial counts. Seawater was collected in 2-l diver-deployed Niskin
bottles at approximately 10 m depth within 30 cm of the benthos on coral reefs
across the Pacific and Atlantic O ceans
47
. Samples were fixed with 2% final concen-
tration paraformaldehyde within four hours of collection. Pacific Ocean samples
were filtered and stained with SYBR Gold (Life Technologies, USA), mounted on
slides and analysed by epifluorescence microscopy
47
. Atlantic Ocean samples were
flash frozen and stored in liquid nitrogen unt il analysis on a BD FACSCalibur flow
cytometer
48
. Investigators were blinded when conducting all counts in this study
(environmental or experimental), with sites or incubation samples imaged and
analysed in a random order and identified only after analysis.
Predator–prey modelling. Steady state solutions to the dynamic model of Weitz
and Dushoff (2007)
8
were calculated under varying carrying capacities (K). The
chemostat model of Thingstad et al. (2014)9 was run for varying K, and the final
point in the evolution of the system plotted.
A standard lytic model49 that incorporates a logistic or trophic-state dependence
for the microbe growth rate r is given by the equations
δN/δt = r • (1 N/K) • N (dN) (φNV)
δV/δt = (βφN/KNV) (mV)
where d and m are, respectively, the trophic-independent death rates for microbes
and phage, N and V are, respectively, microbial host and viral abundances, β is
the burst size, and φ is the adsorption coefficient. This corresponds to the Weitz–
Dushoff model8 with their parameter a (the fractional reduction of lysis at carrying
capacity term) set equal to 0.
In this case the specific viral production rate per microbe is given by the product
βφ. In the new PtW model of viral–host interactions proposed here we replace
this product with the quantity βφN/K, suppressing viral production as the
system moves away from K (that is, N/K becomes smaller) to simulate augmenta-
tion of lysogeny in eutrophic conditions. In this case βφ has the interpretation
as the maximum value for the specific viral production rate per microbe. Steady
state solutions of host and viral densities in the PtW model generated herein were
calculated across a range of K (Fig. 1b). All models are available as Matlab scripts
from https://github.com/benjaminwilliamknowles/Piggyback-the-Winner.
Meta-analysis of cell and viral abundances. The relationships between published
VLP and cell abundances from disparate environments were probed from 22
studies17,28,44,50–68. When abundances were not available, we used the
WebPlotDigitizer tool to recover data from graphs (http://arohatgi.info/
WebPlotDigitizer/app/). Samples were grouped by habitat: animal-associated,
polar lakes, coastal/estuarine, coral reefs, deep ocean, drinking water, open ocean,
sediment, soil, soil water and temperate lake/river. We similarly extracted data from
published studies and tested the relationship between cell abundance and the fre-
quency of lysogenic cells as studied by mitomycin C induction in previous studies
from the Adriatic Basin, Arctic Shelf, Mid Atlantic Ridge and Tampa Bay4,28–30.
Metaviromic sampling and processing. Viral metagenomic samples were col-
lected at 24 reefs (Extended Data Table 2), a subset of sites sampled for counts
as previously described
47
. Pacific viral concentrates were treated with 250 μl of
chloroform per 50 ml of concentrate to destroy microbes and purified using CsCl
step gradient ultracentrifugation47. Viral DNA was extracted using the formamide/
phenol/chloroform isoamyl alcohol technique47 and amplified using the Linker
Amplified Sequencing Library approach69 and sequenced on an Illumina MySeq
platform (Illumina, USA). Atlantic viral concentrates were passed through a
0.22 μm filter and 250 μl of chloroform per 50 ml of concentrate was added to
remove microbes, followed by ult racentrif ugation for further concentration. DNA
from Atlantic sites was extracted by the phenol/chloroform/is oamyl alcohol tech-
nique, amplified using multiple displ acement amplification
20
and sequenced on an
Ion Torrent sequencer (Life Sciences, USA). Microbial metagenomes were prepared
by DNA extraction from the >0.22 μm fraction of the microbial community using
Nucleospin Tissue Extraction kits (Macherey Nagel, Germany)47 and sequencing
on an Illumina MySeq platform (Illumina, USA).
Bioinformatics. Sequences less than 100 bp and with mean quality scores less
than 25 were removed using PrinSeq
70
. Acceptable sequences were then derepli-
cated with TagCleaner71 and potential contaminants matching lambda or human
DNA sequences removed with DeconSeq72. Focusing on microbial reads, micro-
bial metagenomes were taxonomically annotated based on k-mer similarity
using FOCUS
73
. Rank-abundance tables were then used to calculate microbial
species-level Shannon (base e) taxonomic diversity. For the virome analysis, protein
sequences of all integrase, excisionase, and competence gene sequences on the
NCBI RefSeq database (http://www.ncbi.nlm.nih.gov/refseq/) were downloaded
and made into BLAST databases (makeblastdb command; BLAST version 2.2.29+,
ftp://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/). The Virulence
Factors of Pathogenic Bacteria Database (http://www.mgc.ac.cn/VFs/main.htm)
was used as a protein database for virulence genes. The percentage of each
sequence library composed of integrase, excisionase, competence, or virulence
genes was computed as the number of sequences with >60 bp match at a 40%
identity to database sequences identified using BLASTx, normalized by the total
number of sequences in the virome. CRISPRs were identified in microbiomes
using the CRISPR Recognition Tool (https://github.com/ajmazurie/CRT) and hits
normalized to parts per million (p.p.m.) against tot al re ads. The fraction of known
prophage-like reads in the viromes, normalized by total sequences, was assessed by
a stringent (e-value 10
10
) BLAST against known prophages in cultured bacteria
downloaded from NCBI (hosts (number of prophage): Escher ichia coli (36), Shigella
flexneri (31), Salmonella enterica (16), Staphylococcus aureus (14), Xylella fastidiosa
(12), Yersinia pseudotuberculosis (11), Yersinia pestis (9), Shewanella baltica (8),
Streptococcus pyogenes (7), Pseudomonas syringae (7), Salmonella typhimurium
(6), Xanthomonas campestris (5), Mycobacterium tuberculosis (4), Yersinia enter-
ocolitica (3), Streptococcus agalactiae (3), Stenotrophomonas maltophilia (3),
Pseudomonas putida (3), Staphylococcus haemolyticus (2), Streptomyces avermitilis
(1), Streptococcus uberis (1), Listeria monocytogenes (1), Caulobacter sp. (1)). For
functional diversity analysis, reads of each virome were assembled using MIRA74
followed by ORF calling using FragGenScan
75
and ORF clustering at 85% identit y
using CD-HIT76 to build protein cluster databases. We then performed BLASTx
of reads against clusters databases to assess the number of reads assigned to each
protein cluster. An OTU-like table was built using each cluster as a rank unit and
read counts as abundance. Shannon (base e) indexes were calculated using the
VEGAN package in R (http://cran.r-project.org/web/packages/vegan/index.html).
Average viral genome size estimates were performed using GAAS
77
and virome
clustering was performed using crAss78.
Bioinformatic code availability. The following codes and parameters were used
for each step of the viral functional diversity analysis:
Assembly parameters used in Mira:
minimum overlap = 30 and minimum relative score = 90.
FragGeneScan code: ./run_FragGeneScan.pl -genome=[seq_file_name] -out=
[output_file_name] -complete=0 -train=illumina_10.
CD-HIT code: cd-hit –i [input fastafilename.faa] -o [outputfi lename]_85 -c 0.85 -n 5
CD-HIT output was used as database for BLASTx with virome reads, and output
format 6 was parsed with the following python script to create rank-abundance tables:
f="BlastOutput.txt"
myfile=open(f)
h={};temp=""
for line in myfile:
line=line.split()
if temp!=line[0]:
if line[1] not in h:
h[line[1]]=0
h[line[1]]+=1
temp=line[0]
myfile.close()
Incubation experiments. Water samples were collected at Palmyra Atoll, a pristine
coral reef in the central Pacific, and Mission Bay, a degraded embayment in San
Diego, CA. Samples were twice filtered t hrough 0.8 μm pre-combusted GF/F filters
to remove protists. Palmyra water was subsampled in 100-ml aliquots and distrib-
uted in 12 Whirl-Pak bags (Cole-Parmer, IL, USA), divided into two experimental
groups and one control, each one containing four randomly chosen replicate bags.
For the two experimental groups we added a DOC cocktail containing 48 different
labile carbon sources79 at the final concentration of 500 μM or 60 μM (+DOC
treatment; Extended Dat a Fig. 3a), while no DOC was added to the control group
(DOC treatment; Extended Data Fig. 3a). Viral decay in microbe-f re e incubation
bags was monitored as an additional control with 0.22 μm double-filtered water
samples (Extended Data Fig. 3b). 1 ml samples were taken at times 0 h, 24 h, 48 h,
72 h, and 120 h from each bag for cell and viral counts. Mission Bay water was
filtered and separated in three groups as above. 250 ml aliquots were distributed
in each bag and incubated with 0 μM, 1 μM or 100 μM final concentrations by
DOC addition. Samples were taken at times 12 h, 24 h, 48 h, and 72 h for counts. All
incubations were performed in the dark at 25 °C. Samples were fixed and analysed
by epifluorescence microscopy as described above.
Statistical analysis. No statistical methods were used to predetermine sample
size. Significance was determined using an alpha of 0.05 when direct counts data
were compared, and using an alpha of 0.1 when analysing counts versus bioinfor-
matic analyses to account for the disparate nature of t hese data sets (although 95%
prediction intervals are also shown). The relationship between microbial density
and microbial diversity, CRISPR sequences, and competence genes were tested
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE RESEARCH
for significant deviation from a slope of 0 by linear regression. The relationship
between VLP and microbial densities in Figs 1a, 2 (al l except the final panel show-
ing VMR), and 3a, c were tested for slopes significantly different to 1 by t-tests that
tested the null hypothesis that the slope is not equal to 1 against the two-sided
alternative; the corresponding P value is given for this test. The fdrtool package
in R was used to provide false discovery rate-corrected (FDR) P values for the
multiple comparisons conducted in Fig. 2 (Extended Data Table 1). Conclusions
were similar between FDR and uncorrected analyses. Experimental data in
Fig. 3 was complemented by average counts taken from previous studies3,31 using
the WebPlotDigitizer tool. Data was taken from the nutrient added treatment of
Hennes et al. (1995)31 as it was described as showing ‘lytic’ dynamics (Fig. 3c, d).
Data from the ‘non-lytic’ 30%, 20%, 10%, and 3% dilutions by Wilcox and Fuhrman
(1994)3 were used as they had encounter rates (the product of viral and host
densities) ~10
12
or less at the beginning of the incubations, described as the cutoff
below which lytic dynamics were not sustained. A thin plate spline was applied to
experimental and literature values for visualization and interpretation (Fig. 3b, d).
While some data sets used to examine alternatives to PtW and published values in
Figs 1c, d, 2, 3c, d, Extended Data Fig. 1a, and Extended Data Fig. 2, violated the
assumptions of linear regression, this analysis was used for comparability. Robust
regressions were used in Fig. 4 and Extended Data Fig. 4a analyses in order to
accommodate high-leverage samples on parametric statistical models, allowing all
samples to be retained in the analysis. Results are presented for robust regression
estimation using Tukey’s biweight and corresponding bootstrapped 90th percentile
and 95th percentile confidence intervals (90% and 95% CIs) for the slope using
1,000 bootstrap replications. It should be noted for Fig. 4a that even though 95% CI
covers 0, the 90% CI does not cover 0 indicating that there is evidence at the 0.1
confidence level that the slope is positive. For subsequent analyses in Fig. 4, 95%
CIs do not straddle 0, showing that there is evidence at the 0.05 confidence level
that the slope is negative (Fig. 4c) or positive (Fig. 4b, d). To account for error in
the y axis we also performed Model II regression analyses with data shown in
Figs 1, 2 and 4 using the package lmodel2 in R (Extended Data Table 3). It should be
noted, however, that these results should be treated with caution, as error variance
and goodness of fit metrics are not obtainable for this analysis.
47. Haas, A. F. et al. Unraveling the unseen players in the ocean - a eld guide to
water chemistry and marine microbiology. J. Vis. Exp. 93, e52131 (2014).
48. Brussaard, C. P. D., Payet, J. P., Winter, C. & Weinbauer, M. G. Quantication
of aquatic viruses by ow cytometry. Man. Aquat. Viral Ecol. 11, 102–109
(2010).
49. Murray, J. D. Mathematical biology: I. an introduction. (Springer, 2002).
50. Amossé, J. et al. The ows of nitrogen, bacteria and viruses from the soil to
water compartments are inuenced by earthworm activity and organic
fertilization (compost vs. vermicompost). Soil Biol. Biochem. 66, 197–203
(2013).
51. Bettarel, Y., Bouvy, M., Dumont, C. & Sime-Ngando, T. Virus-bacterium
interactions in water and sediment of West African inland aquatic systems.
Appl. Environ. Microbiol. 72, 5274–5282 (2006).
52. Bouvier, T. & Maurice, C. F. A single-cell analysis of virioplankton adsorption,
infection, and intracellular abundance in dierent bacterioplankton physiologic
categories. Microb. Ecol. 62, 669–678 (2011).
53. Glud, R. N. & Middelboe, M. Virus and bacteria dynamics of a coastal
sediment: implication for benthic carbon cycling. Limnol. Oceanogr. 49,
2073–2081 (2004).
54. Furlan, M. Viral and microbial dynamics in the human respiratory tract.
(San Diego State Univ., 2009).
55. Hewson, I., O’Neil, J. M., Fuhrman, J. A. & Dennison, W. C. Virus-like particle
distribution and abundance in sediments and overlying waters along
eutrophication gradients in two subtropical estuaries. Limnol. Oceanogr. 46,
1734–1746 (2001).
56. Kim, M.-S., Park, E.-J., Roh, S. W. & Bae, J.-W. Diversity and abundance of
single-stranded DNA viruses in human feces. Appl. Environ. Microbiol. 77,
8062–8070 (2011).
57. Lisle, J. T. & Priscu, J. C. The occurrence of lysogenic bacteria and microbial
aggregates in the lakes of the McMurdo Dry Valleys, Antarctica. Microb. Ecol.
47, 427–439 (2004).
58. Laybourn-Parry, J., Marshall, W. A. & Madan, N. J. Viral dynamics and patterns
of lysogeny in saline Antarctic lakes. Polar Biol. 30, 351–358 (2006).
59. Madan, N. J., Marshall, W. a. & Laybourn-Parry, J. Virus and microbial loop
dynamics over an annual cycle in three contrasting Antarctic lakes.
Freshw. Biol. 50, 1291–1300 (2005).
60. Maurice, C. F., Bouvier, T., Comte, J., Guillemette, F. & Del Giorgio, P. A. Seasonal
variations of phage life strategies and bacterial physiological states in three
northern temperate lakes. Environ. Microbiol. 12, 628–641 (2010).
61. Maurice, C. F. et al. Disentangling the relative inuence of bacterioplankton
phylogeny and metabolism on lysogeny in reservoirs and lagoons. ISME J. 5,
831–842 (2011).
62. Mei, M. L. & Danovaro, R. Virus production and life strategies in aquatic
sediments. Limnol. Oceanogr. 49, 459–470 (2004).
63. Parsons, R. J., Breitbart, M., Lomas, M. W. & Carlson, C. A. Ocean time-series
reveals recurring seasonal patterns of virioplankton dynamics in the
northwestern Sargasso Sea. ISME J. 6, 273–284 (2012).
64. Parsons, R. J. et al. Marine bacterioplankton community turnover within
seasonally hypoxic waters of a subtropical sound: Devil’s Hole, Bermuda.
Environ. Microbiol. 17, 3481–3499 (2015).
65. Rinta-Kanto, J. M., Lehtola, M. J., Vartiainen, T. & Martikainen, P. J. Rapid
enumeration of virus-like particles in drinking water samples using SYBR
green I-staining. Water Res. 38, 2614–2618 (2004).
66. Schapira, M. et al. Distribution of heterotrophic bacteria and virus-like particles
along a salinity gradient in a hypersaline coastal lagoon. Aquat. Microb. Ecol.
54, 171–183 (2009).
67. Payet, J. P., McMinds, R., Burkepile, D. E. & Vega Thurber, R. L. Unprecedented
evidence for high viral abundance and lytic activity in coral reef waters of the
South Pacic Ocean. Front. Microbiol. 5, 493 (2014).
68. Patten, N. L., Harrison, P. L. & Mitchell, J. G. Prevalence of virus-like particles
within a staghorn scleractinian coral (Acropora muricata) from the Great
Barrier Reef. Coral Reefs 27, 569–580 (2008).
69. Duhaime, M. B., Deng, L., Poulos, B. T. & Sullivan, M. B. Towards quantitative
metagenomics of wild viruses and other ultra-low concentration DNA samples:
a rigorous assessment and optimization of the linker amplication method.
Environ. Microbiol. 14, 2526–2537 (2012).
70. Schmieder, R. & Edwards, R. Quality control and preprocessing of
metagenomic datasets. Bioinformatics 27, 863–864 (2011).
71. Schmieder, R., Lim, Y. W., Rohwer, F. & Edwards, R. TagCleaner: Identication
and removal of tag sequences from genomic and metagenomic datasets.
BMC Bioinformatics 11, 341 (2010).
72. Schmieder, R. & Edwards, R. Fast identication and removal of sequence
contamination from genomic and metagenomic datasets. PLoS One 6,
e17288 (2011).
73. Silva, G. G. Z., Cuevas, D. A., Dutilh, B. E. & Edwards, R. A. FOCUS:
an alignment-free model to identify organisms in metagenomes using
non-negative least squares. PeerJ 2, e425 (2014).
74. Chevreux, B., Wetter, T. & Suhai, S. Genome sequence assembly using trace
signals and additional sequence information. in German conference on
bioinformatics 45–56 (1999).
75. Rho, M., Tang, H. & Ye, Y. FragGeneScan: predicting genes in short and
error-prone reads. Nucleic Acids Res. 38, e191 (2010).
76. Fu, L., Niu, B., Zhu, Z., Wu, S. & Li, W. CD-HIT: accelerated for clustering the
next-generation sequencing data. Bioinformatics 28, 3150–3152 (2012).
77. Angly, F. E. et al. The GAAS metagenomic tool and its estimations of viral and
microbial average genome size in four major biomes. PLOS Comput. Biol. 5,
e1000593 (2009).
78. Dutilh, B. E. et al. Reference-independent comparative metagenomics using
cross-assembly: crAss. Bioinformatics 28, 3225–3231 (2012).
79. Sanchez, S. E. et al. Phage Phenomics: physiological approaches to
characterize novel viral proteins. J. Vis. Exp. 100, e52854 (2015).
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE
RESEARCH
Extended Data Figure 1 | The observed decline in virus to microbe
ratio with increasing host density is not supported by horizontal
transfer (for example, of resistance genes) under conditions where
strain diversity is predicted to rise. a, Host competence gene composition
likely does not facilitate the expected rise in resistance to viral infection
(n = 66; m = 0.25, t = 2.40, d.f. = 64, P = 0.02; R2 = 0.08; microbial
abundance log-transformed; linear regression). b, Lysogeny may provide
strain diversification similar to the co-evolutionary diversification
predicted by Thingstad et al. (2014)9 nested-infection chemostat model.
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE RESEARCH
Extended Data Figure 2 | Meta-analysis of the frequency of lysogenic
cells (FLC) from mitomycin C induction experiments yields ambiguous
results. FLC from four published studies is plotted against total cell
abundance. Although a sometimes-significant negative relationship exists
at a within-study level (microbial abundance log-transformed; Muck
et al. (2014)28, n = 9, m = 10.79, t = 1.76, d.f. = 7, P = 0.12; R2 = 0.31;
Bongiorni et al. (2005)29, n = 4, m = 17.23, t = 1.91, d.f. = 2, P = 0.20;
R2 = 0.65; Payet and Suttle (2013)4, n = 9, m = 48.31, t = 4.80, d.f. = 7,
P = 1.96 × 103; R2 = 0.77; Williamson et al. (2002)30, n = 5, m = 26.08,
t = 1.08, d.f. = 3, P = 0.36; R2 = 0.28; linear regression of each data set
examined independently), when examined altogether across the full range
of host abundances studied, no significant slope was observed (microbial
abundance log-transformed; n = 27, m = 0.11, t = 0.04, d.f. = 25,
P = 0.97; R2 = 5.94 × 105; linear regression of pooled data).




 !
!





! ! " "!
!!
# $%
&' (
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE
RESEARCH
Extended Data Figure 3 | Decline in virus to microbe ratio (VMR)
observed in incubations with elevated host density over time, contrasted
with published values and viral decay. a, Log-transformed VLP density
in experimental incubations is plotted against microbial host density over
time (dot size) with VMR indicated by dot colour. Data from Mission Bay
(MB) and Palmyra (Pal) water with DOC added (+ DOC) or not ( DOC)
is complemented by the nutrient-added ‘lytic’ system of Hennes et al.
(1995)31 (H + Nutrients) as well as the ‘non-lytic’ dilutions (3%, 10%, 20%,
and 30% final concentration seawater diluted by 0.02 μm filtered seawater)
of Wilcox and Fuhrman (1994)3; WF 3% SW, WF 10% SW, WF 20% SW,
WF 30% SW). n = 1 all incubations and published mean values.
b, Significant viral decay was not observed in cell-free viral decay controls
in incubation experiments (Palmyra: n = 4, m = 1.64 × 103, t = 1.48,
d.f. = 2, P = 0.28; R2 = 0.52; Mission Bay: n = 6, m = 4.53 × 103, t = 1.87,
d.f. = 4, P = 0.14; R2 = 0.47; linear regression with log-transformed viral
density).







 




 !"
##$"%&'
(""'!
##)&'
&!*%'%""
(+$'"'
,
,)
,)
,-)
#,
,
-,
-,
-,
 ./
),
),
),
),
),)
,
,)
-,
 ./
),)
),)
),-)
,
,)
,)
),)
,
,)
 0./
,
,
,
,
,
,0
,
,#
 0./
,
,
),
),
),
),)
,
,)
-,
1%234
-,#
-,#)
,
,)
-,
-,
-,0
-,
-,)
1%234
-,#
,
,
,
-,
-,
-,
!%234
,
-,
-,
-,
-,
,)
,)
,-)
-,
1+234
-,#
,
,
,
,0
,
-,
-,)
-,)
-,-)
!+234
-,
-,)
,
,)
#,
,)
-,
-,)
,
!+234
-,
-,)
,
,)
#,
,)
-,
-,)
,












!%234
-,
-,)
,
,)
,)
-,
-,)
,
!
-,)
-,
,#
-,
)0 5"!'"

"1!&!&!
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE RESEARCH
Extended Data Figure 4 | Temperateness of viral communities increases
with host density and viral functional composition change. a, The
relative composition of provirus-like reads, normalized by total sequences
in each sample, increases with host density in viral metagenomes
(host density log-transformed; n = 24 independent measures). The linear
equations and line of best fit from robust regression and bootstrapped 95%
and 90% confidence intervals (CIs) for the slope are shown. Goodness
of fit metrics are inappropriate for robust regression and are omitted.
b, Viromes clustered by functional similarity (crAss cross-assembly),
showing higher host density Pacific viromes (*) grouped away from lower
host density Atlantic viromes (†); site names coloured by host density.














 !"
#$%&'()
#$%&' (**)
+
,-
+-
+./
+0
+1
+1
+
+
+
+
+23
+43
,563
,532
,789
,7:;
+7
+7
,<
,563
+-3
-=
+5>
.
+-3
-=




© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE
RESEARCH
Extended Data Table 1 | Summary of linear regression analyses of published microbial and viral counts
Data shown in Fig. 2. Slope, intercept and R2 are reported for each ecosystem, followed by the P value for two-tailed t-tests testing for slopes dierent from 1 and false discovery
rates (FDR). Signicant values are highlighted in bold. FDR-correction yielded similar results to uncorrected linear regressions.
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE RESEARCH
Extended Data Table 2 | Summary information on the post-quality control viromes analysed
Site, region sampled, the year samples were taken, sequencing platform, and the number of reads and base pairs (bp) in each virome are shown.
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE
RESEARCH
Extended Data Table 3 | Summary of model II OLS, MA, and SMA regression analyses
The rst column indicates the variables tested and the corresponding gures in the main text. Slopes, intercepts, condence intervals and P values are shown. Rows with
condence intervals not covering 1 for Figs 1a and 2, or not covering 0 for Figs 1c, d, and 4, are signicant.
© 2016 Macmillan Publishers Limited. All rights reserved
ARTICLE RESEARCH
Web An analysis of 24 coral reef viromes challenges the view that lytic phages are believed to
predominate when the density of their hosts increase and shows instead that lysogeny is more
important at high host densities; the authors also show that this model is consistent with virus–
host dynamics in a range of other ecosystems, such as animal-associated, sediment and soil
systems.
summary
... Phage-prokaryote relationships are complex with far-reaching consequences beyond particular pairwise interactions (Diaz-Munoz and , on which a lot of researchers have done in-depth research, including the relationship between phage and prokaryotic community diversity (Thingstad and Lignell, 1997;Benmayor et al., 2008), horizontal gene transfer (HGT) between phage and prokaryote (Chen et al., 2021;Zhang et al., 2021), phage survival strategy (Winter et al., 2010;Knowles et al., 2016). Phages are viruses, the most abundant biological entities on earth (Suttle, 2007), which can infect prokaryotes and top-down control microbial abundance by lysis of prokaryotic cells and subsequent release of cellular contents during lytic infection to maintain high biodiversity in marine systems (Breitbart et al., 2018). ...
... Similarly, phage and prokaryotic communities fluctuate the KtW model in stable ecosystems and both persist over time in aquatic ecosystems (Rodriguez-Brito et al., 2010). Rohwer and colleagues proposed the Piggyback-the-Winner (PtW) model in an analysis of coral reefs, and lysogeny became increasingly important in high prokaryotic densities ecosystems (Knowles et al., 2016), which was also reflected in the murine gut (Kim and Bae, 2018). These studies suggest that different phage survival strategies could be favored depending on different ecosystem conditions. ...
Article
Full-text available
Emerging evidence supports that the phage-prokaryote interaction drives ecological processes in various environments with different phage life strategies. However, the knowledge of phage-prokaryote interaction in the shrimp culture pond ecosystem (SCPE) is still limited. Here, the viral and prokaryotic community profiles at four culture stages in the intestine of Litopenaeus vannamei and cultural sediment microhabitats of SCPE were explored to elucidate the contribution of phage-prokaryote interaction in modulating microbial communities. The results demonstrated that the most abundant viral families in the shrimp intestine and sediment were Microviridae, Circoviridae, Inoviridae, Siphoviridae, Podoviridae, Myoviridae, Parvoviridae, Herelleviridae, Mimiviridae, and Genomoviridae, while phages dominated the viral community. The dominant prokaryotic genera were Vibrio, Formosa, Aurantisolimonas, and Shewanella in the shrimp intestine, and Formosa, Aurantisolimonas, Algoriphagus, and Flavobacterium in the sediment. The viral and prokaryotic composition of the shrimp intestine and sediment were significantly different at four culture stages, and the phage communities were closely related to the prokaryotic communities. Moreover, the phage-prokaryote interactions can directly or indirectly modulate the microbial community composition and function, including auxiliary metabolic genes and closed toxin genes. The interactional analysis revealed that phages and prokaryotes had diverse coexistence strategies in the shrimp intestine and sediment microhabitats of SCPE. Collectively, our findings characterized the composition of viral communities in the shrimp intestine and cultural sediment and revealed the distinct pattern of phage-prokaryote interaction in modulating microbial community diversity, which expanded our cognization of the phage-prokaryote coexistence strategy in aquatic ecosystems from the microecological perspective and provided theoretical support for microecological prevention and control of shrimp culture health management.
... Lysogeny is generally favored in environments where bacterial hosts are less abundant and/or live in conditions unfavorable to their growth, whereas lytic infections are usually high in productive ecosystems (Bongiorni et al. 2005, Laybourn-Parry et al. 2007, where bacteria are abundant and physiologically active, sustaining the rapid proliferation of phage parasites according to the "Kill-the-Winner" paradigm (Thingstad andLignell 1997, Winter et al. 2010). However, this relationship is undoubtedly more complex and is increasingly contested: e.g. by the recent theory of "Piggybackthe-Winner," which postulates that lysogeny could be a dominant strategy in environments where bacterial hosts are present in high concentrations (Knowles et al. 2016). ...
... Such results are in line with previous findings showing that lysogeny is a dominant reproduction pathway in low productive environments, where the abundance of potential hosts is reduced (Williamson et al. 2001, Weinbauer et al. 2003, Paul 2008, Bettarel et al. 2011b, Maurice et al. 2010, Brum et al. 2016). However, since the "Piggyback-the-Winner" hypothesis was put forward (Knowles et al. 2016), recent findings have strongly questioned this view. That model, developed essentially from data obtained on animal microbiomes, suggests that lysogeny could also be a relevant strategy at high bacterial host density, by providing many ecological benefits to infected bacterial communities, in particular by allowing them to escape the lytic pressure of phages and conferring superinfection exclusion, and thus continue to fulfill their functions (defence and nutrition) for the host (Silveira and Rohwer 2016, Howard-Varona et al. 2017, Silveira, Luque and Rohwer 2021. ...
Article
Full-text available
Ecological traits of aquatic microorganisms have been poorly investigated in tropical latitudes, especially in lagoons, which are often subjected to strong anthropogenic influence, conducive to microbial development. In this study, we examined the abundance of both viral and bacterial communities, as well as their interactions (lytic and lysogenic infections) in the water and sediment of seven main stations of the Ebrié Lagoon (Ivory Coast) with contrasting levels of eutrophication. The highest bacterial and viral concentrations in both planktonic and benthic samples were found in the most eutrophicated stations, where viral lytic infections also exhibited their highest values. Conversely, the highest fractions of inducible lysogens were measured in the most oligotrophic stations, suggesting that these two main viral life strategies are mutually exclusive in this lagoon. Our findings also revealed the importance that nutrients (especially ammonium) play as drivers of the interactions between viruses and their bacterial hosts in tropical lagoons.
... Lysogeny is generally highly prevalent in environments with anaerobic (low aeration) conditions, such as the gut, deep soil, and deep-sea hydrothermal fluids [51][52][53]. Phage-mediated host lysis is usually not a major factor for microbial mortality in these environments [52,54]. Culture-based experiments also suggest a potential link between aeration conditions and phage life cycles. ...
... KtW is a long-standing theory supported by many studies regarding the natural microbial community [72,73,106] and prophage induction assays [11,33,107,108]. In contrast, the PtW theory proposes that an increase in host density leads to a decrease in VBR, which results in the persistence of dominant populations, thereby reducing microbial diversity [51]. PtW is a young theory that numerous studies have challenged since it was proposed in 2017 [109]. ...
Article
Full-text available
Phages are viruses that infect bacteria. They affect various microbe-mediated processes that drive biogeochemical cycling on a global scale. Their influence depends on whether the infection is lysogenic or lytic. Temperate phages have the potential to execute both infection types and thus frequently switch their infection modes in nature, potentially causing substantial impacts on the host-phage community and relevant biogeochemical cycling. Understanding the regulating factors and outcomes of temperate phage life cycle transition is thus fundamental for evaluating their ecological impacts. This review thus systematically summarizes the effects of various factors affecting temperate phage life cycle decisions in both culturable phage-host systems and natural environments. The review further elucidates the ecological implications of the life cycle transition of temperate phages with an emphasis on phage/host fitness, host-phage dynamics, microbe diversity and evolution, and biogeochemical cycles.
... Similarly, integrase-encoding phages have been found at higher abundances in locations associated with more environmental stresses [52]. Given the unique challenges of the soil environment -extraordinary spatial heterogeneity, lack of mixing, microniches of low nutrient availability [53] -some hypothesize that lysogeny may be a predominant lifestyle of soil phages [12,46,50,[54][55][56]. In soil, environmental conditions that drive lysogeny and prophage induction largely remain unknown [45,57]. ...
Preprint
Full-text available
Soil viruses are highly abundant and ubiquitous, yet their impact on soil microbiome structure and function remains essentially unknown. We quantified how viruses and their hosts respond to the first rainfall of the year in seasonally dry soils – a moment commonly referred to as "wet-up", when resident soil microbes are both resuscitated and lysed, with a disproportionately large effect on annual carbon turnover. We applied time-resolved metagenomics, viromics, and quantitative stable isotope probing to track spatiotemporal virus-host trends, and quantify cell death (and by proxy, the amount of biomass carbon released). Dry soil is a sparse yet diverse reservoir of putative virions, of which only a subset thrives following wet-up. A massive degradation of distinct viruses occurs within 24h post wet-up, but after one-week, viral biomass increased by up to seven-fold. Thriving viruses were not induced temperate phages. Our measures of viral-mediated microbial host death indicate that viruses drive a consistent rate of cell lysis after wet-up. These calculations – the first to demonstrate the quantitative impact of viral effects on soil microbiomes and carbon cycling – evidence that viruses significantly impact microbial community assembly. However, viruses do not appear to serve as top-down controls on soil microbial communities following wet-up.
... One is the killthe-winner model, and the other is the piggyback-thewinner model. 76 When bacterial density increases in an ecosystem, bacteria are called winners. As bacterial density increases in an ecosystem, so does the number of phages that infect those bacteria. ...
Article
Full-text available
List of abbreviations: EMBL-EBI The European Bioinformatics Institute; E. coli Escherichia coli; E. faecalis Enterobacter faecalis; B. fragilis Bacteroides fragilis; B. vulgatus Bacteroides vulgatus; SaPIs Staphylococcus aureus pathogenicity islands; ARGs Antibiotic resistance genes; STEC Shiga toxigenic E. coli; Stx Shiga toxin; BLAST Basic Local Alignment Search Tool; TSST-1 Toxic shock toxin 1; RBPs Receptor-binding proteins; LPS lipopolysaccharide; OMVs Outer membrane vesicles; PT Phosphorothioate; BREX Bacteriophage exclusion; OCR Overcome classical restriction; Pgl Phage growth limitation; DISARM Defense island system associated with restrictionmodification; R-M system Restriction-modification system; BREX system Bacteriophage exclusion system; CRISPR Clustered regularly interspaced short palindromic repeats; Cas CRISPR-associated; PAMs Prospacer adjacent motifs; crRNA CRISPR RNA; SIE; OMPs; Superinfection exclusion; Outer membrane proteins; Abi Abortive infection; TA Toxin-antitoxin; TLR Toll-like receptor; APCs Antigen-presenting cells; DSS Dextran sulfate sodium; IELs Intraepithelial lymphocytes; FMT Fecal microbiota transfer; IFN-γ Interferon-gamma; IBD Inflammatory bowel disease; AgNPs Silver nanoparticles; MDSC Myeloid-derived suppressor cell; CRC Colorectal cancer; VLPs Virus-like particles; TMP Tape measure protein; PSMB4 Proteasome subunit beta type-4; ALD Alcohol-related liver disease; GVHD Graft-versus-host disease; ROS Reactive oxygen species; RA Rheumatoid arthritis; CCP Cyclic citrullinated protein; AMGs Accessory metabolic genes; T1DM Type 1 diabetes mellitus; T2DM Type 2 diabetes mellitus; SCFAs Short-chain fatty acids; GLP-1 Glucagon-like peptide-1; A. baumannii Acinetobacter baumannii; CpG Deoxycytidylinate-phosphodeoxyguanosine; PEG Polyethylene glycol; MetS Metabolic syndrome; OprM Outer membrane porin M.
... Interestingly, with increasing spacer abundance the ratio of protospacer to spacer abundance decreased (slope = − 0.48, R 2 = 0.32, Fig. 4B), suggesting that highly abundant spacers are effective in decreasing protospacer abundance. A similar correlation has previously been described for the viral fraction outside of the cells measured by virusmicrobe-ratios (VMR) [36,83]. ...
Article
Full-text available
Background Phages are key drivers of genomic diversity in bacterial populations as they impose strong selective pressure on the evolution of bacterial defense mechanisms across closely related strains. The pan-immunity model suggests that such diversity is maintained because the effective immune system of a bacterial species is the one distributed across all strains present in the community. However, only few studies have analyzed the distribution of bacterial defense systems at the community-level, mostly focusing on CRISPR and comparing samples from complex environments. Here, we studied 2778 bacterial genomes and 188 metagenomes from cheese-associated communities, which are dominated by a few bacterial taxa and occur in relatively stable environments. Results We corroborate previous laboratory findings that in cheese-associated communities nearly identical strains contain diverse and highly variable arsenals of innate and adaptive (i.e., CRISPR-Cas) immunity systems suggesting rapid turnover. CRISPR spacer abundance correlated with the abundance of matching target sequences across the metagenomes providing evidence that the identified defense repertoires are functional and under selection. While these characteristics align with the pan-immunity model, the detected CRISPR spacers only covered a subset of the phages previously identified in cheese, providing evidence that CRISPR does not enable complete immunity against all phages, and that the innate immune mechanisms may have complementary roles. Conclusions Our findings show that the evolution of bacterial defense mechanisms is a highly dynamic process and highlight that experimentally tractable, low complexity communities such as those found in cheese, can help to understand ecological and molecular processes underlying phage-defense system relationships. These findings can have implications for the design of robust synthetic communities used in biotechnology and the food industry.
Article
In wastewater treatment plants (WWTPs), the stable operation of biological wastewater treatments is strongly dependent on the stability of associated microbiota. Bacteriophages (phages), viruses that specifically infect bacteria and archaea, are highly abundant and diverse in WWTPs. Although phages do not have known metabolic functions for themselves, they can shape functional microbiota via various phage-host interactions to impact biological wastewater treatment. However, the developments of phage-host interaction in WWTPs and their impact on biological wastewater treatment are overlooked. Here, we review the current knowledge regarding the phage-host interactions in biological wastewater treatment, mainly focusing on the characteristics of different phage populations, the phage-driven changes in functional microbiota, and the potential driving factors of phage-host interactions. We also discuss the efforts required further to understand and manipulate the phage-host interactions in biological wastewater treatment. Overall, this review advocates more attention to the phage dynamics in WWTPs.
Article
Over the past 20 years, our knowledge of virus diversity and abundance in subsurface environments has expanded dramatically through application of quantitative metagenomic approaches. In most subsurface environments, viral diversity and abundance rival viral diversity and abundance observed in surface environments. Most of these viruses are uncharacterized in terms of their hosts and replication cycles. Analysis of accessory metabolic genes encoded by subsurface viruses indicates that they evolved to replicate within the unique features of their environments. The key question remains: What role do these viruses play in the ecology and evolution of the environments in which they replicate? Undoubtedly, as more virologists examine the role of viruses in subsurface environments, new insights will emerge.
Article
Full-text available
Forest soil microbiomes have crucial roles in carbon storage, biogeochemical cycling and rhizosphere processes. Wildfire season length, and the frequency and size of severe fires have increased owing to climate change. Fires affect ecosystem recovery and modify soil microbiomes and microbially mediated biogeochemical processes. To study wildfire-dependent changes in soil microbiomes, we characterized functional shifts in the soil microbiota (bacteria, fungi and viruses) across burn severity gradients (low, moderate and high severity) 1 yr post fire in coniferous forests in Colorado and Wyoming, USA. We found severity-dependent increases of Actinobacteria encoding genes for heat resistance, fast growth, and pyrogenic carbon utilization that might enhance post-fire survival. We report that increased burn severity led to the loss of ectomycorrhizal fungi and less tolerant microbial taxa. Viruses remained active in post-fire soils and probably influenced carbon cycling and biogeochemistry via turnover of biomass and ecosystem-relevant auxiliary metabolic genes. Our genome-resolved analyses link post-fire soil microbial taxonomy to functions and reveal the complexity of post-fire soil microbiome activity.
Article
Full-text available
The ISME Journal: Multidisciplinary Journal of Microbial Ecology is the official Journal of the International Society for Microbial Ecology, publishing high-quality, original research papers, short communications, commentary articles and reviews in the rapidly expanding and diverse discipline of microbial ecology.
Article
Full-text available
We measured microbial heterotrophic activity, bacteria, and virus-like particle (VLP) abundance in homogenized, undiluted, and anoxic enclosures of sediment collected at a coastal station. The bacterial growth rate and VLP net production increased along with the respiratory activity in response to temperature. This suggests that VLPs represent a dynamic component of benthic microbial communities and that the net production of viriobenthos is regulated by the metabolic activity of bacteria. The abundance, net production, and decay rate of VLPs were significantly higher than those encountered in most pelagic systems. However, the rates were lower than the very few available potential rates (three studies) of viriobenthic activity, which all were obtained applying different slurry approaches. Our measurements support the general observation that virus abundance and production correlate with the trophic status of the environment and show that microbial activity can regulate the viriobenthic production in undiluted, homogenized marine sediments. The virus-induced bacterial mortality corresponded to similar to20% of bacterial net production and similar to2% h(-1) of the total bacterial population. This is moderate compared with the results of most pelagic studies, and the associated leakage of lysates (dissolved organic carbon) only amounted to 4-8% of the produced dissolved inorganic carbon. Despite high standing stocks and relatively high turnover rates, VLP-induced bacterial lysis represented only a minor shunt in the benthic carbon cycle at the investigated site.
Chapter
Full-text available
Silent viral infections occur in all forms of life, from bacteria to humans, as indicated from genomic sequencing. Temperate phages can infect bacteria and establish a symbiotic relationship termed lysogeny, enabling the phage genome to be propagated in host daughter cells. The expression of prophage genes often results in an altered bacterial phenotype, often turning benign bacteria into virulent pathogens. The most widely used method to detect lysogens is to chemically induce their prophages and detect these via microscopy or flow cytometry. Although chemical induction is the gold standard in prophage detection, not all prophages can be detected by it. This review gives two methods for prophage induction in heterotrophic bacterioplankton, a method for induction of Synechococcus populations, and a method for isolating temperate phage, as well as a simple method to recognize prophage-like elements in bacterial genomes.
Article
Full-text available
Rapid warming in the highly productive western Antarctic Peninsula (WAP) region of the Southern Ocean has affected multiple trophic levels, yet viral influences on microbial processes and ecosystem function remain understudied in the Southern Ocean. Here we use cultivation-independent quantitative ecological and metagenomic assays, combined with new comparative bioinformatic techniques, to investigate double-stranded DNA viruses during the WAP spring-summer transition. This study demonstrates that (i) temperate viruses dominate this region, switching from lysogeny to lytic replication as bacterial production increases, and (ii) Southern Ocean viral assemblages are genetically distinct from lower-latitude assemblages, primarily driven by this temperate viral dominance. This new information suggests fundamentally different virus-host interactions in polar environments, where intense seasonal changes in bacterial production select for temperate viruses because of increased fitness imparted by the ability to switch replication strategies in response to resource availability. Further, temperate viral dominance may provide mechanisms (for example, bacterial mortality resulting from prophage induction) that help explain observed temporal delays between, and lower ratios of, bacterial and primary production in polar versus lower-latitude marine ecosystems. Together these results suggest that temperate virus-host interactions are critical to predicting changes in microbial dynamics brought on by warming in polar marine systems.The ISME Journal advance online publication, 21 August 2015; doi:10.1038/ismej.2015.125.
Article
Full-text available
Current investigations into phage-host interactions are dependent on extrapolating knowledge from (meta)genomes. Interestingly, 60-95% of all phage sequences share no homology to current annotated proteins. As a result, a large proportion of phage genes are annotated as hypothetical. This reality heavily affects the annotation of both structural and auxiliary metabolic genes. Here we present phenomic methods designed to capture the physiological response(s) of a selected host during expression of one of these unknown phage genes. Multi-phenotype Assay Plates (MAPs) are used to monitor the diversity of host substrate utilization and subsequent biomass formation, while metabolomics provides bi-product analysis by monitoring metabolite abundance and diversity. Both tools are used simultaneously to provide a phenotypic profile associated with expression of a single putative phage open reading frame (ORF). Representative results for both methods are compared, highlighting the phenotypic profile differences of a host carrying either putative structural or metabolic phage genes. In addition, the visualization techniques and high throughput computational pipelines that facilitated experimental analysis are presented. Video Link The video component of this article can be found at http://www.jove.com/video/52854/
Article
Full-text available
Prochlorococcus is an abundant marine cyanobacterium that grows rapidly in the environment and contributes significantly to global primary production. This cyanobacterium coexists with many cyanophages in the oceans, likely aided by resistance to numerous co-occurring phages. Spontaneous resistance occurs frequently in Prochlorococcus and is often accompanied by a pleiotropic fitness cost manifested as either a reduced growth rate or enhanced infection by other phages. Here, we assessed the fate of a number of phage-resistant Prochlorococcus strains, focusing on those with a high fitness cost. We found that phage-resistant strains continued evolving toward an improved growth rate and a narrower resistance range, resulting in lineages with phenotypes intermediate between those of ancestral susceptible wild-type and initial resistant substrains. Changes in growth rate and resistance range often occurred in independent events, leading to a decoupling of the selection pressures acting on these phenotypes. These changes were largely the result of additional, compensatory mutations in noncore genes located in genomic islands, although genetic reversions were also observed. Additionally, a mutator strain was identified. The similarity of the evolutionary pathway followed by multiple independent resistant cultures and clones suggests they undergo a predictable evolutionary pathway. This process serves to increase both genetic diversity and infection permutations in Prochlorococcus populations, further augmenting the complexity of the interaction network between Prochlorococcus and its phages in nature. Last, our findings provide an explanation for the apparent paradox of a multitude of resistant Prochlorococcus cells in nature that are growing close to their maximal intrinsic growth rates.
Article
Full-text available
Shifts from coral to algae dominance of corals reefs have been correlated to fish biomass loss and increased microbial metabolism. Here we investigated reef benthic and planktonic primary production, benthic DOC release, and bacterial growth efficiency in the Abrolhos Bank, South Atlantic. Benthic DOC release rates are higher at impacted reefs, and bacterial growth efficiency on algae-derived carbon is lower compared with water column carbon. A trophic model based on the benthic and planktonic primary was able to predict the observed relative fish biomass in healthy reefs. In contrast, in impacted reefs, the observed omnivorous fish biomass is higher, while that of the herbivorous/coralivorous fish is lower than predicted by the primary production-based model. Incorporating recycling of benthic-derived carbon in the model through microbial and sponge loops explains the difference and predicts the relative fish biomass in both reef types. Increased benthic carbon release rates and bacterial carbon metabolism, but decreased bacterial growth efficiency could lead to carbon losses through respiration and account for the uncoupling of benthic and fish production in phase-shifting reefs. Carbon recycling by microbial and sponge loops seems to promote an increase of small-bodied fish productivity in phase-shifting coral reefs. This article is protected by copyright. All rights reserved.
Article
Full-text available
Viral lysis of microbial hosts releases organic matter that can then be assimilated by nontargeted microorganisms. Quantitative estimates of virus-mediated recycling of carbon in marine waters, first established in the late 1990s, were originally extrapolated from marine host and virus densities, host carbon content and inferred viral lysis rates. Yet, these estimates did not explicitly incorporate the cascade of complex feedbacks associated with virus-mediated lysis. To evaluate the role of viruses in shaping community structure and ecosystem functioning, we extend dynamic multitrophic ecosystem models to include a virus component, specifically parameterized for processes taking place in the ocean euphotic zone. Crucially, we are able to solve this model analytically, facilitating evaluation of model behavior under many alternative parameterizations. Analyses reveal that the addition of a virus component promotes the emergence of complex communities. In addition, biomass partitioning of the emergent multitrophic community is consistent with well-established empirical norms in the surface oceans. At steady state, ecosystem fluxes can be probed to characterize the effects that viruses have when compared with putative marine surface ecosystems without viruses. The model suggests that ecosystems with viruses will have (1) increased organic matter recycling, (2) reduced transfer to higher trophic levels and (3) increased net primary productivity. These model findings support hypotheses that viruses can have significant stimulatory effects across whole-ecosystem scales. We suggest that existing efforts to predict carbon and nutrient cycling without considering virus effects are likely to miss essential features of marine food webs that regulate global biogeochemical cycles.