Molecular Ecology (2012) 21, 3852–3868 doi: 10.1111/j.1365-294X.2012.05659.x

Different biogeographic patterns of prokaryotesand microbial eukaryotes in epilithic biofilms

MARIE RAGON,*‡ MICHAEL C. FONTAINE,*†‡ DAVID MOREIRA* and PURIFICACION LOPEZ-

GARCIA*

*Unite d’Ecologie, Systematique et Evolution, CNRS UMR 8079, Universite Paris-Sud, batiment 360, 91405 Orsay Cedex,

France, †Eco-anthropologie et Ethnobiologie, UMR 7206, Museum National d’Histoire Naturelle, CNRS, Universite Denis

Diderot, 75005 Paris, France

Corresponde

Fax: +33 1 69

‡These autho

Abstract

Microbial biogeography studies expend much effort in determining whether environ-

mental selection or stochastic processes related to dispersal are more important in

shaping community composition. While both types of factors are possibly influential, it

is tacitly assumed that protists, or microbial eukaryotes in general, behave biogeograph-

ically as prokaryotes because of their small physical size. However, direct evidence for

this in exactly the same environment and at the same phylogenetic depth is lacking. In

this study, we compared the structure of both prokaryotic and eukaryotic components of

microbial communities forming biofilms on mineral substrates in different geographic

locations at the level of small-subunit (SSU) rRNA-based operational taxonomic units

(OTUs). These microbial communities are subjected to strong environmental selection

and contain significant proportions of extremophilic microorganisms adapted to

desiccation and UV radiation. We find that the nature of the substrate as well as

climatic variables and geography influences microbial community structure. However,

constrained correspondence analyses and distance–decay curves showed that, whereas

the substrate type was the most significant factor structuring bacterial communities,

geographic location was the most influential factor for microbial eukaryote communities.

Biological explanations implying a higher dispersal success for bacteria combined with

more mobile lifestyles for predatory protists may underlie these different prokaryote

versus microbial eukaryote biogeographic patterns.

Keywords: algae, desiccation, dispersal, distance-decay, fungi, microbial biogeography, photo-

trophic biofilm, protist, radioresistance

Received 11 November 2011; revision received 4 April 2012; accepted 24 April 2012

Introduction

The massive application of molecular tools to describe

prokaryotic and eukaryotic microorganisms in natural

environments has revitalized interest in microbial bioge-

ography, that is, the study of spatial patterns of variation

in microbial community structure. This has led to multiple

discussions on whether or not a true biogeography exists

for free-living microorganisms and whether microbial bio-

geography follows patterns and processes that are already

known for macroorganisms in classical ecology theory. As

nce: Purificacion Lopez-Garcıa,

15 46 97; E-mail: puri.lopez@u-psud.fr

rs contributed equally to this work.

summarized in a number of recent reviews (Green &

Bohannan 2006; Martiny et al. 2006; Ramette & Tiedje

2007; Lindstrom & Langenheder 2011), nonrandom distri-

butions exist both for prokaryotes (bacteria and archaea)

and for microbial eukaryotes (essentially protists). Fur-

thermore, similar to macrobial communities, taxa–area

relationships have been found for bacteria (Horner-Devine

et al. 2004) and for eukaryotes (Green et al. 2004). How-

ever, the factors driving those patterns, whether they are

fundamentally deterministic (governed by environmental

selection) or stochastic (governed by dispersal history),

remain to be established.

The long-standing debate about which factors drive

microbial diversity patterns finds its roots in the

� 2012 Blackwell Publishing Ltd

MICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3853

development of microbiology as a science (O’Malley

2007). The fact that particular culture media enriched

the same particular microbial species strongly influ-

enced the idea promoted by Beijerinck, one of the foun-

ders of the so-called Delf school of microbiology, that

microorganisms, especially bacteria, dispersed ubiqui-

tously, but were ‘selected out’ by ecology (Beijerinck

1913). This idea was skilfully summarized by Baas-Bec-

king in his assertion ‘everything is everywhere but the envi-

ronment selects’ (Baas-Becking 1934; de Wit & Bouvier

2006). Ubiquitous distribution sorted out by ecological

determinism soon reached the status of a general princi-

ple for microbiology, something that distinguished it

from macrobial ecology (van Niel 1949). This view

dominated microbial ecology until the onset of the

twenty-first century, when claims of global dispersal of

small protists and other small microbes based on classi-

cal phenotypic properties were put forward (Finlay

2002). However, higher-resolution molecular analyses

soon began to reveal more complex processes, leading

to dissenting views (Whitaker et al. 2003; Foissner 2006;

Pedros-Alio 2006; Whitaker 2006).

Since then, the debate on the forces shaping micro-

bial biogeography has been reinvigorated with the

addition of many specific studies focusing on either

bacteria or protists (Martiny et al. 2006; Lindstrom &

Langenheder 2011). Some examples suggest that there

are barriers for microbial dispersal that isolate endemic

populations. This would be the case of some sea ice

bacteria (Staley & Gosink 1999; Vyverman et al. 2007),

hot spring hyperthermophilic archaea (Whitaker et al.

2003) or hyperhalophilic bacteria (Rossello-Mora et al.

2008). However, demonstrating that any given micro-

bial species is endemic is unrealistic because of the

impossibility of exhaustive sampling. A way to over-

come under-sampling is to compare community com-

positions (Martiny et al. 2006). Nevertheless, in spite of

some apparent similarity, especially in extreme envi-

ronments, sampled habitats are complex and display

varying local physico-chemical conditions that influ-

ence community structure. As a consequence, microbial

community composition would be indeed the result of

local environmental selection. For instance, the bacterial

taxa–area relationship found in salt marsh sediments

was primarily explained by environmental heterogene-

ity (which usually increases with distance) and not by

geographic distance (Horner-Devine et al. 2004),

although large environmental differences may mask

distance effects (Bissett et al. 2010). Similarly, some

studies showed that the diversity and biogeography of

soil bacterial communities are controlled by edaphic

variables (Fierer & Jackson 2006), including those of

soils under strong environmental pressure, such as the

Arctic soils (Chu et al. 2010).

� 2012 Blackwell Publishing Ltd

These apparently contradictory examples suggest

that, possibly, both alternatives are true (Pedros-Alio

2006; Whitaker 2006). Indeed, some examples for global

dispersal and probable endemicity of both bacteria and

protist species exist, suggesting a complex interplay

between dispersal limitation and local environmental

effects (Bass et al. 2007; Pommier et al. 2007; Schauer

et al. 2010). Unfortunately, the controversy about the

factors governing microbial biogeography is not only

due to the difficulties of disentangling deterministic

from chance effects, but also to a variety of difficult-

to-resolve issues, including the enormous phylogenetic

breadth of microbial taxa, the lack of a unifying species

concept for prokaryotes (Staley 2006; Achtman & Wag-

ner 2008; Fraser et al. 2009), as well as for protists, the

technical difficulties of fully exploring microbial diver-

sity in natural ecosystems and a lack of knowledge

regarding the relationship between dispersal rates and

rates of evolution (Caron 2009).

Disentangling deterministic from random effects

may be extremely difficult as, as recently suggested,

the balance between the two may depend on the spa-

tial (Martiny et al. 2011) or temporal (Nolte et al. 2010)

scale and also on the phylogenetic depth considered

(Bissett et al. 2010). From such dependence on spatial,

temporal and phylogenetic scales, a similar behaviour

of microbial and macrobial communities could be con-

cluded. However, large population sizes and, most

particularly, dormancy (Locey 2010) may play a much

more determinant role in structuring microbial com-

munities than macrobial communities, leading to dif-

ferent biogeographic patterns. In prokaryotes,

horizontal gene transfer across very different phyloge-

netic scales is an additional process operating in adap-

tation and evolution (Fraser et al. 2009; Kloesges et al.

2011). Horizontal gene transfer affects more frequently

adaptive, functional genes, which might justify a func-

tional microbial biogeography based on traits (Green

et al. 2008; Parnell et al. 2010), although this could be

different from a phylogeny-based microbial biogeogra-

phy. This would actually constitute an outstanding dif-

ference not only between macrobial and microbial

communities, but also between eukaryotic and pro-

karyotic microbial biogeographies. However, while

microbial phylogeny-based biogeography studies are

scarce and fragmentary, functional-trait-based studies

are especially rare (Green et al. 2008). Furthermore,

most microbial biogeography studies concentrate on just

bacteria ⁄ prokaryotes or protists but, to the best of our

knowledge, comparative studies analysing simulta-

neously and at the same phylogenetic depth prokaryotic

and eukaryotic components of microbial communities

to test whether they behave similarly have never been

performed.

3854 M. RAGON ET AL.

In this study, we compare the structure of both pro-

karyotic and eukaryotic components of microbial com-

munities forming epilithic biofilms in different

geographic locations in Northern Ireland and France.

We first show that these communities are subject to

strong environmental selection, containing important

proportions of extremophilic organisms adapted to des-

iccation and UV radiation. We then show that the nat-

ure of the substrate as well as climatic variables and

geography influences microbial community structure.

However, whereas the substrate type is the factor influ-

encing most bacterial communities, geography explains

the microbial eukaryotic component.

Materials and methods

Sampling and DNA purification

Microbial biofilms growing on the same control sub-

strate exposed to the open air for a period ranging from

2 to 8 (average 4.7) years were collected from various

sites located at three regions in France (Charente-Mari-

time, Lyon and Paris area) and from Northern Ireland

(Ballyclare). The mineral composite substrate consisted

of a mixture of siliceous sand and a cement-agglutinat-

ing agent, which produced a rock-like regular surface

having a consistency and porosity comparable to those

of concrete. Composite mineral substrates were initially

sterile. In addition, biofilms growing on concrete, and

naturally occurring limestone and granite on the Bally-

clare site, were also collected (Table 1). In all cases, the

sampling sites corresponded to suburban areas, either

close to large inland cities (Paris, Lyon) or to the sea-

side (Charente, Ballyclare). Biofilm samples were col-

lected by scraping the surface of the corresponding

mineral substrate (5–10 cm2) with a sterile spatula. The

material collected was stored at )20 �C until use. DNA

extraction was carried out from approximately 250 lL

of scratched powder (measured in labelled eppendorf

tubes, weight varied between 0.2–0.4 mg) using the

Power soil DNA extraction kit from Mobio (Carlsbad,

CA USA) following the instructions of the manufac-

turer. DNA was eluted in 80 lL of Tris-HCl, pH 8, and

conserved at )20 �C.

PCR amplification of SSU rDNA regions, libraryconstruction and sequence analysis

SSU rDNA genes were amplified using specific primers

for each domain of life. For bacteria, we also amplified

the adjacent internal transcribed spacer (ITS) region,

using the primers 27F (AGAGTTTGATCCTGGCTCAG)

and 23S-1R (GGGTTTCCCCATTCGGAAATC). Eukary-

otic SSU rDNA was amplified using specific primers

82F (GAAACTGCGAATGGCTC) and 1520R (CYG-

CAGGTTCACCTAC). In order to maximize the chances

to amplify archaeal SSU rDNA genes, we used combi-

nations of the following archaea-specific primers with the

reverse primer 1492R (GGTTACCTTGTTACGACTT):

Ar109F (AC(G ⁄ T)GCTGCTCAGTAACACGT), ANMEF

(GGCTCAGTAACACGTGGA), W36 (TCCAG

GCCCTACGGGG) and 21F (TTCCGGTTGATCCT

GCCGGA). Despite the use of archaea-specific primers,

archaeal genes failed to amplify. PCRs were carried out

in 25 lL of reaction buffer, containing 1 lL of the eluted

DNA, 1.5 mM MgCl2, dNTPs (10 nmol each), 20 pmol of

each primer and 0.2 U Taq platinum DNA polymerase

(Invitrogen). PCRs were performed under the following

conditions: 35 cycles (denaturation at 94 �C for 15 s,

annealing at 50 to 55 �C for 30 s, extension at 72 �C for

2 min) preceded by 2 min denaturation at 94 �C and

followed by a 7-min extension at 72 �C. In all cases,

PCRs with no DNA were carried out in parallel as neg-

ative controls. Clone libraries were constructed using

the TopoTA cloning kit (Invitrogen, Carlsbad, CA,

USA) according to the manufacturer’s instructions.

Clone inserts were PCR-amplified using flanking vector

primers, and SSU rDNA were partially sequenced with

either 1492R or 1520R. Representative SSU rDNA clones

were fully sequenced by using forward primers. Com-

plete sequences were assembled using Code Aligner

(CodonCode Corporation; http://www.codoncode.com)

prior to phylogenetic analyses.

Phylogenetic analyses

Our sequences were compared with those deposited in

GenBank by BLAST (Altschul et al. 1997). We retrieved

the closest environmental sequences identified in this

way and included them in an alignment also containing

sequences from the closest cultivated members and

some representative sequences of the major taxa found.

Sequences were aligned using MUSCLE (Edgar 2004),

and the multiple alignment was then manually edited

using the program ED from the MUST package (Phi-

lippe 1993). Preliminary neighbour-joining (NJ) trees

were constructed for the different prokaryotic and

eukaryotic taxa. Clusters of sequences sharing at least

98% nucleotide identity were considered to belong to

the same operational taxonomic unit (OTU, see below).

For further phylogenetic analyses, we selected subsets

of sequences belonging to the different major taxa iden-

tified. Each subset of sequences contained representa-

tive OTU sequences, as well as their closest

environmental sequences in GenBank and a variety of

reference sequences of described species belonging to

the taxon under consideration. The number of

sequences belonging to each OTU retrieved from each

� 2012 Blackwell Publishing Ltd

Tab

le1

Sam

ple

su

sed

inth

isst

ud

yan

dd

etai

lsab

ou

tth

eir

asso

ciat

edb

acte

rial

and

euk

ary

oti

cd

iver

sity

esti

mat

es

Sam

ple

Nam

e

Geo

gra

ph

ic

ori

gin

Ty

pe

of

sub

stra

teC

oo

rdin

ates

Bac

teri

aE

uk

ary

ote

s

No

.o

fS

SU

rDN

A

seq

uen

ces

anal

yse

d

No

.o

f

OT

Us

Sim

pso

nS

han

no

nC

hao

1

% Co

ver

age

No

.o

f

SS

U

rDN

A

seq

uen

ces

anal

yse

d

No

.o

f

OT

Us

Sim

pso

nS

han

no

nC

hao

1

% Co

ver

age

A3.

1L

yo

n,

Fra

nce

Min

eral

com

po

site

46¢’1

2¢09

.44

¢’N5¢

’11¢

23.3

3¢’E

201

470.

063.

212

8.2

9165

100.

331.

3712

90.8

A3.

3L

yo

n,

Fra

nce

Min

eral

com

po

site

46¢’1

2¢09

.44

¢’N5¢

’11¢

23.3

3¢’E

172

470.

102.

968

.187

.268

160.

142.

2840

.590

.9

A2.

1aC

har

ente

-Mar

itim

e,

Fan

ce

Min

eral

com

po

site

45¢’5

6¢38

.01

¢’N1¢

’18¢

24.3

3¢’W

186

310.

102.

983

94.6

5012

0.29

1.7

22.3

90

A2.

3C

har

ente

-Mar

itim

e,

Fan

ce

Min

eral

com

po

site

46¢’0

1¢57

.48

¢’N1¢

’22¢

44.5

0¢’W

588

0.32

1.3

6.5

100

475

0.54

1.04

1895

.7

A2.

4bC

har

ente

-Mar

itim

e,

Fan

ce

Min

eral

com

po

site

46¢’0

9¢28

.59

¢’N1¢

’09¢

13.9

4¢’W

254

400.

133.

113

5.9

94.1

468

0.13

2.09

1297

.8

RP

2.6

Po

rtC

erg

y,

Par

is

reg

ion

,F

ran

ce

Min

eral

com

po

site

49�2

.368

¢N2�

3.66

6¢E

185

290.

272.

154

.693

.518

530

0.32

1.97

91.7

594

.05

B1.

3B

ally

clar

e,

No

rth

ern

Irel

and

Min

eral

com

po

site

54�

45.1

12¢N

6�0.

156¢

W

849

0.41

1.18

1497

.611

419

0.13

2.5

5194

.7

B1.

4B

ally

clar

e,

No

rth

ern

Irel

and

Lim

esto

ne

54�

45.1

12¢N

6�0.

156¢

W

121

330.

058

3.1

89.5

93.4

128

280.

22.

448

.194

.5

B1.

6B

ally

clar

e,

No

rth

ern

Irel

and

Co

ncr

ete

54�

45.1

12¢N

6�0.

156¢

W

112

200.

102.

633

95.5

126

150.

41.

3482

92.8

B1.

7B

ally

clar

e,

No

rth

ern

Irel

and

Gra

nit

e54

�45

.112

¢N6�

0.15

6¢W

126

230.

092

3.1

84.6

94.4

127

180.

281.

857

94.5

n.d

.,n

ot

det

erm

ined

.

M ICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3855

� 2012 Blackwell Publishing Ltd

3856 M. RAGON ET AL.

environmental sample is indicated in the figures show-

ing the corresponding phylogenetic trees. Final phyloge-

netic trees were then reconstructed using final data sets

by maximum likelihood (ML) using TREEFINDER (Jobb

et al. 2004) applying a general time-reversible model of

sequence evolution (GTR), and taking among-site rate

variation into account by using a four-category discrete

approximation of a C distribution. Gaps and ambigu-

ously aligned positions were excluded from our analy-

sis. The number of positions retained for each

phylogenetic tree is indicated in the corresponding fig-

ure legends. ML bootstrap proportions were inferred

using 1000 replicates. Phylogenetic trees were viewed

using the program FIGTREE (http://tree.bio.ed.ac.uk/

software/figtree/). The alignments used to build each

phylogenetic tree (alignments.NEX) as well as the phy-

logenetic trees obtained (Phylogenetic trees.NWC) are

provided in NEXUS format as supplementary material.

The sequences reported in this study have been depos-

ited in GenBank with accession numbers JQ627396-

JQ627584.

Statistical methods for community analyses

Distance matrices were generated from high-quality

partial sequences (sequences longer than 600 bp having

a minimum Phred score of 20 and then revised manu-

ally) for each clone library using ClustalX (Larkin et al.

2007) and used as input for DOTUR (Schloss & Han-

delsman 2005). Diversity indices (Table 1) and accumu-

lation curves for OTUs were calculated using DOTUR

at a cut-off value of 0.02, that is, ‡98% sequence iden-

tity. Coverage values were calculated using the Good

estimator (Good 1953) following the equation

C = (1)n ⁄ N) · 100, where C is the percentage of cover-

age of the library, n the number of singletons and N the

total number of clones analysed. To compare commu-

nity compositions, OTU abundance data were squared-

root transformed, and pairwise dissimilarities among

samples were calculated using the Bray–Curtis index

(Bray & Curtis 1957). Ordination methods were used to

structure high-dimensional community composition

data along simple axes expressing overall compositional

similarity and dissimilarity between sites. We used non-

metric multidimensional scaling (NMDS) as an uncon-

strained ordination method to depict major

compositional variations among sites. To assess to

which extent community composition was related to

eco-environmental variation, we retained the following

variables as putative significant descriptors: latitude

and longitude (continuous geographic descriptors), and

16 parameters related to temperature, precipitation,

humidity and dew point levels (minimum, maximum

and averaged over, respectively, the last month and the

last year prior to sampling). As these variables were

highly correlated, we used a principal component anal-

ysis (PCA) to summarize eco-environmental variation

into synthetic independent variables (principal compo-

nents, PCs). The first 3 PCs accounted for 99.9% of the

total variation and were retained as variables. In addi-

tion, we considered the type of substrate and the loca-

tion of each sample as categorical variables. The

contribution of eco-environmental descriptors in shap-

ing the community composition was tested first by fit-

ting each variable with the unconstrained ordination

variables and tested using 105 permutations. Then, we

performed constrained correspondence analysis (CCA)

to explore the variation explained by the environmental

variables (Legendre & Legendre 1998). The significance

of all descriptors considered simultaneously, individu-

ally or in partial models was assessed using 105 permu-

tations. To construct distance–decay similarity curves,

pairwise community similarity between the samples

was calculated based on the presence or absence of each

OTU using the Sørensen’s index (Sørensen 1957). Dis-

tance–decay curves were constructed separately for bac-

teria and eukaryotes by plotting the Sorensen’s index as

a function of the geographic distance between sites.

Least-squared regression was then fitted to the data

and the slope determined. The 95% confidence interval

was estimated using 105 bootstrap resampling. All the

data analyses were conducted within the R statistical

environment (R Development Core Team 2011) using

the package vegan (Oksanen et al. 2011).

Results and discussion

Prokaryotic and eukaryotic diversity in sunlight-exposed microbial biofilms

We first characterized by SSU rRNA gene-based

approaches the diversity of prokaryotic and eukaryotic

microorganisms associated with epilithic biofilms grow-

ing on mineral composite samples and other mineral

substrates (concrete, limestone and granite) in different

sites (Table 1). Biofilms appeared mature, exhibiting

more or less mixed coloration patterns going from

green or red to black.

We aimed at studying the diversity of microorgan-

isms belonging to the three domains of life, Archaea,

Bacteria and Eucarya, by amplifying SSU rRNA genes

using domain-specific primers. However, despite the

use of several archaeal-specific primer combinations

and nested PCR attempts, we failed to obtain archaeal

amplicons. This strongly suggests that archaea are

absent in this type of biofilms or occur in too small

amounts to be detected by molecular methods. This

result is in agreement with recent observations in

� 2012 Blackwell Publishing Ltd

MICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3857

sunlight-exposed biofilms growing on similar substrates

(Ragon et al. 2011). We succeeded in amplifying bacte-

rial and eukaryotic SSU rRNA genes. In total, we analy-

sed 1499 sequences for bacteria and 997 for eukaryotes,

which were classed in different operational taxonomic

units (OTUs) (Table 1). OTUs were defined at a cut-off

level of 0.02 (‡98% sequence identity), which appears

to be appropriate for the species-level distinction in bac-

teria and also among most protists (Caron et al. 2009).

Bacterial and eukaryotic gene libraries from the differ-

ent samples were explored until sequence accumulation

curves reached a plateau indicating near-saturation

(Figs S1 and S2, Supporting information). Accordingly,

coverage values for each gene library exceeded 90% in

all cases (Table 1), which suggests a comprehensive

sampling of the dominant microbial groups present in

the different biofilms. These data were used to calculate

different richness estimators (Table 1). OTU richness

varied between sites for both bacteria and eukaryotes.

Similarly, the relative contribution of each OTU to the

total diversity identified in each sample fluctuated. In

some cases, few abundant OTUs dominated prokaryotic

and eukaryotic libraries (e.g. A2.3); in others, many dif-

ferent OTUs at intermediate frequencies contributed to

both prokaryotic and eukaryotic diversity (e.g. B1.4)

(Fig. S3, Supporting information).

The relative proportions of different major taxa iden-

tified in the bacterial and eukaryotic gene libraries

Bacteria

AlphaproteobacteriaacteriaacteriaacteriaBetaproteobacteriaGammaproteobacterteriaterterDeltaproteobactecteriactecteActinobacteriaeriaeriaeriaAcidobacteriaBacteroidetesDeinococcalesFirmicutesCyanobacteriaSpirochaetaUncertain classification

Viridiplantae; ChlorophytaViridiplantae; StrepthophytaFungi; AscomycotaFungi; BasidiomycotaFungi; ChytridiomycotaStramenopilesAlveolataAlveolataAlveolataAlveolataAmoebozoaCryrr ptophytaMetazoa

0000

BBacccttteree ia

Eukaryrr otes

BBBaaaaaallyclarellyclarararllycllyc

Paris

Charentemaritime

Lyon

Fig. 1 Taxonomic distribution of bacterial and eukaryotic SSU rDNA

strates in different geographic locations in Northern Ireland and Fran

� 2012 Blackwell Publishing Ltd

from biofilms are shown in Fig. 1. The relative abun-

dance of SSU rRNA gene sequences is currently used

as a proxy for the relative abundance of microbial

OTUs in environmental studies. However, this can

only be considered as a semiquantitative measurement

because of potential biases linked to different factors,

including differential cell lysis, gene copy number and

PCR amplification. Although the phylogenetic distribu-

tion of sequences varied among samples and sites,

some taxa were clearly dominant in all or most of

them. Thus, Deinococcales, Actinobacteria and, in some

cases, Acidobacteria dominated control mineral bio-

films, with collective proportions reaching 70–80%. The

major compositional difference observed between con-

trol mineral samples and the other substrates was the

high proportion of Deinococcales in the former and

their scarcity in the latter. Bacteroidetes exhibited an

opposite trend. The proportion of Alphaproteobacteria

was remarkably constant in all samples (between 5% and

20%). Other taxa usually associated to phototrophic

biofilms, such as Cyanobacteria, were present in almost

all bacterial gene libraries, albeit in relative low

proportions.

In the case of eukaryotes, all the samples were domi-

nated by green algae (Chlorophyta) or, in some sam-

ples, their close Streptophyta relatives, and ascomycete

fungi (Ascomycota). These groups represented collec-

tively more than 80–90% of sequences in most gene

BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBacteria Eukaryrr otes

Eukaryrr oooottttes

B1.3_33 Mineral composite

B1.4_ Limestone

B1.6_ 66 Concrete

B1.7_ Granite

RP2.6

A2.1a

A2.3

A2.4b

A3.1

A3.3

0

sequences from phototrophic biofilms growing on mineral sub-

ce.

3858 M. RAGON ET AL.

libraries. The only exception was the sample A3.1 from

the Lyon area, where metazoan sequences represented

a relative high proportion. As metazoans contribute

many cells and, consequently, more DNA per individ-

ual, their relative high counts in this library should be

considered with caution. To explore at finer phyloge-

netic scale the microbial diversity in the biofilms, we

constructed phylogenetic trees with near full-length

sequences representative of the different OTUs and

0.2

Deinococcus radiophilus (Y11333)

Benthic hotspring cyanobacteriu

Antarctica McMurdo desert quartz AN2_3 (

Rivularia atra BIRMGR1 (AM

Myxosarcina sp. PCC732

Coleodesmium sp. ANTLH52

A2.1a_1C_17

A3.3_1C_28

Unc. bacterium AKIW

Unc. Deinococcus AV_8S_

B.1_7_1B_98

RP2.6_1B_91

A2.1a_3B_96

Unc. Deinococcus sp

A2.3_2B_46

Spacecraft cleanroom bacterium JSC2_

Chroococcidiopsis sp. CC1 (DQ9148

B1.3_1B_71

Leptolyngbya sp. FI5_2HA3 (HM0186

Hypolithic cyanobacterium C22 (FJ23

B1.6_1B_108

Gloeobacter violaceus (AF132791)

Chroococcidiopsis th

Leptolyngbya frigida ANT_LH52B3 (AY493612

Phormidium priestleyi ANTLG24 (A

Prochlorococcus marinus NATL2A (CP00

Petalonema sp. ANT_GENT

A3.3_3B_137

Microcystis aeruginosa SPC77

Alps_endolithic bacter

Truepera radiovictrix (DQ022077)

Rivularia sp. MU24_UAM_305 (Tolypothrix sp. (FJ660999)

B1.6_1B_88

Leptolyngbya badia CRS_1 (EF429)

A2.1a_1C_5

Deinococcus 100M2_

Unc.Chroococcidiopsis sp. AR2_

A2.1A_3B_59

A2.1a_2B_6

Anabaena variabilis KCTCAG10064

Loriellopsis cavernicola (HM74

A2.1a_3B_65

A2.4b_1C_33

A2.4B_2B_24

Gloeothece sp. KO68DGAs (

Pleurocapsa minor SAG4

Deglaciated soil bacterium AK4AB1_0

Unc. Chroococcidiopsis sp. (FJ805928)

Dolomite rock cyanobacterium DO

Alps endolithic bacterium (

Chroococcus sp. 2T05h (F

Chroococcidiopsis

A3.1_1C_43

Gloeocapsa sp. PCC73106 (AB

Dermocarpella incrassa

Leptolyngbya sp. LLi18 (DQ786166)

B1.3_1B_81

B.1_7_1B_78 Spacecraft cleanroom bacter

Leptolyngbya appalachiana (EF429286)

A3.1_3B_63

A3.1_3B_118

Pleurocapsa sp. CALU_1

A3.3_1C_62

A3.3_3B_54

A3.3_3B_139

Deinococcus radiodurans (M21413

Nostoc commune M13 (AB0884

Cyanobium sp. JJ25_2 (AM710370)

A3.1_1C_7

Unc. Deinococcus 20E7

Endolithic cyanobacterium EPLS0

Pseudophormidium sp. ANT_GE

Limestone cliff cyanobacterium O

Unc. dust bacterium BF000

A3.1_1C_45

Ruidera stromatolite cyano

Deglaciated Andes soil bacterium 500

Thermus thermo

Thermosynechococcus elongatus (BA000039)

RP2.6_1B_24

Annapurna soil cyanobacteriu

Calothrix desertica PCC7102 (AF13

Gloeotrichia echinulata URA3 (AM2

B1.3_1B_72

Chroococcus sp. JJCM Ch

100

52

53

98

100

85

59

56

95

75

100

96

92

77

62

97

70

74

66

90

81

65

56

89

82

64

67

93

53

82

100

61

99

98

98

80

74

87

71

67

93

98

9868

79

98

68

91

55

99

72

85

64

94

65

96

98

98

77

57

76

90

94

92

70

100

Fig. 2 Phylogenetic tree of SSU rDNA sequences belonging to the Th

reconstructed by maximum likelihood using 664 nonambiguously al

nodes. Coloured circles indicate the number of sequences affiliating

biofilm samples. Accession numbers are given in parentheses. The sc

branch length. Unc., uncultured.

including their closest relatives in GenBank and refer-

ence sequences of cultivated species (Figs 2–5 and

Figs S4–S5, Supporting information). In general, our

most abundant phylotypes were very closely related to

organisms well known for their adaptation to extreme

conditions, such as Deinococcales and Actinobacteria

(Fig. 1). We identified a series of OTUs that were very

closely related among them and to environmental

Deinococcus sequences retrieved from endolithic

m LLi67 (DQ786165)

FJ805912)

230675)

5 (AJ344562)

B5 (AY493596)

875 (DQ129342)

B11 (EU341277)

. J34H5 (DQ366015)

A6 (DQ532167)

63)

84)

0787)

ermalis PCC7203 (AB039005)

)

Y493580)

0095)

NER2_8 (AY493624)

7 (EF121241)

ium (AB374378)

EU009149)

A8 (DQ514091)

1 (FJ805854)

(DQ23482)

8318)

AB067580)

_99 (AJ344564)

5A (GQ396895)

L63 (HM224428)

AB473897)

R798926)

sp. SAG_2025 (AM709635)

039000)

ta (AJ344559)

ium JSC8_C11 (DQ532230)

126 (DQ293994)

)

05)

(DQ366013)

31 (EF522220)

NTNER2_1 (AY493615)

U_20 (GQ162224)

2D090 (AM697579)

bacterium D1E09 (EU753646)

M2_C2 (DQ514209)

philus (X07998)

m B107212D (HQ189094)

2779)

30705)

rooM (AM710384)

B1.3RP2.6A2.1

aA2.3A2.4

bA3.1A3.3 B1.4B1.6B1.7

44 27

6 23 2

1 2 4 3

5 7 1

319 7 6 21 38

331 1 2 4

10 112 2 4 13

24

2 4 1

31 1

2 4 3 1

1

1 1

23

2 4 1

1 11

6

4 2 2

Ther

mus

/Dei

noco

ccus

Cya

noba

cter

ia

1

Paris Lyon Charente Ballyclare

ermus ⁄ Deinococcus group and the Cyanobacteria. The tree was

igned positions. Bootstrap values higher than 50% are given at

to the different OTUs found in gene libraries from the different

ale bar indicates the number of changes per nucleotide per unit

� 2012 Blackwell Publishing Ltd

0.06

A2.4B_3B_55

A3_3_3B_125

Andes altitude volcano soil G11_SB3A (FJ592818)

A3.3_3B_58

B1.4_1B_73

Arthrobacter sp. D42008 (FJ535483)

Soil Rubrobacteridae Gsoil_1167 (AB245333)

Kineococcus sp.1P02MC (EU977818)

Geodermatophilus sp. (X92364)

Atacama desert actinobacterium E1B_A4_114 (EF016798)

A2.1A_3B_58

A2_4B_3B_139

Demetria terragena HK1_0089 (NR_026425)

A1.A_3B_33

Nocardia asteroides (Z36934)

Phycicoccus dokdonensis DS_8 (EF555583)

B1.6_1B_93

Actinobacterium from anoxic bulk soil

Pseudonocardia sp. TFS_1235 (EF216360)

Frigoribacterium sp. NMC14 (GU321353)

A2.1A_3B_119

A2.4B_3B_57

RP1.5_1B_140

Ancient wall painting A8 (AM746686)

B1.4_1B_109

Actinomycetales Ellin143 (AF408985)

Pseudonocardia petroleophila (GU083570)

Pseudonocardia alaniniphila (FR733722)

B1.4_1B_68

Unc. bacterium ncd1968e04c1 (JF171627)

Cellulomonas chitinilytica (AB268586)

A3.3_3B_99

Conexibacter woesei (NR 028979)

A3.1_3B_98

Marmoricola sp. BAC242 (EU180564)

B1_7_1B_62B1_4_1B_80

Geodermatophilaceae bacterium CIBE_G12 (AY903278)

A3.1_3B_67

B1.7_1B_110

Frankia sp. strain_AVN17s (L40613)

Unc. bacterium 2-9A (EU289511)

Pseudonocardia sp. YIM_63233 (FJ817379)

Uranium mine actinobacterium Sliv_136 (FM877677)

B1.3_1B_79

Blastococcus saxobsidens (FN600641)

Frigoribacterium sp. Enf57 (DQ339618)

Antarctica actinobacterium UMAB_cl_47 (FN811231)

Curtobacterium flaccumfaciens (EU977762)

Volcanic deposit Franckia KVD_unk_16 (DQ490442)

Antarctica actinobacterium UMAB cl_46 (FN811230)

A2.1A_3B_97

McMurdo cryptoendolithic bacterium FBP218 (AY250865)

Pseudonocardia sp. RI44_RCaA106 (AB546273)

Humicoccus sp. KCTC_19426 (EU939310)

Blastococcus aggregatus (AJ430193)RP1.5_1B_101

Unc bacterium AKIW658 (DQ129587)

Blastococcus sp. CNJ868_PL04 (DQ448697)

A2.4B_3B_31

A3.3_3B_105

Marmoricola sp. Sco_A36 (FN386723)

Antarctic cryptoendolithic bacterium FBP460 (AY250884)

Frankiaceae isolate G10Namibia (X92365)

A2.1A_3B_64

A3_3_3B_118

Unc bacterium 1_1E (EU289419)

B1.4_1B_105

Unc Microbacteriaceae B04_04A (FJ542985)

Rubrobacter xylanophilus (X87135)

Kineosporia aurantiaca (X87110)

Amycolatopsis sp. 232068 (GU130127)

A2.4B_3B_4

Humicoccus flavidus (DQ321750)

A3.1_3B_120

B1.6_1B_73

Antarctic bacterium R_8287 (AJ440992)

Modestobacter sp. BMG5755 (FJ966173)

A3.3_3B_110

Antarctica actinobacterium UMAB_cl_13 (FN811197)

Lapillicoccus jejuensis (AM398397)

Clavibacter sp. Ens76 (DQ339594)

Mycobacterium chlorophenolicum (X79094)

Unc. bacterium BF0001B03

Antarctica soil actinobacterium P23 (DQ351736)

A2.4B_3B_36

Falkland soil actinobacterium Fl 1F C10 (EF220492)

Dust-associated bacterium BF0002B089 (AM697139)

Solirubrobacter soli (AB245334)

A3.1_2B_38

Atacama desert actinobacterium E1B_B11_114 (EF016803)

Unc. bacterium ncd1969g03c1 (JF171716)

A2.1A_3B_98Solirubrobacter sp. Gsoil 917 (AB245335)

B1.4_1B_104

Actinomycetales bacterium Gsoil_1632 (AB245397)

A3.3_3B_91

A3.1_3B_45

A3.1_3B_40

Patulibacter sp. P4_5 (EU710748)

Unc. bacterium ncd1355d01c1 (JF116511)

B1.4_1B_92

A3_1_3B_61

A3.1_3B_70

A2_1a_2B_9

Rubrobacter taiwanensis LS 293 (AF465803)

Pseudonocardia hydrocarbonoxydans (GU083569)

A3.1_3B_78_x

Marmoricola aequoreus (AM295338)

A3.1_3B_116

Rathayibacter iranicus ICPB70005 (FJ595101)

A2.1A_3B_84

Phycicoccus aerophilum 5516T_20 (EF493847)

B1.3_1B_55

Actinotelluria brasiliensis Tu6233 (DQ029102)

Marmoricola aurantiacus (Y18629)

B1.3_1B_58

B1.6_1B_77

A2.1A_3B_101

Pseudonocardia sp. MN08_A0270 (AB521671)

RP2.6_1B_121

Amycolatopsis sp 13676C (EU741234)

Antarctic soil actinobacterium UMAB_cl_68 (FN811252)

Rubrobacter radiotolerans P1 (NR_029191)

B1.4_1B_62

Nocardioides plantarum (X69973)

Semi-coke Estonia 4_C16_66 (EF540514)

Cliff cactus soil Blastococcus sp. OS1_29 (FN178402)

A2.1A_3B_116

Uranium mine actinobacterium Sliv_78 (FM877676)

B1.7_1B_118

Rock varnish bacterium DRV B001 (AY923078)

A2.3_2B_45

B1.6_1B_128

Arthrobacter sp. VTT_E (EF093123)

A2.4B_3B_132

Antarctica soil Rubrobacter sp. 354H (AY571811)

Atacama Desert soil bacterium E1B D2 114 (EF016816)

Dust-associated bacterium FA04E02 (FM872640)

Blastococcus sp. YIM_68236 (GQ494034)

Arthrobacter sp. 210_41 (GQ199743)

93

78

71

90

68

77

80

78

88

97

87

90

58

56

99

99

71

61

54

85

74

94

68

67

79

98

94

99

67

88

99

81

83

62

87

87

100

54

57

88

75

87

87

76

55

53

62

96

84

100

90

80

99

100

58

94

56

96

91

58

99

98

89

93

50

83

75

81

97

70

99

68

99

98 53

86

70

55

96

58

99

69

61

99

83

83

100

75

57

66

76

99

97

100

100

75

77

82

98

97

93

64

82

55

96

71

92

93

8383

77

1

1 7

7 12

22

3 1 3 2

3

127

75

11478 11

76 9 116

11131

17 3111 1

15 2 112

1

2

21411

13 1111

429

3195122

11

11 3

11 331

11 5

4

349

1422

11

19

1 5 1121 113 1

273 116

Rub

roba

cter

ales

B1.3RP2.6A2.1

aA2.3A2.4

bA3.1A3.3 B1.4B1.6B1.7

Paris Lyon Charente Ballyclare

selatecymonit c

A

3

115

3 1

Fig. 3 Phylogenetic tree of SSU rDNA sequences belonging to the Actinobacteria. The tree was reconstructed by maximum likeli-

hood using 759 nonambiguously aligned positions. Bootstrap values higher than 50% are given at nodes. Coloured circles indicate

the number of sequences affiliating to the different OTUs found in gene libraries from the different biofilm samples. Accession num-

bers are given in parentheses. The scale bar indicates the number of changes per nucleotide per unit branch length. Unc., uncultured.

MICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3859

� 2012 Blackwell Publishing Ltd

0.04

B1.3_1E_68

B1.6_1E_95

RP2.6_1E_124

RP2_6_1E_42

Collophora rubra STE_U_6110 (GQ154629)

RP2_6_1E_146

Xanthoria candelaria (AM498768)

RP1.5_1E_47

Rock Chaetothyriales sp. TRN1 (FJ358319)

Knufia cryptophialidica (EF137364)

RP2.6_1E_54

RP2.6_1E_190Chaetothyriales sp. TRN115 (FJ358324)

Crinula caliciiformis (AY544729)

A2.4b_1E_30

Bensingtonia miscanthi JCM5733 (D38236)

Cucurbitaria berberidis (U42481)

Aspergillus nidulans CBS100_20 (X78539)

Bullera nica (D78330)

RP2.6_1E_113

RP1.5_1E_23

Synchaetomella lunatospora (AY342013)

Capnobotryella sp. MA_3612 (AM746202)

Porpidia macrocarpa (AY456687)

Phoma destructiva (AB454203)

A3_3_1E_52

Verrucaria mucosa AFTOL_ID 2264 (EF689877)

Xanthoria sorediata (AM495020)

A3_3_1E_22

RP2_6_1E_15

Taphrina pruni subcordatae (AB000957)

RP2.6_1E_166

RP2.6_1E_122

Sporobolomyces foliicola (AB021671)

RP1.5_1E_42

Rhodotorula marina (AB126645)

Coniosporium apollinis (Y11713)

Endoconidioma populi (AY604526)

B1.6_1E_77

Pleospora herbarum (U43448)

B1.3_1E_49

RP2_6_1E_148

A3.3_1E_59

Scleroconidioma sphagnicola (AY220610)

Pleosporales sp. CBS_536

Sporobolomyces coprosmae (D66880)

Lepolichen coccophorus (AF274110)

RP2.6_1E_27

B1_6_1E_56

Thelebolus stercoreus CBS_717_69 (AY942193)

Sarcinomyces petricola (Y18702)

Cryptococcus dimennae (AB032627)

Cyphellophora laciniata AFTOL_ID 1033 (EF413618)

Unc. Tremellomycetes (FM178233)

B1.4_1E_57

Phaeosphaeriaceae sp. 11F (FJ744122)

Coniosporium sp. MA4597 (AJ972863)

RP2.6_1E_144

Norrlinia peltigericola (AY779280)

A2_1a_1E_33

Leptosphaeria maculans (U04238)

RP1.5_1E_185

RP1.5_1E_163

RP1.5_1E_150

Bensingtonia phylladus JCM_7476 (D38237)_

B1.6_1E_101

B1.6_1E_116

RP 1_5_1E_142

A2.2a_1E_2

Acrospermum compressum (AF242258)

RP1.5_1E_147

RP1.5_1E_146

Cochliobolus sativus (U42479)

B1.3_1E_142

A2_1a_1E_28

Phoma sp. CCF3818 (FJ430776)

Rock Chaetothyriales sp. TRN4 (FJ358320)

A2.2a_1E_40

Rhodotorula aurantiaca (AB030354)

B1.4_1E_121

B1.7_1E_116

RP1_5_1E_21

Alternaria alternata (U05194)

A2.4b_1E_33

Geomyces destructans MmyotGER_1 (GU999983)

A3.3_1E_12

A2.2a_1E_17

Filobasidium floriforme (D13460)

A2.1a_1E_41

B1.3_1E_79

Cryptococcus sp. CBS6578 (AB085801)

RP1.5_1E_130

Bullera globispora (D31650)

Kondoa malvinella (D13776)

B1.3_1E_130

Mycosphaerella fragariae CBS_719 (EU167605)

B1.3_1E_143

Polyblastia melaspora AFTOL_ID 1356 (EF413600)

A2_4b_1E_13

Taphrina californica (D14166)

A2.2a_1E_51

Kurtzmanomyces nectairei (D64122)

A3.3_1E_3

Alternaria sp. CBS AFTOL_ID_1579 (DQ678016)

RP1.5_1E_10

Lecania cyrtella (AF091589)

Leucosporidium scottii (AY707092)

Cercospora beticola CPC11557 (AY840527)

A2_1a_1E_21

Bipolaris sp. JF2 (FJ666899)

Cladosporium cladosporioides (EF392680)

A3.1_1E_9

RP1.5_1E_15

Aureobasidium pullulans P2A1 (EU682927)

Capronia dactylotricha AJ232943

RP1.5_1E_55

Phaeosphaeria sp. sn23_1 (EU189215)

RP2.6_1E_157

A3.3_1E_13

B1.4_1E_69

Unc. eukaryote Th111207 9 (HM030914)

Cryptoendolithic fungus Arthoniales sp. CCFEE_5303 (GU250331)

Verrucaria dolosa AFTOL_ID_2253 (EF689867)

Alternaria raphani AR6 (U05199)

Rhinocladiella sp. MA4765 (AJ972862)

Capnobotryella sp. MA_3615 (AM746203)

Capnobotryella sp. MA_4701 (AJ972856)

B1.7_1E_96

RP.1_5_1E_9

Trichoglossum hirsutum AFTOL_ID_64 (AY544697)

B1.7_1E_113

RP1.5_1E_3

Physcia aipolia (AF241542)

A3.1_1E_6

Cladophialophora sp. CBS_985.96 (AJ232953)

Teratosphaeria microspora 36i (EU343025)

Davidiella tassiana 104m (EU343102)

A2_1a_1E_27

Coniosporium perforans (Y11714)

B1.3_1E_65

Ascomycote sp. CCFEE_5493 (GU250354)

B1.3_1E_55

B1.4_1E_126

RP2.6_1E_173

74

98

96

80

45

100

63

75

72

82

100

95

100

96

98

54

70

99

99

61

71

98

84

93

57

68

98

60

6391

98

96

99

95

68

86

83

76

100

100

64

53

100

93

78

100

100

83

59

64

100

68

99

68

92

73

96

87

81

78

98

99

92

99

67

61

64

98

96

91

67

99

100

79

95

65

79

100

52

87

93

100

99

54

99

87

99

100

81

82

10066

100

99

98

91

85

8279

99

87

99

99

99

92

73

56

99

Basidiomycota

Ascomycota

1

2

2

126 5

21

2

2

5

22

4 5 21

111

211

12 1

4

7

114 21

23 171

33

9

4

21 59

1 2 1

2

1 7

23

7

4

5 16 515

1 4

214

2 31

B1.3

RP2.6A2.1aA2.3A2.4bA3.1A3.3

B1.4B1.6B1.7

Paris Lyon Charente Ballyclare

Fig. 4 Phylogenetic tree of fungal SSU rDNA sequences from sunlight-exposed biofilms growing on mineral substrates. The tree

was reconstructed by maximum likelihood using 1403 nonambiguously aligned positions and rooted with four chytrid sequences.

Bootstrap values higher than 50% are given at nodes. Coloured circles indicate the number of sequences affiliating to the different

OTUs found in gene libraries from the different biofilm samples. Accession numbers are given in parentheses. The scale bar indicates

the number of changes per nucleotide per unit branch length. Unc., uncultured.

3860 M. RAGON ET AL.

� 2012 Blackwell Publishing Ltd

0.2

Stichococcus bacillaris CCAP379 (FR717539)

B 1 4 1E 144

Botryidiopsidaceae sp. PAB 485 (AM695636)

A 3 3 1E 55

B 1 6 1E 70

RP 2 6 1E 186

Oxytricha granulifera (X53486)

A 2 1a 1E 2

TTibet desert tuff eukaryote VE52 (FJ790587)

A 3 1 1E 52

Klebsormidium flaccidum SAG 7.91 (EU434019)

B 1 4 1E 135

Stylonychia pustulata (X03947)Tetrahymena thermophila (X56165)

Didymium nigripes (AF239230)

Stichococcus bacillaris SAG 335 (EU434029)

Pyramimonas propulsa (AB017123)

Trebouxiophyte sp. UR75 (AY762607)

Trebouxia impressa (Z21551)

A 2 1a 1E 10

Unc marine eukaryote NA2 1H8 (EF526889)

B 1 7 1E 108

Ostreococcus tauri (Y15814)

A 2 1a 1E 32

Cry Anapurna alpine soil E108 09D (HQ188966)

A 3 1 1E 14

A 3 1 1E 30

Dunaliella salina (M84320)

Unc Dunaliellaceae Amb 606 (EF023293)

Parietochloris alveolaris UTEX 8 (EU878373)

RP 2 6 1E 12

RP 2 6 1E 170

Propallene sp. longiceps (AB292196)

B 1 4 1E 134

Leptus sp. AP 2010 (HM070355)

B 1 4 1E 119

RP 2 6 1E 176

Hyperamoeba flagellata (AF411289)

Chlorella saccharophila/Watanabea reniformis (X73991)

Dicksoniaceae clone Amb 629 (EF023312)

Isohypsibius prosostomus (EF620404)

B 1 4 1E 136

Trebouxia arboricola SAG 219 1a (Z68705)

A 3 3 1E 1

Chlorella sp. MBIC10057 (AB058305)

B 1 6 1E 109

Trebouxia erici (AB080310)

RP 2 6 1E 10

RP 2 6 1E 73

B 1 4 1E 117P truncata95935

B 1 4 1E 139

Trebouxia usneae 019A1 (Z68702)

Ibalia sp. JMH 2010 (GQ410645)

Eniochthonius minutissimus (EF091428)

Phyllosiphon arisari (JF304483)

A 3 1 1E 46

A 3 3 1E 42

Coenocystis inconstans (AB017435)

A 2 4b 1E 10

RP 2 6 1E 142

B 1 4 1E 132

Trebouxiophyte sp UR474 (AY762604)

Airborn eukaryote Th090408 47 (HM030919)

B 1 4 1E 66

Chlorella ellipsoidea partial (FM946018)

B 1 4 1E 131

B 1 6 1E 115

Microcaeculus JCO 1999 (AF287232)

B 1 4 1E 56

B 1 3 1E 99

Elliptochloris sp SAG 2117 (FJ648515)

B 1 7 1E 141

Discophrya collini (L26446)

B 1 3 1E 88

Ephelota sp. QD 5 (DQ834370)

Heliophrya erhardi 3 (AY007447)

B 1 3 1E 62

Uronema sp. CCAP 384 1 (FR717538)

B 1 4 1E 101

B 1 7 1E 107cher 2 1E 89

Physarum didermoides (AY183449)

B 1 7 1E 58

A 2 4b 1E 35

Myrmecia sp. H1VF1 (AF513369)

B 1 7 1E 133

Stichococcus bacillaris (AB055864)

cher 2 1E 115

Limulus polyphemus (U91490)

Oxytrichidae clone Amb 1444 (EF023975)

Printzina lagenifera (DQ399586)

Botrydiopsis pyrenoidosa (AJ579337)

B 1 4 1E 103

Acrochaete repens E093db (FJ715684)

Heterococcus caespitosus (AM490820)

Unc eukaryote Th090408 11 (HM030915)

B 1 4 1E 53

Trebouxia asymmetrica encoding (Z21553)

A 3 1 1E 27

Chlamydomonas reinhardtii (M32703)

Eimeriidae env sample Amb 1393 (EF023933)

B 1 7 1E 85

Botrydiopsis intercedens (U41647)

Chlorella sp UTEX 318 (EF159951)

Phaeothamnion confervicola (AB365203)

cher 6 1E 117

A 3 1 1E 18

Klebsormidium subtilissimum (EF372517)

Unc Dunaliellaceae Amb 930 (EF023670)

85,

55

100

100

52

65

98

72

96

92

64

66

56

93

60

58

88

99

100

100

92

100

76

100

52

91

100

61

72

100

91

51

100

9367

90

80

53

94

100

57

79

56

100

90

99

99

87

73

75

100

96

99

99

52

100

100

93

99

75

100

58

88

100

95

87

92

100

75

72

99

90

100

96

Virid

ipla

ntae

Alveolata

Amoebozoa

Met

azoa

Stramenopiles

2

1

28 4

3 1

1

513 2

78

1 2 5

6 11 25 11

1

4 11

24 36 3

1

7

92 4 11 1

5

1111

11 2

82812

43 1

7 41 11

32

98 2972115

44 312

2 1 1 1

(1/2)

B1.3RP2.6A2.1

aA2.3A2.4

bA3.1A3.3 B1.4B1.6B1.7

Paris Lyon Charente Ballyclare

3 1

Stre

ptop

hyta

Chl

orop

hyta

Fig. 5 Phylogenetic tree of nonfungal eukaryotic SSU rDNA sequences identified in epilithic biofilms. The tree was reconstructed by

maximum likelihood using 1287 nonambiguously aligned positions. Bootstrap values higher than 50% are given at nodes. Coloured

circles indicate the number of sequences affiliating to the different OTUs found in gene libraries from the different biofilm samples.

Accession numbers are given in parentheses. The scale bar indicates the number of changes per nucleotide per unit branch length.

Unc., uncultured.

MICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3861

� 2012 Blackwell Publishing Ltd

3862 M. RAGON ET AL.

communities at high-altitude or recently deglaciated

soils (Nemergut et al. 2007) as well as from dust parti-

cles found in buildings (Rintala et al. 2008), airplane

cabins (Osman et al. 2008) and urban aerosols (Brodie

et al. 2007) (Fig. 2). This suggests that these organisms

have wide-dispersal properties, which are most likely

due to the high intrinsic tolerance of Deinococcus to des-

iccation, UV and ionizing radiation (Daly 2009). Deino-

coccales are frequently retrieved from hot and cold

desert samples (e.g. Chanal et al. 2006; de la Torre et al.

2003; Rainey et al. 2005). These properties of extreme

resistance to UV and ionizing radiation are also shared

by members of the actinobacterial genus Rubrobacter

(Suzuki et al. 1988), which were particularly abundant

in our biofilm samples (Fig. 3). Rubrobacterales were

indeed very abundant in recently studied sunlight-

exposed biofilms in the highly irradiated Chernobyl

area (Ragon et al. 2011). In addition to Rubrobacter, a

wide diversity of other Actinobacteria was present in

our biofilms (Fig. 3). The most abundant actinobacterial

sequences had as closest relatives sequences identified

in desert endolithic communities or, as in the case of

Deinococcus, dust particles (Fig. 3). Similarly, the cyano-

bacterial sequences identified in our biofilms were clo-

sely related to sequences retrieved from endolithic or

epilithic communities growing on rocks at high-altitude

or soils and in desert areas (Fig. 2), sharing similar hab-

itats with Deinococcales and Actinobacteria (de la Torre

et al. 2003; Nagy et al. 2005; Chanal et al. 2006). Acido-

bacteria, which are important members of soil commu-

nities (Quaiser et al. 2003), were highly represented in

some of our biofilm samples as well (Fig. S4, Support-

ing information). The most abundant phylotype was

very closely related to sequences retrieved from a

recently deglaciated soil (Nemergut et al. 2007). Acido-

bacteria can also colonize rock surfaces and have been

found to dominate epilithic communities in ancient cat-

acombs (Zimmermann et al. 2006). Finally, among the

Alphaproteobacteria, the most represented OTUs were

related to the genus Sphingomonas (Fig. S5, Supporting

information), whose members are known by their abil-

ity to degrade recalcitrant natural and anthropogenic

compounds, including cyanobacterial exudates and

many xenobiotic compounds (Edwards & Lawton 2009;

Stolz 2009). The ability to degrade complex compounds

is also shared by many Betaproteobacteria (Madigan

et al. 2002), which were also present in some biofilm

samples to a moderate extent (Fig. S5, Supporting infor-

mation).

As expected, microbial eukaryotes were less diverse

than bacteria also at the OTU level (Table 1). Ascomy-

cete fungi were the most diverse group. Basidiomycete

fungi were moderately represented by sequences related

to the widely distributed yeast genera Rhodotorula,

Bensingtonia and Kondoa (Fig. 4). The most abundant

ascomycete OTUs were related to the lichen-forming

genera Polycladia and Verrucaria, often found in endo-

lithic communities (Garvie et al. 2008), and Xanthoria.

Other fungi that are potential facultative plant and ani-

mal parasites such as Phoma or Phaeosphaeria were

detected as well. Verrucaria-forming lichens have been

seen to very rapidly colonize stones in restored habitats

(Nascimbene et al. 2009). Together with fungi, green

algae dominated mineral-associated biofilms. The only

exception corresponded to one Lyon sample where

sequences related to the moss Pottia truncata were rela-

tively abundant (Figs 1 and 5). Mosses, like animals,

may bias gene libraries by contributing much DNA per

individual. Nonetheless, their presence is indicative of

mature, well-established communities. Among green

algae, the most abundant sequences corresponded to

OTUs closely related among them and ascribing to the

genus Trebouxia, which was found to dominate endo-

lithic Antarctic communities (Banerjee et al. 2000). Tre-

bouxia very often forms lichens (usually with Xanthoria),

although not necessarily (Bubrick et al. 1984). Other

highly represented OTUs affiliated to the related genera

Stichococcus, Chlorella and Printzina. Trebouxiophyceae

are well adapted to light exposure and, consequently,

resistant to UV and ionizing radiation, being often

found in building facades and also detected in Cher-

nobyl concrete-associated biofilms (Ragon et al. 2011).

This resistance is partly due to the production of

mycosporines, very efficient protective pigments (Karsten

et al. 2005, 2007). The rest of eukaryotic sequences iden-

tified corresponded to yellow-green algae (Xanthophyta,

Stramenopiles), some ciliates (Alveolata), slime-mould

amoeba and, within the animals, mites and tardigrades.

The latter are also known for their extreme resistance to

various stresses, including desiccation, ionizing radiation

and hot and cold temperatures (Mobjerg et al. 2011). The

presence of typical protozoan grazers such as ciliates or

amoeba, which have less mobility than animals, can be

taken as a further indication that these biofilms are mature

and sustain more or less complex trophic interactions.

In summary, the general microbial diversity found in

biofilms growing on sunlight-exposed mineral surfaces

was dominated by microorganisms having wide-dis-

persal ability and being highly resistant to desiccation

and UV radiation, (Deinococcales, Actinobacteria, green

algae and melanized spore-forming Ascomycetes). This

implies that these biofilm communities develop under

strong environmental selection (Gorbushina & Brough-

ton 2009). Strong selective pressures limit diversity. In

fact, these communities are much less diverse than soils

or freshwater systems. Because of their lower biological

complexity, they represent easier models to explore

questions related to biogeography.

� 2012 Blackwell Publishing Ltd

MICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3863

Comparison of community composition in epilithicmicrobial biofilms

To compare community structure in the different micro-

bial biofilms, we applied nonmetric multidimensional

scaling (NMDS). This unconstrained ordination method

permits the distribution of high-dimensional commu-

nity composition data along simple axes to yield infor-

mation about the major compositional variations among

sites (Legendre & Legendre 1998). We obtained NMDS

ordination plots for bacteria and for eukaryotes sepa-

rately. Interestingly, the distribution of biofilm samples

and OTUs was qualitatively different for bacteria and

for eukaryotes (Fig. 6).

First, in the case of bacteria, the ordination of samples

along the first axis (NMDS1) clearly discriminated the

samples from different substrates. Samples correspond-

ing to different substrates collected in the same geo-

graphic site were distributed from the extreme right

(B1.4 and B1.7, limestone and granite) to the extreme

left (B1.3, control composite substrate) of the axis, B1.6

(concrete) being intermediate (Fig. 6). This implies that

the variation related to the differences in substrates is

larger than that related to geographic variation. By con-

trast, in the case of eukaryotes, community composition

appeared to be more similar among different substrates

from the same locality than among different sites for

the same substrate, although with some variation in

terms of the OTUs present (Figs 2–5 and Figs S4–S5,

–0.4 –0.2 0.0 0.2 0.4 0.6

–0.4

–0.2

0.0

0.2

0.4

NMDS1

NM

DS

2

+

+

++++

+++++++++++++++

++

++++++

+

+++++++++++++

+

++

+++

++++++

+

++

+

++

++++++

+ ++

++++++++++++++

+++ ++++

++

Act4Act27

Act28

Act51

Act22

Bact15

Alpha26

Bact21

Act5

Bact18

Bact23

Aci10

Act7

Dei5

Cya2

Dei3

Bact2

Aci3

Aci4

Dei1

Bact19

Alpha28

Alpha6

Beta2

Act6

Alpha1

Alpha22

Bact16

Act15Bact11

Dei2

Act19

Cya9

Alpha23Act25

Beta5

Act40

Aci9Bact14Bact17

Beta9

Beta6

Alpha13

Bact12Alpha12

Act1

Bact9

Act48

Act32

Aci2

Act12

Cya13

Cya12

Cya14

Cya6

Alpha18Bact8

Delta1B1.3 B1.4

B1.6

B1.7RP2.6

A2.1a

A2.4bA3.1

A3.3

A2.3

Bacteria

Fig. 6 Unconstrained ordination analyses (NMDS) describing the ma

communities among samples. Samples are in bold black and OTUs in

indicating an area of high OTU density.

� 2012 Blackwell Publishing Ltd

Supporting Information). Thus, samples from Northern

Ireland and Paris area were placed on the left of the

NMDS1 axis, whereas samples from Charente-Maritime

were located on the right of the axis and those from

Lyon at an intermediate position (Fig. 6). Thus, differ-

ences linked to the geographic position appeared to be

more important in eukaryotes than those due to the

substrate nature.

Second, not only the distribution of samples along

axes was different in bacteria and eukaryotes, but also

that of OTUs among samples. In the case of bacteria,

OTUs appeared more disperse, whereas in eukaryotes,

OTUs tended to aggregate around samples (Fig. 6).

This implies that many bacterial OTUs were shared

between samples, whereas eukaryotic OTUs appeared

more specific to each sample (see also Figs 2–5 and

Figs S4–S5Supporting Information). In other words, this

suggests that bacteria effectively disperse more widely

than microbial eukaryotes under exactly the same envi-

ronmental constraints.

Determinants of prokaryotic and eukaryoticcommunity composition in epilithic biofilms

To understand which ecological factors determine

microbial community structure on epilithic biofilms and

to test whether they affected differently prokaryotic and

eukaryotic communities, we carried out a series of sta-

tistical analyses considering the following environmen-

–1.0 –0.5 0.0 0.5 1.0

–0.5

0.0

0.5

1.0

NMDS1

NM

DS

2

A3.3

+++

+

++

+++

+++++++

+

++

+++++++++++

++++++++++

Chlo1

Asc8

Asc16

Chlo17

Meta1

Chlo3

Chlo9

Stram2

Asc31

Asc2Asc26

Asc1

Chlo18

Asc17

Bas3

Asc30

Asc36Asc32

Asc9

Asc24

Chlo21

Meta2Asc18

Strep3

Alve2

Asc19

Asc21Asc4

Chlo19

Strep1

Chlo8

Asc3

Asc13

Alve1

Chlo6Asc34

Chlo20Asc35

Chlo16

B1.3

B1.4

B1.6

RP2.6

A2.1a

A2.3

A2.4b

A3.1

B1.7

Eukaryotes

jor compositional variation in bacterial and microbial eukaryote

grey. When many OTU labels overlap, a ‘+’ symbol was used,

3864 M. RAGON ET AL.

tal descriptors: substrate type, location, latitude, longi-

tude and a collection of environmental parameters

related to temperature, precipitation, humidity and dew

point levels. These environmental parameters correlated

among them and were summarized into 3 synthetic

independent variables following a principal component

analysis (PCA). The first principal component (PC)

accounted for 80% of the total variation, and together,

the 3 PCs selected accounted for 99.9% of it (Fig. S6,

Supporting Information).

First, we tested the linear univariate relationship

between the community composition summarized using

the NMDS ordination and each of the putative descrip-

tors considered (i.e. the 3 first PCs, substrate type and

location) using 105 permutations (Table S1, Supporting

Information). Only the substrate revealed to be a signifi-

cant descriptor of variation in bacterial community

composition (r2 = 0.55, P = 0.022), but not in the eukary-

ote community composition (P = 0.470). By contrast, the

variation in eukaryote community composition was sig-

nificantly related to the geographic location (r2 = 0.69,

P = 0.007), with marginal significant effects of the PC1

(r2 = 0.53, P = 0.080) and PC2 (r2 = 0.58, P = 0.054).

These results confirm previous qualitative observations

in NMDS plots (Fig. 6) and are in agreement with the

distribution of high-level taxa across samples (Fig. 1).

Thus, in the case of bacteria, Deinococcales were very

abundant in the control substrate, but Acidobacteria

and, especially, Bacteroidetes appeared to increase their

abundance in the rest of substrates. This might be

related to differences in porosity, and hence desiccation

–1 0 1 2

–10

12

CCA1

CC

A2

+

+

+

+

+

++

+

+

+

+

+

+

+++

++

+

+

+

+

++

+

+

+

+

+++

+

+

++++

+

+

+

++

+

+

+

+

+

++

+

+

++

+

+

+

+

+

+

+

+

+

+

+++

+

++

++

+

+

+ +

++

++

+

+

++

+

+

++

+

+

+

+++

+

++

+

++

+

+

+

+

+

+

+

++

+

+

++

+

+

+

+

+

+

+

+

+

+

+

+

+

++

+

+

+

++

+

+

+

+

++

+

+

+

+

+

+

+

+

+

++

++

PC1

PC2

PC3

Subst. (granite)Subst. (limestone)Subst.

(control)

Location (CM)

Location(Lyon)

–10

1

x

xxx

x

x

x

x

B1.3

B1.4

B1.6

B1.7

RP2.6

A2.1a

A2.3

A2.4b

A3.1A3.3

CC

A2

Bacteria

Fig. 7 Constrained correspondence analysis (CCA) describing the var

ronmental and geographic descriptors. CM, Charente-Maritime; Subst

potential, in the control substrate. However, qualitative

differences in the presence of eukaryotic phyla were not

apparent, suggesting a different behaviour of prokary-

otes and microbial eukaryotes with respect to substrate

type.

Second, to quantify the variation explained by the

selected environmental variables, we performed a con-

strained correspondence analysis (CCA) (Fig. 7). When

considering all the descriptors in the model, the varia-

tion in community composition explained collectively

by environmental factors (3 PCs, substrate and location)

was 93% of the total variation for bacteria (P = 0.024)

and 92% for eukaryotes (P = 0.023) (Table 2). When

considering each variable separately, we observed again

significant differences between bacteria and eukaryotes.

In both cases, the PC1 had a significant effect in

explaining community composition, 13% of the vari-

ance (P = 0.009) for bacteria and 14% of the variance

(P = 0.033) for eukaryotes. However, whereas for bacte-

rial communities, the substrate explained 38% of the

total variation (P = 0.034) and location was not found

as a significant descriptor (P = 0.470), for eukaryotes,

location was a significant descriptor and explained 39%

of the variance (P = 0.008), but not the substrate

(P = 0.258). Furthermore, we also tested the interaction

between the different variables and, although the

design of this study did not allow to retrieve significant

results in most cases, the only significant effect found

was that of the substrate on the variation in bacterial

community composition, after having removed the

effect of the PC1 (the geo-environmental variables)

–1 0 1 2

–1.5

–1.0

–0.5

0.0

0.5

1.0

1.5

CCA1

+

+

++

+

+

+

+

+

+

+

+ ++

+

+++

++ +

+

+

+

+

+

+

+ +

+

+

++

+

+

+

+

+

+

+

+

++++

+

+

+

+

+++

+

+++

+

+

+

+

++

+

+

+

++

+

++

+

+

+

+

+

+

+

+

+

PC1

PC2

PC3

Subst. (granite)

Subst.(limestone)

Subst. (control)

Location (CM)

Location (Lyon)

0

xx

x

xx

x

x

x

B1.3

B1.4

B1.6B1.7

RP2.6

A2.1a

A2.3

A2.4b

A3.1

A3.3

Eukaryotes

iation in bacterial and eukaryote community explained by envi-

., substrate.

� 2012 Blackwell Publishing Ltd

Table 2 Computational procedure for the partitioning model using constrained correspondence analyses

Active components

Bacteria Eukaryotes

v2 % P v2 % P

Inertia 5.142 4.693

0 PC1 & PC2 & PC3 & Substrate & Location 4.770 0.93 0.024* 4.320 0.92 0.023*

1 Substrate 1.968 0.38 0.034* 1.617 0.34 0.258

2 Location 1.719 0.33 0.470 1.823 0.39 0.008**

3 PC1 0.691 0.13 0.009** 0.662 0.14 0.033*

4 PC2 0.528 0.10 0.805 0.620 0.13 0.087

5 PC3 0.501 0.10 0.819 0.542 0.12 0.354

6 PC1 & PC2 & PC3 1.720 0.33 0.412 1.824 0.39 0.007**

7 PC1 & PC2 & PC3 & Substrate 3.641 0.71 0.033* 3.380 0.72 0.003**

8 PC1 & PC2 & PC3 & Location 2.848 0.55 0.505 2.764 0.59 0.039*

9 Substrate& Location 3.641 0.71 0.034* 3.380 0.72 0.005**

% refers to the proportion of inertia explained by the model. The significance of each step (P-value) was tested using 105

permutations. Significant values are shown in bold; *P < 0.05; **P < 0.01.

●●

●

● ●

●

●

●●●●● ●

●

●

●●

●

●

●

●

0 2 4 6 8 10

0.0

0.2

0.4

0.6

0.8

Ln (distance + 1)

LN (S

oren

son

+ 1)

● BacteriaEukaryotes

Fig. 8 Distance–decay curves for bacteria and microbial

MICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3865

(P = 0.013) (Table S2, Supporting information). This

confirms that the substrate is the dominant factor shap-

ing the variation in the bacterial community composi-

tion, with a contribution of environmental variables. On

the contrary, the geographic location appears more

important to shape microbial eukaryotic communities.

Finally, to further explore whether the geographic

location is actually more influential for eukaryotic

microbes than for bacteria, we constructed distance–

decay curves for both (Fig. 8). Eukaryote and bacteria

curves showed a decreasing trend in community simi-

larity as a function of the geographic distance

(bEuk = )0.04, 95% confidence interval (CI):[)0.07,

)0.02], bBact = )0.02, 95% CI:[)0.04, 0.00]. However,

only the distance–decay curve of eukaryotes displayed

a slope significantly different from 0. These results con-

firm that geographic distance significantly influences

eukaryotic, but not prokaryotic, community composi-

tion in the same biofilm communities.

eukaryotes in biofilm communities. The Sorensen’s index is

plotted as a function of the geographic distance between sam-

ples.

Concluding remarks

Microbial biogeography studies are on the rise. Deter-

mining which factors are more important to shape com-

munity composition and whether environmental factors

acting through selection or stochastic factors related to

dispersal are the focus of intense research efforts. Some

studies favour selection and others chance, but both of

them probably contribute (Green & Bohannan 2006;

Martiny et al. 2006; Lindstrom & Langenheder 2011).

However, it is tacitly assumed that protists, or microbial

eukaryotes in general, behave as bacteria due to their

small size. This is largely based on ideas by Finlay and

Fenchel, according to whom, free-living protists behave

as cosmopolitan metapopulations and respond to the

� 2012 Blackwell Publishing Ltd

‘everything is everywhere, but the environment selects’

maxim (Finlay 2002; Finlay & Fenchel 2004). When

counterexamples showing differences in protist commu-

nities with geographic distance are seen, they are intui-

tively associated to observed differences of prokaryote

community composition with distance because they are

microscopic as well. However, in addition to size, other

life history traits differentiate prokaryotes and eukary-

otes. Do prokaryotes and protists actually behave the

same from a biogeography perspective?

To try to answer this question, we used sunlight-

exposed microbial biofilms growing on different

3866 M. RAGON ET AL.

mineral substrates in different locations as models.

These microbial communities were composed of micro-

organisms adapted to desiccation and UV radiation and

were therefore subject to strong environmental selec-

tion. This model was advantageous for two reasons.

First, the microbial composition is less complex in these

than in other ecosystems, simplifying the description

and comparison of communities. Second, under such

strong environmental selection, if the ‘everything is

everywhere, but the environment selects’ applied similarly

to prokaryotes and eukaryotes, we should expect to see

the same biogeographic trends in prokaryote and micro-

bial eukaryote communities. We were able to show that

the nature of the substrate as well as climatic variables

and geographic location influences both microbial com-

munity structures. However, whereas the structure of

bacterial communities was significantly explained by

the nature of the substrate, that of eukaryotes was bet-

ter explained by geographic distance.

These significant differences in prokaryotic and

eukaryotic community patterns suggest that the under-

lying processes are different as well, with more deter-

ministic processes (environmental selection) prevailing

for bacteria. Explanations for this might be related to

dissimilar dispersal and survivability as well as to their

lifestyle. The dominant bacterial groups detected are

extremely resistant. This increases the probability of

successful cosmopolitan dispersal via dust particles in

atmospheric aerosols (Brodie et al. 2007; Osman et al.

2008), where they are exposed to high irradiation and

long desiccation periods. By contrast, although some

eukaryotes have highly resistant dispersal structures,

for example, ascomycetes, and might then have a bioge-

ographic pattern more akin to bacteria, others do not or

their dispersive structures (e.g. green algae, ciliate or

amoeba cysts) may be less resistant to long-term desic-

cation and irradiation. Another possible explanation

resides in lifestyle. Although prokaryotes and some

eukaryotes share sedentary activities (phototrophy or

osmotrophy), many protists are motile predators. There-

fore, it could be hypothesized that their presence would

be more dependent on other biological variables (e.g.

presence and nature of prey) than on abiotic environ-

mental selection. All these different properties and life-

styles may contribute to make the biogeography of

microbial eukaryotes appear different from that of bac-

teria, at least at the phylogenetic resolution of the

domain level. Biogeographic patterns are also likely to

change at lower phylogenetic scales as a function of the

dominant lifestyle and dispersal properties, as has been

observed in soil bacteria (Bissett et al. 2010). Finally, if

protist biogeography is more influenced by distance

under strong environmental selection, it could be pre-

dicted that this trend would be even stronger in

situations where environmental selection is less chal-

lenging. Further studies on protist versus prokaryote

biogeography in other ecosystems will be needed to test

this prediction.

Acknowledgements

M.R. and this work were financed by a French Convention

Industrielle de Formation par la Recherche (CIFRE). We thank Jean

Marc Ghigo for helpful discussions.

References

Achtman M, Wagner M (2008) Microbial diversity and the

genetic nature of microbial species. Nature Reviews

Microbiology, 6, 431–440.

Altschul SF, Madden TL, Schaffer AA et al. (1997) Gapped

BLAST and PSI-BLAST: a new generation of protein

database search programs. Nucleic Acids Research, 25, 3389–

3402.

Baas-Becking LGM (1934) Geobiologie of inleiding tot de

milieukunde W.P. Van Stockum & Zoon, The Hague, the

Netherlands.

Banerjee M, Whitton BA, Wynn-Williams DD (2000)

Phosphatase activities of endolithic communities in rocks of

the Antarctic Dry valleys. Microbial Ecology, 39, 80–91.

Bass D, Richards TA, Matthai L, Marsh V, Cavalier-Smith T

(2007) DNA evidence for global dispersal and probable

endemicity of protozoa. BMC Evolutionary Biology, 7, 162.

Beijerinck MW (1913) De infusies en de ontdekking der bakterien

Jaarboek van de Koninklijke Akademie van Wetenschappen.

Bissett A, Richardson AE, Baker G, Wakelin S, Thrall PH

(2010) Life history determines biogeographical patterns of

soil bacterial communities over multiple spatial scales.

Molecular Ecology, 19, 4315–4327.

Bray JR, Curtis JT (1957) An ordination of upland forest

communities of southern Wisconsin. Ecological Monographs,

27, 325–349.

Brodie EL, DeSantis TZ, Parker JP et al. (2007) Urban aerosols

harbor diverse and dynamic bacterial populations.

Proceedings of the National Academy of Sciences, USA, 104, 299–

304.

Bubrick P, Galun M, Frensdorff A (1984) Observations on free-

living Trebouxia Depuymaly and Pseudotrabouxia Archibald,

and evidence that both symbionts from Xanthoria parietina

(L) th fr can be found free-living in nature. New Phytologist,

97, 455–462.

Caron DA (2009) Past President’s address: protistan

biogeography: why all the fuss? Journal of Eukaryotic

Microbiology, 56, 105–112.

Caron DA, Countway PD, Savai P et al. (2009) Defining DNA-

based operational taxonomic units for microbial-eukaryote

ecology. Applied and Environment Microbiology, 75, 5797–5808.

Chanal A, Chapon V, Benzerara K et al. (2006) The desert of

Tataouine: an extreme environment that hosts a wide

diversity of microorganisms and radiotolerant bacteria.

Environmental Microbiology, 8, 514–525.

Chu H, Fierer N, Lauber CL et al. (2010) Soil bacterial diversity

in the Arctic is not fundamentally different from that found

in other biomes. Environmental Microbiology, 12, 2998–3006.

� 2012 Blackwell Publishing Ltd

MICROBIAL BI OGEOGRAPHY IN E PILI THIC BIOFILMS 3867

Daly MJ (2009) A new perspective on radiation resistance

based on Deinococcus radiodurans. Nature Reviews

Microbiology, 7, 237–245.

Edgar RC (2004) MUSCLE: multiple sequence alignment with

high accuracy and high throughput. Nucleic Acids Research,

32, 1792–1797.

Edwards C, Lawton LA (2009) Bioremediation of cyanotoxins.

Advances in Applied Microbiology, 67, 109–129.

Fierer N, Jackson RB (2006) The diversity and biogeography of

soil bacterial communities. Proceedings of the National

Academy of Sciences, USA, 103, 626–631.

Finlay BJ (2002) Global dispersal of free-living microbial

eukaryote species. Science, 296, 1061–1063.

Finlay BJ, Fenchel T (2004) Cosmopolitan metapopulations of

free-living microbial eukaryotes. Protist, 155, 237–244.

Foissner W (2006) Biogeography and dispersal of micro-

organisms: a review emphasizing protists. Acta

Protozoologica, 45, 111–136.

Fraser C, Alm EJ, Polz MF, Spratt BG, Hanage WP (2009) The

bacterial species challenge: making sense of genetic and

ecological diversity. Science, 323, 741–746.

Garvie LA, Knauth LP, Bungartz F, Klonowski S, Nash TH 3

rd (2008) Life in extreme environments: survival strategy of

the endolithic desert lichen Verrucaria rubrocincta.

Naturwissenschaften, 95, 705–712.

Good IJ (1953) The population frequencies of species and

the estimation of population parameters. Biometrika, 40, 237–

264.

Gorbushina AA, Broughton WJ (2009) Microbiology of the

atmosphere-rock interface: how biological interactions and

physical stresses modulate a sophisticated microbial

ecosystem. Annual Review of Microbiology, 63, 431–450.

Green J, Bohannan BJ (2006) Spatial scaling of microbial

biodiversity. Trends in Ecology & Evolution, 21, 501–507.

Green JL, Holmes AJ, Westoby M et al. (2004) Spatial scaling

of microbial eukaryote diversity. Nature, 432, 747–750.

Green JL, Bohannan BJ, Whitaker RJ (2008) Microbial

biogeography: from taxonomy to traits. Science, 320, 1039–

1043.

Horner-Devine MC, Lage M, Hughes JB, Bohannan BJ (2004) A

taxa-area relationship for bacteria. Nature, 432, 750–753.

Jobb G, von Haeseler A, Strimmer K (2004) TREEFINDER: a

powerful graphical analysis environment for molecular

phylogenetics. BMC Evolutionary Biology, 4, 18.

Karsten U, Friedl T, Schumann R, Hoyer K, Lembcke S (2005)

Mycosporine-like amino acids and phylogenies in green

algae: Prasiola and its relatives from the Trebouxiophyceae

(Chlorophyta). Journal of Phycology, 41, 557–566.

Karsten U, Lembcke S, Schumann R (2007) The effects of

ultraviolet radiation on photosynthetic performance, growth

and sunscreen compounds in aeroterrestrial biofilm algae

isolated from building facades. Planta, 225, 991–1000.

Kloesges T, Popa O, Martin W, Dagan T (2011) Networks of

gene sharing among 329 proteobacterial genomes reveal

differences in lateral gene transfer frequency at different

phylogenetic depths. Molecular Biology and Evolution, 28,

1057–1074.

Larkin MA, Blackshields G, Brown NP et al. (2007) Clustal W

and Clustal X version 2.0. Bioinformatics, 23, 2947–2948.

Legendre P, Legendre L (1998) Numerical ecology, 2nd English

edn. Elsevier Science BV, Amsterdam.

� 2012 Blackwell Publishing Ltd

Lindstrom ES, Langenheder S (2011) Local and regional factors

influencing bacterial community assembly. Environmental

Microbiology Reports, 4, 1–9.

Locey KJ (2010) Synthesizing traditional biogeography with

microbial ecology: the importance of dormancy. Journal of

Biogeography, 37, 1835–1841.

Madigan MT, Martinko JM, Parker J (2002) Brock Biology of

Microorganisms, 10th edn. Prentice-Hall, Inc., New Jersey.

Martiny JB, Bohannan BJ, Brown JH et al. (2006) Microbial

biogeography: putting microorganisms on the map. Nature

Reviews Microbiology, 4, 102–112.

Martiny JB, Eisen JA, Penn K, Allison SD, Horner-Devine MC

(2011) Drivers of bacterial {beta}-diversity depend on spatial

scale. Proceedings of the National Academy of Sciences, USA,

108, 7850–7854.

Mobjerg N, Halberg KA, Jorgensen A et al. (2011) Survival in

extreme environments – on the current knowledge of

adaptations in tardigrades. Acta Physiologica, 202, 409–420.

Nagy ML, Perez A, Garcia-Pichel F (2005) The prokaryotic

diversity of biological soil crusts in the Sonoran Desert

(Organ Pipe Cactus National Monument, AZ). FEMS

Microbiology Ecology, 54, 233–245.

Nascimbene J, Thus H, Marini L, Nimis PL (2009) Early

colonization of stone by freshwater lichens of restored

habitats: a case study in northern Italy. Science of the Total

Environment, 407, 5001–5006.

Nemergut DR, Anderson SP, Cleveland CC et al. (2007)

Microbial community succession in an unvegetated, recently

deglaciated soil. Microbial Ecology, 53, 110–122.

van Niel CB (1949) The ‘Delft School’ and the rise of general

microbiology. Bacteriological Reviews, 13, 161–174.

Nolte V, Pandey RV, Jost S et al. (2010) Contrasting seasonal

niche separation between rare and abundant taxa conceals

the extent of protist diversity. Molecular Ecology, 19, 2908–

2915.

Oksanen J, Blanchet G, Kindt R et al. (2011)

Vegan: Community Ecology Package. R package version

1.17-9, Available from http://CRAN.R-project.org/package=

vegan.

O’Malley MA (2007) The nineteenth century roots of

‘everything is everywhere’. Nature Reviews Microbiology, 5,

647–651.

Osman S, La Duc MT, Dekas A, Newcombe D, Venkateswaran

K (2008) Microbial burden and diversity of commercial

airline cabin air during short and long durations of travel.

ISME Journal, 2, 482–497.

Parnell JJ, Rompato G, Latta LCt et al. (2010) Functional

biogeography as evidence of gene transfer in hypersaline

microbial communities. PLoS ONE, 5, e12919.

Pedros-Alio C (2006) Marine microbial diversity: can it be

determined? Trends in Microbiology, 14, 257–263.

Philippe H (1993) MUST, a computer package of Management

Utilities for Sequences and Trees. Nucleic Acids Research, 21,

5264–5272.

Pommier T, CanbACk B, Riemann L et al. (2007) Global

patterns of diversity and community structure in marine

bacterioplankton. Molecular Ecology, 16, 867–880.

Quaiser A, Ochsenreiter T, Lanz C et al. (2003) Acidobacteria

form a coherent but highly diverse group within the

bacterial domain: evidence from environmental genomics.

Molecular Microbiology, 50, 563–575.

3868 M. RAGON ET AL.

R Development Core Team (2011) R: A Language and

Environment for Statistical Computing. R Foundation for

Statistical Computing, Vienna, Austria.

Ragon M, Restoux G, Moreira D, Moller AP, Lopez-Garcia P

(2011) Sunlight-exposed biofilm microbial communities are

naturally resistant to Chernobyl ionizing-radiation levels.

PLoS ONE, 6, e21764.

Rainey FA, Ray K, Ferreira M et al. (2005) Extensive diversity

of ionizing-radiation-resistant bacteria recovered from

Sonoran Desert soil and description of nine new species of

the genus Deinococcus obtained from a single soil sample.

Applied and Environment Microbiology, 71, 5225–5235.

Ramette A, Tiedje JM (2007) Biogeography: an emerging

cornerstone for understanding prokaryotic diversity, ecology,

and evolution. Microbial Ecology, 53, 197–207.

Rintala H, Pitkaranta M, Toivola M, Paulin L, Nevalainen A

(2008) Diversity and seasonal dynamics of bacterial community

in indoor environment. BMC Microbiology, 8, 56.

Rossello-Mora R, Lucio M, Pena A et al. (2008) Metabolic

evidence for biogeographic isolation of the extremophilic

bacterium Salinibacter ruber. ISME Journal, 2, 242–253.

Schauer R, Bienhold C, Ramette A, Harder J (2010) Bacterial

diversity and biogeography in deep-sea surface sediments of

the South Atlantic Ocean. ISME Journal, 4, 159–170.

Schloss PD, Handelsman J (2005) Introducing DOTUR, a

computer program for defining operational taxonomic units

and estimating species richness. Applied and Environment

Microbiology, 71, 1501–1506.

Sørensen T (1957) A method of establishing groups of equal

amplitude in plant sociology based on similarity of species

and its application to analyses of the vegetation on Danish

commons. Biologiske Skrifter ⁄ Kongelige Danske Videnskabernes

Selskab, 5, 1–34.

Staley JT (2006) The bacterial species dilemma and the

genomic-phylogenetic species concept. Philosophical

Transactions of the Royal Society of London. Series B, Biological

sciences, 361, 1899–1909.

Staley JT, Gosink JJ (1999) Poles apart: biodiversity and

biogeography of sea ice bacteria. Annual Review of Microbiology,

53, 189–215.

Stolz A (2009) Molecular characteristics of xenobiotic-degrading

sphingomonads. Applied Microbiology and Biotechnology, 81,

793–811.

Suzuki K, Collins MD, Iijima E, Komagata K (1988)

Chemotaxonomic characterization of a radiotolerant

bacterium, Arthrobacter radiotolerans – description of Rubrobacter

radiotolerans gen. nov., comb. nov. FEMS Microbiology Letters,

52, 33–39.

de la Torre JR, Goebel BM, Friedmann EI, Pace NR (2003)

Microbial diversity of cryptoendolithic communities from the

McMurdo Dry Valleys, Antarctica. Applied and Environment

Microbiology, 69, 3858–3867.

Vyverman W, Verleyen E, Sabbe K et al. (2007) Historical

processes constrain patterns in global diatom diversity.

Ecology, 88, 1924–1931.

Whitaker RJ (2006) Allopatric origins of microbial species.

Philosophical Transactions of the Royal Society of London. Series

B, Biological sciences, 361, 1975–1984.

Whitaker RJ, Grogan DW, Taylor JW (2003) Geographic

barriers isolate endemic populations of hyperthermophilic

archaea. Science, 301, 976–978.

de Wit R, Bouvier T (2006) ‘Everything is everywhere, but, the

environment selects’; what did Baas Becking and Beijerinck

really say? Environmental Microbiology, 8, 755–758.

Zimmermann J, Gonzalez JM, Saiz-Jimenez C (2006) Epilithic

biofilms in Saint Callixtus Catacombs (Rome) harbour a

broad spectrum of Acidobacteria. Antonie van Leeuwenhoek,

89, 203–208.

M.R. was a PhD student at the time when this work was done;

she is interested in microbial diversity in biofilms and extreme

environments. M.F. is a postdoctoral researcher with expertise

in statistics applied to population genetics and genomics of

very different model species, from fungi and cetaceans to

humans. D.M. and P.L.-G. are senior researchers with broad

interests in prokaryotic and protist diversity and evolution.

Data accessibility

DNA sequences have been deposited in GenBank with acces-

sion numbers JQ627396-JQ627584. Final DNA sequence align-

ments and phylogenetic trees in NEXUS format are provided

as online supplemental material.

Supporting Information

Additional supporting information may be found in the online

version of this article.

Appendix S1. Alignments and phylogenetic trees in NEXUS

format. Environmental parameters and PCA axes.

Fig. S1 Accumulation curves for bacterial SSU rDNA

sequences in biofilms from different geographic areas. The

accumulation curves have been done at different SSU rDNA

similarity cut-off values.

Fig. S2 Accumulation curves for eukaryotic SSU rDNA

sequences in biofilms from different geographic areas. The

accumulation curves have been done at different SSU rDNA

similarity cut-off values.

Fig. S3 Species (OTU) rank–abundance distributions expressed

as the cumulative frequencies.

Fig. S4 Phylogenetic tree of SSU rDNA sequences of Acidobac-

teria and Bacteroidetes identified in epilithic biofilms.

Fig. S5 Phylogenetic tree of proteobacterial SSU rDNA

sequences identified in sunlight-exposed biofilms growing on

mineral substrates.

Fig. S6 Principal component analysis (PCA) on the 18 environ-

mental variables used to explain community compositions.

Table S1 Relationship between the NMDS and the environ-

mental descriptors.

Table S2 Computational procedure for the partitioning model

using constrained correspondence analyses with covariables.

Please note: Wiley-Blackwell are not responsible for the content

or functionality of any supporting information supplied by the

authors. Any queries (other than missing material) should be

directed to the corresponding author for the article.

� 2012 Blackwell Publishing Ltd