Genes under positive selection in a model plant pathogenic fungus, Botrytis.

Abstract

The rapid evolution of particular genes is essential for the adaptation of pathogens to new hosts and new environments. Powerful methods have been developed for detecting targets of selection in the genome. Here we used divergence data to compare genes among four closely related fungal pathogens adapted to different hosts to elucidate the functions putatively involved in adaptive processes. For this goal, ESTs were sequenced in the specialist fungal pathogens Botrytis tulipae and Botrytis ficariarum, and compared with genome sequences of Botrytis cinerea and Sclerotinia sclerotiorum, responsible for diseases on over 200 plant species. A maximum likelihood-based analysis of 642 predicted orthologs detected 21 genes showing footprints of positive selection. These results were validated by resequencing nine of these genes in additional Botrytis species, showing they have also been rapidly evolving in other related species. Twenty of the 21 genes had not previously been identified as pathogenicity factors in B. cinerea, but some had functions related to plant-fungus interactions. The putative functions were involved in respiratory and energy metabolism, protein and RNA metabolism, signal transduction or virulence, similarly to what was detected in previous studies using the same approach in other pathogens. Mutants of B. cinerea were generated for four of these genes as a first attempt to elucidate their functions.

Full text

Page 1

Genes under positive selection in a model plant pathogenic fungus, Botrytis
Gabriela Aguiletaa,b,c, Juliette Lengellec, Hélène Chiapelloc, Tatiana Girauda,b, Muriel Viaudd,
Elisabeth Fournierf, François Rodolphec, Sylvain Martheyc, Aurélie Ducassed, Annie Gendraultc,
Julie Poulaine, Patrick Winckere, Lilian Goutd,g,⇑
aEcologie, Systématique et Evolution, Université Paris-Sud UMR8079, F-91405 Orsay Cedex, France
bEcologie, Systématique et Evolution, CNRS UMR8079, 91405 Orsay Cedex, France
cMIG – INRA UR1077, Domaine de Vilvert, 78352 Jouy en Josas Cedex, France
dBIOGER – INRA UR1290, Avenue Lucien Brétignières, 78850 Thiverval-Grignon, France
eGenoscope – Centre National de Séquençage, UMR CNRS 8030, 2 Gaston Crémieux, CP 5706, 91507 Evry, France
fBGPI – INRA UMR385, TA A 54/K, Campus International de Baillarguet, 34398 Montpellier Cedex 5, France
gBIOGER – AgroParisTech UR1290, 75231 Paris Cedex 05, Paris, France
a r t i c l ei n f o
Article history:
Received 6 December 2011
Received in revised form 15 February 2012
Accepted 23 February 2012
Available online 3 March 2012
Keywords:
Botrytis
Molecular evolution
Fungi
Natural selection
Coevolution
a b s t r a c t
The rapid evolution of particular genes is essential for the adaptation of pathogens to new hosts and new
environments. Powerful methods have been developed for detecting targets of selection in the genome.
Here we used divergence data to compare genes among four closely related fungal pathogens adapted to
different hosts to elucidate the functions putatively involved in adaptive processes. For this goal, ESTs
were sequenced in the specialist fungal pathogens Botrytis tulipae and Botrytis ficariarum, and compared
with genome sequences of Botrytis cinerea and Sclerotinia sclerotiorum, responsible for diseases on over
200 plant species. A maximum likelihood-based analysis of 642 predicted orthologs detected 21 genes
showing footprints of positive selection. These results were validated by resequencing nine of these genes
in additional Botrytis species, showing they have also been rapidly evolving in other related species.
Twenty of the 21 genes had not previously been identified as pathogenicity factors in B. cinerea, but some
had functions related to plant–fungus interactions. The putative functions were involved in respiratory
and energy metabolism, protein and RNA metabolism, signal transduction or virulence, similarly to what
was detected in previous studies using the same approach in other pathogens. Mutants of B. cinerea were
generated for four of these genes as a first attempt to elucidate their functions.
? 2012 Elsevier B.V. All rights reserved.
1. Introduction
There is considerable interest in finding the genes that underlie
the capacity of pathogens to evolve and adapt to new habitats and
hosts, especially in the case of devastating emergent diseases
(Archie et al., 2009). Global trade and the worldwide transport of
goods and people increasingly contribute to the spread of patho-
gens to new habitats and potential hosts (Lebarbenchon et al.,
2008). Fungi are among the most devastating plant pathogens,
causing important losses in agriculture, decimating natural popu-
lations, and they have been associated with several cases of emer-
gent plant diseases in new hosts and novel environments
(Anderson et al., 2004; Desprez-Loustau et al., 2007; Blaustein
and Johnson, 2010). In this context, it is highly desirable to deter-
mine the genes that evolved rapidly, in order to understand how
new diseases originated.
Recent studies have indeed shown that the rapid evolution of
particular genes can play a significant role in favoring the capacity
of a pathogen to adapt to a new environment or to infect new host
species (Matute et al., 2008; Sacristán and García-Arenal, 2008; de
Crecy et al., 2009). Pinpointing the relevant genes in the lab is often
a daunting and costly task. A faster and cheaper alternative, in the
absence of a priori candidates for genes involved in pathogenicity,
is to look for evidence of genes evolving rapidly or that have been
subject to positive selection. For a pathogen, advantageous substi-
tutions that contribute to the capacity of infecting a host or adapt-
ing a new environment will in principle be rapidly fixed in the
population. Genes showing footprints of positive selection can
additionally be involved in the arms race with their host (Aguileta
et al., 2009) and these are also interesting genes to be identified.
Such a blind, computational approach has already identified the
genes and gene functions likely involved in the origin of new path-
ogenic species, specialized on closely related plant hosts (Aguileta
et al., 2010).
The genus Botrytis Persoon (Ascomycota, Leotiomycetes, Sordar-
iaceae) encompasses 22 pathogenic species with different host
1567-1348/$ - see front matter ? 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.meegid.2012.02.012
⇑Corresponding author at: BIOGER – INRA UR1290, Avenue Lucien Brétignières,
78850 Thiverval-Grignon, France. Tel.: +33 1 30 81 54 34; fax: +33 1 30 81 53 06.
E-mail addresses: gout@versailles.inra.fr, lilian.gout@agroparistech.fr (L. Gout).
Infection, Genetics and Evolution 12 (2012) 987–996
Contents lists available at SciVerse ScienceDirect
Infection, Genetics and Evolution
journal homepage: www.elsevier.com/locate/meegid

Page 2

ranges. All Botrytis species are necrotrophic pathogens, meaning
that they feed only on dead tissues; they therefore possess the abil-
ity to kill the attacked plant cells and to develop on dead tissues.
The genus is separated into two clades (Staats et al., 2005), one
with four species attacking only Eudicot plants, the other including
the 18 remaining species, found mainly on Monocot plants. The
Eudicot clade includes Botrytis cinerea (teleomorph Botryotinia
fuckeliana), the agent of gray mold on numerous plants (>200), that
has recently been separated into two cryptic species (Fournier
et al., 2005; Walker et al., 2011). The wide variety of symptoms
on different organs and plants suggests that the B. cinerea genome
possesses a large ‘‘arsenal of weapons’’ to attack its hosts (Choquer
et al., 2007; Williamson et al., 2007). Apart from the gray mold
agent, most Botrytis species have a narrow host range, restricted
to few or a single species. For example, in the clade associated with
Monocot hosts, Botrytis tulipae attacks several Tulipae spp., and
Botrytis ficariarum has been described only on Ficaria verna. The
closely related species Sclerotinia sclerotiorum is also necrotrophic
pathogen with a wide host range (>400; Bolton et al., 2006).
Searching for rapidly evolving genes between these four species
may therefore help to elucidate the genetic basis of host range
determination and adaptation to different hosts or environments,
or genes involved in the rapid coevolution with their host plants.
B. cinerea is also one of the most studied necrotrophic fungi, for
which numerous reverse genetic and transcriptomic tools are
available (van Kan, 2006; Choquer et al., 2007), making it possible
to test the functions of particular candidate genes.
In this study, we used a genome-wide approach to compare as
many orthologs as possible from the species B. cinerea, B. tulipae,
B. ficariarum and S. sclerotiorum, four closely related fungal phyto-
pathogens adapted to different hosts and with narrow or wide host
ranges. The goal was to find genes evolving under positive selec-
tion that might thus be involved in host range determination and
adaptation to the host. Rapidly evolving genes may alternatively
be involved in coevolution with the current hosts (Aguileta et al.,
2009) or genetic incompatibilities between species (Presgraves
et al., 2003), which are also important phenomena to understand
in pathogens. For this goal, EST libraries were sequenced for B. tuli-
pae and B. ficariarum, and compared with full genome sequences of
B. cinerea and S. sclerotiorum (Amselem et al., 2011). The best can-
didate genes were sequenced in additional species of the Botrytis
genus to validate the evolutionary pressure. Finally, we also at-
tempted to validate using functional genetics that these candidate
genes were involved in pathogenicity and adaptation to the host,
by disrupting them in one B. cinerea strain and testing mutants
in host plant-infection studies.
2. Experimental procedures
2.1. Strains, culture conditions, RNA isolation, cDNA library
construction and sequencing
The B. cinerea T4 strain was isolated in a glasshouse of tomatoes
(Lycopersicum esculentum) in Avignon, France (Levis et al., 1997).
The B. ficariarum CBS17663 and the B. tulipae Bt9901 strains were
obtained from culture collections of micro-organisms of the Cen-
traalbureau voor Schimmelcultures (CBS, Utrecht, The Nether-
lands) and of the Applied Plant Research (PPO) organization
(Lisse, the Netherlands), respectively. The origins of B. ficariarum
and B. tulipae strains were described previously (Staats et al.,
2005). These isolates were routinely grown on solid NY medium
(2 g l?1malt extract, 2 g l?1yeast extract, 15 g l?1agar) at 21 ?C
with 16 h daylight a day.
The genomic sequence of B. cinerea T4 isolate was obtained from
theURGI(http://urgi.versailles.inra.fr/index.php/urgi/Species/Botrytis)
and that of S. sclerotiorum were obtained from the Broad Institute
(http://www.broadinstitute.org/annotation/genome/sclerotinia_
sclerotiorum/MultiHome.html). Two EST libraries were obtainedforB.
ficariarum CBS17663 isolate and B. tulipae Bt9901 isolate. For RNA
extraction, mycelia and spores of 4-day-old cultures were harvested
by flooding the plate with 1 ml of sterile dH2O, inoculated into
100 mlofliquidNYmediumandgrownat21 ?Cfor48 hwithaeration
(shaking incubator, 120 r.p.m.). For each culture, the fungal cell mass
wasthenfilteredthroughasterileNylonmembrane,washedoncewith
sterile dH2O and transferred to 100 ml of V8-based medium. This rich
plant-based medium was also supplemented with 20% homogenized
leavesoftulip(Tulipaspp.)orbuttercup(F.verna),whichareknownhost
plants of B. tulipae and B. ficariarum, respectively. Such enrichment in
host plant compounds was successfully used in previous proteomic
studiestoinducetheexpressionofB.cinereagenesrelatedtotheinfec-
tion(Espinoetal.,2010;Shahetal.,2009).After2 hincubationat21 ?C,
fungaltissueswereharvestedbyfiltration,frozeninliquidnitrogenand
stored at ?80 ?C until RNA extraction. Total RNA was extracted from
mycelialsamplesgroundinliquidnitrogenwithTRIzol(Invitrogen)fol-
lowing the manufacturer’s instructions. Thereafter, cDNA was synthe-
sizedfrom2lgoftotalRNAandcDNAlibrarieswerepreparedbyusing
the Creator SMART cDNA Library Construction kit (ClonTech), accord-
ing to the manufacturers’ instructions. For each Botrytis species, the
cDNA library was robotically arrayed and 20,000 clones were submit-
ted for systematic 50end sequencing performed by Genoscope (CEA
Evry,France)followingstandardprotocolsinordertoidentifyasample
of genes having orthologues in the four species.
2.2. Sequence cleaning, assembly and annotation
Sequences are available in Genbank (accessions numbers
FO084242 to FO113725). Raw sequence data were cleaned from
vector and adaptor sequences. Contaminating plasmid sequences
were removed from the analyses. The SURF (SeqUence Repository
and Feature detection) package (Lannuccelli, 2005) was used for
sequence base-calling, cleaning, and for detection of any contami-
nation in putative inserts. This analysis involved three steps
including the use of PHRED (Ewing and Green, 1998; Ewing
et al., 1998), which detected bad quality regions, of RepeatMasker,
which masked low complexity regions, and of Crossmatch, which
found putative contaminated sequences of the UNIVEC database
and single nucleotide repetitions. Only sequences with a SURF
score over 20 on at least 100 bp were released in the EST division
of the EMBL-EBI Nucleotide Sequence Database.
ESTs were aligned and assembled into contigs using the CAP3
software (Huang and Madan, 1999) when the criterion of a mini-
mum identity of 95% over 50 bp was met. When an EST could not
be assembled with others in a contig, it was retained as a ‘‘singlet’’.
The contigs and the singlets should thus correspond to sequences of
unique genes, and will be called hereafter ‘‘unisequences’’.
For the annotation of the predicted ESTs, the consensus se-
quences of the contigs and the sequences of the singlets were
compared to B. cinerea sequences in the Genbank database and
in the Uniprot database (The Uniprot Consortium, 2007) using
the tBLASTx and the BLASTx algorithms (Altschul et al., 1997).
Unisequences showing significant similarity (E-value 6 10?4) to
database entries were annotated using their most significant
match. Unisequences were also classified into gene ontology
functional categories (http://www.geneontology.org) based on
BLAST similarities to known genes of the NCBI nr (non-redun-
dant) protein database and using the Blast2GO annotation tool
(Conesa et al., 2005).
The EST data obtained for B. ficariarum and B. tulipae were inte-
gratedinto a databasecalled FunyEST,which is part of the FunyBase
resource (Marthey et al., 2008) and freely accessible through the
web site http://genome.jouy.inra.fr/funybase. FunyEST structure
988
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996

Page 3

and content is similar to one developed for the MICROBASE re-
source, previously described in Aguileta et al. (2009) and dedicated
to Microbotryum violaceum EST analysis. The FunyEST database in-
cludes information on EST sequences, contigs, annotations, gene
ontology functional categories and search programs to compare
similarities of any sequence against the database.
2.3. Unisequence CDS predictions and clustering
The pipeline of the Prot4EST software (Wasmuth and Blaxter,
2004) was used to predict unisequence CDS positions and to trans-
late coding regions into protein sequences. In a first step, all puta-
tive ribosomal sequences in the dataset were identified through a
BLASTn search against the rRNA sequence database (Ribosomal
Database II) and the sequences whose BLASTn E-value P 1e-65
were discarded. The second and third steps use the BLASTx algo-
rithm to detect any similarity between unisequences and se-
quences from both the mitochondrial protein database (NCBI ftp
site) and the Uniprot database (The Uniprot Consortium, 2007).
Unisequences showing a significant BLAST result (i.e. cutoff of
e-08) against the mitochondrial database were annotated as mito-
chondrial genes to be translated subsequently with the relevant
genetic code. We removed all sequences corresponding to trans-
posable elements, as well as those containing internal stop codons.
Sequences that matched the Uniprot-Swissprot database signifi-
cantly (i.e. cutoff of e-08) were considered as CDS and a HSP tile
path was constructed. This means that Prot4EST then considers
that the nascent translation of these sequences can be extended
at either end in the same reading frame. Only sequences that
yielded no sequence similarity were then submitted to the fourth
step of the pipeline which aims at identifying coding regions using
hidden Markov models implemented in the ESTscan software (Iseli
et al., 1999). For this step, a transition matrix was created from the
genome ORF sequences ofB. cinerea and CDS sequences ofB. tulipae
and Botrytis elliptica, obtained from the EMBL. Predicted polypep-
tides satisfying a given length threshold criteria (CDS of at least
30 codons in length and covering at least 10% of the input se-
quence) then underwent the extension process (like for HSP tiling).
In a fifth step, the DECODER program (Fukunishi and Hayashizaki,
2001) is used to predict the CDS and polypeptide translations for
the remaining sequences. DECODER exploits the quality scores of
the sequences produced from base calling software (such as PHRED
used in the SURF package) and additional text base information
(such as optimal codon usage). DECODER computes a likelihood
score for each possible CDS, and the one with the lowest score is
chosen as the correct CDS. Finally, a last attempt is performed to
provide a putative polypeptide translation based on the longest
string of amino acids uninterrupted by stop codons from a six-
frame translation of the sequence. In spite of all the caution
exerted at the previously described stages for predicting CDSs,
potential problems could arise from the incorporation of an intron
that does not destroy the reading frame. In this case, the predicted
CDS would contain a region of neutral sequence that could bias
analyses of selective pressure. This phenomenon however is unli-
kely to be a problem here as gaps were removed from alignments.
Clustering of unisequence CDSs into groups of orthologs
included three main steps for which we employed Ortho-MCL (Li
et al., 2003) and custom-made Perl scripts. The first step detected
thesingle-copyunisequencesfromeachCDSlibrary.Todothis,each
library was aligned against itself by using a BLASTn algorithm. All
CDS sequences having exactly one significant hit (e-value < 1e-10)
wereconsideredassingle-copyunisequences(thusavoidinghidden
paralogy) and were kept for the detection of orthologs among the
four libraries, using a derivative of the Best Bidirectional Hit for n
sequences. The single-copy unisequences from all four libraries
were combined in a single file and aligned against themselves using
the BLASTn algorithm. All single-copy unisequences with a hit
(e-value < 0.1) with a unisequence of another library were consid-
ered to have an ortholog in the corresponding species and were
therefore kept for the last step. The last script compiled alignment
results and built clusters of putative orthologs, including four
sequences, each of them belonging to a different species.
2.4. Orthologous gene alignment, filtering and sorting by alignment
length
The predicted protein unisequences of orthologs were aligned
using T-coffee (Notredame et al., 2000) with default settings. The
corresponding nucleotide alignments were performed by using the
protein alignments as guide, as implemented in the tranalign pro-
gram of the EMBOSS package (http://embossgui.sourceforge.net/
demo/manual/tranalign.html). In order to keep only reliable align-
ments, which are crucial for the subsequent detection of selection,
the alignments were then filtered using different criteria. First, we
required a level of protein sequence identity of at least 70% for all
alignments of putative orthologs. Second, the alignments were
post-processed to remove gaps and keep only unambiguously
aligned blocks of sequence. This step was performed using Gblocks
(Castresana, 2000) with the maximum number of non-conserved
positions set to 8, and the minimal block size set to 5 (for all other
parameters, default settings were used). Finally, we used the length
of the final alignments to classify the resulting ortholog clusters for
subsequent analysis: clusters with at least 300 nucleotides were
analyzed individually.
2.5. Detection of positive selection
Positive selection was tested using the CODEML program of the
PAML4 package (Yang, 2007). Selective pressure was measured
by using the nonsynonymous/synonymous substitution rate ratio
(dN/dS), also referred to as x. An x < 1 suggests purifying selection,
x = 1 is consistent with neutral evolution, and x > 1 is indicative
of positive selection (Yang and Bielawski, 2000). Nested codon
models implementing the x ratio can be compared by means of
a likelihood ratio test (LRT) (Anisimova et al., 2001). We used the
null model M1a, which assumes two site classes with 1 > x > 0,
and x1= 1, which therefore implicitly supposed that no site is
under positive selection, and compared it with the alternative
model M2a, which adds an extra class of sites that allows x to take
values >1. We also compared the null model M7, which assumes a
beta distribution of x across sites, with the alternative model M8,
which adds an extra class of sites to M7 where x can take values
>1. Thereby positive selection can be detected if a model allowing
for positive selection is significantly more likely (as estimated by
the LRT) than a null model without positive selection.
2.6. Functional annotation
In order to assign functional annotation to cluster orthologs
exhibiting evidence for positive selection, we first used the avail-
able FunyEST annotation of individual unisequences contained in
each cluster of B. ficariarum and B. tulipae. For B. cinerea and S. scle-
rotiorum BLAST searches against their whole genomes at NCBI were
conducted. As described previously, this annotation was obtained
from the BLASTp best hit and the corresponding GO terms.
We then performed three complementary analyses in order to
collect maximal functional information contained in the clusters
of interest. In a first step, we looked for all possible motifs, signal
and domains in the individual sequences using the Interproscan
software (assuming default settings) of the Interpro database
(Mulder et al., 2007). In the second and third steps, we tried to
identify distant homologs for each individual sequences using
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996
989

Page 4

two complementary methods: (1) Interproscan with default set-
tings using the two EST banks of B. tulipae and B. ficariarum and
the two genomic banks of B. cinerea and S. sclerotiorum. Different
databases were searched: Pfam (Finn et al., 2010) and Panther
(Thomas et al., 2003) for motif search, Superfamily (Gough, 2002)
for protein family assignment, CATH (Cuff et al., 2011) for 3D struc-
ture, SignalP (Bendtsen et al., 2004) for signal peptides and an-
chors, and Tmhmm (Krogh et al., 2001) for transmembrane
domain identification. (2) PSI-BLAST (Position-Specific Iterated
BLAST) algorithm (Altschul et al., 1997), which is an efficient meth-
od to detect weak but biologically relevant sequence similarities
against the Uniprot database.
2.7. Gene inactivation of candidate genes
Fourcandidategenesfromthe blind in silicoapproachwereinac-
tivated by the construction of Knock-Out (KO) cassettes and proto-
plast transformation. The hygromycin resistance gene hph was
amplified from the plasmids pCB1004 (Carroll et al., 1994) or
pCSN44 (Colot et al., 2006) (primers in Table S1). The pairs of prim-
ers 5F/5R and 3F/3R were used to amplify regions of about 1 kb in 50
and in 30of each of the target genes (Table S1). Then, KO cassettes
consisting of the 50region of gene, the resistance gene and the 30re-
gion of gene were generated by using two alternative strategies.
Either the KO cassettes were generated by using the double-joint
PCR strategy described by Yu et al. (2004) or the KO cassettes were
generated by homologous recombination in yeast as described in
Colot et al. (2006). The Gene Knock-out kit was obtained from the
Fungal Genetic Stock Centre (http://www.fgsc.net/clones.html).
Protoplasts from B05.10 were prepared and transformed as
described previously (Levis et al., 1997) using 2 lg of linear DNA.
Transformed protoplasts were plated in molten osmotically
stabilized medium agar containing 100 lg ml?1
(Invitrogen). Transformants were selected after 6–8 days at 23 ?C,
sub-cultured twice on selective media and single spore cultures
were made to get genetically pure transformants. The screenings
for gene inactivation event were done by PCR using the primer
verif-5 located upstream the 50flanking region and Hyg-F located
inside the hph gene.
hygromycin
2.8. Infection assays
Infection assays of B. cinerea WT strain and mutants were done
on French bean (Phaseolus vulgaris cv. Caruso) and tomato leaves
(Solanum lycopersicum cv. Moneymaker) by inoculating detached
leaves with young unsporulating mycelium or conidial suspensions
from cultures on NY. Bean leaves were harvested from 2 weeks-old
plants and placed in a transparent plastic box lined with tissue
moistened with sterile water. Leaves were inoculated with
1.8 mm diameter-plugs of 3 days old mycelium. Alternatively, con-
idia were collected from 10 days-old plates and suspended in su-
crose phosphate buffer (10 mM sucrose, 10 mM KH2PO4) to a final
concentration of 105conidia per ml. Droplets of 10 ll were applied
tothe leaves.Storage boxescontaininginoculatedleaves wereincu-
batedinagrowthcabinetat21 ?Cwith16 hdaylight.Diseasedevel-
opment on leaves was recorded daily as radial spread from the
inoculation point to the lesion margin. Pathogenicity assays on
leaves wererepeatedthreetimesusingat leastfive leaves perassay.
3. Results
3.1. Sequence analysis: gene finding, assembly and clustering
We obtained 15,014 ESTs for B. ficariarum and 19,006 ESTs for B.
tulipae after the cleaning step. Next, unisequences (i.e., contigs or
singlets corresponding to sequences of unique genes) were re-
trieved and assembled for each species, which resulted in 2855
and 3715 unisequences for B. ficariarum and B. tulipae, respectively.
The following step involved the prediction of coding sequences
(CDS) for each unisequence to obtain the coding frame required
for detecting synonymous and nonsynonymous substitutions.
After carefully filtering incorrect results, a total of 2798 and 3661
CDSs were predicted for B. ficariarum and B. tulipae, respectively.
Finally, ortholog detection from comparisons between the two
EST libraries of species B. ficariarum and B. tulipae with the two
available genomes of B. cinerea and S. sclerotiorum yielded 979
clusters of orthologs, including four sequences from the four spe-
cies, all of which were at least 300 nucleotides long, the minimal
length for avoiding stochastic sampling problems during the detec-
tion of positive selection. These 979 alignments were verified for
their quality and a total of 642 ortholog clusters were further ana-
lyzed to detect signals of positive selection.
3.2. Detection of positive selection
Positive selection is detected when a model of evolution allow-
ing for positive selection appears significantly more likely than a
null model without positive selection, as indicated by likelihood
ratio tests (LRTs). The individual analysis of the 642 ortholog clus-
ters detected 21 genes as likely candidates to be subjected to posi-
tive selection, as revealed by the significant LRTs comparing
models M1a vs M2a and M7 vs M8. Table 1 presents the parameter
estimates of each model and Table 2 shows the results of the two
LRTs performed. The orthologs found to be under selection are:
B152, B161, B194, B248, B24, B266, B387, B398, B402, B417,
B431, B541, B549, B57, B605, B695, B814, B821, B907, B967 (LRTs
of M7–M8, P < 0.05) and B897, for which the LRT comparing M7
and M8 was marginally significant (P = 0.054).
3.3. Functional annotation of the putative genes detected as being
under positive selection
We performed the functional annotation of the 21 genes in a
two-step approach. In a first step, we assigned functional catego-
ries to genes using both the BLASTp best hit obtained for each of
the coding sequences included in the clusters (Table S2). The 21
genes were classified into eight GO (Gene Ontology) categories
according to their function: (1) regulation of gene expression [1
cluster], (2) respiratory and energy metabolism [6 clusters], (3)
protein degradation [3 clusters], (4) protein folding [1 cluster],
(5) cellular development [3 clusters], (6) cell wall modification [1
cluster], (7) binding [1 cluster] and (8) unknown function [5 clus-
ters]. We compared the proportions of genes in the different GO
classes in the whole set of 642 orthologs and in the 21 genes under
positive selection. Some GO classes appeared to include a higher
proportion of genes under positive selection, as compared to the
whole set of orthologs (especially the ‘‘molecular function class’’).
However, the differences between the two distributions were not
significant (Chi2 test, data not shown).
We performed a second step of annotation, by using three com-
plementary methods on each sequence of the 21 genes. First, motif
and domain predictions were obtained by using the Interpro data-
base (Table S3). These genes under positive selection and the list of
642 orthologous genes were also analyzed by TargetP in order to
detect signal peptides indicating their subcellular localization.
Interestingly, 33% of the genes under positive selection show a sig-
nal peptide suggesting that they are secreted or localized in the
membrane. This corresponds to a 2-fold enrichment compared to
the 642 genes analyzed in this study (13% have a predicted signal
peptide) and to the putative secreted and membrane genes of the
whole genome (15% have a predicted signal peptide; Amselem
990
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996

Page 5

Table 1
Parameter estimates of site-specific CODEML analyses.
Cluster M1a (neutral) M2a (selection)M7 (beta) M8 (beta&w)
?lnL Parameter estimates
?lnL Parameter estimates
?lnL Parameter
estimates
?lnL Parameter estimates
B1521309.21
p0 = 0.5 (p1 = 0.5), w0 = 0.0, w1 = 1.01275.62
p0 = 0.43, p1 = 0.35 (p2 = 0.22), w0 = 0.0 (w1 = 1.0),
w2 = 623.98
p0 = 0.92, p1 = 0.07 (p2 = 0.007), w0 = 0.07 (w1 = 1.0),
w2 = 35.11
p0 = 0.99, p1 = 0.0 (p2 = 0.007), w0 = 0.11 (w1 = 1.0),
w2 = 52.45
p0 = 0.99, p1 = 0.0 (p2 = 0.007), w0 = 0.0 (w1 = 1.0),
w2 = 10.04
p0 = 0.95, p1 = 0.0 (p2 = 0.05), w0 = 0.01 (w1 = 1.0),
w2 = 3.56
p0 = 0.95, p1 = 0.05 (p2 = 0.005), w0 = 0.08 (w1 = 1.0),
w2 = 12.94
p0 = 0.89, p1 = 0.09 (p2 = 0.02), w0 = 0.0 (w1 = 1.0),
w2 = 12.9
p0 = 0.98, p1 = 0.0 (p2 = 0.01), w0 = 0.01 (w1 = 1.0),
w2 = 4.72
p0 = 0.99, p1 = 0.0 (p2 = 0.005), w0 = 0.03 (w1 = 1.0),
w2 = 157.35
p0 = 0.73, p1 = 0.2 (p2 = 0.07), w0 = 0.06 (w1 = 1.0),
w2 = 1.0
p0 = 0.92, p1 = 0.06 (p2 = 0.02), w0 = 0.05 (w1 = 1.0),
w2 = 8.78
p0 = 0.98, p1 = 0.01 (p2 = 0.009), w0 = 0.04 (w1 = 1.0),
w2 = 10.91
p0 = 0.84, p1 = 0.0 (p2 = 0.16), w0 = 0.0 (w1 = 1.0),
w2 = 3.16
p0 = 0.98, p1 = 0.0 (p2 = 0.02), w0 = 0.02 (w1 = 1.0),
w2 = 8.03
p0 = 0.96, p1 = 0.0 (p2 = 0.04), w0 = 0.0 (w1 = 1.0),
w2 = 2.72
p0 = 0.98, p1 = 0.01 (p2 = 0.01), w0 = 0.0 (w1 = 1.0),
w2 = 29.11
p0 = 0.88, p1 = 0.12 (p2 = 0.008), w0 = 0.0 (w1 = 1.0),
w2 = 19.03
p0 = 0.91, p1 = 0.0 (p2 = 0.09), w0 = 0.02 (w1 = 1.0),
w2 = 3.05
p0 = 0.78, p1 = 0.20 (p2 = 0.02), w0 = 0.03 (w1 = 1.0),
w2 = 6.98
p0 = 0.81, p1 = 0.17 (p2 = 0.01), w0 = 0.11 (w1 = 1.0),
w2 = 18.65
p0 = 0.99, p1 = 0.0 (p2 = 0.006), w0 = 0.005 (w1 = 1.0),
w2 = 99.75
1309.21
p = 0.005,
q = 0.005
p = 0.17, q = 0.93
1275.81
p0 = 0.79, (p1 = 0.22), p = 0.005, q = 0.007.0,
w = 577.84
p0 = 0.99 (p1 = 0.007), p = 0.26, q = 1.77,
w = 31.11
p0 = 0.99 (p1 = 0.007), p = 0.16, q = 1.2,
w = 53.16
p0 = 0.99 (p1 = 0.007), p = 0.005, q = 7.66,
w = 10.04
p0 = 0.95 (p1 = 0.05), p = 1.41, q = 99.0, w = 3.57
B161 1984.61
p0 = 0.91 (p1 = 0.09), w0 = 0.006,
w1 = 1.0
p0 = 0.88 (p1 = 0.12), w0 = 0.0,
w1 = 1.0
p0 = 0.99 (p1 = 0.01), w0 = 0.0,
w1 = 1.0
p0 = 0.91 (p1 = 0.09), w0 = 0.0,
w1 = 1.0
p0 = 0.92 (p1 = 0.08), w0 = 0.06,
w1 = 1.0
p0 = 0.89 (p1 = 0.11), w0 = 0.0,
w1 = 1.0
p0 = 0.95 (p1 = 0.05), w0 = 0.0,
w1 = 1.0
p0 = 0.95 (p1 = 0.05), w0 = 0.0,
w1 = 1.0
p0 = 0.73 (p1 = 0.27), w0 = 0.06,
w1 = 1.0
p0 = 0.9 (p1 = 0.1), w0 = 0.04,
w1 = 1.0
p0 = 0.97 (p1 = 0.03), w0 = 0.04,
w1 = 1.0
p0 = 0.79 (p1 = 0.21), w0 = 0.0,
w1 = 1.0
p0 = 0.94 (p1 = 0.06), w0 = 0.003,
w1 = 1.0
p0 = 0.95 (p1 = 0.05), w0 = 0.0,
w1 = 1.0
p0 = 0.97 (p1 = 0.03), w0 = 0.0,
w1 = 1.0
p0 = 0.88 (p1 = 0.12), w0 = 0.0,
w1 = 1.0
p0 = 0.85 (p1 = 0.15), w0 = 0.0,
w1 = 1.0
p0 = 0.78 (p1 = 0.22), w0 = 0.03,
w1 = 1.0
p0 = 0.79 (p1 = 0.21), w0 = 0.08,
w1 = 1.0
p0 = 0.98 (p1 = 0.02), w0 = 0.0,
w1 = 1.0
1980.821985.50 1980.37
B194 641.61 637.72
641.65
p = 0.005, q = 0.05 637.75
B2482130.09 2124.032136.79
p = 0.01, q = 0.232124.03
B24 970.69967.45 970.78
p = 0.005, q = 0.05967.45
B266 1669.131666.68
1669.78
p = 0.1, q = 0.591666.60
p0 = 0.995 (p1 = 0.005), p = 0.3, q = 2.15,
w = 12.54
p0 = 0.98 (p1 = 0.02), p = 0.01, q = 0.16,
w = 11.89
p0 = 0.98 (p1 = 0.02), p = 1.42, q = 99.0, w = 4.72
B387630.35626.47 630.39
p = 0.005, q = 0.05626.45
B398 969.15967.69 970.70
p = 0.01, q = 0.16 967.69
B402882.05
874.83882.95
p = 0.005, q = 0.05874.83
p0 = 0.99 (p1 = 0.005), p = 3.31, q = 99.0,
w = 157.95
p0 = 0.98 (p1 = 0.02), p = 0.27, q = 0.82,
w = 226.8
p0 = 0.97 (p1 = 0.03), p = 0.25, q = 2.34, w = 7.97
B417923.88923.88 923.72
p = 0.22, q = 0.59919.49
B4311847.671842.76 1850.05
p = 0.1, q = 0.571842.59
B541 1001.61
999.89 1002.27
p = 0.14, q = 1.82 999.34
p0 = 0.99 (p1 = 0.008), p = 0.38, q = 6.61,
w = 10.88
p0 = 0.84 (p1 = 0.16), p = 0.005, q = 99.0, w = 3.16B549707.01 703.27 708.30
p = 0.005, q = 0.01 703.27
B571037.83 1034.671038.47
p = 0.006, q = 0.11034.68
p0 = 0.98 (p1 = 0.02), p = 2.48, q = 99.0, w = 8.04
B605857.20855.22858.53
p = 0.005, q = 0.05 855.22
p0 = 0.96 (p1 = 0.04), p = 0.005, q = 99.0, w = 2.72
B6951024.511016.321028.40
p = 0.005, q = 0.051016.90
p0 = 0.98 (p1 = 0.02), p = 0.005, q = 0.16,
w = 13.69
p0 = 0.99, (p1 = 0.01), p = 0.005, q = 0.05,
w = 14.56
p0 = 0.91 (p1 = 0.09), p = 2.59, q = 99.0, w = 3.05
B8141017.531013.81
1017.73
p = 0.005, q = 0.05 1013.88
B821 1433.531430.641434.74
p = 0.005, q = 0.031430.65
B8971014.15 1012.011013.68
p = 0.10, q = 0.371010.85
p0 = 0.98 (p1 = 0.02), p = 0.15, q = 0.65, w = 5.9
B907695.28 691.44
695.69
p = 0.14, q = 0.35691.39
p0 = 0.99 (p1 = 0.01), p = 0.33, q = 1.01, w = 18.27
B967882.33879.68 885.13
p = 0.01, q = 0.24879.70
p0 = 0.99 (p1 = 0.006), p = 0.005, q = 0.13,
w = 99.86
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996
991

Page 6

et al., 2011). Among the genes that are under positive selection and
predicted to be secreted, two have experimental evidence for
secretion as their products were identified in the secretome of B.
cinerea on plant containing media (B431 and B897; Espino et al.,
2010). Then, using PSI-BLAST comparisons (Table S4), allowing
detecting weak but biologically relevant sequence similarities with
the Uniprot database, annotations were found for three genes:
B266, B417 and B907. Finally, CATH search, to look for compatible
fold in the PDB structure database, resulted in annotations for 10
genes: B161, B248, B24, B398, B431, B541, B549, B57, B695, B821
(Table S5). Overall, owing to the use of multiple annotation meth-
ods, the total 21 clusters of putatively interesting orthologs could
be annotated. Among the putative functions of the genes showing
significant signal of positive selection, many could be related to
interaction with the host plants (see Section 4).
Among the 21 genes exhibiting evidence of positive selection,
B417 and B907 were identified as Leotiomycetes specific following
an orthology analysis made on nine sequenced fungal genomes
(Amselem et al., 2011). This list of genes under positive selection
was also compared to the list of B. cinerea genes that are over-ex-
pressed during sunflower leaves infection (Amselem et al., 2011).
Only the B161 gene, which encodes a putative dehydrogenase is
present in both lists.
3.4. Validation of selected candidate genes showing signals of positive
selection by sequencing in additional Botrytis species
On the basis of their putative function, their elevated x ratio,
their secretion, their expression during plant infection and their
lineage specificity (Amselem et al., 2011; Espino et al., 2010), nine
out of the 21 genes showing signals of positive selection were se-
lected (B161, B24, B266, B417, B431, B57, B897, B907, B967). These
genes were sequenced in 12 additional species of Botrytis special-
ized on different host plants (Table 3) in order to validate that they
have indeed evolved under positive selection during the diversifi-
cation of the Botrytis genus. We designed primers in flanking re-
gions containing conserved sequenced across B. cinerea and S.
sclerotiorum to amplify and determine the complete sequence of
the nine genes in the 12 Botrytis species. Most of these genes could
be specifically amplified and sequenced in a number of species
ranging from 7 to 12 (B161, B24, B431, B57, B897, B967) but the
sequence of the B417 gene could be determined only in five of
the 12 selected species (Table 3). Analysis using the CODEML pro-
gram (Yang, 2007) of alignments containing sequences of all spe-
cies (the four initial species plus the additional Botrytis species)
detected sites under positive selection in all but the B24 gene
and LRTs were significant for all these eight genes (Table 4). To
have an independent sample, analysis of alignments containing
only sequences of the additional Botrytis species (without the four
initial species) further confirmed that seven of these genes (B161,
B417, B431, B57, B897, B907 and B967; Table 4) evolved under po-
sitive selection, with very similar parameter estimates. The ex-
tended alignments enabled detection of additional sites evolving
under positive selection in five of the genes (B161, B417, B431,
B897 and B907) (Table 4). For the B266 gene, footprints of positive
selection were identified only when the sequence of the closely
related Sclerotinia species was included in the alignment. These
results provide strong support for the validity of our approach to
detect genes under positive selection without a priori candidates,
showing that candidate genes under selection also show signature
of positive selection in other closely related pathogen species.
3.5. Experimental validation of the blind in silico approach for
detecting genes under positive selection
Among the 21 genes showing significant signal of positive selec-
tion, four genes (B897, B417, B431 and B161), for which the signa-
ture of positive selection was validated in other closely related
Botrytis species, were chosen for functional characterization by re-
verse genetics on the basis of their putative function, their elevated
x ratio and/or their level of expression during plant infection.
These candidate genes were inactivated in order to test whether
they were involved in aggressiveness of B. cinerea. Gene replace-
ment cassettes conferring resistance to hygromycin were con-
structed (see Section 2 and Table S1), and protoplasts of the
strain B05.10 were transformed with the K.O. cassettes. Protoplast
regeneration and further purification on selective media led to the
isolation of 10 B897, 4 B417, 11 B431, and 10 B161 and transfor-
mants. The expected gene replacement events were screened by
PCR. Using one primer located upstream of the 50region of the tar-
get gene and one primer located inside the hph (see Section 2), we
selected 3 B897D, 3 B417D, 3 B431D and 3 B161D null mutants.
On synthetic media, all mutants obtained for the four candidate
genes had similar growth and conidiation rates as the WT strain
(data not shown). Aggressiveness of the mutants on both hosts
was not significantly different from that of the WT (data not
shown). Therefore, the genes B161, B431 and B897 and B417 do
not appear to be essential for the infectious process of the B05.10
strain on the two standard host plants commonly tested in vitro
to check B. cinerea pathogenicity.
4. Discussion
In this study, we used a genome-wide approach based on the
comparative analysis of complete genome sequences of B. cinerea
and S. sclerotiorum and newly generated EST datasets for B. tulipae
and B. ficariarum to identify rapidly evolving orthologous genes
that might be involved in host range determination, adaptation
and coevolution with host plants in the Botrytis genus. This evolu-
tionary genomics approach provided a list of 21 genes evolving un-
der positive selection that thus represent new candidate genes in
Botrytis, which could be pathogenicity factors. So far, pathogenicity
genes in B. cinerea were identified either by a ‘‘candidate gene’’
strategy (Choquer et al., 2007), by random mutagenesis (Giesbert
et al., 2012) or by profiling fungal gene expression during host
plant infection (Gioti et al, 2006). This latter approach has greatly
Table 2
Likelihood ratio tests of codeml site-specific analyses.
ClusterM1a vs M2a M7 vs M8
2d d.f.
P-value2dd.f.
P-value
B152
B161
B194
B24
B248
B266
B387
B398
B402
B417
B431
B541
B549
B57
B605
B695
B814
B821
B897
B907
B967
67.187
7.576
7.782
6.495
12.129
4.892
7.759
2.922
14.445
0.000
9.817
3.436
7.469
6.347
3.951
16.391
7.441
5.774
4.566
7.679
5.299
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
<0.001
0.023
0.020
0.039
0.002
0.087
0.021
0.232
0.001
1.000
0.007
0.179
0.024
0.042
0.139
<0.001
0.024
0.056
0.102
0.022
0.071
66.813
10.264
7.795
6.642
25.538
6.358
7.883
6.018
16.243
8.463
14.929
5.861
10.055
7.582
6.606
23.007
7.699
8.185
5.813
8.599
10.868
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
<0.001
0.006
0.020
0.036
<0.001
0.042
0.019
0.049
<0.001
0.015
0.001
0.053
0.007
0.023
0.037
<0.001
0.021
0.017
0.055
0.014
0.004
992
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996

Page 7

enhanced our understanding of fungal phytopathogenicity of B.
cinerea and other plant pathogens. The analysis of gene expression
during infection has indeed led to the identification of a number of
fungal genes, that other methods have failed to uncover, related to
pathogenicity and adaptation to the environment in planta or in-
volved in the mechanisms underlying adaptation to plants and
control of host ranges (Gioti et al., 2006; Wise et al., 2007; Genin,
2010). However, genes involved in these functions are not neces-
sarily all overexpressed during infection. Interestingly, 20 of the
21 genes identified in this study as evolving under positive selec-
tion had not been identified as overexpressed during infection of
sunflower by B. cinerea (Amselem et al., 2011). Also, 19 of them
are conserved among filamentous fungi while two (B417 and
B907) appear specific to the Leotiomycetes (Amselem et al.,
2011). So far, only the B897 gene was known to act as a phytotoxin
in B. cinerea and other fungal plant pathogens (Frías et al., 2011).
The known role of the B897 protein in pathogenicity confirms that
the evolutionary approach we used is valuable for the identifica-
tion of fungal determinants of the host/pathogen interaction.
Among the 20 remaining genes, some have functions that are re-
lated to plant–fungus interactions.
4.1. Putative functions of genes under positive selection in Botrytis
The functional features of the 21 genes detected as having
evolved under positive selection were analyzed to gain additional
insights into the evolutionary pressures acting on Botrytis species.
For most of the genes (17 out of 21), putative functions can be pos-
tulated on the basis of their structural, biochemical, or physiologi-
cal characteristics (Tables S1–S3). Compared with the set of 642
orthologs analyzed, this list of 21 proteins is also enriched in pre-
dicted secreted proteins that are more likely involved in plant/
pathogen interactions and may contribute to the infection process
of their host plant by Botrytis species. These 21 candidate genes
have a wide variety of predicted functions and include genes
known to be involved in various aspects of the respiratory and en-
ergy metabolism, protein and RNA metabolism, signal transduction
or virulence. Several genes under positive selection have putative
functions that can be related to pathogenicity or recognition of
the pathogen by the host plant.
The B431 gene encodes a putative pectin methylesterase (PME),
which hydrolyzes pectin, the major component of plant cell wall,
andisinvolvedinmacerationandsoft-rottingofplanttissuebyfun-
gal pathogens. B. cinerea secretes numerous pectinolytic enzymes
(Amselem et al., 2011), including two PMEs and many polygalactu-
ronases, to promote cell wall degradation. The other PME, BcPME1
was shown to be involved in the infection process (Valette-Collet
et al., 2003). It may be expected that cell wall degrading enzymes
evolve under positive selection as they lie at the interface between
the host and the pathogen. For instance, the B. cinerea endopolyga-
lacturonaseBcPG1,knownasbothavirulencefactor(tenHaveetal.,
1998) and as an elicitor of defense responses in grape (Poinssot
et al., 2003), was recently shown to be under positive selection
(Cettul et al., 2008).
The B897 gene encodes a protein with a conserved domain
related to the fungal cerato-platanin phytotoxic proteins. Cerato-
plataninproteinsare involvedin the interactionwith the host-plant
and induce both cell necrosis and phytoalexin synthesis, which is
one of the first plant defense-related events (Pazzagli, 1999; Scala
et al., 2004). Frías et al. (2011) have shown that this protein (that
they named BcSpl1) is involved in virulence on tobacco leaves and
is able to induce necrosis. Interestingly, the two sites found here
to be under positive selection in the BcSpl1/B897 protein were situ-
atedpreciselyintheregion(residues59and99,Table4)withnecro-
sis-inducing activity on tobacco leaves.
Table 3
List of Botrytis and Sclerotinia species, typical host plant species, and DNA sequences determined in additional species used in this study for nine genes under positive selection.
Species
Collection
number
Host plant
speciesa
Host plant
familya
B161
B24
B266
B417
B431
B57
B897
B907
B967
S. sclerotiorum
1980
>400 plant species
Polyphagous
SS1G_06287.1
SS1G_04988.1
SS1G_10560.1
SS1G_09799.1
SS1G_10165.1
SS1G_13070.1
SS1G_10096.1
SS1G_10923.1
SS1G_08505.1
B. cinerea
T4
>235 plant
species
Polyphagouson eudicotyledons
BofuT4_
P134980.1
BofuT4_
P159210.1
BofuT4_
P115660.1
BofuT4_
P091190.1
BofuT4_
P089390.1
BofuT4
_P154400.1
BofuT4_
P011930.1
BofuT4_
P009160.1
BofuT4_
P001080.1
B. ficariarum
CBS17663
F. verna
Ranunculaceae
U
U
U
U
U
U
U
U
U
B. tulipae
Bt9901
Tulipa spp.
Liliaceae
U
U
U
U
U
U
U
U
U
Additional species
B. pelargonii
CBS497.50
Pelargonium spp.
Geraniaceae
U
U
na
U
U
U
U
U
U
B. fabae
MUCL98
Vicia spp., Pisum spp,
Lens spp., Phaseolus spp.
Fabacae
nab
U
U
na
U
U
U
U
na
B. aclada
MUCL 8415
Allium spp.
Alliaceae
U
U
U
na
U
U
U
U
U
B. paeoniae
MUCL16084
Paeoniae spp.
Paeoniaceae
U
U
ns
U
U
U
U
ns
U
B. elliptica
Be9714
Lilium spp.
Liliaceae
U
U
U
na
U
U
U
U
U
B. galanthina
MUCL435
Galanthus spp.
Amaryllidaceae
U
U
ns
U
na
U
U
ns
U
B. hyacinthi
MUCL442
Hyacinthus spp.
Hyacinthaceae
U
U
U
U
U
U
U
na
U
B. croci
MUCL436
Crocus spp.
Iridaceae
U
U
U
na
U
U
U
ns
U
B. narcissicola
MUCL2120
Narcissus spp.
Amaryllidaceae
U
U
U
ns
ns
U
U
U
U
B. polyblastis
MUCL21492
Narcissus spp.
Amaryllidaceae
U
ns
U
U
U
U
U
ns
U
B. convoluta
MUCL11595
Iris spp.
Iridaceae
U
ns
ns
U
U
na
U
U
U
B. porri
MUCL3234
Allium spp.
Alliaceae
U
na
U
na
ns
U
U
U
U
aHost plants species according to Staats et al. (2005).
bns, non specific amplification; na, no amplification.
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996
993

Page 8

The B161 gene encoding a putative oxydoreductase is a putative
ortholog to RED1, a gene involved in the T-Toxin synthesis in Coch-
liobolus heterostrophus, the agent of Southern Corn Leaf Blight
(Inderbitzin et al., 2010). The T-toxin polyketide is a determinant
of high virulence to maize carrying Texas male sterile cytoplasm.
As in C. heterostrophus, B161 gene is localized at a genomic locus
together with a polyketide synthase encoding gene and other
genes possibly involved in the synthesis of secondary metabolites.
Interestingly, B161 was identified as overexpressed during sun-
flower infection (Amselem et al., 2011).
Several genes (B387, B398, B605 and B967) were annotated as
plasma membrane vacuolar type H+-ATPases, which are proton
pumps playing a key role in the physiology of fungi. These vacuolar
ATPasescontrolessentialfunctionssuchasnutrientuptake,osmotic
balance, ion homeostasis, and stress tolerance (Portillo 2000) and
were shown to be essential for the growth of Saccharomyces cerevi-
siae in stressful environmental conditions, including alkaline condi-
tions (Finnigan et al., 2011). Genes putatively involved in the ability
to survive environmental stresses, such as those found in the host
plants, were also well represented. B248 and B821 were similar to
enzymes (6-phosphogluconate dehydrogenase and transaldolase,
respectively) involved in the pentose phosphate pathway, which
is critical for the ability of fungi to resist and adapt to oxidative
stress (Juhnke et al., 1996). B24and B541have putativeroles in pro-
tein folding or protein catabolism and may rapidly initiate protein
productionuponinfection.Forinstance,B24isapredictedcyclophi-
lin.AnotherB.cinereacyclophilinhaspreviouslybeenshowntoplay
a critical role in plant tissues colonization (Viaud et al., 2003).
Othergenes identifiedas under positiveselection are involved in
signal transduction pathways (B695, GTP-binding nuclear protein
GSP1/Ran) and in transcription (B402, ribosomal protein S8E;
B814, 40S ribosomal protein S7). Signatures of positive selection
found in these latter genes indicate that a plasticity in gene expres-
sion may also have an important role in adaptation of organisms to
environmental changes, along with variability in gene coding se-
quences, as reported in a number of studies on adaptive responses
of species to their environment (Roelofs et al., 2010). Furthermore,
recent studies have shown that stress-related genes are particularly
prone to tuned expression (Lopez-Maury et al., 2008).
4.2. Functional experiment
To investigate whether genes detected using a blind approach as
evolving under positive selection do have functions potentially
important in the pathogenesis of Botrytis species, we made reverse
genetics analysis of four of the detected genes (B431, B897, B161,
andB417).FourmutantsofB.cinereaweregenerated,inwhichthese
genes were inactive, and then assayed for defects in pathogenicity
on bean and tomato leaves. Overall, the null mutant generated for
the four genes were not significantly affected in virulence and no
particular phenotype was observed in infection assays. However
only two hosts (bean and tomato) were used, and in vitro conditions
maynotreflectnaturalconditions.ThismayalsosuggestthatB.cine-
rea has evolved a number of backup biological processes, such as
functionalredundancyandcompensatoryprocesses,inordertopro-
tectinfectionprocessfrombeingimpaired.Forinstance,theabsence
of significant virulence defect in the B431 mutantmay be due to the
functional redundancy of these enzymes, as two putative pectin
methylesterases were identified in the genome of B. cinerea (Amse-
lem et al., 2011). These genes under positive selection therefore
remain interesting candidates to be involved in host adaptation.
Alternatively, they may be involved in coevolution with their
current host, or in genetic conflicts, and may cause genetic incom-
patibilities between species as a consequence of their rapid evolu-
tion (Presgraves et al., 2003). Further experiments on the mutants
generated in this study and on gene expression should reveal the
functions involved.
4.3. Comparison with previous studies on other pathogens
It is remarkable that previous studies using the same approach
of scans of genes under positive selection on different pathogens
Table 4
Predicted positively selected sites identified in nine genes evolving under positive selection in the initial Botrytis and Sclerotinia species and in the additional Botrytis species.
B161 B24 B266B417B431B57B897B907B967
Gene length in B. cinerea
Length of alignment used for the initial analysis
Sites detected under positive selection in the initial species
1050
915
170 Aa
1116
483
20 K**
151 I*
1068
855
10 S*
375
345
102 A*
984
816
88 E*
263 E*
264 S**
269 Q*
930
594
559
34 A*
114 T*
414
384
58 K**
1596
324
496 F**
597
467
118 T*
Length of alignment used for the validation
Sites detected under positive selection in
All species
99011131056381591 4411260555
170 A**
288 S**
335 A*
(20 K*)b
10 S*
54 G*
102 A*
47 T*
88 E**
263 E*
264 S**
269 Q**
310 A*
21 S*
47 T*
88 E**
237 K*
264 S**
310 A*
88 E**
237 K*
264 S**
310 A*
34 A**
58 K**
99 G**
20 P*
262 T*
397 E**
435 A*496 F**
118 T*
Additional species only288 S**
(–)c
–c
(54 G*)107 R*
58 K**
99 G**
262 T**
397 E**
435 A*
496 F**
118 T*
Botrytis species 288 S**
(–)c
–c
54 G**
102 A*
107 R*
114 T*
58 K**
99 G**
262 T**
397 E**
118 T*
aBayes Empirical Bayes (BEB) analysis of sites under positive selection.
bSignificant sites detected by BEB analysis and associated with non-significant M7–M8 likelihood ratio tests are indicated within brackets.
c–, no site.
*P > 95%.
**P > 99%.
994
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996

Page 9

have detected genes with similar functions. Li et al. (2009) com-
pared three strains of the baker’s yeast S. cerevisiae, one of them
being the commonly used s288c lab strain extracted from a rotten
fig about 70 years ago, the pathogenic strain YJM789 from a patient
with pneumonia, and the wild strain RM11-1a, isolated from a
vineyard and used in labs since 1996. As outgroup they used the
genomes of Saccharomyces paradoxus and Saccharomyces mikatae.
The different yeast species and strains used correspond to different
life-styles and environments. In another study, Aguileta et al.
(2010) analyzed four species of Microbotryum, each specialized
pathogen of a Caryophyllaceae plant species. In all these cases,
even if the number and identity of the genes under positive selec-
tion are not strictly the same, the common functions annotated
provide an important clue as to which are the most relevant mech-
anisms and cellular components that are rapidly evolving and
maybe associated with infection or specialization. In the compari-
son of S. cerevisiae strains, of the 76 orthologs under positive selec-
tion six were related to cell wall, four to metal ion transport, three
were transmembrane proteins, and two were cellular bud mem-
brane; the rest were not annotated or corresponded to unknown
functions (Li et al., 2009). In the case of Microbotryum, of the 42
genes subject to positive selection, six were transmembrane pro-
teins involved in transport activity, four were related to protein
and RNA metabolism, two to protein folding, two to respiration
and energy metabolism, one was a secreted protein involved in
extracellular communication, and one was related to cell regula-
tion (Aguileta et al., 2010). Finally, in our study on four Botrytis/
Sclerotinia species, the 21 positively selected genes were annotated
as follows: six were related to protein and RNA metabolism, three
were transmembrane proteins with transporter activity, six partic-
ipate in respiration and energy metabolism, two were secreted
proteins possibly acting in extracellular communication, one in-
volved in cell signaling, and two in protein folding. As in the other
studies, the remaining orthologous clusters were either not anno-
tated or with unknown functions. The fraction of genes detected
as evolving under positive selection in the three studies (76/3300
[2.3%] in S. cerevisiae; 42/372 [11.2%] in Microbotryum; and 23/
642 [3.6%] in Botrytis sp.) differed, but the species sampling and
the type of data varied among studies (either cDNA libraries or
whole genome sequences), being thus not directly comparable.
Therefore, they should not be taken as estimates of the proportion
of positively selected genes across the genome.
In all cases, a large fraction of the annotated genes under posi-
tive selection are transmembrane proteins, putative secreted pro-
teins or proteins located in the cell wall that presumably
participate in transporter activities and establish communication
with the host (host recognition) cell and the external environment.
As Li and colleagues (2009) also pointed out, positive selection ap-
pears to respond to external, environmental conditions. In the con-
text of emerging fungal diseases, this is all the more relevant as
rapid adaptation to new hosts is generally observed. Also, functions
related to rapid cellular growth appear to evolve under positive
selection among these pathogen species, indicating that part of
the resources is allocated to growing invading tissues. Although
in smaller numbers, proteins involved in respiration under stress-
ful conditions, cell signaling and regulation as well as protein fold-
ing appear as being under positive selection in all the studied
cases. Overall, the same relevant functions have been found in
the present work and in the studies of Li and colleagues (2009)
and Aguileta et al. (2010), adding support to their involvement in
host adaptation, coevolution and/or genetic conflicts.
Acknowledgments
We are grateful to the INRA MIGALE bioinformatics platform
(http://migale.jouy.inra.fr) for providing help and computational
resources. We would also like to acknowledge Adeline Simon, Guil-
laume Morgant and Pascal Le Pêcheur (INRA BIOGER, Grignon) for
their help in gene annotation, transformation and pathogenicity
tests, respectively. This work was funded by the Grants ANR-06-
BLAN-0201 and ANR 07-BDIV-003, by a post-doctoral Grant from
the French Ile-de-France Region, and by the ‘Consortium National
de Recherche en Génomique’ for sequencing the cDNA libraries.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.meegid.2012.02.012.
References
Aguileta, G., Lengelle, J., Marthey, S., Chiapello, H., Rodolphe, F., Gendrault, A.,
Yockteng, R., Vercken, E., Devier, B., Fontaine, M.C., Wincker, P., Dossat, C.,
Cruaud, C., Couloux, A., Giraud, T., 2010. Finding candidate genes under positive
selection in non-model species: examples of
specialization in pathogens. Mol. Ecol. 19, 292–306.
Aguileta, G., Refrégier, G., Yockteng, R., Fournier, E., Giraud, T., 2009. Rapidly
evolving genes in pathogens: methods for detecting positive selection and
examples among fungi, bacteria, viruses and protists. Infect. Genet. Evol. 9, 656–
670.
Altschul, S.F., Madden, T.L., Schäffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J.,
1997. Gapped BLAST and PSI-BLAST: a new generation of protein database
search programs. Nucleic Acids Res. 25, 3389–3402.
Amselem, J., Cuomo, C.A., Van Kan, J.A.L., Viaud, M., Benito, E.P., Couloux, A., et al.,
2011. Genomic analysis of the necrotrophic fungal pathogens Sclerotinia
sclerotiorum and Botrytis cinerea. PLoS Genet. 7, e1002230.
Anderson, P.K., Cunningham, A.A., Patel, N.G., Morales, F.J., Epstein, P.R., Daszak, P.,
2004. Emerging infectious diseases of plants: pathogen pollution, climate
change and agrotechnology drivers. Trends Ecol. Evol. 19, 535–544.
Anisimova, M., Bielawski, J.P., Yang, Z., 2001. Accuracy and power of the likelihood
ratio test in detecting adaptive molecular evolution. Mol. Biol. Evol. 18, 1585–
1592.
Archie, E.A., Luikart, G., Ezenwa, V.O., 2009. Infecting epidemiology with genetics: a
new frontier in disease ecology. Trends Ecol. Evol. 24, 21–30.
Bendtsen, J.D., Nielsen, H., von Heijne, G., Brunak, S., 2004. Improved prediction of
signal peptides: SignalP 3.0. J. Mol. Biol. 340, 783–795.
Blaustein, A.R., Johnson, P.T.J., 2010. Conservation biology: when an infection turns
lethal. Nature 465, 881–882.
Bolton, M.D., Thomma, B.P.H.J., Nelson, B.D., 2006. Sclerotinia sclerotiorum (Lib.) de
Bary: biology and molecular traits of a cosmopolitan pathogen. Mol. Plant
Pathol. 7, 1–16.
Carroll, B.J., Sweigard, J.A., Valent, B., 1994. Improved vectors for selecting resistance
to hygromycin. Fungal Genet. Newsl. 41, 22.
Castresana, J., 2000. Selection of conserved blocks from multiple alignments for
their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552.
Cettul, E., Rekab, D., Locci, R., Firrao, G., 2008. Evolutionary analysis of
endopolygalacturonase-encoding genes of Botrytis cinerea. Mol. Plant Pathol.
9, 675–685.
Choquer, M., Fournier, E., Kunz, C., Levis, C., Pradier, J.-M., Simon, A., Viaud, M., 2007.
Botrytis cinerea virulence factors: new insights into a necrotrophic and
polyphageous pathogen. FEMS Microbiol. Lett. 277, 1–10.
Colot, H.V., Park, G., Turner, G.E., Ringelberg, C., Crew, C.M., Litvinkova, L., Weiss,
R.L., Borkovich, K.A., Dunlap, J.C., 2006. A high-throughput gene knockout
procedure for Neurospora reveals functions for multiple transcription factors.
Proc. Natl. Acad. Sci. USA 103, 10352–10357.
Conesa, A., Götz, S., García-Gómez, J.M., Terol, J., Talón, M., Robles, M., 2005.
Blast2GO: a universal tool for annotation, visualization and analysis in
functional genomics research. Bioinformatics 21, 3674–3676.
de Crecy, E., Jaronski, S., Lyons, B., Lyons, T.J., Keyhani, N.O., 2009. Directed evolution
of a filamentous fungus for thermotolerance. BMC Biotechnol. 9, 74.
Cuff, A.L., Sillitoe, I., Lewis, T., Clegg, A.B., Rentzsch, R., Furnham, N., Pellegrini-
Calace, M., Jones, D., Thornton, J., Orengo, C.A., 2011. Extending CATH:
increasing coverage of the protein structure universe and linking structure
with function. Nucleic Acids Res. 39, D420–D426.
Desprez-Loustau, M.-L., Robin, C., Buée, M., Courtecuisse, R., Garbaye, J., Suffert, F.,
Sache, I., Rizzo, D.M., 2007. The fungal dimension of biological invasions. Trends
Ecol. Evol. 22, 472–480.
Espino, J.J., Gutiérrez-Sánchez, G., Brito, N., Shah, P., Orlando, R., González, C., 2010.
The Botrytis cinerea early secretome. Proteomics 10, 3020–3034.
Ewing, B., Green, P., 1998. Base-calling of automated sequencer traces using Phred.
II. Error probabilities. Genome Res. 8, 186–194.
Ewing, B., Hillier, L., Wendl, M.C., Green, P., 1998. Base-calling of automated
sequencer traces using Phred. I. Accuracy assessment. Genome Res. 8, 175–185.
Finn, R.D., Mistry, J., Tate, J., Coggill, P., Heger, A., Pollington, J.E., Gavin, O.L.,
Gunasekaran, P., Ceric, G., Forslund, K., Holm, L., Sonnhammer, E.L.L., Eddy, S.R.,
Bateman, A., 2010. The Pfam protein families database. Nucleic Acids Res. 38,
D211–D222.
genesinvolved in host
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996
995

Page 10

Finnigan, G.C., Ryan, M., Stevens, T.H., 2011. A genome-wide enhancer screen
implicatessphingolipid composition
Saccharomyces cerevisiae. Genetics 187, 771–783.
Fournier, E., Giraud, T., Albertini, C., Brygoo, Y., 2005. Partition of the Botrytis cinerea
complex in France using multiple gene genealogies. Mycologia 97, 1251–1267.
Frías, M., González, C., Brito, N., 2011. BcSpl1, a cerato-platanin family protein,
contributes to Botrytis cinerea virulence and elicits the hypersensitive response
in the host. The New Phytol. 192, 483–495.
Fukunishi, Y., Hayashizaki, Y., 2001. Amino acid translation program for full-length
cDNA sequences with frameshift errors. Physiol. Genomics 5, 81–87.
Genin, S., 2010. Molecular traits controlling host range and adaptation to plants in
Ralstonia solanacearum. The New Phytol. 187, 920–928.
Giesbert, S., Schumacher, J., Kupas, V., Espino, J., Segmüller, N., Haeuser-Hahn, I.,
Schreier, P.H., Tudzynski, P., 2012. Identification of pathogenesis-associated
genes by T-DNA-mediated insertional mutagenesis in Botrytis cinerea: a type 2A
phosphoprotein phosphatase and an SPT3 transcription factor have significant
impact on virulence. Mol. Plant Microbe Interact. 25, 481–495.
Gioti, A., Simon, A., Pêcheur, P.Le., Giraud, C., Pradier, J.M., Viaud, M., Levis, C., 2006.
Expression profiling of Botrytis cinerea genes identifies three patterns of up-
regulation in planta and an FKBP12 protein affecting pathogenicity. J. Mol. Biol.
358, 372–386.
Gough, J., 2002. SUPERFAMILY: HMMs representing all proteins of known structure.
SCOP sequence searches, alignments and genome assignments. Nucleic Acids
Res. 30, 268–272.
ten Have, A., Mulder, W., Visser, J., van Kan, J.A., 1998. The endopolygalacturonase
gene Bcpg1 is required for full virulence of Botrytis cinerea. Mol. Plant Microbe
Interact. 11, 1009–1016.
Huang, X., Madan, A., 1999. CAP3: a DNA sequence assembly program. Genome Res.
9, 868–877.
Inderbitzin, P., Asvarak, T., Turgeon, B.G., 2010. Six new genes required for
production of T-toxin, a polyketide determinant of high virulence of
Cochliobolus heterostrophus to maize. Mol. Plant Microbe Interact. 23, 458–472.
Iseli, C., Jongeneel, C.V., Bucher, P., 1999. ESTScan: a program for detecting,
evaluating, and reconstructing potential coding regions in EST sequences. Proc.
Int. Conf. Intell. Syst. Mol. Biol., 138–148.
Juhnke, H., Krems, B., Kötter, P., Entian, K.D., 1996. Mutants that show increased
sensitivity to hydrogen peroxide reveal an important role for the pentose
phosphate pathway in protection of yeast against oxidative stress. Mol. Gen.
Genet. 252, 456–464.
van Kan, J.A.L., 2006. Licensed to kill: the lifestyle of a necrotrophic plant pathogen.
Trends Plant Sci. 11, 247–253.
Krogh, A., Larsson, B., von Heijne, G., Sonnhammer, E.L., 2001. Predicting
transmembrane protein topology with a hidden Markov model: application to
complete genomes. J. Mol. Biol. 305, 567–580.
Lannuccelli, E., 2005. SURF (SeqUence Repository and Feature detection) package.
Lebarbenchon, C., Brown, S.P., Poulin, R., Gauthier-Clerc, M., Thomas, F., 2008.
Evolution of pathogens in a man-made world. Mol. Ecol. 17, 475–484.
Levis, C., Fortini, D., Brygoo, Y., 1997. Transformation of Botrytis cinerea with the
nitrate reductase gene (niaD) shows a high frequency of homologous
recombination. Curr. Genet. 32, 157–162.
Li, Y.-D., Liang, H., Gu, Z., Lin, Z., Guan, W., Zhou, L., Li, Y.-Q., Li, W.-H., 2009.
Detecting positive selection in the budding yeast genome. J. Evol. Biol. 22,
2430–2437.
Li, L., Stoeckert, C.J., Roos, D.S., 2003. OrthoMCL: identification of ortholog groups for
eukaryotic genomes. Genome Res. 13, 2178–2189.
Lopez-Maury, L., Marguerat, S., Bahler, J., 2008. Tuning gene expression to changing
environments: from rapid responses to evolutionary adaptation. Nat. Rev.
Genet. 9, 583–593.
Marthey, S., Aguileta, G., Rodolphe, F., Gendrault, A., Giraud, T., Fournier, E., Lopez-
Villavicencio, M., Gautier, A., Lebrun, M.-H., Chiapello, H., 2008. FUNYBASE: a
FUNgal phYlogenomic dataBASE. BMC Bioinformatics 9, 456.
Matute, D.R., Quesada-Ocampo, L.M., Rauscher, J.T., McEwen, J.G., 2008. Evidence for
positive selection in putative virulence factors within the Paracoccidioides
in vacuolarATPase function in
brasiliensis species complex. In: Taylor, J.W. (Ed.), PLoS Negl. Trop. Dis., vol. 2, p.
e296.
Mulder, N.J., Apweiler, R., Attwood, T.K., Bairoch, A., Bateman, A., Binns, D., et al.,
2007. New developments in the InterPro database. Nucleic Acids Res. 35, D224–
D228.
Notredame, C., Higgins, D.G., Heringa, J., 2000. T-Coffee: a novel method for fast and
accurate multiple sequence alignment. J. Mol. Biol. 302, 205–217.
Pazzagli, L., 1999. Purification, characterization, and amino acid sequence of Cerato-
platanin, a new phytotoxic protein from Ceratocystis fimbriata f. sp. platani. J.
Biol. Chem. 274, 24959–24964.
Poinssot, B., Vandelle, E., Bentéjac, M., Adrian, M., Levis, C., Brygoo, Y., Garin, J.,
Sicilia, F., Coutos-Thévenot, P., Pugin, A., 2003. The endopolygalacturonase 1
from Botrytis cinerea activates grapevine defense reactions unrelated to its
enzymatic activity. Mol. Plant Microbe Interact. 16, 553–564.
Portillo, F., 2000. Regulation of plasma membrane H+-ATPase in fungi and plants.
Biochim. Biophys. Acta Biomembr. 1469, 31–42.
Presgraves, D.C., Balagopalan, L., Abmayr, S.M., Orr, H.A., 2003. Adaptive evolution
drives divergence of a hybrid inviability gene between two species of
Drosophila. Nature 423, 715–719.
Roelofs, D., Morgan, J., Stürzenbaum, S., 2010. The significance of genome-wide
transcriptional regulation in the evolution of stress tolerance. Evol. Ecol. 24,
527–539.
Sacristán, S., García-Arenal, F., 2008. The evolution of virulence and pathogenicity in
plant pathogen populations. Mol. Plant Pathol. 9, 369–384.
Scala, A., Pazzagli, L., Comparini, C., Santini, A., Tegli, S., Cappugi, G., 2004. Cerato-
platanin, an early-produced protein by Ceratocystis fimbriata f. sp. platani, elicits
phytoalexin synthesis in host and non-host plants. J. Plant Pathol. 86, 27–33.
Shah, P., Atwood, J.A., Orlando, R., El Mubarek, H., Podila, G.K., Davis, M.R., 2009.
Comparative proteomic analysis of Botrytis cinerea secretome. J. Proteome Res.
8, 1123–1130.
Staats, M., van Baarlen, P., Van Kan, J.A.L., 2005. Molecular phylogeny of the plant
pathogenic genus Botrytis and the evolution of host specificity. Mol. Biol. Evol.
22, 333–346.
The UniProt Consortium, 2007. The Universal Protein Resource (UniProt). Nucleic
Acids Res. 35, D193–D197.
Thomas, P.D., Campbell, M.J., Kejariwal, A., Mi, H., Karlak, B., Daverman, R., Diemer,
K., Muruganujan, A., Narechania, A., 2003. PANTHER: a library of protein
families and subfamilies indexed by function. Genome Res. 13, 2129–2141.
Valette-Collet, O., Cimerman, A., Reignault, P., Levis, C., Boccara, M., 2003.
Disruption of Botrytis cinerea pectin methylesterase gene Bcpme1 reduces
virulence on several host plants. Mol. Plant Microbe Interact. 16, 360–367.
Viaud, M., Brunet-Simon, A., Brygoo, Y., Pradier, J.-M., Levis, C., 2003. Cyclophilin A
and calcineurin functions investigated by gene inactivation, cyclosporin A
inhibition and cDNA arrays approaches in the phytopathogenic fungus Botrytis
cinerea. Mol. Microbiol. 50, 1451–1465.
Walker, A.-S., Gautier, A., Confais, J., Martinho, D., Viaud, M., Pêcheur, P.Le., Dupont,
J., Fournier, E., 2011. Botrytis pseudocinerea, a new cryptic species causing grey
mould in French vineyards in sympatry with Botrytis cinerea. Phytopathology
101, 1433–1445.
Wasmuth, J.D., Blaxter, M.L., 2004. Prot4EST: translating expressed sequence tags
from neglected genomes. BMC Bioinformatics 5, 187.
Williamson, B., Tudzynski, B., Tudzynski, P., van Kan, J.A.L., 2007. Botrytis cinerea:
the cause of grey mould disease. Mol. Plant Pathol. 8, 561–580.
Wise, R.P., Moscou, M.J., Bogdanove, A.J., Whitham, S.A., 2007. Transcript profiling in
host–pathogen interactions. Annu. Rev. Phytopathol. 45, 329–369.
Yang, Z., 2007. PAML 4: phylogenetic analysis by maximum likelihood. Mol. Biol.
Evol. 24, 1586–1591.
Yang, Z., Bielawski, J.P., 2000. Statistical methods for detecting molecular
adaptation. Trends Ecol. Evol. 15, 496–503.
Yu, J.-H., Hamari, Z., Han, K.-H., Seo, J.-A., Reyes-Domínguez, Y., Scazzocchio, C.,
2004. Double-joint PCR: a PCR-based molecular tool for gene manipulations in
filamentous fungi. Fungal Genet. Biol. 41, 973–981.
996
G. Aguileta et al./Infection, Genetics and Evolution 12 (2012) 987–996