Downloaded from http://rsbl.royalsocietypublishing.org/ on June 16, 2017
Community ecology
rsbl.royalsocietypublishing.org
Research
Cite this article: Pansu J, Winkworth RC,
Hennion F, Gielly L, Taberlet P, Choler P. 2015
Long-lasting modification of soil fungal diversity associated with the introduction of rabbits
to a remote sub-Antarctic archipelago. Biol.
Lett. 11: 20150408.
http://dx.doi.org/10.1098/rsbl.2015.0408
Received: 13 May 2015
Accepted: 6 August 2015
Subject Areas:
ecology, molecular biology
Keywords:
environmental DNA, invasive species,
Kerguelen Islands, metabarcoding,
soil communities
Author for correspondence:
Johan Pansu
e-mail: [email protected]
Electronic supplementary material is available
at http://dx.doi.org/10.1098/rsbl.2015.0408 or
via http://rsbl.royalsocietypublishing.org.
Long-lasting modification of soil fungal
diversity associated with the introduction
of rabbits to a remote sub-Antarctic
archipelago
Johan Pansu1,2, Richard C. Winkworth3,4, Françoise Hennion5,
Ludovic Gielly1,2, Pierre Taberlet1,2 and Philippe Choler1,2
1
LECA, Université de Grenoble Alpes, Grenoble 38000, France
LECA, CNRS, Grenoble 38000, France
3
Department of Natural Sciences, Unitec Institute of Technology, Auckland, New Zealand
4
Institute of Fundamental Sciences, Massey University, Palmerston North, New Zealand
5
UMR ECOBIO, Université de Rennes 1, CNRS, Rennes 35042, France
2
During the late nineteenth century, Europeans introduced rabbits to many of
the sub-Antarctic islands, environments that prior to this had been devoid of
mammalian herbivores. The impacts of rabbits on indigenous ecosystems are
well studied; notably, they cause dramatic changes in plant communities and
promote soil erosion. However, the responses of fungal communities to such
biotic disturbances remain unexplored. We used metabarcoding of soil extracellular DNA to assess the diversity of plant and fungal communities at sites
on the sub-Antarctic Kerguelen Islands with contrasting histories of disturbance
by rabbits. Our results suggest that on these islands, the simplification of plant
communities and increased erosion resulting from the introduction of rabbits
have driven compositional changes, including diversity reductions, in indigenous soil fungal communities. Moreover, there is no indication of recovery at
sites from which rabbits were removed 20 years ago. These results imply that
introduced herbivores have long-lasting and multifaceted effects on fungal
biodiversity as well as highlight the low resiliency of sub-Antarctic ecosystems.
1. Introduction
The distribution of organismal diversity in soils is linked to abiotic conditions
and the nature of the aboveground plant community [1,2]. Mammalian herbivores can therefore affect soil communities via direct physical disturbance or
indirectly via effects on plant cover and soil properties that induce changes in
habitat and resource availability [2]. The direction and magnitude of such effects
may differ greatly depending on ecosystem type and disturbance intensity [2,3].
The sub-Antarctic islands are extremely isolated. Their ecosystems are typically species-poor and characterized by an absence of indigenous mammalian
herbivores [4]. Following their discovery in 1772, the Kerguelen Islands were
subject to numerous alien plant and animal introductions that have resulted
in substantial changes to ecosystem structure and functioning [4]. The most
notable of these introductions is that of rabbits (Oryctolagus cuniculus) in
1874. They now occur at high densities both on the main island, Grande
Terre, and several of the smaller ones [5]. Where rabbits occur, their grazing
and burrowing activities have significantly affected the native vegetation,
resulting in the decreased abundance of certain native species (e.g. Azorella
selago, Pringlea antiscorbutica) and the over-dominance of others (e.g. Acaena
magellanica). Moreover, rabbit-induced increases in erosion and changes to
soil characteristics have tended to favour introduced plant species [5].
& 2015 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rsbl.royalsocietypublishing.org/ on June 16, 2017
(a)
(c)
2
30
Australia
40
50
Vegetation type
IB03
AU08
Australia (AU)
Guillou (GU)
60
dense vegetation
fellfield
Isthme Bas (IB)
Antarctica
(b)
Acaena magellanica
Location
Kerguelen
Islands
Kerguelen Islands
rsbl.royalsocietypublishing.org
20
Africa
IB01
Biol. Lett. 11: 20150408
GU13
IB04
GU15
GU14
AU10
Port-auxFrançais
IB02
AU05
lle Australia
lle Guillou
AU09
Isthme Bas
GU11
AU07
AU06
N
0
10
20
Other native plants
30 km
IB16
GU12
(no vegetation)
Introduced plants
Figure 1. Location and floristic composition of study sites. (a) Map showing the location of the Kerguelen Islands. (b) Map of the Kerguelen Islands showing sampling
locations. (c) Ternary plot describing the composition of plant communities based on vegetation surveys. Site IB16 was highly eroded with no vegetation recorded.
While the effects of alien herbivores on Kerguelen’s native
vegetation are well described [5], their impacts on soil communities—in particular soil fungi—are poorly known.
Fungal communities play a central role in terrestrial ecosystem functioning (e.g. in nutrient cycling) [6]. It has been
assumed that the sub-Antarctic islands host limited fungal
diversity and that the relative simplicity of these communities
may make them particularly sensitive to disturbance [7].
However, the response of sub-Antarctic fungal communities
to biotic disturbance-induced changes to plant communities
has not been explored.
We hypothesized that (i) rabbit-induced changes to the
structure of aboveground plant communities and to soil
characteristics has altered the diversity and composition of
soil fungal communities and (ii) such changes are long lasting. To test these hypotheses, we studied communities from
sites on the Kerguelen Islands with different rabbit disturbance histories. We used soil metabarcoding to evaluate the
composition of both fungal and plant communities [8].
These data were used to assess the relative importance of
biotic and abiotic factors in shaping soil fungal assemblages.
2. Material and methods
Study sites on the Kerguelen Islands were located on the Isthme
Bas (Grande Terre) where rabbits occur at high density, Ile Guillou where a successful eradication programme was undertaken
in 1994 [5] and Ile Australia that has remained rabbit free
(figure 1). In each area, we established five or six 10 10 m
plots within which we collected soil samples—two for DNA
metabarcoding and a third for soil physico-chemical analyses
(e.g. organic matter content, acidity; electronic supplementary
material, table S1)—and conducted vegetation surveys (e.g.
plant assemblage and cover; figure 1c).
Using a previously described method, one or two extracellular
DNA extractions were performed per soil sample [8]. Standard
methods were used to PCR amplify metabarcodes for vascular
plants (P6-loop of the chloroplast trnl intron [9]) and fungi
(nuclear ribosomal ITS1 [10]) from each extraction. In each case,
a pair of PCR replicates along with extraction and PCR controls
were conducted. Plant and fungal amplicons were paired-end
sequenced on Illumina HiSeq (2 100 bp) and MiSeq (2 250 bp) platforms, respectively. Sequences were then filtered
and clustered using an established workflow. Taxonomic assignments were made by comparison to reference sets derived by
in silico PCR from public databases (EMBL, UNITE) as well as
an exhaustive database for Kerguelen Island vascular plants.
We used non-metric multidimensional scaling (NMDS) to
ordinate fungal and plant communities. Vector fitting and multivariate analysis of variance (PERMANOVA [11]) were used to
examine the effects of soil conditions and, in the case of fungal
communities, biotic parameters (i.e. plant assemblages characteristics) on community composition. Ordinations for different
datasets were compared using Procrustes analyses. All statistical
analyses were conducted in R using the vegan package [12,13].
Full methodological details are given in the electronic
supplementary material.
3. Results
Sequencing of plant metabarcodes generated 1 328 674 reads,
which after filtering resulted in 25 unique sequence types.
Ordinations of plant communities inferred from metabarcoding and botanical surveys were highly similar (table 1;
electronic supplementary material, figures S5 and S6). Moreover, there was a strong linear correlation between the
relative abundances of plant species in both datasets ( p ,
0.001; electronic supplementary material, figure S4). This
suggests that on the Kerguelen Islands, current plant cover
is well represented by soil DNA. Our results imply two distinct situations. Undisturbed plots on Ile Australia that are
characterized by species-rich, native-dominated plant
Downloaded from http://rsbl.royalsocietypublishing.org/ on June 16, 2017
(a)
(b)
6.5
Acaena
magellanica
AU10
GU12
2
OMC
plant
cover
GU15
AU09
AU10 GU13
IB04
IB01
0
6.0
GU11
IB16
AU07
AU09
5.5
GU12
AU05
AU06
GU11
IB04
IB16
AU08
IB02 GU14
pH
axis 2
1
IB03
IB02
AU08
IB01
GU13
GU14
IB03
5.0
AU06
C/N
–2
other
natives Shannon
index
–3
–2
AU07
pH
introduced
species
0
GU15
4.5
2
0
4
axis 1
0.5
1.0
1.5
2.0
plant community diversity (Shannon index)
Figure 2. Environmental parameters determining the structure of soil fungal communities. (a) Ordination of soil fungal communities with vector fitting of biotic and
abiotic variables (NMDS stress ¼ 0.134). Grey and black arrows represent non-significant and significant (at p , 0.01) variables, respectively. Acaena magellanica,
per plot frequency of A. magellanica; other natives, per plot frequency of other native plants; introduced species, per plot frequency of introduced plants; OMC, soil
organic matter content; C/N, carbon/nitrogen ratio; pH, soil acidity, Shannon index, plant community diversity; plant cover, aboveground plant cover. (b) Per plot
fungal MOTU richness relative to soil pH and plant community diversity (Shannon index). Circle size proportional to the number of fungal MOTUs. See the electronic
supplementary material for details.
Table 1. Results of Procrustes analyses for pairwise comparisons of ordinations of plant and fungal communities inferred from surveys and metabarcoding.
Significant correlations indicate that ordinations were more similar to one another than would be expected by chance. See the electronic supplementary material
for details.
Procrustes sum of
square
correlation in a symmetric Procrustes
rotation
significance
botanical survey-based plant communities/fungal
0.62
0.62
0.004
communities
DNA-based plant communities/fungal communities
0.49
0.71
0.001
0.23
0.98
0.001
botanical survey-based plant communities/DNA-based
plant communities
communities and disturbed—either naturally on Ile Australia
or as a result of rabbits on Isthme Bas and Ile Guillou—plots
with species-poor communities dominated by Acaena or
introduced Taraxacum spp. (figure 1c; electronic supplementary material, figures S2 and S5). Vector fitting indicates a
correlation between plant community composition and
organic matter content ( p , 0.05; electronic supplementary
material, figure S6).
For fungi, 8 602 068 reads were obtained; a total of 547
molecular operational taxonomic units (MOTUs) were identified with per plot diversity varying markedly between sites
(electronic supplementary material, figure S7). Average
MOTU diversity was higher at undisturbed sites on Ile Australia than on the Isthme Bas and Ile Guillou (electronic
supplementary material, table S3). Fungal communities also
showed striking compositional differences with axes of discrimination related to (i) the density of plant cover, soil
organic matter content and carbon–nitrogen ratio and (ii)
plant community diversity, that distinguishes rich and undisturbed native communities (AU05-AU07, AU09) from the
others (vector fitting, p , 0.01; figure 2a). Analyses using
PERMANOVA confirm that each of the parameters identified
by vector fitting explains a significant proportion of the variance ( p , 0.05; electronic supplementary material, Results).
Ordinations of fungal and plant communities were more
similar to one another than would be expected by chance
(table 1; electronic supplementary material, figure S6). This
result suggests that the composition and richness of fungal
and plant communities are co-structured across plots.
Full details of results are provided in the electronic
supplementary material.
4. Discussion
We identified striking differences between soil fungal communities at our three study locations. Specifically, fungal
communities on rabbit-free Ile Australia are broadly distinct—both in terms of MOTU richness and composition—
from those on the Isthme Bas or Ile Guillou, locations that
Biol. Lett. 11: 20150408
AU05
introduced
–1
rsbl.royalsocietypublishing.org
3
3
Downloaded from http://rsbl.royalsocietypublishing.org/ on June 16, 2017
Data accessibility. Filtered DNA sequences and soil physico-chemical
data are provided in the electronic supplementary material. Unfiltered DNA sequences, as well as vegetation survey data, are
available from the Dryad Digital Repository: http://dx.doi.org/10.
5061/dryad.t8534.
Authors’ contributions. P.C. and P.T. designed the study. P.C., R.C.W. and
F.H. performed sampling. J.P. and L.G. performed DNA and soil analyses. J.P. processed and analysed the data. J.P., R.C.W. and P.C.
interpreted the results. J.P. drafted the manuscript with contributions
from the other authors. All the authors approved the final version.
Competing interests. L.G. and P.T. hold patents related to use of the P6
loop of the chloroplast trnl (UAA) intron for plant identification
from degraded template DNA as well as the primers used to amplify
this region. These patents restrict commercial applications and have
no impact on the use of this locus by academic researchers.
Funding. We thank the Institut Polaire Paul Emile Victor (IPEV) Programme 136 (led by M. Lebouvier) for financial support and
fieldwork coordination.
Acknowledgements. We thank Marine Pouvreau, Marion Lombard, Julie
Vingère and Françoise Cardou for field assistance; Nigel Gilles
Yoccoz for providing the Kerguelen plant DNA database; Cindy
Arnoldi and Jérôme Poulenard for help with soil analyses.
References
1.
2.
3.
Ettema CH, Wardle DA. 2002 Spatial soil ecology.
Trends Ecol. Evol. 17, 177– 183. (doi:10.1016/
S0169-5347(02)02496-5)
Bardgett RD, Wardle DA. 2003 Herbivore-mediated
linkages between aboveground and belowground
communities. Ecology 84, 2258 –2268. (doi:10.
1890/02-0274)
Zhou X, Wang J, Hao Y, Wang Y. 2010 Intermediate
grazing intensities by sheep increase soil bacterial
diversities in an Inner Mongolian steppe. Biol.
Fertil. Soils 46, 817–824. (doi:10.1007/s00374-0100487-3)
4.
5.
6.
Frenot Y, Chown SL, Whinam J, Selkirk PM, Convey
P, Skotnicki M, Bergstrom DM. 2005 Biological
invasions in the Antarctic: extent, impacts and
implications. Biol. Rev. 80, 45 –72. (doi:10.1017/
S1464793104006542)
Chapuis JL, Frenot Y, Lebouvier M. 2004 Recovery of
native plant communities after eradication of rabbits
from the subantarctic Kerguelen Islands, and
influence of climate change. Biol. Conserv. 117,
167 –179. (doi:10.1016/S0006-3207(03)00290-8)
Wall DH. 2005 Biodiversity and ecosystem
functioning in terrestrial habitats of Antarctica.
7.
8.
9.
Antarct. Sci. 17, 523– 531. (doi:10.1017/
S0954102005002944)
Steiman R, Frenot Y, Sage L, Seigle-Murandi F,
Guiraud P. 1995 Contribution to the knowledge of
the mycoflora of Kerguelen Islands. Cryptogamie
Mycol. 16, 277– 291.
Taberlet P et al. 2012 Soil sampling and isolation of
extracellular DNA from large amount of starting material
suitable for metabarcoding studies. Mol. Ecol. 21,
1816–1820. (doi:10.1111/j.1365-294X.2011.05317.x)
Taberlet P et al. 2007 Power and limitations of the
chloroplast trnl (UAA) intron for plant DNA
4
Biol. Lett. 11: 20150408
Rabbits were eradicated from Ile Guillou 20 years ago [5] but
our analyses indicate that soil fungal communities on this island
are yet to recover (figure 2a; electronic supplementary material,
figure S7). Indeed, both plant and fungal communities on Ile
Guillou are more similar to the rabbit-disturbed communities
on Isthme Bas than they are to the undisturbed ones on Ile
Australia (figures 1c and 2a; electronic supplementary material,
figures S5–S7). We suggest that the current status of fungal
communities on Ile Guillou reflects slow recovery of the
native vegetation following rabbit eradication. That the vegetation on this island has not yet begun to show signs of
recovery has been linked to (i) slow growth rates in dominant
species (e.g. Azorella) and (ii) changes in abiotic conditions
(e.g. soil, climate) that may have induced a stable state more
favourable to introduced species (e.g. Taraxacum sp.) [4,5].
The introduction of rabbits to Kerguelen has led to changes
in plant communities and soil characteristics that have modified
the availability of both microhabitats and resources to soil
fungi. Our study illustrates the multifaceted and long-lasting
effects of introduced herbivores on fungal diversity as well as
the overall sensitivity of sub-Antarctic ecosystems to such introductions. Rabbit-induced changes to soil fungal communities
may have had flow-on effects in terms of lost functionality.
The functional characterization of these communities will
provide new insights into the resistance and resilience of
threatened sub-Antarctic ecosystems.
rsbl.royalsocietypublishing.org
are or were rabbit-affected, respectively (figure 2; electronic
supplementary material, figure S7). We found no consistent
differences in soil parameters nor have any reason to assume
that limited dispersal explains these differences (electronic
supplementary material, Results, table S2 and figure S3).
Instead, the co-structuring of soil fungal and plant communities (table 1; electronic supplementary material, figure S6)
suggests that differences between fungal communities are the
result of contrasting rabbit disturbance histories. Similar
trends in soil community composition have been observed in
other ecosystems when pressure from mammalian grazers
was high [2,3]. We hypothesize that simplification of plant
communities as a result of rabbit activity reduces the
number of soil microhabitats, and thereby soil fungal diversity
(figure 2b). The exclusion of the cushion plant A. selago from
areas affected by rabbits may be particularly significant in
terms of explaining differences between soil fungal communities. Specifically, the cushion growth form has been shown
to modify local abiotic conditions and promote soil microhabitat diversity [14]. This hypothesis is also consistent with our
results for individual plots on the Isthme Bas where more
diverse fungal communities were found under mixed species
cover than mono-specific Acaena cover (figure 2b; electronic
supplementary material, figures S2 and S7).
Although there are broad differences in fungal communities
between study sites, similarities between individual plots
suggest additional processes are at work. At two naturally disturbed plots on Ile Australia (AU08 and AU10), soil fungal
communities are more similar to those on Isthme Bas and Ile
Guillou (figure 2; electronic supplementary material, figure
S7). In this case, we suggest that disturbance-induced erosion
has promoted a loss of soil organic matter, and thereby lowered
soil fungal diversity. If so then rabbits may also be mediating
differences in soil fungal diversity via a direct impact on
resource availability [15]. Burrowing activities would promote
increased soil erosion; consistent with this, the plot with the
lowest soil fungal richness is also substantially eroded
(IB16; electronic supplementary material, figure S7). We also
identified relationships between soil parameters and soil
fungal community diversity with no potential link to rabbitinduced disturbance history. Specifically, plots with waterlogged, highly acidic soils are associated with lower soil
fungal diversity on both Ile Australia (AU07) and Ile Guillou
(GU15) (figure 2b; electronic supplementary material, figure
S7). In this case, our results imply that soil acidity limits
fungal richness directly.
Downloaded from http://rsbl.royalsocietypublishing.org/ on June 16, 2017
Ecol. 26, 32 –46. (doi:10.1111/j.1442-9993.2001.
01070.pp.x)
12. R Core Team. 2013 R: a language and
environment for statistical computing.
Vienna, Austria: R Foundation for Statistical
Computing.
13. Oksanen J et al. 2015 vegan: Community Ecology
Package. R package v. 2.2 –1. See http://CRAN.Rproject.org/package=vegan.
14. Roy J, Albert CH, Ibanez S, Saccone P, Zinger L,
Choler P, Clément JC, Lavergne S, Geremia RA. 2013
Microbes on the cliff: alpine cushion plants structure
bacterial and fungal communities. Front. Microbiol.
4, 64. (doi:10.3389/fmicb.2013.00064)
15. Waldrop MP, Zak DR, Blackwood CB, Curtis CD, Tilman
D. 2006 Resource availability controls fungal diversity
across a plant diversity gradient. Ecol. Lett. 9, 1127–
1135. (doi:10.1111/j.1461-0248.2006.00965.x)
5
rsbl.royalsocietypublishing.org
barcoding. Nucleic Acids Res. 35, e14. (doi:10.1093/
nar/gkl938)
10. Epp LS et al. 2012 New environmental
metabarcodes for analysing soil DNA: potential for
studying past and present ecosystems. Mol. Ecol.
21, 1821 –1833. (doi:10.1111/j.1365-294X.2012.
05537.x)
11. Anderson MJ. 2001 A new method for nonparametric multivariate analysis of variance. Austral
Biol. Lett. 11: 20150408