The ISME Journal (2012) 6, 542–553

& 2012 International Society for Microbial Ecology All rights reserved 1751-7362/12
www.nature.com/ismej

ORIGINAL ARTICLE
Diversity and dynamics of rare and of resident
bacterial populations in coastal sands
Angélique Gobet1,2,7, Simone I Böer1,8, Susan M Huse3, Justus EE van Beusekom4,
Christopher Quince5, Mitchell L Sogin3, Antje Boetius1,6 and Alban Ramette1
1
HGF-MPG Group for Deep Sea Ecology and Technology, Max Planck Institute for Marine Microbiology,
Bremen, Germany; 2Jacobs University Bremen GmbH, Bremen, Germany; 3Josephine Bay Paul Center,
Marine Biological Laboratory, Woods Hole, MA, USA; 4Alfred Wegener Institute for Polar and Marine
Research, List/Sylt, Germany; 5School of Engineering, University of Glasgow, Glasgow, UK and
6
HGF-MPG Group for Deep Sea Ecology and Technology, Alfred Wegener Institute for Polar and Marine
Research, Bremerhaven, Germany

Coastal sands filter and accumulate organic and inorganic materials from the terrestrial and marine
environment, and thus provide a high diversity of microbial niches. Sands of temperate climate
zones represent a temporally and spatially highly dynamic marine environment characterized by
strong physical mixing and seasonal variation. Yet little is known about the temporal fluctuations of
resident and rare members of bacterial communities in this environment. By combining community
fingerprinting via pyrosequencing of ribosomal genes with the characterization of multiple
environmental parameters, we disentangled the effects of seasonality, environmental heterogeneity,
sediment depth and biogeochemical gradients on the fluctuations of bacterial communities of
marine sands. Surprisingly, only 3–5% of all bacterial types of a given depth zone were present at all
times, but 50–80% of them belonged to the most abundant types in the data set. About 60–70% of the
bacterial types consisted of tag sequences occurring only once over a period of 1 year. Most
members of the rare biosphere did not become abundant at any time or at any sediment depth, but
varied significantly with environmental parameters associated with nutritional stress. Despite the
large proportion and turnover of rare organisms, the overall community patterns were driven by
deterministic relationships associated with seasonal fluctuations in key biogeochemical parameters
related to primary productivity. The maintenance of major biogeochemical functions throughout the
observation period suggests that the small proportion of resident bacterial types in sands perform
the key biogeochemical processes, with minimal effects from the rare fraction of the communities.
The ISME Journal (2012) 6, 542–553; doi:10.1038/ismej.2011.132; published online 6 October 2011
Subject Category: microbial population and community ecology
Keywords: 454 pyrosequencing; coastal seas; bacterial diversity; multivariate analysis; rare biosphere

Introduction advected by currents and winds, including dis-

solved and particulate organic matter derived from
Marine coasts represent highly dynamic ecosystems living and dead biomass of terrestrial or marine
where the atmosphere, continents and the ocean origin (de Beer et al., 2005). Sandy sediments are
interact. Permeable sands constitute the dominant also constantly subjected to biotic (e.g., bioturba-
sediment type on continental shelves (Emery, 1968; tion) and abiotic disturbances (e.g., mixing by
Boudreau et al., 2001) and have a central role for currents, seasonal and tidal temperature fluctuation,
global carbon and nutrient cycles. They act as anoxia; Boudreau et al., 2001). They may also host
biocatalytic filters for various types of materials human pathogens, depending on the human impact
by, for example, recreational use of beaches, and
temperature anomalies (Ruppert et al., 2004; Dins-
Correspondence: A Ramette, HGF-MPG Group for Deep Sea dale et al., 2008).
Ecology and Technology, Max Planck Institute for Marine The high diversity of niches provided by the
Microbiology, Celsiusstrasse 1, 28359 Bremen, Germany.
E-mail: aramette@mpi-bremen.de sand habitat supports a rich community of bacteria
7
Current address: Evolution of Plankton and PaleoOceans (EPPO) fluctuating with environmental variations (e.g.,
team, UMR 7144 CNRS – UPMC, Station Biologique de Roscoff, Hunter et al., 2006; Sørensen et al., 2007; Mills
Roscoff, France. et al., 2008; Rusch et al., 2009; Gaidos et al., 2011).
8
Current address: Federal Institute of Hydrology, Koblenz,
Germany. Yet, the dynamics of environmental parameters in
Received 2 May 2011; revised 29 July 2011; accepted 25 August coastal sands could still be a challenge to many
2011; published online 6 October 2011 microbial populations, selecting for few, but tolerant
 Bacterial biosphere of coastal sands
A Gobet et al
543
and well-adapted resident types (Gaidos et al., sands over six sampling dates within a year (2005–
2011). Some of these produce extracellular poly- 2006), and over different sediment depths (0–5 cm,
meric substances to attach to sand grain surfaces. 5–10 cm, 10–15 cm). For this purpose, we combined
The resulting biofilm habitat can be considered as several data sets from the Long Term Ecological
an active bioreactor accumulating diverse microbial Research (LTER) Station, Sylt, Germany, including
populations, their extracellular enzymes, nutrients 454 massively parallel tag sequencing (MPTS) target-
and organic matter from various origins (Flemming ing the V6 region of the 16S rRNA gene (Gobet et al.,
and Wingender, 2010). The pore volume between 2010) and concurrent environmental surveys (e.g.,
the sand grains represents a place of constant nutrients, enzymatic activities, bacterial production,
particle exchange due to turbulent flow of currents weather and water parameters; Böer et al., 2009a, b), to
and pore-water advection, which may induce signi- obtain as complete a description as possible of the
ficant fluxes of particulate organic matter (Stoodley changes in the bacterial community in their environ-
et al., 2005). It remains yet unknown whether sand mental context. We falsified the following hypotheses:
grain-associated biofilms trap microbes from terres- (1) bacterial communities in coastal sands have the
trial sources and from the water column flushing same composition across sediment depth and in
through the sand, and if they are a source of comparison with the overlying water column, (2) as
diversity to the sea around them. As they represent a biofilm community, their temporal fluctuation is
a transition zone between land and ocean character- low, and (3) the rare microbial biosphere in sands
ized by high mixing rates, the presence of many rare does not respond to environmental changes.
bacterial types could be expected. In coarse-grained
reef sediments, it has been shown that the typical
patterns of rare microbial types found in DNA-based Materials and methods
libraries (i.e., the characteristic long tail in rank- Study site, sampling procedures and contextual
abundance curves; Pedrós-Alió, 2007) could not be parameters
observed in RNA-based libraries, suggesting that the Detailed sample processing and environmental mea-
rare biosphere may be inactive (Gaidos et al., 2011). surements have been published elsewhere (Böer et al.,
So far, few studies have examined the structure 2009a). Briefly, sediment samples were repeatedly
and ecology of microbial communities in coastal collected every 2–3 months on a shallow (0.5–2.5 m),
sediments, and these were mostly based on low- subtidal sand flat of the LTER site at the island of Sylt
resolution community fingerprinting approaches in the North Sea (551 000 47.700 N, 81 250 59.300 E)
(Urakawa et al., 2000; Bertics and Ziebis, 2009; Böer between February 2005 and March 2006. Sandy cores
et al., 2009b). Although those molecular techniques of 15 cm were directly sectioned into three 5 cm
have provided a first description of abundant layers. Subsamples were stored at 20 1C for DNA
populations, they generally did not describe rare extraction or used for the measurement of contextual
populations that might represent considerable parameters (extracellular enzymatic activities, nutri-
diversity (Acinas et al., 2004). High-throughput ents, pigments, carbohydrates, bacterial cell counts;
sequencing strategies such as shotgun sequencing Böer et al., 2009a). Water column data consisting of
(Venter et al., 2004) or pyrosequencing (Sogin et al., chlorophyll a, pH and water temperature, were
2006) allow a deeper exploration of microbial obtained from the Sylt LTER time series (van
diversity. Pyrosequencing of ribosomal genes has Beusekom et al., 2009), and data on wind speed was
already permitted the description of the rare micro- obtained from the German Weather Service (Deutsche
bial biosphere in several types of ecosystems, such Wetterdienst; weather station of List at Sylt).
as surface and deep-sea water (Sogin et al., 2006), To compare the communities of sediments, pore
extreme environments (Huber et al., 2007; Brazelton water and overlying water, at one time, additional
et al., 2010), or the human hand surface (Fierer samples were taken at Sylt Königshafen in April 2008
et al., 2008). Patterns of the rare biosphere in the (sediments 551 20 2800 N, 81 240 2600 E; seawater 551 10
Arctic Ocean have been related to differences in 4100 N, 81 260 1000 E). Seawater of 1 l was successively
water masses and attributed to dispersal limitation filtered through 10 mm and 0.2 mm filters (type GTTP;
(Galand et al., 2009). Here we focus on temporal diameter 47 mm) before DNA extraction. Pore water
fluctuations of resident and of rare microbial types of the upper 5 cm of the sediment was separated from
in coastal sands, and on their link with fluctuations the sand grains (with bacterial biofilms) by low-speed
in environmental parameters. centrifugation (10 min at 1801 g, at 4 1C) through
Including the rare microbial biosphere in ecolo- GF-C filters (47 mm). The resulting ‘dry’ sediment
gical analyses is challenging when considering the (about 8 g) and 2-ml pore water (filtered through
potential for technical noise in pyrosequencing data 0.2-mm GTTP-type filters; diameter 47 mm) were
(Quince et al., 2009; Kunin et al., 2010). Here we subsequently used for DNA extraction.
apply a new analytical method to assess the effects
of rarity on community structure and ecological
interpretation (Gobet et al., 2010). The main objec- DNA extraction and 454 MPTS
tive of this study was to explore the putative factors For samples taken in 2008, DNA was extracted using
structuring bacterial diversity in temperate coastal an UltraClean Soil DNA Isolation Kit (MoBio

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
544
Laboratories Inc., Carlsbad, CA, USA) from (1) ca. Acinetobacter, Rickettsiella, Pseudomonas and
10 g of homogenized wet sediment separated in the Ralstonia)) were standardized by Hellinger transfor-
dry sand and the pore water, and (2) 1 l of water mation (Legendre and Gallagher, 2001). A forward
filtered through 0.2-mm filters (pore water and selection (based on a canonical redundancy analysis
water column) cut into pieces with a sterilized algorithm and 999 Monte Carlo permutation tests)
cutter. DNA was further stored in a final volume of of the environmental factors was done to find the set
50–100 ml of Tris-EDTA buffer. To study temporal of parameters that could best explain the variation
fluctuations and vertical community patterns in in the community table. The best-fitting models
sediments, DNA from samples collected between were chosen using the Akaike Information Criterion.
2005–2006 were similarly extracted from 1 g of Canonical variation partitioning (Borcard et al.,
sediment, as described previously (Böer et al., 1992; Ramette and Tiedje, 2007b) was then applied
2009b). on the community data to disentangle the specific
DNA was deposited within the International effects of environmental variables (pigments, nutri-
Census of Marine Microbes project (http://icomm. ents, extracellular enzymatic activities, cell abun-
mbl.edu/), where it was amplified using primers dance) selected previously and of their covaria-
targeting the V6 region of the bacterial 16S rRNA tion on microbial community structure. Statistical
gene, and which included 454 Life Science’s A or analyses were carried out using the R statistical
B sequencing adapter according to Sogin et al. environment (R version 2.10.0, R Development Core
(2006). Pyrosequencing was performed on a Genome Team, 2009), using the vegan (Oksanen et al., 2009),
Sequencer 20 system (Roche, Basel, Switzerland) gplots (Bolker et al., 2009) packages, and custom R
at 454 Life Sciences (Branford, CT, USA) by primer scripts.
extension (Margulies et al., 2005), yielding on
average 60-bp long fragments after removing primer
sequences. Results and discussion
The bacterial community composition of coastal sands
Sequence analysis The three compartments, sand, pore water and the
Data from the 454 MPTS were retrieved from the overlying water column, may be expected to share a
publicly available ‘Visualization and Analysis of large proportion of the same microbial assemblages,
Microbial Populations Structure (VAMPS)’ website because the pore space of permeable sands is
(http://vamps.mbl.edu/; projects AB_SAND_Bv6 constantly flushed by the overlying water, trapping
and ICM_FIS_Bv6) in the framework of the Inter- detritus and living cells from the water column
national Census of Marine Microbes project. (Boudreau et al., 2001). We compared bacterial
Sequences were taxonomically assigned by an auto- communities from one large, well-mixed sample of
matic annotation pipeline (Sogin et al., 2006), using each compartment (sand, pore water and water
several known databases (Entrez Genome, RDP, column) at one time, to investigate the similarity
SILVA), allowing a taxonomic annotation of up to in community composition between those coastal
94% of the data set. In our study, analyses were compartments, which are constantly mixed by wind
based on a definition of operational taxonomic units and waves. However, microscopic observations
(OTU) as unique (i.e. two sequences belong to two (Figure 1a) confirm that sand grains are covered by
different OTUunique, when they differ by at least one a biofilm consisting of cells embedded in extra-
base) to keep a consistent definition throughout. cellular polymeric substances, and that there is
Other classifications used in the study are indicated potentially little exchange with cells flowing
as follows: OTUall represents all, un-annotated through the sediment (Rusch et al., 2001). At any
sequences, PyroNoise0% and PyroNoise3% corre- time, the pore water contains less than 0.2% of the
spond to the PyroNoise-corrected data sets at 0% cell abundance associated with the sand grains (this
and 3% sequence identity, respectively (Quince study and Rusch et al., 2003). Accordingly, at that
et al., 2009), OTUall—x% rare, which represent the specific sampling time, we also found a rather low
OTUall data set truncated of its x% rarer OTUunique proportion of shared bacterial populations between
according to Gobet et al. (2010). the sand and the water column, at the level of both
community composition and evenness (Figures
1b and c, Supplementary Figure S1). Only 2–3% of
Taxa-environment relationships all unique OTUunique (sequences from the original
To investigate taxa-environment relationships, most OTUall data set that have at least one nucleotide
of the parameters (except pH, water temperature, difference) were shared between the three compart-
wind speed and salinity) were log10-transformed, ments, confirming previous findings (Epstein et al.,
whereas the community matrices (OTUall data 1997; Llobet-Brossa et al., 1998).
set, and the resident single-sequence OTU rela- The Sylt (North Sea) water column was mainly
tive (SSOrel) data sets or the (potential) patho- dominated by Bacteroidetes and by the Alpha- and
gen sequence abundance matrix (including the Gamma-subdivisions of the Proteobacteria, confirm-
genera Parachlamydia, Arcobacter, Francisella, ing previous results on the basis of fluorescence

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
545

Figure 1 Microbial community distribution in the sand and the water column. (a) From top to bottom: acridine orange staining of
bacteria in the water column, the pore water and on the surface of a sand grain (scale bar ¼ 50 mm). (b) Relative number of sequences in
different compartments for the top 5-cm sand layer and in the overlying water column in April 2008. Here, the phylum level was chosen
for illustrative purposes. (c) Sequence distribution in the sand over time, in which each bar represents an OTUunique (only OTUunique
occurring more than 100 times in the whole data set are shown). The Proteobacteria phylum was further split into its corresponding
classes, for example, Alpha, Gamma, Delta; Cy, Cyanobacteria; Ba, Bacteroidetes; Aci, Acidobacteria; Others: Actinobacteria, NA, not
annotated Proteobacteria, Planctomycetes, Chloroflexi, Verrucomicrobia, WS3, Firmicutes, Lentisphaerae, Deferribacteres and
Gemmatimonadetes.

in situ hybridization and 16S rRNA gene-based and Spirochaetes (Schöttner et al., 2011) in the
clone libraries (Glöckner et al., 1999; Eilers et al., community. Acidobacteria sequences were abun-
2000; Zubkov et al., 2002). Sequences of the phyla dant in temperate (Figure 1b, Supplementary
Verrucomicrobia and Actinobacteria were also Figure S2) and tropical sands (Gaidos et al., 2011;
abundant (Figure 1b). The top 5 cm of Sylt sands Schöttner et al., 2011). At the phylum level, the pore
were dominated by Bacteroidetes, Gammaproteo- water microbial composition resembled the sand
bacteria, Deltaproteobacteria and Planctomycetes, community with a dominance of Bacteroidetes,
also in accord with previous clone libraries Gammaproteobacteria, Deltaproteobacteria and
obtained from coastal sediments of the Wadden Acidobacteria (Figure 1b). At the OTUunique level,
Sea (Llobet-Brossa et al., 1998; Musat et al., 2006), or substantial differences were detected between the
other permeable sediments from shelf sands pore water and the sand-associated bacterial com-
from South Atlantic Bight and coastal sediment in munity or the water column (Supplementary Figure
Hawaii (Hunter et al., 2006; Sørensen et al., 2007). In S1C). The further analysis of temporal fluctuations
contrast, tropical coral reef sands were less consis- was restricted to wet sands containing both the
tent, with a dominance of Proteobacteria, Actino- biofilm community, as well as the pore water types.
bacteria (Gaidos et al., 2011), or Alphaproteobacteria Certainly, temporal variation and the role of the

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
546
environment in the exchange of microbial commu- shared between sediment layers or any two sam-
nities between land and sea, intertidal and subtidal pling dates. Only about 20% of sequences were
flats with their different physical compartments shared between the deeper layers and the upper
deserve further studies. layer (Supplementary Figure S3A), the communities
The characterization of coastal sand communities of the two deeper layers were slightly more similar
over six sampling dates (Figure 1) provided a total of (Figure 2a). Moreover, resident OTUs seemed to be
197 684 rRNA gene sequences, corresponding to characterized by a high and relatively stable number
27 630 OTUunique. Per sample, sequences ranged of sequences (Figure 3b). About 70% of the resident
from about 5000–19 000, corresponding to 496– OTUs had abundances of more than 10 sequences
2993 OTU3% (i.e., OTU defined at 3% sequence per sample (Table 1, Figure 3b), suggesting that they
difference after denoising the V6 sequences by belonged to archetypal sand microbial communities.
PyroNoise; Supplementary Table S1). Noticeably, When time was considered alone, only 18–37% of
the high bacterial richness estimates for coastal OTUunique were found to be shared between any two
sands were comparable to 454 MPTS-based esti- sampling times (Figure 2b, Supplementary Figure
mates reported for coarse carbonate sands from coral S3A). This indicates that a very large fraction of the
reefs (Gaidos et al., 2011), soils (Roesch et al., 2007), community may be constantly replaced. Yet, the fact
marine sediments and crusts (Sogin et al., 2006; that the turnover rate did not increase with sampling
Huber et al., 2007; Brazelton et al., 2010). A high, yet time suggests that some populations vanished and
unexplored diversity in these habitats was already reappeared during the investigated time period
previously proposed based on clone libraries (Supplementary Figure S3). Some of the most
(Hunter et al., 2006; Mills et al., 2008; Rusch et al., sequence-abundant groups were found to be posi-
2009). In addition, here we found that total sequence tively correlated with time (Figure 4); for example,
and OTU numbers systematically increased with Gammaproteobacteria and Planctomycetes, which
depth from 0–15 cm at most of the sampling dates followed the seasonal fluctuations in cell abun-
(Supplementary Table S1). A positive relationship dances observed in Sylt sediment (Musat et al.,
between sediment depth and bacterial diversity was 2006). Remarkably, the fluctuation within a month
previously observed using different fingerprinting (beginning and end of March 2006) was almost as
techniques (Urakawa et al., 2000; Böer et al., 2009b), high as within a year. This could be explained by
and explained by increasing physico-chemical the observed large environmental variations that
stability of the habitat. The Delta- and Beta- occurred in March 2006, starting with cold temp-
subdivisions of the Proteobacteria, Deferribacteres, eratures and high nutrient levels at the end of the
Spirochaetes and Nitrospira, all showed an increas- winter period, and ending with spring blooms and
ing OTU richness with depth. In contrast, Cyano- stormy conditions (Böer et al., 2009b).
bacteria and Bacteroidetes decreased with depth An interesting aspect is the abundance and role of
(Supplementary Figure S2). the resident, potentially sand-biofilm associated
populations versus rare and transient populations
getting advected through the sands. Indeed, in our
Turnover of bacterial diversity with sediment depth data set, more than 50% of the OTUunique present at
and time all times in the three depth layers were members of
The phylum to class levels are commonly used to the Gammaproteobacteria and Deltaproteobacteria,
describe the diversity of bacterial communities on two classes representing up to 23% and up to 10%,
the basis of whole-cell fluorescence in situ hybridi- respectively, of the total cell counts in such sands
zation or 16S-based clone libraries (Llobet-Brossa (Musat et al., 2006). But also, 40% of all OTUunique
et al., 1998; Musat et al., 2006). At those levels, appearing only once in a given sample or in the
microbial community patterns in sands mostly whole OTUall data set belonged to these classes.
remained unchanged over time and sediment depth Noticeably, the rare members of dominant classes
(Supplementary Figure S2). However, drastic never became abundant at any sampling time,
changes in bacterial community composition were similarly to observations made in Arctic waters
evidenced when finer taxonomic resolutions were (Galand et al., 2009).
used; overall, only 152 OTUunique (0.55% of the total Observing a high turnover for a substantial
number of 27 630 OTUunique) were present in all fraction of the bacterial community in marine sands
samples at all times, and their sequence abundance raises several questions about the ecological sig-
ranged from 54–6550. Also, only 3–5% of all nificance of these dynamics. Large turnover in
OTUunique within a sampling depth layer were community composition could be due to various
present at all times (Table 1). The majority (77%) phenomena, which are not mutually exclusive, such
of these OTUunique, which we define as resident as: (1) migration of non-resident bacterial popula-
OTU, consisted of the phyla Acidobacteria, Actino- tions into and out of an ecosystem (Sloan et al.,
bacteria, Bacteroidetes, Cyanobacteria, Alpha-, Delta- 2006). This could be due to the rapid physical
and Gammaproteobacteria and Verrucomicrobia. transport processes prevailing in marine coastal
For the OTUall data set (i.e., data set with all sand ecosystems. (2) High adaptability of micro-
OTUunique), a low percentage of OTUunique were bial communities to ever-changing and complex

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
547
Table 1 Temporal patterns of OTU variationa

OTUall Sequence similarity thresholds for PyroNoise-corrected data

PyroN0% PyroN3% PyroN5%

Number of OTU (%)b

Total 19 785 (100) 9036 (100) 7668 (100) 5927 (100)
SSOabs 13 676 (69) 5272 (58) 3970 (52) 2624 (44)
SSOrel 3735 (19) 2136 (24) 2268 (30) 2087 (35)
DSOabs 1108 (6) 697 (8) 536 (7) 386 (6)
TSOabs 236 (1) 187 (2) 170 (2) 133 (2)
With X5 sequences minimum in any sample 203 (1) 124 (1) 125 (1) 135 (2)
With X10 sequences minimum in any sample 106 (0.5) 67 (0.7) 67 (0.9) 69 (1)

Temporal patternsc
Feb Apr Jul Nov Mar 1 Mar 2
’ ’ ’ ’ ’ ’ 654 (3) 394 (4) 412 (5) 443 (7)
’ ’ ’ 29 (0.2) 11 (0.1) 12 (0.2) 10 (0.2)
’ ’ ’ 157 (0.8) 110 (1) 116 (1) 106 (2)
’ ’ ’ 34 (0.2) 21 (0.2) 23 (0.3) 23 (0.4)
’ ’ ’ 32 (0.2) 18 (0.2) 18 (0.2) 19 (0.3)
’ 1998 (10/9)d 685 (8/6) 515 (7/5) 304 (5/4)
’ 1371 (7/6) 622 (7/6) 492 (6/5) 345 (6/5)
’ 1807 (9/8) 704 (8/7) 554 (7/6) 349 (6/5)
’ 3657 (18/17) 1571 (17/15) 1171 (15/13) 821 (14/11)
’ 3539 (18/16) 1706 (19/15) 1301 (17/13) 915 (15/12)
’ 2833 (14/13) 1018 (11/9) 767 (10/8) 501 (8/7)

Type of temporal relationshipse

Linear (total) 490 (2) 287 (3) 284 (4) 261 (4)
Linear increase 375 (2) 231 (2) 230 (3) 215 (4)
Linear decrease 115 (0.6) 56 (0.6) 54 (0.7) 46 (0.8)
Quadratic (total) 227 (1) 113 (1) 114 (1) 103 (2)
Quadratic positive 208 (1) 100 (1) 97 (1) 90 (1)
Quadratic negative 19 (0.1) 13 (0.1) 17 (0.2) 13 (0.2)

Abbreviations: DSOabs, double sequence OTU absolute; OTU, Operational Taxonomic Units; PyroN, PyroNoise; SSOabs, single sequence OTU
absolute, SSOrel, single sequence OTU relative; TSOabs, triple sequence OTU absolute; VAMPS, Visualization and Analysis of Microbial
Populations Structure.
a
Sequences from the first top 10-cm sediment layers pooled were processed with the VAMPS pipeline and with the PyroNoise algorithm to
remove pyrosequencing and PCR amplification artifacts. Thresholds from 0–5% sequence similarity were used to define OTU.
b
Numbers of sequences are given for single-sequence OTU occurring in the whole data set (SSOabs) or at least in one sample (SSOrel), double
sequence OTU (DSOabs), or triple sequence OTU (TSOabs) in the whole data set. Figures in parentheses correspond to the percentage of the total
number of sequences in the first top 10-cm sediment layers.
c
Occurrence patterns were defined as the presence (black square) or absence of OTU at specific sampling dates from February 05 (Feb), April 05
(Apr), July 05 (Jul), November 05 (Nov), beginning of March 06 (1 Mar), to end of March 06 (2 Mar).
d
The second value in parentheses indicates the respective percentage of SSOabs.
e
Different patterns of OTU abundance were examined by specifying linear or quadratic models of change with time. A positive quadratic
relationship with time implies a U-shape relationship, which is associated with high abundance at the beginning and at the end of the study, and
low abundance observed at the third and fourth sampling times. A negative quadratic relationship would be conversely described by an inverted
U-shape abundance function over time.

environments, as previously observed in the labora- column and sand described above may indicate that
tory (Rosenzweig et al., 1994). (3) Emergence of the described bacterial community is typical of the
latent ‘rare’ prokaryotic stages that may become sand ecosystem, and that a large effect of extra-
dominant when appropriate conditions are met cellular DNA on our results is not likely. To further
(Pedrós-Alió, 2006). This fluctuation from rare to test hypotheses 1–3, we investigated the contribu-
dominant types may be supported by the ‘seed bank’ tion of the rare biosphere to the overall community
hypothesis (Finlay, 2002). (4) Finally, the presence turnover and determined how the rare biosphere
of extracellular DNA could also have strongly depends on environmental conditions.
impacted our estimates of community turnover. This
could be possible because, first, sands are known to
act as natural filters that concentrate particles and Impact of the rare biosphere on community turnover
DNA in suspension (Naviaux et al., 2005). Second, The rare biosphere has been postulated to consist
our molecular approaches did not differentiate of low-abundance microbial organisms that would
between DNA from living cells and extracellular not be subjected to predation or viral lysis, and
DNA settling from seawater (Corinaldesi et al., 2005). would likely represent a huge proportion of micro-
However, the OTUunique dissimilarity between water bial communities, as generally indicated by long

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
548

Figure 2 Bacterial community turnover between two consecu-

tive depth layers (a) or sampling times (b). The percentage of OTU
shared between a sampling depth (or date) and the previous one
was calculated after PyroNoise correction and OTU clustering of
the 454 MPTS data set at several levels of sequence dissimilarity.
Bars correspond to s.d. calculated over 4–6 sampling dates (a) or
three depth layers (b), except for July and November 2005 in
which two depth layers were considered. The first depth layer
and sampling date (February 2005) are indicated by the grey point
as 100% of OTU. OTUall represents the original data set with all
OTUunique. It is used here as a reference data set to study the effects
of various correction levels on the observed dynamics of the
bacterial community composition.

distribution tails in rank-abundance curves (Pedrós-

Alió, 2007). For example, among the rare bacteria in Figure 3 Distribution of the maximum abundance of (a) SSOrel
coastal sands were the potential human pathogens (i.e., those OTU0% that, at least in one sample, consisted of only
one sequence) and (b) resident OTU0% (i.e., OTU0% present at all
(including Parachlamydia, Arcobacter, Francisella, times) in the top 10-cm layer. In (a), panels 1, 2 and 3 are
Acinetobacter, Rickettsiella, Pseudomonas and examples of cases, in which particular high fluctuations from the
Ralstonia) with a total of 16–88 sequences corre- single-sequence case (white arrow) to higher-sequence abun-
sponding to 2–54 OTUunique in the OTUall data set dances were observed. In (b), panels 1, 2, 3 illustrate cases of
particularly high fluctuations in relative abundances (no absolute
(0.22% of all sequences in OTUall). To understand abundances were calculated in this case). All data were initially
what fraction of the community may be associated processed to remove pyrosequencing noise. Rel. abundance,
with the large diversity turnover observed in the relative abundance to the total number of sequences per sampling
sands, we gradually removed increasing fractions of time: February 2005 (13 285), April 2005 (7902), July 2005
the rare sequences in the data set starting from the (10 910), November 2005 (21 931), 1 March 2006 (17 302), 2 March
2006 (17 897).
rarest ones (Gobet et al., 2010); the large turnover
previously described for the complete data set over
both sediment depth and time was no longer obser-
ved when up to 50% of the rare sequences were representing about 20% of all PyroNoise-corrected
removed (Supplementary Figure S3C), indicating OTU0%), or as OTU appearing only once in the
that most of the community turnover was due to whole data set (i.e., SSO absolute (SSOabs), repre-
changes in the rare tail of the data set. senting about 58% of the OTU0%), as compared with
The rare biosphere could be identified as OTU the 3–5% of the OTU0% that were resident (Table 1).
appearing only once in a given sample (i.e., SSOrel, Moreover, when only the SSOrel fraction of the data

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
549

Figure 4 Environmental factors associated with variations of the bacterial community structure at the phylum level. Pearson’s
r indicates correlations between phylum sequence abundance and several environmental parameters. For example, a red square
between sediment depth and Chloroflexi indicates a higher number of sequences with increasing depth; nonsignificant relationships
between water temperature and sequence variation in any of the bacterial groups are indicated by white squares; a blue
square between chlorophyll a and Chloroflexi denotes a decrease in sequences as chlorophyll a concentration gets higher (the
latter being concordant with the relationship between increasing depth and Chloroflexi sequences number). The Proteobacteria phylum
level was separated into its corresponding classes to obtain a higher resolution. NA-Proteobacteria are the Proteobacteria
with missing class annotation. The total number of sequences in each phylum is indicated in parentheses. SiO2, silicate; PO4,
phosphate; NO2, nitrite; NO3, nitrate; NH4, ammonium; Chl a, chlorophyll a; Pheo, pheophytin; BCC, bacterial abundance; Bprod,
bacterial carbon production; Chit, chitinase; a-glu, a-glucosidase; b-glu, b-glucosidase; Lip, lipase; Amin, aminopeptidase; Phos,
phosphatase.

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
550
set was retained, very similar fractions of the total Table S2), with all three parameters being highly
explained-biological variation were identified as for correlated with sediment depth. Cell abundance was
the total data set (Supplementary Figure S4). This also identified as an important factor influencing the
study therefore reports that not only a large fraction variation in the different OTU data sets (Supple-
of bacterial communities consists of rare types, as mentary Table S2). Noticeably, community diversity
previously found in many 454 MPTS-based studies and composition varied much more than microbial
in other environmental samples (e.g., Gilbert et al., functions (e.g., biomass, benthic oxygen consump-
2009; Bolhuis and Stal, 2011), which may undergo tion and extracellular enzymatic activities). This
substantial replacement over few centimeters of may suggest that the few continuously abundant,
sediment depth, as previously observed using lower resident microbes perform these main microbial
resolution molecular techniques (Hewson et al., functions in sandy ecosystems, and that the rare
2007), but may also dramatically change over few types being replaced at high rates have little effects
months of time. In addition, rare bacterial types on bulk functions. Furthermore, with regard to
seemed to remain rare in the bacterial community heterotrophic degradation of organic matter, a sub-
over time and sediment depth; 94% of the SSOrel stantial level of functional redundancy may exist in
had a maximum abundance of 10 sequences only the dominant taxa of the Sylt sands, comprising both
(Figure 3a). In conclusion, both the presence of this resident and transient organisms.
large proportion of singleton OTU0% in sandy Our statistical analyses showed that the temporal
sediments and the large turnover in community dynamics of bacterial communities in temperate
composition could be explained by the dispersal of sandy sediments were clearly distinct from the
OTU from other sand locations by advective trans- seasonally reoccurring, cyclic bacterial patterns that
port and physical mixing (Boudreau et al., 2001). were observed in water column samples offshore
Further research would be needed to examine more Southern California (Fuhrman et al., 2006), in the
in detail the biogeography of rare and of resident English channel (Gilbert et al., 2009), or in the Baltic
bacterial types in coastal sands. Sea (Andersson et al., 2010). Those patterns also
differed from the observed stability of the Arctic
bacterioplankton community through seasons
Ecological interpretation of overall microbial diversity (Kirchman et al., 2010). This indicates that contrast-
patterns ing patterns of temporal variation may be associated
To determine whether observed temporal patterns with different microbial habitats and ocean regions,
may be attributed to either deterministic (niche- thus inviting further studies on the main drivers
based) or stochastic processes, or both (Ramette and of temporal fluctuation or stability in marine
Tiedje, 2007a), bacterial community structure was ecosystems.
interpreted in concert with previously published
contextual data (Böer et al., 2009a) for the investi-
gated samples, to test the effects of time, sediment The rare biosphere responds to environmental drivers
depth, cell abundance and biogeochemical gradients The biogeography (Galand et al., 2009) and the
(i.e., pigments, nutrients and extracellular enzymes, temporal dynamics over thousands of years (Brazel-
Figure 4). A multivariate variation partitioning ton et al., 2010) of the rare microbial biosphere
approach (Borcard et al., 1992; Ramette and Tiedje, have previously been described in different marine
2007b) showed that biogeochemical gradients habitats, but not much is known about factors
(pigments, nutrients and extracellular enzymes), influencing the composition of the rare microbial
cell abundance and their covariation were directly biosphere. To better understand whether patterns of
related to the major changes in community structure the rare fraction were random or environmentally
(Supplementary Figure S4), as previously assessed driven, the proportion of SSOrel (Supplementary
on the basis of a lower-resolution fingerprinting Table S3) in each sample was correlated with
technique for the most abundant OTUs (Böer et al., environmental parameters. Notably, fluctuations in
2009b). This study, including also the rare types, SSOrel were not random (as determined by model
showed that significant biogeochemical variables selection using Monte Carlo permutation test and by
included in the most parsimonious multivariate selecting the lowest Akaike Information Criterion
models were qualitatively almost the same for the model values) and seemed to be negatively corre-
resident OTU, the phylum, OTUall and SSOrel levels lated with pigments (chlorophyll a, pheopigments)
(Supplementary Figure S4 and Supplementary Table and bacterial carbon production in the sediments
S2). Also, a greater amount of biological variation (Supplementary Figure S5). For OTU0%, the corre-
could be explained when a lower taxonomic resolu- sponding Pearson correlation coefficients between
tion was used, leading to a less complex data set the proportions of SSOrel and chlorophyll a, pheo-
(Gobet et al., 2010). pigments and bacterial carbon production were
The main parameters significantly influencing the 0.730, 0.695 and 0.626, respectively, and were
variation for several OTU definition levels were all highly significant. However, neither sediment
chlorophyll a, activity of the extracellular enzyme depth nor sampling time could significantly explain
phosphatase and cell abundance (Supplementary the variation in the proportion of rare types. Because

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
551
pigment concentration and bacterial carbon produc- massive parallel 16S rRNA gene tag sequencing. ISME
tion may indicate food availability in sand ecosys- J doi:10.1038/ismej.2011.52.
tems (Rusch et al., 2003), an increased proportion of Böer S, Arnosti C, van Beusekom J, Boetius A. (2009a).
rare types that is concomitant with a decrease of Temporal variations in microbial activities and carbon
turnover in subtidal sandy sediments. Biogeosciences
those parameters would indicate that rarity becomes
6(7): 1149–1165.
more prevalent when the environmental conditions Böer SI, Hedtkamp SIC, van Beusekom JEE, Fuhrman JA,
become harsher for microbial life. Those hypotheses Boetius A, Ramette A. (2009b). Time- and sediment
are concordant with the fact that primary and depth-related variations in bacterial diversity and
secondary productions in Sylt sands reach a mini- community structure in subtidal sands. ISME J 3:
mum in late winter, when temperatures are close to 780–791.
the freezing point and strong winds dominate (van Bolker B, Bonebakker L, Gentleman R, Liaw WHA,
Beusekom, personal communication). Therefore, Lumley T, Maechler M et al. (2009) gplots: Various R
our previous hypotheses 2 (microbial adaptability) programming tools for plotting data. R package version
and 3 (seed bank hypothesis) to explain the 2.7.4.
Borcard D, Legendre P, Drapeau P. (1992). Partialling out
observed high turnover in sandy sediments may be the spatial component of ecological variation. Ecology
both supported by this finding of significant 73: 1045–1055.
environmental control on the fluctuations of rare Boudreau BP, Huettel M, Forster S, Jahnke RA, McLachlan
microbial types. Future structural and functional A, Middelburg JJ et al. (2001). Permeable marine
community analyses are needed to examine whether sediments: overturning an old paradigm. EOS 82:
the large proportion of single OTU can be linked to 133–136.
specific environmental conditions in other marine Brazelton WJ, Ludwig KA, Sogin ML, Andreishcheva EN,
coastal sandy habitats and hence typical for this Kelley DS, Shen CC et al. (2010). Archaea and
microbial realm dominated by physical transport. bacteria with surprising microdiversity show shifts
in dominance over 1000-year time scales in hydro-
thermal chimneys. Proc Nat Acad Sci USA 107:
1612–1617.
de Beer D, Wenzhofer F, Ferdelman TG, Boehme SE,
Acknowledgements Huettel M, van Beusekom JEE et al. (2005). Transport
and mineralization rates in North Sea sandy intertidal
We acknowledge Martina Alisch, Marianne Jacob, sediments, Sylt-Romo Basin, Wadden Sea. Limnol
Susanne Menger and Shalin Seebah for great assistance Oceanogr 50: 113–127.
with Sylt sampling and help with sample processing. We Corinaldesi C, Danovaro R, Dell’Anno A. (2005). Simulta-
thank Rafael Stiens for cell microscopy. This work was neous recovery of extracellular and intracellular DNA
supported by the Marie Curie Early Stage Training suitable for molecular studies from marine sediments.
fellowship in Marine Microbiology (MarMic EST contract Appl Environ Microbiol 71: 46–50.
MEST-CT-2004-007776 to AG) and by the International Dinsdale EA, Pantos O, Smriga S, Edwards RA,
Max Planck Research School of Marine Microbiology Angly F, Wegley L et al. (2008). Microbial ecology of
(AG), as well as by the Max Planck Society (AR, AB), four coral atolls in the Northern Line Islands. PloS
Helmholtz Association (AB, JB) and Leibniz Program of ONE 3: e1584.
the DFG (AB). CQ is supported by an Engineering and Eilers H, Pernthaler J, Glockner FO, Amann R. (2000).
Physical Sciences Research Council Career Acceleration Culturability and in situ abundance of pelagic bacteria
Fellowship (EP/H003851/1). The sequencing and bioinfor- from the North Sea. Appl Environ Microbiol 66:
matic infrastructure (VAMPS) were supported by grants 3044–3051.
from the Alfred P Sloan and the William M Keck Emery K. (1968). Relict sediments on continental
Foundations to MLS. This is a contribution to the shelves of the world. Am Assoc Pet Geol Bull 52:
International Census of Marine Microbes (ICoMM). 445–464.
Epstein SS, Alexander D, Cosman K, Dompe A, Gallagher
S, Jarsobski J et al. (1997). Enumeration of sandy
sediment bacteria: Are the counts quantitative or
References relative? Mar Ecol Prog Ser 151: 11–16.
Fierer N, Hamady M, Lauber CL, Knight R. (2008). The
Acinas SG, Klepac-Ceraj V, Hunt DE, Pharino C, Ceraj I, influence of sex, handedness, and washing on the
Distel DL et al. (2004). Fine-scale phylogenetic diversity of hand surface bacteria. Proc Nat Acad Sci
architecture of a complex bacterial community. Nature USA 105: 17994–17999.
430: 551–554. Finlay BJ. (2002). Global dispersal of free-living microbial
Andersson AF, Riemann L, Bertilsson S. (2010). Pyrose- eukaryote species. Science 296: 1061–1063.
quencing reveals contrasting seasonal dynamics of Flemming H-C, Wingender J. (2010). The biofilm matrix.
taxa within Baltic Sea bacterioplankton communities. Nat Rev Micro 8: 623–633.
ISME J 4: 171–181. Fuhrman JA, Hewson I, Schwalbach MS, Steele JA,
Bertics VJ, Ziebis W. (2009). Biodiversity of benthic Brown MV, Naeem S. (2006). Annually reoccurring
microbial communities in bioturbated coastal sedi- bacterial communities are predictable from ocean
ments is controlled by geochemical microniches. conditions. Proc Nat Acad Sci USA 103: 13104–13109.
ISME J 3: 1269–1285. Gaidos E, Rusch A, Ilardo M. (2011). Ribosomal tag
Bolhuis H, Stal LJ. (2011). Analysis of bacterial and pyrosequencing of DNA and RNA from benthic coral
archaeal diversity in coastal microbial mats using reef microbiota: community spatial structure, rare

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
552
members and nitrogen-cycling guilds. Environ Micro- Pedrós-Alió C. (2006). Marine microbial diversity: can it
biol 13: 1138–1152. be determined? Trends Microbiol 14: 257–263.
Galand PE, Casamayor EO, Kirchman DL, Lovejoy C. (2009). Pedrós-Alió C. (2007). Dipping into the rare biosphere.
Ecology of the rare microbial biosphere of the Arctic Science 315: 192–193.
Ocean. Proc Nat Acad Sci USA 106: 22427–22432. Quince C, Lanzen A, Curtis TP, Davenport RJ, Hall N,
Gilbert JA, Field D, Swift P, Newbold L, Oliver A, Smyth T Head IM et al. (2009). Accurate determination of
et al. (2009). The seasonal structure of microbial microbial diversity from 454 pyrosequencing data. Nat
communities in the Western English Channel. Environ Methods 6: 639–641.
Microbiol 11: 3132–3139. R Development Core Team (2009). R: A language and
Glöckner FO, Fuchs BM, Amann R. (1999). Bacterioplank- environment for statistical computing. R Foundation
ton compositions of lakes and oceans: a first compar- for Statistical Computing: Vienna, Austria.
ison based on fluorescence in situ hybridization. Appl Ramette A, Tiedje JM. (2007a). Biogeography: an
Environ Microbiol 65: 3721–3726. emerging cornerstone for understanding prokaryotic
Gobet A, Quince C, Ramette A. (2010). Multivariate Cutoff diversity, ecology, and evolution. Microb Ecol 53:
Level Analysis (MultiCoLA) of large community data 197–207.
sets. Nucleic Acids Res 38: e155. Ramette A, Tiedje JM. (2007b). Multiscale responses of
Hewson I, Jacobson/Meyers ME, Fuhrman JA. (2007). microbial life to spatial distance and environmental
Diversity and biogeography of bacterial assemblages heterogeneity in a patchy ecosystem. Proc Nat Acad
in surface sediments across the San Pedro Basin, Sci USA 104: 2761–2766.
Southern California Borderlands. Environ Microbiol 9: Roesch LF, Fulthorpe RR, Riva A, Casella G, Hadwin
923–933. AKM, Kent AD et al. (2007). Pyrosequencing enumer-
Huber JA, Mark Welch D, Morrison HG, Huse SM, ates and contrasts soil microbial diversity. ISME J 1:
Neal PR, Butterfield DA et al. (2007). Microbial 283–290.
population structures in the deep marine biosphere. Rosenzweig RF, Sharp RR, Treves DS, Adams J. (1994).
Science 318: 97–100. Microbial evolution in a simple unstructured environ-
Hunter EM, Mills HJ, Kostka JE. (2006). Microbial ment—genetic differentiation in Escherichia coli.
community diversity associated with carbon and Genetics 137: 903–917.
nitrogen cycling in permeable shelf sediments. Appl Ruppert J, Panzig B, Guertler L, Hinz P, Schwesinger G,
Environ Microbiol 72: 5689–5701. Felix SB et al. (2004). Two cases of severe sepsis
Kirchman DL, Cottrell MT, Lovejoy C. (2010). The due to Vibrio vulnificus wound infection acquired
structure of bacterial communities in the western in the Baltic Sea. Eur J Clin Microbiol Infect Dis 23:
Arctic Ocean as revealed by pyrosequencing of 16S 912–915.
rRNA genes. Environ Microbiol 12: 1132–1143. Rusch A, Forster S, Huettel M. (2001). Bacteria, diatoms
Kunin V, Engelbrektson A, Ochman H, Hugenholtz P. and detritus in an intertidal sandflat subject to
(2010). Wrinkles in the rare biosphere: pyrosequen- advective transport across the water-sediment inter-
cing errors can lead to artificial inflation of diversity face. Biogeochemistry 55: 1–27.
estimates. Environ Microbiol 12: 118–123. Rusch A, Hannides A, Gaidos E. (2009). Diverse commu-
Legendre P, Gallagher ED. (2001). Ecologically meaningful nities of active Bacteria and Archaea along oxygen
transformations for ordination of species data. gradients in coral reef sediments. Coral Reefs 28:
Oecologia 129: 271–280. 15–26.
Llobet-Brossa E, Rossello-Mora R, Amann R. Rusch A, Huettel M, Reimers CE, Taghon GL, Fuller CM.
(1998). Microbial community composition of (2003). Activity and distribution of bacterial popula-
Wadden Sea sediments as revealed by fluorescence tions in Middle Atlantic Bight shelf sands. FEMS
in situ hybridization. Appl Environ Microbiol 64: Microbiol Ecol 44: 89–100.
2691–2696. Schöttner S, Pfitzner B, Grünke S, Rasheed M, Wild C,
Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Ramette A. (2011). Drivers of bacterial diversity
Bemben LA et al. (2005). Genome sequencing in dynamics in permeable carbonate and silicate coral
microfabricated high-density picolitre reactors. Nature reef sands from the Red Sea. Environ Microbiol 13:
437: 376–380. 1815–1826.
Mills HJ, Hunter E, Humphrys M, Kerkhof L, Sloan WT, Lunn M, Woodcock S, Head IM,
McGuinness L, Huettel M et al. (2008). Characteriza- Nee S, Curtis TP. (2006). Quantifying the roles
tion of nitrifying, denitrifying, and overall bacterial of immigration and chance in shaping prokar-
communities in permeable marine sediments of the yote community structure. Environ Microbiol 8:
northeastern Gulf of Mexico. Appl Environ Microbiol 732–740.
74: 4440–4453. Sogin ML, Morrison HG, Huber JA, Welch DM, Huse SM,
Musat N, Werner U, Knittel K, Kolb S, Dodenhof T, Neal PR et al. (2006). Microbial diversity in the deep
van Beusekom JEE et al. (2006). Microbial community sea and the underexplored ‘rare biosphere’. Proc Nat
structure of sandy intertidal sediments in the North Acad Sci USA 103: 12115–12120.
Sea, Sylt-Romo Basin, Wadden Sea. Syst Appl Micro- Sørensen KB, Glazer B, Hannides A, Gaidos E.
biol 29: 333–348. (2007). Spatial structure of the microbial community
Naviaux RK, Good B, McPherson JD, Steffen DL, in sandy carbonate sediment. Mar Ecol Prog Ser 346:
Markusic D, Ransom B et al. (2005). Sand DNA—a 61–74.
genetic library of life at the water’s edge. Mar Ecol Prog Stoodley P, Dodds I, De Beer D, Scott HL, Boyle JD. (2005).
Ser 301: 9–22. Flowing biofilms as a transport mechanism for
Oksanen J, Kindt R, Legendre P, O’Hara B, Simpson GL, biomass through porous media under laminar and
Solymos P et al. (2009) vegan: Community Ecology turbulent conditions in a laboratory reactor system.
Package. R package version 1.15-2. Biofouling 21: 161–168.

The ISME Journal

 Bacterial biosphere of coastal sands
A Gobet et al
553
Urakawa H, Yoshida T, Nishimura M, Ohwada K. (2000). Zubkov MV, Fuchs BM, Tarran GA, Burkill PH,
Characterization of depth-related population variation Amann R. (2002). Mesoscale distribution of dominant
in microbial communities of a coastal marine sedi- bacterioplankton groups in the northern North
ment using 16S rDNA-based approaches and quinone Sea in early summer. Aquatic Microb Ecol 29:
profiling. Environ Microbiol 2: 542–554. 135–144.
van Beusekom JEE, Loebl M, Martens P. (2009). Distant
riverine nutrient supply and local temperature drive
the long-term phytoplankton development in a This work is licensed under the Creative
temperate coastal basin. J Sea Res 61: 26–33.
Commons Attribution-NonCommercial-No
Venter JC, Remington K, Heidelberg JF, Halpern AL,
Rusch D, Eisen JA et al. (2004). Environmental genome Derivative Works 3.0 Unported License. To view a
shotgun sequencing of the Sargasso Sea. Science 304: copy of this license, visit http://creativecommons.
66–74. org/licenses/by-nc-nd/3.0/

Supplementary Information accompanies the paper on The ISME Journal website (http://www.nature.com/ismej)

The ISME Journal