332
Because our cloning-sequencing studies in Lake Pavin
were designed to characterize the taxonomic diversity of
‘small unidentifi ed’ HNF and their related functions, their
potential trophic strategies were inferred from the trophic
behaviors displayed by their phylogenetically closest
described representatives (Lefèvre et al. 2007 , 2008 ). This
exercise yielded the following relative trophic mode affi lia-
tions for our genetic sequences: 19–31 % for parasites,
15–20 % for saprophytes and (or) parasites, and 4–26 % for
bacterivores. In addition, there were 6–12 % novel clades of
cercozoan-affi liated for which we were not able to assign
any trophic role.
Because Fungi had rarely been taken into account in
microbial food web processes in pelagic systems, the pres-
ence of a high percentage of Fungi in our retrieved sequences
was our most unexpected result. Therefore we decided to
direct our efforts towards the characterization of the pelagic
fungal community of the Lake Pavin using two approaches:
(i) the cloning/sequencing of the 18S, ITS1, 5.8S, ITS2 and
28S regions using primers specifi cally targeting Fungi, and
(ii) the pyrosequencing of the 18S rDNA hypervariable V2,
V3 and V5 regions using both universal eukaryotic primers
and Fungi-specifi c primer (Monchy et al. 2011 ). The
cloning- sequencing approach yielded 146 complete SSU
rDNA gene sequences that clustered into 46 OTUs at 99 %
similarity level. Half of these OTUs belonged to Fungi with
15, 7 and 1 representatives for Chytridiomycota , Ascomycota
and Basidiomycota , respectively. The other half of our
sequences affi liated to Viridiplantae (11), Cryptophyta (5),
Alveolata (4), Telonemida (1), Ichtyosporea (1) and
Stramenopile (1). Sample rarefaction curves , however, did
not reach saturation, indicating that the biodiversity was
underestimated. Our pyrosequencing approach greatly
improved this issue and a better diversity coverage was
obtained in our samples. When using eukaryotic and fungal
primers, a total of 23 519 and 18 545 reads were obtained,
respectively. Using eukaryotic primers, the most represented
groups were Stramenopile (27 % of reads), Chlorophyta
(19 %), Metazoa (15 %), Fungi (12 %), Alveolata (10 %),
Telonemida (8 %), and Cercozoa (7 %). Although the use of
fungal primers resulted in an enrichment of fungal sequences
(15 % of the pyrosequence reads) and a decrease in non-
targeted reads such as metazoan and Viridiplantae, untar-
geted groups such as Katablepharidophyta and Cryptophyta,
still represented 60 % and 10 % of the total reads, respec-
tively (Fig. 20.2 ). Primer specifi city is one of the biggest
issue in molecular diversity surveys and although the use of
a fungal primer set resulted in a substantial enrichment of
fungal reads, including 4 new chytrid OTUs (and 3 other
fungi), the issue related to primer specifi city towards pelagicFig. 20.2 Proportion of taxonomic groups identifi ed in Lake Pavin
using pyrosequencing of 18S rRNA gene hypervariable regions ampli-
fi ed by primer sets targeting two different taxonomic levels (Eukaryotes
and Fungi). The pie diagram displayed the proportion of reads, obtained
from two sampling stations and using the two primers sets (eukaryote
and fungus), belonging to a particular phylum. ‘No hit’ corresponds to
reads having no homologous sequence in database. ‘Not assigned’ cor-
responds to reads having a match in database but without a precise taxo-
nomic phylum assignment. For more details, see the main text and
Monchy et al. ( 2011 )T. Sime-Ngando et al.