Content uploaded by Jean-michel Drezen
Author content
All content in this area was uploaded by Jean-michel Drezen on Aug 31, 2020
Content may be subject to copyright.
Molecular Ecology. 2020;00:1–18. wileyonlinelibrary.com/journal/mec
|
1© 2020 John Wiley & Sons Ltd
Received: 20 May 2020
|
Revised: 17 July 2020
|
Accepted: 20 July 2020
DOI: 10.1111/mec.15567
ORIGINAL ARTICLE
Quantitative trait loci involved in the reproductive success of a
parasitoid wasp
Romain Benoist1 | Claire Capdevielle-Dulac1 | Célina Chantre1 | Rémi Jeannette1 |
Paul-André Calatayud1,2 | Jean-Michel Drezen3 | Stéphane Dupas1 | Arnaud Le Rouzic1 |
Bruno Le Ru1 | Laurence Moreau4 | Erwin Van Dijk5 | Laure Kaiser1 | Florence Mougel1
Funding information This work was su pported by the French Na tional Resear ch Agency (AB C Papogen proje ct ANR-12-ADAP-0001 and pro ject Cotebio AN R-17-CE32-0015), and by
the authors’ op erating gran ts from IRD, CNR S and icipe. R. Be noist is funde d by the Ecole docto rale 227 MNHN-UPMC Sc iences de la Natu re et de l’Homme : évolution et éco logie. The
funde rs had no role in st udy design, dat a collection a nd analysis, decision to publi sh or preparati on of the manuscr ipt.
1Université Paris-Saclay, CNRS, IRD, UMR
Évolution, Génomes, Comportement et
Écologie, Gif-sur-Yvette, France
2icipe, International Center of Insect
Physiology and Ecology, Nairobi, Kenya
3Institut de Recherche sur la Biologie de
l'Insecte, UMR CNRS 7261, Université Tours,
Tours, France
4Université Paris-Saclay, INR AE, CNRS,
AgroParisTech, UMR GQE - Le Moulon, Gif-
sur-Yvette, France
5Université Paris-Saclay, CNRS, CEA , UMR
Institut de Biologie Intégrative de la Cellule,
Gif-sur-Yvette, France
Correspondence
Florence Mougel, Laboratoire Évolution,
Génomes, Comportement et Écologie, UMR
CNRS 9191, IRD 247, Université Paris-
Saclay, 91198 Gif-sur-Yvette, France.
Email: Florence.mougel-imbert@egce.
cnrs-gif.fr
Funding information
Agence Nationale de la Recherche, Grant/
Award Number: ANR-12-ADAP-0001 and
ANR-17-CE32-0015; Ecole doctorale 227
MNHN-UPMC Sciences de la Nature et de
l’Homme: évolution et écologie
Abstract
Dissecting the genetic basis of intraspecific variations in life history traits is essential
to understand their evolution, notably for potential biocontrol agents. Such varia-
tions are observed in the endoparasitoid Cotesia typhae (Hymenoptera: Braconidae),
specialized on the pest Sesamia nonagrioides (Lepidoptera: Noctuidae). Previously,
we identified two strains of C. typhae that differed significantly for life history traits
on an allopatric host population. To investigate the genetic basis underlying these
phenotypic differences, we used a quantitative trait locus (QTL) approach based on
restriction site-associated DNA markers. The characteristic of C. typhae reproduc-
tion allowed us generating sisters sharing almost the same genetic content, named
clonal sibship. Crosses between individuals from the two strains were performed to
generate F2 and F8 recombinant CSS. The genotypes of 181 clonal sibships were de-
termined as well as the phenotypes of the corresponding 4,000 females. Informative
markers were then used to build a high-quality genetic map. These 465 markers
spanned a total length of 1,300 cM and were organized in 10 linkage groups which
corresponded to the number of C. typhae chromosomes. Three QTLs were detected
for parasitism success and two for offspring number, while none were identified for
sex ratio. The QTLs explained, respectively, 27.7% and 24.5% of the phenotypic vari-
ation observed. The gene content of the genomic intervals was investigated based
on the genome of C. congregata and revealed 67 interesting candidates, as potentially
involved in the studied traits, including components of the venom and of the symbi-
otic virus (bracovirus) shown to be necessary for parasitism success in related wasps.
KEYWORDS
linkage map, offspring number, parasitism success, parasitoid, polydnavirus, Quantitative trait
loci, venom
1 | INTRODUCTION
Biotic interactions exert a strong selection pressure on living or-
ganisms and constitute a major evolutionary strength. Among
them, intimate relationships such as host–parasite associations
may lead to specific evolutionary patterns, namely coevolution
(Thompson, 2009; Woolhouse, Webster, Domingo, Charlesworth, &
Levin, 2002). The process of coevolution corresponds to adaptive
2
|
BENOIST ET al.
changes in the two partners under reciprocal influence. It requires
genetic variation within the parasite and host populations influenc-
ing the outcome of the interaction. Deciphering the components of
this genetic variation, that is the number of genes involved, the inter-
actions between them, their potential pleiotropic effects and their
impact on the organism's fitness, is of prime importance to elucidate
the genomic bases of coadaptation.
Among parasites, insect parasitoids represent an interesting case
study of environmental adaptation both to improve our knowledge
of biotic interactions and in the applied perspective of biocontrol
(Wang, Liu, Shi, Huang, & Chen, 2019). These species spend their
larval stage as parasites and live freely at the adult stage. They be-
have in an intermediate way between parasites and predators be-
cause they need to kill their host to develop. The larval environment
is thus mainly biotic, and parasitoid effective reproduction relies on
fine adaptation to their host (Godfray, 1994).
The host adaptation encompasses the ability to detect and par-
asitize efficiently the host. Host detection is based on localizing the
host as well as recognizing it s su itability. Once a suitable host is found,
parasitoids have to optimize their oviposition behaviour to maximize
their fitness. This is illustrated by the occurrence of different ovipo-
sition strategies (Godfray, 1994; Waage & Greathead, 1986). Usually,
parasitoids are classified as solitar y or gregarious depending upon the
offspring number produced per host, with one for solitary (not neces-
sarily implying the injection of only one egg) and several for gregari-
ous. In haplodiploid Hymenoptera, oviposition strategy also includes
sex ratio of the progeny, which corresponds to the proportion of fer-
tilized eggs that a female lays (Heimpel & de Boer, 2008). Both the
number of eggs injected (and therefore the offspring number) and
the sex ratio are under adaptive constraints. For example, the hosts’
carrying capacity limits the number of parasitoid larvae that could de-
velop and thus impacts the behaviour a species will adopt in the distri-
bution of eggs among hosts encountered by a female (Godfray, 1994;
Le Masurier, 1987, 1991). Environmental factors are also perceived
and taken into account by parasitoid females to adjust their clutch size
and in some cases their sex ratio (Charnov & Skinner, 1985; Waage
& Ming, 1984). Those two traits (offspring number and sex ratio) are
therefore genetically and environmentally determined.
Following oviposition, parasitoid larval development depends on
the parasitoid ability to inhibit host immune defences and alter its
growth to the parasitoid benefit. This is especially important for en-
doparasitoid species which develop in host haemocoele, where host
haemocytes are recruited to encapsulate parasitoid eggs or larvae
(Quicke, 2014). The virulence arsenal developed by parasitoids to en-
sure the development of their progeny, hereafter defined as parasit-
ism success, is quite phenomenal and combines factors derived from
maternal and embryonic origins. Depending on species, the maternal
factors encompass venom, ovarian proteins, polydnavirus and vi-
ru s- like par ticle s and are in jec ted alo ng wi th eg gs in the host (A sgar i &
Ri vers , 2011; Herni ou et al., 2013 ; Pen nacc hio & Stran d , 2006; Pich on
et al., 201 5; Stran d & Burke, 2015 ). Th e emb r yonic facto rs encomp ass
teratocytes, which are cells derived from the membrane surrounding
the parasitoid embryo and parasitoid larvae itself (Strand, 2014). The
most studied factors are probably venom and polydnavirus. Venom
is mainly devoted to host metabolism manipulation, ensuring para-
sitoid development (Mrinalini & Werren, 2016). Within hymenoptera,
the venom gland is a conserved organ but venom composition seems
to be highly variable depending on species (Moreau & Asgari, 2015).
Polydnaviruses are double-stranded DNA viruses associated with
few groups of hymenopteran parasitoids (Webb, 1998). The polyd-
navirus associated with Braconidae is called Bracoviruses, whereas
those associated with Ichneumonidae are called Ichnoviruses. They
derive from the integration viruses into the genome of ancestral
parasitoids and are now vertically transmitted (Bézier, Annaheim,
et al., 2009; Bézier, Herbinière, Lanzrein, & Drezen, 2009; Drezen
et al., 2017; Volkoff et al., 2010). In species harbouring polydnavi-
ruses, viral particles contain viral DNA circles bearing virulence genes
which are expressed in host tissues and involved in the inactivation
of the host immune system and alteration of host growth (Beckage
& Drezen, 2012; Beckage & Gelman, 2004; Edson, Vinson, Stoltz, &
Summers, 1981; Marti, 2003; Wyler & Lanzrein, 2003).
Besides the interspecific variation in virulence arsenal and more
generally in reproductive success traits, intraspecific divergence in
life history traits is also obser ved in parasitoid species (Chassain &
Bouletreau, 1987; Dubuffet et al., 2009; Henter, 1995; Kaiser, Couty,
& Perez-Maluf, 2009; Legner, 1987; Orzack & Gladstone, 1994).
This polymorphism underlies the ability of parasitoids to evolve in
response to host selection. Identifying the polymorphic genes in-
volved in parasitoid success is thus of prime importance to identify
key components and understand the physiological and behavioural
basis of host adaptation.
The identification of genes underlying such complex traits
is hampered by their multifactorial determinism and high plas-
ticity. Quantitative trait loci (QTL) approaches are optimal to over-
come such complex situations (Broman & Sen, 2009; Lander &
Botstein, 1989; Mackay, Stone, & Ayroles, 2009). These approaches
are also particularly relevant in Hymenopteran species because of
their haplodiploid sex determination system. Indeed, in these spe-
cies males are haploid and produce spermatozoids all bearing iden-
tical genetic content. So, when a male is crossed with a completely
homozygous female, all daughters produced are genetically identi-
cal. This allows us to repeat phenotypic measures on females with
the same genotype, reducing the impact of environmental variation
(Dupas, Frey, & Carton, 1998; Pannebakker, Watt, Knott, West, &
Shuker, 2011; Velthuis, Yang, Van Opijnen, & Werren, 2005). To our
knowledge, in parasitoids, the QTL approach has been performed
only once on sex ratio and offspring number in Nasonia vitripennis
(Pannebakker et al., 2011) and never performed to study parasitism
success. In this paper, we developed a QTL approach to determine
the genetic basis of these traits in a parasitoid species: Cotesia typhae
(Fernández-Triana; Hymenoptera, Braconidae).
Cotesia typhae is an African gregarious endoparasitoid, parasitiz-
ing exclusively larvae of the crop pest Sesamia nonagrioides (Lefebvre;
Lepidoptera, Noctuidae). It was formerly undistinguished from the
generalist species Cotesia sesamiae composed of several populations
with different host ranges, but was recently recognized as a distinct
|
3
BENOIS T ET al.
species (Kaiser, Dupas, et al., 2017; Kaiser, Fernandez-Triana, et al.,
2017; Kaiser et al.., 2015). The ability of C. typhae to parasitize a pop-
ulation of S. nonagrioides, an invasive species in France, was previ-
ously studied in a biological control perspective (Benoist et al., 2017).
This study highlighted differences between two strains of C. typhae
named Kobodo and Makindu specifically in their reproductive suc-
cess on this allopatric host population. Indeed, Kobodo females had a
higher rate of parasitism success and produced more offsprings than
Makindu females under laboratory conditions. The marked differ-
ence in the reproductive success of the two strains sets the stage for
studying the genetic basis of these traits. Furthermore, an annotated
genome of the close relative Cotesia congregata recently became
available, which can be used as a reference for the identification of
genes within QTL (Gauthier et al., 2020).
The aim of this work was to decipher the genetic architecture
involved in the variation of the reproductive success of C. typhae. To
do this, we developed a QTL approach based on restriction site-as-
sociated DNA markers (RAD-tags) to detect genomic regions associ-
ated with parasitism success, offspring number and sex ratio traits.
The gene content in the identified QTL region was investigated
based on the annotated genome of C. congregata.
2 | MATERIALS AND METHODS
2.1 | Biological material
The Kobodo and Makindu C. typhae parasitoid strains were obtained
from adults emerged from naturally parasitized S. nonagrioides cat-
erpillars collected in the field at two localities in Kenya: Kobodo
(0.679S, 34.412E; West Kenya; 3 caterpillars collected in 2013)
and Makindu (2.278S, 37.825E; South-East Kenya; 10 caterpillars
collected in 2010–2011). These strains were reared separately at
the International Centre of Insect Physiology and Ecology (ICIPE,
Nairobi, Kenya). Isofemale lines were produced and maintained from
these rearings from 2015 at the Evolution, Génome, Comportement
et Ecologie laboratory (EGCE, Gif-sur-Yvette, France), where cross
experiments and phenotyping were performed.
Two host strains of S. nonagrioides were used: a Kenyan strain
initiated from caterpillars collected at Makindu and a French strain
initiated from individuals collected in maize fields (Longage-Berat
area in Haute-Garonne district, 43.368N, 1.192E and within a 10 km
distance) which was renewed yearly. The French strain was used for
phenotyping, whereas the Kenyan strain was used for C. typhae rear-
ing to prevent any adaptation of the parasitoid to the French host.
The rearing protocol of C. typhae and S. nonagrioides is detailed
in Benoist et al. (2020).
2.2 | Genetic cross-design
Cotesia typhae is a haplodiploid species. Females are produced from
fertilized eggs and are diploid, whereas males are produced from
unfertilized eggs and are haploid. We combined two cross-schemes
to produce F2 and F8 recombinant individuals (Figure 1). The pro-
duction of F2 recombinant individuals was used to build a genetic
map. This data set was completed by F8 recombinant individuals to
increase the number of recombination events for QTL detection.
Three generations of sib-mating were first realized for both Kobodo
and Makindu parental lines. These sib-mating crosses following nu-
merous generations (around 50 for Kobodo and 20 for Makindu) of
rearing in small populations led to highly inbred parental lines. Virgin
males and females from the parental strains were then crossed to
generate F1 individuals. Part of the F1 females were isolated and kept
virgin (Figure 1a). All other F1 individuals were mixed in a single pop-
ulation, which was maintained to reach F8 generation (Figure 1b). At
each gene rat ion , all adults were stor ed in the same cage to allow ran-
dom mating and randomly chosen females were used to produce the
next generation to avoid overlapping between generations. Several
F7 females were isolated and kept virgin. Virgin females from F1 or
F7 generation were allowed to oviposit to produce, respectively, F2
and F8 recombinant males. For F2 males, each recombinant male
was backcrossed with a single female from Kobodo or Makindu
parental strains. For F8 males, only the Makindu strain was used,
because a dominance of Kobodo alleles on these traits was suscep-
tible to mask phenotypic variation as indicated by a previous study
(Benoist et al., 2017). Since no meiosis occurs in males and parental
strains are highly inbred, the offsprings from one recombinant male
and one female from a parental line contain recombinant females
with almost the same recombinant genotype at all loci (they were
considered as genetically identical). Each female progeny was called
a clonal sibship—CSS in the following text (Pannebakker et al., 2011).
In total, 181 CSSs were produced: 47 from F2 Kobodo backcross,
45 from F2 Makindu backcross and 89 from F8 Makindu backcross.
This experimental design allows replicated phenotyping for each re-
combinant genotype and thus the measurement of proportion traits
on one CSS as well as parasitism success (see below), which requires
several individuals to be assessed for accurate estimation, because
of its variability.
2.3 | Phenotyping
Between 15 and 20 females were tested for each CSS. In order to
allow mating, sibling females were left at least one day with their
brothers in the same cage before the experiments (mating was not
controlled). The host caterpillars were placed individually under a
2-cm-diameter plastic top with one female until the ovipositor inser-
tion was observed. A female that refused to oviposit within 3 min
was discarded. Very few females were removed as host acceptance is
very high (>90%) for the two parasitoid strains (Benoist et al., 2020).
After exposure, parasitized host caterpillars were kept in rearing
conditions until observation of either the formation of parasitoid
cocoon mass (following the emergence of the parasitoid larvae from
the host), the death of the host without parasitoid emergence or the
formation of host pupa. After parasitoid emergence, each cocoon
4
|
BENOIST ET al.
mass was placed in a tube to obtain the adults. For each CSS, the par-
asitism success was calculated as the proportion of parasitized hosts
from which parasitoid larvae emerged. The two other traits were
estimated, taking into account successful parasitism only. The mean
offspring number was calculated as the mean number of parasitoid
larvae that emerged from the host. The sex ratio for each CSS was
determined by counting the number of adult females, on one side,
and the number of adult males, on the other, emerging from each
successful parasitism. To avoid bias, progenies of unmated females
identified from their only male content were not taken into account
FIGURE 1 Cross-schemes used to generate clonal sibship, adapted from Pannebakker et al. (2011): (a): F2 cross-scheme; (b) F8 cross-
scheme. C. typhae is a haplodiploid species. A haploid genome set is represented by 5 squares, with one set for males and two sets for
females. Kobodo and Makindu genetic contributions are represented in black and white, respectively. Inbred parental strains are considered
homozygous. Kobodo and Makindu individuals from parental strains were crossed to generate F1 individuals. A subset of F1 hybrid females
were kept virgin to produce F2 recombinant males (a). All other F1 individuals were mixed in a single population which was maintained
until F8 generation (b). F2 (a) and F8 (b) recombinant males were then backcrossed with females from the parental strains to produce clonal
sibship (CSS) females (i.e., females considered as having identical genotypes). Each rectangle inside one stack represents a female in a
CSS. F8 males were backcrossed only with Makindu females. In total, 181 CSS were produced: 47 from F2 Kobodo backcross, 45 from F2
Makindu backcross and 89 from F8 Makindu backcross. For each CSS, between 15 and 20 females were used to measure the phenotypic
traits of the CSS and all females were pooled for RAD-sequencing
|
5
BENOIS T ET al.
because the offspring number varies significantly between mated
and unmated females (Benoist et al., 2017). All phenotypic data are
presented in Table S1. The broad-sense heritability in the CSS popu-
lations (H2) could not be estimated for parasitism success due to the
lack of repetitions: only one measure is obtained for each CSS based
on the number of success and failure of the parasitism. By contrast,
each success led to the estimation of an offspring number and of a
number of males and females. For these two traits, a generalized
linear model (GLM) was built, taking into account CSS and backcross
type to explain the trait variation. We used the “quasi-Poisson” error
family for the offspring number and the “quasi-binomial” error fam-
ily for the sex ratio. Genetic variance was estimated from the sum of
squares associated with the CSS and it was divided by the total sum
of squares to approach H2.
2.4 | RAD-sequencing
A RAD-seq approach was performed in order to obtain genetic
markers widely distributed in the genome. For each CSS, DNA was
extracted from a pool of females using the NucleoSpin Tissue kit
(Macherey-Nagel) and a RNase treatment was performed (Roche).
All CSS DNA samples were digested using PstI restriction endonu-
clease. The expected number of restriction sites was approximately
18,500 based on C. sesamiae kitale genome analysis (Gauthier et al.,
2020). A P1 adapter containing an Illumina adapter sequence, a
10-bp barcode (to identify CSS) and a sticky-end extremity, corre-
sponding to the PstI site, was ligated to the PstI-digested fragments.
F2 and F8 samples were pooled separately, and adapter-ligated
fragments were randomly sheared to obtain an average size of
600 bp. The two Illumina libraries were prepared by ligating a P2
adapter with a divergent end to DNA fragment s to ensure that only
fragments with P1 and P2 adapters would be fully amplified. These
libraries were then amplified by PCR with P1 and P2 primers and
paired-end sequenced on an Illumina NextSeq 500 instrument. The
read length was 75 bp (including 10 bp of the barcode and 6 bp of
the restriction site).
2.5 | Identification of RAD locus and genotypes
with Stacks
RAD-seq reads were trimmed to remove adaptors with cutadapt
v1.9.1 (Martin, 2011). RAD loci and the associated genotypes were
determined using the s tacks v1.48 software package (Catchen,
Hohenlohe, Bassham, Amores, & Cresko, 2013; Rochette &
Catchen, 2017). The trimmed reads were de-multiplexed, and the
barcodes were removed from reads using process radtags, discard-
ing reads with an uncalled base and/or low-quality score. F2 and F8
reads were treated separately. For each sample, reads were grouped
in “stacks” to build loci using two approaches: de novo (ustacks)
and reference-based (pstacks, in this case a reference genome is
used to build loci). This double approach maximized the number of
detected markers as the available reference genome is a draft ver-
sion and is from a different species. Prior to de novo analyses, PCR
duplicates were identified based on sequence identity and removed
using home-made software. Preliminary tests were performed to
optimize Stacks parameters: minimum depth coverage (-m) between
3 and 5, maximum distance between stacks (-M) between 2 and 3
and the number of mismatches allowed to build the catalog (-n) be-
tween 0 and 2 were tested. The selected parameters (m = 3, M = 2,
n = 2) were those that maximized the number of loci and minimized
the variance between samples. For reference-based analyses, read
pairs were mapped to the closely related C. sesamiae kitale genome
(Gauthier et al., 2020) using bwa v0.7.17 (Li et al., 2009), and PCR
duplicates sharing strict identical coordinates were removed with
samtool s v1.9. The error rate upper bound was fixed at 0.01 for F2
(used to define the locus catalog, see below) and at 0.1 for F8. The
deleveraging algorithm was used for de novo analyses. For all other
parameters, default settings were used. A de novo and a reference-
based catalogue were generated from all F2 stacks (choosing popu-
lation option) using the cstacks program with default parameters.
Both F2 and F8 stacks were then matched to these catalogs to infer
genotypes using the sstacks program with default parameters. We
thus obtained two sets of genotypes for each CSS (F2 and F8): one
from de novo and one from reference-based analysis.
2.6 | Locus selection and genotype correction
The selection of markers from available loci was performed on
F2 data. De novo and reference-based data were treated sepa-
rately. Loci respecting the following conditions were selected: (a)
diverging at 1 to 3 nucleotide sites between Kobodo and Makindu
strains; (b) showing invariance within each parental line; (c) found
in at least 70 CSS; and (d) showing no segregation distortion. For
all retained markers, some genotypes of F2 and F8 CSS could be
corrected by taking advantage of the backcross design. For each
backcross type, only two genotypes are expected among CSS:
homozygous for parental allele or heterozygous. For a CSS from
the Makindu backcross, if a genotype was inferred as homozygous
for Kobodo alleles, it was corrected for heterozygosity, assuming
that the second allele was not detected because of low coverage.
The same procedure was applied for CSS from the Kobodo back-
cross. After locus selection and genotype correction, de novo and
reference-based data were merged as follows. Genotypes were
compared for the markers shared between both analyses. If in-
ferred genotypes differed between de novo and reference-based
analyses, the heterozygous ones were retained: we assumed that
it was more likely to miss a second allele than to detect a false
positive one.
Only four markers showed segregation distortion. In total,
102,446 genotypes were inferred (181 samples × 566 markers)
and fewer than 1% were corrected based on in silico analysis.
The comparison of de novo and reference-based approaches led
to the detection of 400 genotype discrepancies between the two
6
|
BENOIST ET al.
approaches, which contributed to the improvement of the geno-
typic data.
2.7 | Linkage map construction and correction
The construction of the linkage map was performed with ca rtha g-
ene v1.3 (de Givry, Bouchez, Chabrier, Milan, & Schiex, 2005) and
the F2 data. The LOD and genetic distance thresholds used for
linkage group identification were, respectively, 7 and 0.5 Morgan.
Marker ordering was done with the default set of algorithms
(Defalgo option). For each CSS, genotypes were organized follow-
ing the linkage map order to visualize recombination events and to
correct some genotypes. When crossovers were detected on both
sides of a marker, the corresponding genotype was questioned. In
this case, genotypes were encoded as missing values when cover-
age was below 10 (159 individual marker genotypes) or manually
sequenced (86 individual marker genotypes). The final data set
comprised 3,143 missing genotypes (3% of the data set). A new
map was then generated from corrected data. When several mark-
ers were at null genetic distance on F2 and F8 data, the marker
with more data was retained.
2.8 | QTL analysis
All QTL analysis were performed with R Software (R Core Team,
2018). For each CSS, the probability of the genotypic states (KK, MM
and KM for homozygous Kobodo genotypes, homozygous Makindu
genotypes and heterozygotes, respectively) at every cM map po-
sition were estimated using the package R/qtl v1.44-9 (Broman &
Sen, 2009). At each position, additive and dominance indices were
determined with the following formulas: additive index = 2PKK +
PKM; dominance index = PKM, with PKK and PKM corresponding to the
probability that the genotype was homozygous for Kobodo alleles
or heterozygous.
To identify QTL for each trait, multiple regressions using GLM
were performed. Due to the cross-scheme, data could be classified
into three backcross types which may influence their genotype: F2
backcross with Makindu strain, F2 backcross with Kobodo strain or
F8 backcross with Makindu strain. As a consequence, phenotypic
variation was analysed in several steps. In a first step, a GLM was
built for each trait, taking into account only the backcross type. They
were based on the “quasi-binomial” error family for parasitism suc-
cess and sex ratio and on the “Gaussian” error family for the mean
number of offsprings.
In a second step, all positions were scanned with a full GLM
using residuals of the firs t model as response variable and additive
and dominance indices of the given position as explanatory vari-
ables. Position LOD scores (LODPosition) were calculated using the
formula: LODPosition = n/2 × log10 (RSSnull/RSSfull), where n is the
sample size, RSSfull is the residual sum of squares of the full model
with additive and dominance indices, and RSSnull is the residual
sum of squares of the null model (also based on residuals but with-
out explanatory variables). The significant LOD score threshold
was estimated by performing 2,000 permutations between phe-
notypes and genotypes within cross types and taking the 5% cut-
off of the maximum LOD scores obtained as significant threshold
value (Churchill & Doerge, 1994). The thresholds obtained were
3.06, 3.03 and 3.01 for offspring number, parasitism success and
sex ratio, respectively.
In the ne x t ste p, addit ive and domin anc e ind ices at the position
with the highest significant LOD score were included to build a
new model comprising this fixed position, and the genome was
rescanned for an additional QTL. The process was repeated until
no more significant position was detected. Such a process is es-
pecially powerful to detect genetically linked QTLs and QTLs in
epistatic relationship. To test for interactions between QTLs,
interaction terms for all QTL pairs were added to the model, in-
cluding all QTLs detected. Each interaction term was then tested
separately by an analysis of deviance comparing the model with
and without interaction terms. The interaction was declared sig-
nificant if the p-value was inferior to 0.05. For each trait, the final
model that comprises all fixed QTLs and their significant pairwise
interactions is called “the complete model” in the following text.
In these models, F8 backcross was the reference to calculate the
intercept.
Additive and dominance coefficients are the estimated coef-
ficients in these complete models. Additive effect corresponds to
the variation associated with one Kobodo allele. Dominance ef-
fect corresponds to the variation resulting from the interaction
between the parental alleles at one locus. Due to the “quasi-bino-
mial” error family used for parasitism success models, estimated
coefficients for this trait are given in logits scale. These coeffi-
cients were transformed using the inverse logit transformation to
estimate the parasitism success variation associated with additive
and dominance effect (Crawley, 2012). The percentage of total
phenotypic variance explained by each QTL was determined from
the complete models with the formula SSQTL/TSS, where SSQTL
is the sum of squares associated to QTL and TSS the total sum
of squares. The confidence interval of each QTL position corre-
sponded to all positions around the QTL for which LOD score was
above LOD score max – 2.
2.9 | Candidate gene identification
Gene identification was performed by genomic comparison with
the available genome of C. congregata. We first assembled RAD-
seq data from all samples with spade s v3.11.1 (Nurk et al., 2013). All
markers included in the QTL confidence intervals were mapped on
the scaffolds obtained using bl astn v2.6.0+ (Boratyn et al., 2012).
Scaffolds containing markers were then mapped to the annotated
genome of C. congregata using BL AST to identify genes in QTLs. For
this genome, an automatic annotation was performed and refined
with manual inspection for some gene families (Bracovirus, Venom,
|
7
BENOIS T ET al.
Immunity, Detoxification and Chemodetection; Gauthier et al.,
2020).
2.10 | Gene Ontology enrichment analysis
Gene Ontology (GO) enrichment was tested using the R package
topGO (Alexa, Rahnenfuhrer, & Lengauer, 2006). Annotations were
those derived from blast2GO analyses of the Cotesia congregata ge-
nome performed by Gauthier et al. (2020). Two set of genes were
built for each trait based of the gene content of QTL confidence in-
terval. We compared the results of four algorithms implemented in
topGO, namely “classic,” “elim,” “weight” and “weight01” that inte-
grate the hierarchical structure of Gene Ontology in different ways.
Retaining only nodes larger than 5 we computed Fisher tests for the
three categories of “Biological Process,” “Molecular Function” and
“Cellular Compartment.”
3 | RESULTS
3.1 | Descriptive statistics
The two parental strains showed a substantial difference for para-
sitism success and offspring number, but they did not differ for sex
ratio (Table 1). The Kobodo strain was more efficient with a higher
parasitism success and a higher offspring number. The parasitism
success of the clonal sibships (CSS) fr om the Kobodo ba ckcross was
equivalent to those of the Kobodo parental strain. By contrast, the
parasitism success of the CSS from Makindu backcrosses (F2 and
F8) was higher than the Makindu parental strain, although below
the Kobodo. The offspring number of CSS from the Kobodo back-
cross was far higher than the Kobodo parental strain, and the off-
spring number of the CSS from the Makindu backcrosses was close
to those of the Kobodo parental strain. Data showed a highly sig-
nificant correlation (Spearman's ρ = 0.505, p-value = 5.932 × 10–
11, calculated with all CSS) between the parasitism success and
offspring number, suggesting a relationship between both traits.
The broad-sense heritability was 29.82% for offspring number and
14.46% for sex ratio.
3.2 | Marker selection and genotype identification
The number of loci identified by Stacks varied considerably between
the two approaches: 119,176 and 33,906 for de novo and reference-
based, respectively (Table 2). However, the number of remaining
loci after the first filter step was equivalent between the two ap-
proaches, around 30,000, which was similar to the 37,000 expected
loci (based on the genome of the sister species C. sesamiae). This
suggests that a large proportion of loci are split into different stacks
with the de novo approach. Among these 30,000, only a small frac-
tion was polymorphic, suggesting that the two strains are genetically
close. The majority of the identified markers were shared between
the two approaches, but more than 20% (123/566) were specific to
one of them, which highlighted their complementarity. The median
coverage was above 30, which is satisfying for reliable genotype
inference.
3.3 | Linkage map
The final linkage map generated from F2 data after removing redun-
dant loci is presented in Figure 2 and in Table S2. It includes ten
linkage groups, which is consistent with the number of chromo-
somes identified in C. typhae (C. Bressac, personal communication),
indicating that the map is saturated. In addition, each linkage group
matched to one chromosome from C. congregata (Gauthier et al.,
2020). The linkage map comprises 465 markers spanning a total
length of 1,300 cM with an average distance of 2.7 cM between two
consecutive markers.
3.4 | QTL mapping for parasitism
success and offspring number
Three QTLs were identified for parasitism success (PS) and two for of f-
spring number (ON), localized on linkage groups 4 and 5 and on the
linkage groups 5 and 6, respectively (Figure 3). No QTL was detected
for sex ratio. One peak for parasitism success was not ret ained (linkage
group 1, two QTL scans) because it was observed in the first scan only
and was locate d in a region with low mar ker densi ty. Th e QTL obser ved
TABLE 1 Phenotypes of the parental strains and the clonal sibships (CSS) from the different backcrosses
Data type Parasitism success Offspring number Sex ratio (% male)
Parental Kobodoa 78.70% n = 108 71.07 ± 24 .06 n = 71 68.80 ± 20.22 n = 71
Parental Makindua 36.20% n = 80 36.48 ± 16.68 n = 21 74.90 ± 16. 50 n = 21
CSS from Kobodo F2 81.57 ± 14.8 0% n = 832 (47) 91.05 ± 17.54 n = 491 (47) 39. 09 ± 21.28 n = 490 (47)
CSS from Makindu F2 70.96 ± 22.55% n = 857 (45) 74.94 ± 24.77 n = 440 (45) 35.89 ± 16.78 n = 437 (45)
CSS from Makindu F8 64.00 ± 20.24% n = 1737 (89) 66.83 ± 19.71 n = 827 (88) 33.01 ± 20.00 n = 801 (88)
H2 = 29. 82% H2 = 14.4 6%
Note: ±Standard deviation; n, number of female progenies analysed; ( ) number of CSS; H2 = Broad-sense heritability.
aData from Benoist et al. (2017).
8
|
BENOIST ET al.
Analysis
De novo Reference-based
Number of loci Identified by stacks 119,176 33,906
Present in at least 70 samples 31,599 2 9,797
Fixed in parental strains and
polymorphic
541 507
With at most 3 SNP 524 491
Without segregation distortion 522 487
De novo/ Reference-based
specific
79 44
Used for linkage map
construction
566
Retained in final linkage map 4 65
Marker coverage
per sample
Mean 35.55 4 9.32
Median 32 34
Note: The locus coverage is given for the 566 markers used for the genetic map construction.
TABLE 2 Number of retained loci at
each step of the loci selection and locus
coverage
FIGURE 2 Cotesia typhae linkage map
based on RAD genotyping of 92 F2 CSS.
The linkage map includes 465 markers,
n = number of markers by linkage group
(LG). Genetic distances are calculated
using Kosambi's map function
FIGURE 3 Result of Cotesia typhae genome scan for QTL detection of (a) offspring number, (b) parasitism success and (c) sex ratio. The
10 linkage groups (LG) and their markers are indicated on x-axis. LOD SCORES are calculated from generalized linear models (GLM). The
black dotted line corresponds to the QTL threshold calculated by 2,000 permutations: it reaches 3.06, 3.03 and 3.01 for offspring number,
parasitism success and sex ratio, respectively. Successive scans were performed for each linkage group. In One QTL scan, we hypothesized
that only one QTL occurred in the genome. The highest peak was then included in the two QTL scans to search for a second QTL. The
process was repeated until no other significant peak was detected. Coloured lines correspond to each QTL scan performed where a
significant peak was detected
|
9
BENOIS T ET al.
10
|
BENOIST ET al.
on the linkage group 5 was common to the two phenotypic traits PS
and ON. The length of the QTL confidence intervals varied from 9 to
38 cM (Table 3). For both PS and ON, the pairwise interactions be-
tween observed QTLs were not significant. An overdominance effect
was observed for all ON QTLs, whereas QTL1-PS and QTL3-PS were
associated with, respectively, underdominance and partial dominance
of Makindu alleles. The QTL2-PS was strictly additive. The percentage
of phenotypic variance explained by each QTL varied between 4% and
14%. The total phenotypic variance explained reached 27.7% for para-
sitism success and 24.5% for offspring number.
3.5 | Identification of genes within QTL intervals
To identify ca n didate genes, marke rs in QTL intervals were mapped on
the annotated genome of C. congregata. Cumulative length of all QTL
intervals represented around 11.7 Mb (Table 4), which corresponds
to ~5.9% of the genome length and comprises 435 genes (Table S3)
with a putative function. This list of genes was inspected based on
literature data. We focused on studies (mainly genomic and transcrip-
tomic studies) that identified genes potentially influencing reproduc-
tive success in other parasitoid species. From these studies, 67 genes
of interest were identified and are listed in Table 5. Among them we
found Bracovirus genes involved in the production of viral particles
and other bracovirus genes similar to genes involved in the inacti-
vation of the host immune system in other host parasitoid models
(Bézier, Annaheim, et al., 2009; Burke, Walden, Whitfield, Robertson,
& Strand, 2014; Chevignon et al., 2014, 2015; Falabella et al., 2007;
Gauthier et al., 2020; Pruijssers & Strand, 2007; Thoetkiattikul, Beck,
& Strand, 2005). Some genes producing venom components were
also identified (Ali, Lim, & Kim, 2015; Arvidson et al., 2019; Burke
& Strand, 2014; Colinet, Mathé-Hubert, Allemand, Gatti, & Poirié,
2013; Danneels, Rivers, & de Graaf, 2010; Moreau & Asgari, 2015;
Sim & Wheeler, 2016). Finally, we reported genes that were highly
expressed in ovaries of C. congregata (Gauthier et al., 2020) and N.
vitripennis (Sim & Wheeler, 2016) or differentially expressed between
resting and ovipositing females of N. vitripennis (Pannebakker, Trivedi,
Blaxter, Watt, & Shuker, 2013).
Phenotypic trait
Offspring number Parasitism success
QTL name QTL1- ON QTL 2-O N QT L1-P S QTL2-PS QTL3-PS
Linkage group LG05 LG06 LG04 LG 05 LG04
LOD score value 6.71 5.91 5.08 5.44 4.68
Peak position 116 cM 11 cM 15 cM 110 cM 50 cM
Confidence interval 108–117 cM 2–17 cM 0–18 c M 98–117 cM 22–6 0 c M
Coefficient estimates
Additive effect 7.7 2 0. 52 0.02 0.16 −0.16
Dominance effect 13.05 14.14 − 0.17 –a 0.08
% Total phenotypic variance explained
Additive component 8.29% 2.55% 4.68% 11.15% 3.50%
Dominance
component
6.60% 7.0 6 % 7.6 2 % –a 0.75%
Total by QTL 14.89% 9. 61% 12.30% 11.15% 4.25%
Tot al 24.5% 27. 7 %
Note: Additive effect corresponds to the effect of one Kobodo allele. Positive dominance effects
indicate dominance of Kobodo alleles, whereas negative values reveal dominance of Makindu ones.
aThe percentage of the total phenotypic variance explained by the dominance component of the
QTL2-PS was not calculated due to the lack of significance of this component in the QTL model.
TABLE 3 Detected QTL position,
confidence interval, coefficient estimates
and percentage of the total phenotypic
variance explained by each QTL detected
QTL name
C. congregata fragment
length (Mb)
Number of genes
Automatically
annotated
Manually
annotated Tot al
QTL1-PS ~1.3 47 15 62
QTL3-PS ~3.7 105 37 142
QTL1-ON + QT L2-PS ~4.2 113 25 138
QTL2-ON ~2.5 90 393
Note: Overlapping QTLs (QTL1-ON and QTL2-PS) was considered as a single interval.
TABLE 4 Summary of gene
identification in quantitative trait loci
(QTL) intervals based on the annotated
genome of C. congregata
|
11
BENOIS T ET al.
3.6 | Gene Ontology enrichment analysis
Gene Ontology enrichment results are provided in Table 6. The GO
terms were ordered following the results obtained with weight01
algorithm and the GO detected as significantly enriched at the 1%
level with the four algorithms are listed. This study revealed an en-
richment in GO terms involved in fine regulation processes. The most
significant term refers to Calpain function, an intracellular calcium-
dependent cysteine protease (GO:0004198). This protease family
is known to modulate the activity of other proteins, a function re-
quired in signal processing (Friedrich & Bozóky, 2005). In Drosophila
melanogaster, Vieira, Cardoso, and Araujo (2017) showed that CalpA,
a member of the calpain family, was involved in regulating the tim-
ing of mitosis during embryonic development. This molecular func-
tion may also favour the proper development of C. typhae lar vae in
its host. Terms involved in tRNA modifications (GO:0008175 and
GO:0002098) were also detected as highly significant. Such post-
transcriptional modifications are well known as modulators of tRNA
activity influencing translation speed and fidelity. Repressing tRNA
methylation decreased growth rates in yeast (Nachtergaele & He,
2017). Interestingly, RNA modifications are also widely used by vi-
ruses to hijack host cell machinery.
4 | DISCUSSION
The QTL approach is used to identify genes differing between
lines diverging for one or several phenotypic traits. It is not de-
signed to detect all the genes involved in the phenotype but
rather the ones that vary and thus are susceptible to evolve.
Applied to fitness-related traits, this approach allows us to iden-
tify the genetic components involved in an organism's adaptation.
In parasitoids, the QTL study of reproductive success provides an
opportunity to identify the key features of host adaptation and
hence the genes submitted to the coevolution processes. Beyond
the interest for understanding parasitism success and reproduc-
tion in parasitoid species, such a study may also be helpful in the
selection process in a biological control perspective, for example
through marker-assisted selection. The QTL strategy is mainly
conducted on model organisms, especially species of agricultural
QTL name Genes of interest
QTL1-PS Venom: alkaline phosphatase-like; disintegrin and metalloproteinase
domain-containing protein 9
Other: btb poz domain-containing adapter for cul3-mediated
degradation protein 3; 60s ribosomal protein l10; 60s ribosomal
protein l23a
QTL3- PS Bracovirus nudiviral cluster: 27b; 35a₁; 35a₂; 38K; GbNVorf19;
HzNVorf106; HzNVorf9_1; HzNVorf9_2; HzNVorf94; Int₁;
K425_438; K425 445; K425 456; K425 459; K425 461; P6.9₁;
P6.9₂; pif3; pif6; PmNVorf87; PmV; ToNVorf29; ToNVorf54₁;
ToNVorfF54₂; vp39
Other: Putative mediator of RNA polymerase II transcription subunit
12; 40s ribosomal protein s15aa
QTL1-ON + QT L2-PS Bracovirus virulence genes: bv20.1.26.4; bv20.2.26.8; ep1.1.1.3;
ep1.2.1.4; ep1.3.1.5; ptp a.26.6; ptp b.1.1; ptp d.1.11 pseudo; ptp
delta.26.1; ptp epsilon.26.7 pseudo; ptp i.1.2; ptp k.1.6; ptp l.1.7;
ptp m.1.8; ptp p.1.9; ptp q.1.10; vank 5.26.2; vank 6.26.3b; vank
9.2 6.5
Venom: calcium-independent phospholipase a2; adenosine
deaminase-like protein; cysteine-rich with EGF-like domain protein
2; serine protease inhibitor; serpin B4-like; serpin B6-like
Other: plasminogen activator inhibitor 1 RNA-binding protein; BTB/
POZ domain-containing protein 7; sorting-nexin 24-like; sorting-
nexin 6; zinc finger and BTB domain-containing protein 41-like; 60S
ribosomal protein l13; 60S ribosomal protein l18a
QTL2-ON Venom: cysteine-rich secretory protein 2-like
Other: guanine nucleotide-binding protein subunit beta-like protein;
60S ribosomal protein l5
Note: Overlapping QTL (QTL1-ON and QTL2-PS) was considered as a single interval. Genes of
interest were classified in four categories. Bracovirus Nudiviral cluster (genes of nudiviral origin
involved in particles bracovirus production): Bracovirus virulence genes (genes packaged in the
particles and expressed during parasitism) and Venom categories refer to parasitoid weapon
arsenal. The “Other” categor y comprises genes selected because they are actively transcribed
in ovaries of C. congregata (Gauthier et al., 2020) and N. vitripennis (Sim & Wheeler, 2016) or
differentially expressed between resting (no contact with host) and ovipositing females of
N. vitripennis (Pannebakker et al., 2013).
TABLE 5 Genes of interest in
quantitative trait loci (QTL) intervals
12
|
BENOIST ET al.
value, to identify genetic components of complex traits and also
to support selection processes (Darvasi, 1998; Tanksley, 1993; Xu
& Crouch, 2008). It was therefore challenging to develop such an
approach on a recently described species.
The first challenge was to build a dense and reliable linkage map
from scratch. The linkage map obtained comprise 10 linkage groups
as the number of chromosomes observed in C. typhae (C. Bressac,
personal communication) each of them corresponding to one of the
10 chromosomes of C. congregata (data not shown, Belle et al., 20 02;
Gauthier et al., 2020). The length of the genetic map is 1,300 cM
with an average recombination rate of 7.8 cM/Mb, in line with es-
timates for other insects (Stapley, Feulner, Johnston, Santure, &
Smadja, 2017). The marker density combined with this recombina-
tion rate (one marker for each 400 kbp) is favourable for QTL detec-
tion and confirms the interest of using RAD-seq to generate markers.
This genetic map will be useful for further studies, for instance for
the ongoing assembly of C. typhae genome at the chromosome scale.
The second challenge was to detect QTL, to characterize their
phenotypic effect and to localize their genomic position as precisely
as possible to obtain candidate genes. Owing to the favourable
density of the map, the main limitation of QTL detection was the
number of progenies to be characterized (Lander & Botstein, 1989).
We phenotyped 181 CSS, each of them being characterized for 20
sibling females (more than 4,000 parasitisms performed and more
than 2,600 successful progenies counted). This phenotyping effort
allowed us to detect four distinct QTLs even with an impact below
5% of the phenotypic variance.
No QTL was detected for sex ratio despite a marked broad-
sense heritability estimated. This negative result probably arises
from a complex determinism of sex ratio by numerous genes of low
effect. In a comparable study on Nasonia vitripennis, Pannebakker
et al. (2011) estimated H2 of 9.5% for sex ratio an d detected one QTL
but explaining only 0.16% of phenotypic variance and 1.56% of ge-
netic variance. These authors argued that sex ratio was likely to re-
sult from a complex architecture with pleiotropic genes influencing
other life history traits such as clutch size. Such trade-off between
traits was observed in C. typhae for which progenies from unmated
females (with only male offsprings) comprised a significantly higher
number of offsprings than progenies from mated females (mixed
progenies with males and females, Benoist et al., 2017). This differ-
ence may result from ovipositing behaviour of the female but also
from the higher survival rate of male larvae compared to females. In
such case, primary sex ratio (i.e., sex ratio at oviposition) should be a
better trait to approach genetic determinism as it is directly imput-
able to the mother behaviour (Ueno & Tanaka, 1997). However, this
index is difficult to estimate, particularly for endoparasitoid species.
TABLE 6 Summary of GO enrichment tests in quantitative trait loci (QTL) intervals based on the annotated genome of C. congregata
Tra it GO,ID Te r m
Observed
number of genes
Expected
number of genes p-Value
Parasitism success Molecular function
GO:0004198 Calcium-dependent cysteine-type
endopeptidase activity
30.21 .00069
GO:0008175 tRNA methyltransferase activity 30.29 .00227
GO:0004674 Protein serine/threonine kinase activity 11 4.17 .0027
Biological process
GO:0002098 tRNA wobble uridine modification 40.25 .000043
GO:0000902 Cell morphogenesis 30.34 .0035
GO:0007169 Transmembrane receptor protein tyrosine
kinase signalling pathway
30.34 .0035
GO:0030488 tRNA methylation 3 0.34 .0035
GO:0120036 Plasma membrane-bounded cell projection
organization
30.42 .0099
Offspring number Molecular function
GO:0004198 Calcium-dependent cysteine-type
endopeptidase activity
20.16 .0093
GO:0030246 Carbohydrate binding 40.85 .0095
Biological process
GO:0 0 42176 Regulation of protein catabolic process 3 0.23 .0 011
GO:0007169 Transmembrane receptor protein tyrosine
kinase signalling pathway
30.26 .0017
GO:0120036 Plasma membrane bounded cell projection
organization
30.33 .006
Note: The list comprises GO detected as significantly enriched at the 1% level with four different algorithms implemented in topGO package: classic,
elim, weight and weight01. p-Values indicated are those obtained with weight01 package.
|
13
BENOIS T ET al.
We identified three QTLs for parasitism success and two QTLs
for offspring number. Comparison of the phenotypic variation ex-
plained by these QTLs with the broad-sense heritability estimated
suggests that our approach succeeded in explaining almost all the
expected genetic effect. This conclusion should be mitigated by
the known bias of QTL detection strategies that result in an over-
estimate of the QTL effect through selection bias (Broman, 2001).
Understanding of the genetic architecture of the host–parasite in-
teraction has been approached mainly through the study of host
genetic variation. In plant–parasite relationships, a gene-for-gene
model has been proposed where a single locus is involved in the
resistance of the host and a single locus involved in the virulence of
the parasite (Thompson & Burdon, 1992). Based on genetic crosses
between virulent and avirulent strains of the parasitoid Leptopilina
boulardi and tests of parasitism success against different hosts of
Drosophila, the same gene-for-gene model was proposed by Dupas
et al. (1998) and Dupas and Carton (1999) to explain the outcome
of the interaction between the two partners. Mochiah, Ngi-Song,
Overholt, and Stouthamer (2002) conducted crosses between
Cotesia sesamiae strains differing in their ability to parasitize Buseola
fusca. They also tested the relative impact of maternal factors and
larval ones through superparasitization experiments. They identi-
fied a higher number of segregating factors and showed that both
mate rna l factors an d lar val com ponents (e.g ., su r fac e pro teins) were
necessary to allow the complete development of the parasitoid and
the emergence of adults. Other theoretical models of coevolution
have been proposed that also involve more than one locus with
strong epistatic relationship between loci (Tellier & Brown, 2007).
The number of loci we identified is consistent with such a model
except we did not find evidence of interactions between QTLs.
However, the number of QTLs has to be analysed cautiously as one
QTL is not synonymous with one gene but may comprise several
genes. The observation of enrichment for some GO terms in the
QTL intervals suggests that several genes contribute to the varia-
tion in phenotype.
Besides the QTL detection, our approach allowed us estimating
the additive and dominance effect for each QTL. Strong overdomi-
nance effects were observed for offspring number QTLs in agree-
ment with results obtained on hybrid F1 (Benoist et al., 2017). Such
effects reveal a positive interaction between alleles, either from
a single gene or from different genes in the QTL interval through
epistatic interactions or pseudo-superdominance. Under the pseu-
do-superdominance hypothesis, the overdominance effect is likely
due to favourable alleles in the repulsion phase at tightly linked
QTLs as observed in maize (Larièpe et al., 2012), suggesting that
both Kobodo and Makindu may carry favourable alleles for off-
spring number. Parasitism success results also reveal the presence
of favourable alleles in the Makindu strain, despite its overall poorer
performance: negative values were estimated for dominance effect
of QTL1-PS and additive effect of QTL3-PS. Taken together, these
results indicate that a recombinant strain between Kobodo and
Makindu, harbouring all favourable alleles, may perform better than
the parental strains and could be useful in biocontrol perspective.
The detection of a QTL shared between parasitism success and
offspring number is consistent with the high correlation detected
between the two traits. In our experiment, the number of oviposi-
tion events was fixed at one. The offspring number thus depends
on (a) the number of eggs laid in the host by the female and (b) the
parasitoid larval survival rate. Both larval mortality and parasitism
success are directly connected to parasitoid ability to inhibit the
host immune system and are therefore naturally correlated. In ad-
dition, some authors have even suggested that parasitoid females
injecting more eggs enhance the survival rate of their larvae through
the saturation of the host immune system (Blumberg & Luck, 1990;
Kapranas, Tena, & Luck, 2012; Rosenheim & Hongkham, 1996). It is
therefore not surprising that the reproductive success is under the
control of at least one common QTL between parasitism success and
offspring number.
In previous studies, we found that the difference in offspring
number observed bet ween Kobodo and Makindu females can be ex-
plained in part by the number of eggs injected into the host (Benoist
et al., 2017, 2020). Thus, we expected to find genes related to the
oviposition behaviour, especially in the QTL specific to offspring
number. However, very few genes influencing such behaviour are
known despite the large number of studies on this topic. This is
probably due to the complexity of behavioural traits that involve a
huge quantity of genes, complicating their study (Flint, Greenspan, &
Kendler, 2010). A refined annotation of genes in the QTL intervals is
therefore required to identify candidate genes related to oviposition
behaviour.
Genes belonging to the bracovirus were found in the common
QTL and in the QTL3-PS. Within the wasp genome, the bracovirus
is organized in two types of regions. The first corresponds to genes
from a nudiviral origin (nudiviral genes) involved in the bracovirus
particle production. Numerous nudiviral genes are clustered in one
region, called the nudiviral cluster (Bézier, Annaheim, et al., 2009;
Bézier, Herbinière, et al., 2009) encoding in particular major cap-
sid components (VP39, 38K, Wetterwald et al., 2010). The second
type corresponds to proviral segments used for the production of
the viral circle contained in viral particles and harbouring virulence
genes. In C. congregata, 35 proviral segments organized in 9 proviral
loci (PL) were identified. Approximately two-thirds of the proviral
segm ent s are loc ali zed at the same geno mic region, known as macro-
locus, and comprising PL1 and PL2, while other are dispersed in the
wasp genome (Bézier et al., 2013). The adaptive role of the braco-
virus in host range evolution was widely documented in the Cotesia
genus. For example, virulence of C. sesamia against Buseola fusca was
linked to allelic variation of the CrV1 bracovirus genes (Branca, Le
Ru, Vavre, Silvain, & Dupas, 2011; Gitau, Gundersen-Rindal, Pedroni,
Mbugi, & Dupas, 2007). Furthermore, several bracovirus genes were
shown to be under positive selection in relation with host adap-
tation, notably in C. sesamiae and C. typhae (Gauthier et al., 2018;
Jancek et al., 2013).
In the QTL common to both traits, we found proviral genes:
BV20 genes and genes belonging to EP1-like, PTP and Vank fam-
ilies for which some members are known as virulence factors of
14
|
BENOIST ET al.
other parasitoid species (Gueguen, Kalamarz, Ramroop, Uribe,
& Govind, 2013; Kwon & Kim, 2008; Pruijssers & Strand, 2007).
These genes are localized in PL5 and PL8 producing circles 1 and
26, respectively. The se two circles are of particular interes t. It was
shown in the Cotesia congregata-Manduca sexta system that they
could integrate into the host genome (Chevignon et al., 2018).
They are produced in higher numbers than others (Chevignon
et al., 2014) and contain genes highly transcribed in the host. For
ins tance, EP1-like genes and Vank9 are among the most expressed
in host haemocytes following parasitism (Chevignon et al., 2014).
Genes on these circles are thus good candidates to explain the
difference in parasitism success and offspring number between
Kobodo and Makindu parasitoid strains. Interestingly, no QTL was
detected in the macrolocus region, which concentrates the major-
ity of proviral genes and no virulence genes known to be under
positive selection were found in our QTL.
Many studies on virulence, with respect to bracovirus, focus
on proviral genes but much less on nudiviral genes encoding parti-
cle structural components and/or involved in particle production.
Interestingly, the QTL3-PS encompasses the whole nudiviral clus-
ter. Among the genes in this QTL, some have a predicted func-
tion based on homology with baculovirus genes: pif3 and pif6 gene
products may play a role in virus entry into host cells, whereas
VP39 and 38K likely produce major components of nucleocap-
sids (Bézier, Annaheim, et al., 2009; Herniou, Olszewski, Cory,
& O’Reilly, 2003; Wetterwald et al., 2010) containing bracovirus
DNA circles. Benoist et al., 2020 obser ved that the amount of viral
particles injected in the host did not explain the difference in par-
asitism success between Kobodo and Makindu parasitoid strains,
which make particle component production unlikely to be involved
in the difference between the two strains. The variation of para-
sitism success induced by this QTL could result from a difference
in particle infectivity between the two C. typhae strains, which in
turn may result from differences in pif3 and/or pif6 copies. It would
be interesting therefore to compare how Kobodo and Makindu
bracovirus infect host cells.
In all QTLs, except in the QTL3-PS, we found genes associated
wi th ven om compo nent s, whose role in host adaptat ion was w ide ly
studied (Cavigliasso et al., 2019). In polydnavirus-associated par-
asitoids, polydnavirus is considered as the main virulence factor.
However, it was shown in many species harbouring polydnavirus
that venom is also required for successful parasitism and could
synergize the effect of polydnavirus (Asgari, 2012; Kitano, 1986;
Moreau & Asgari, 2015; Tanaka, 1987). The presence of venom
genes in the QTLs suggests that their role in virulence might
be significant and would therefore make further investigations
worthwhile.
The goal of this analysis was to identify candidate genes. In total,
67 genes of interest were retained, which is rather high for under-
taking further studies of their individual implication in phenotypic
variation. Complementary approaches, such as comparative tran-
scriptomics or genome wide association study (GWAS), could en-
able us to select the most interesting candidate genes. Compared to
classic QTL approaches, GWAS allows to reach a higher resolution
as soon as the sampling effort in mixed or natural populations and
the marker density are large enough (Hansson et al., 2018; Santure
& Garant, 2018). Focusing on QTL confidence interval, it may help to
better target genes of interest. Once the number of candidates genes
is reduced, we will be able to assess their role using functional anal-
yses available in this model, such as RNA interference, an approach
which was shown to work efficiently to knock down targeted gene
expression in Hymenoptera (Marco Antonio, Guidugli-Lazzarini,
do Nascimento, Simões, & Hartfelder, 2008) and more specifically
in parasitoid wasps (Burke, Thomas, Eum, & Strand, 2013; Colinet
et al., 2014).
5 | CONCLUSION
This work was devoted to the study of genetic components of the
reproductive success of a parasitoid species. It pointed out four
genomic regions involved in the variations of both parasitism success
and offspring number, two traits directly connected to the fitness
of individuals. It allowed the identification of a list of genes of inter-
est, notably including bracovirus and venom genes. The detection of
those well-known components of parasitoid virulence gives strong
support to the strategy presented here. Those genes are particularly
interesting in the topic of coevolution because of their implication in
host adaptation. The number of genes pointed out is quite large but
clearly limited, their location being restricted by the QTL intervals to
well-defined genomic regions. Population studies taking benefit of
linkage disequilibrium at small genomic scale, or comparative tran-
scriptomic studies, will allow to approach closer to the candidate
genes in the future.
ACKNOWLEDGEMENTS
We thank Odile Giraudier and Sylvie Nortier for insect rearing at
Gif, Julius Obonyo for field collection in Kenya, Matthieu Bodet,
Florian Decourcelle, Maxime Villoing and Laurence Signon for their
contribution to the crosses and phenotyping, Maud Silvain, Claude
Thermes, Delphine Naquin and Yan Jaszczyszyn from the I2BC se-
quencing facility for the RAD-sequencing, Florence Prunier for her
methodological advice, Damien Delafoy for his preliminary work on
RAD-seq data, Cécile Courret for her help in QTL analysis, Jérémy
Gauthier for his advice on sequencing data analysis and his help with
the annotated genome of C. congregata, Lionel Saunois, Amandine
Dubois and Virginie Héraudet for maize production and Malcolm
Eden for linguistic editing of the manuscript.
AUTHOR CONTRIBUTIONS
Conceptualization of the study was done by R.B., S.D. and F.M.; meth-
odology for the study was performed by R.B., C.C-D., L.M., A.L.R., L.K.
and F.M.; software analysis was performed by R.B., C.C-D. and F.M.;
validation was done by R.B, C.C-D. and F.M.; formal analysis was done
by R.B., C.C-D., L.M., A.L.R. and F.M.; investigation of the study was per-
formed by R.B., C.C-D., C.C., R.J., P-A.C., E.V.D., L.K. and F.M.; resources
|
15
BENOIS T ET al.
of the study were provided by B.L.R, P-A.C, J-M.D. and L.K.; data cura-
tion was performed by R.B, C.C-D. and F.M.; writing—original draft was
performed by R.B. and F.M.; writing—review & editing was performed
by C.C-D., L.M., A.L.R., P-A.C., E.V.D., S.D., J-M.D., B.L.R and L.K.; super-
vision of the study was done by L.K. an F.M.; project administration was
done by L.K.; and funding acquisition was performed by L.K.
ADDITIONAL INFORMATION
All experimentations were realized under the juridical frame of a
Material Transfer Agreement signed between IRD, icipe and CNRS
(CNRS 072057/IRD 302227/00) and the authorization to import
Cotesia in France delivered by the DRIAAF of Ile de France (IDF
2017-OI-2 6-032).
DATA AVAILAB ILITY STATE MEN T
RAD-Seq raw data sequences are archived at the NCBI SRA in the
BioProject PRJNA622407. Genome database (genomes and an-
notated genes) is available on the website BIPAA (Bioinformatic
Platform for Agrosystem Arthropods) https://bipaa.genou est.org/
is/parwa spdb/.
ORCID
Arnaud Le Rouzic https://orcid.org/0000-0002-2158-3458
Florence Mougel https://orcid.org/0000-0003-0114-288X
REFERENCES
Alexa, A., Rahnenfuhrer, J., & Lengauer, T. (20 06). Improved scoring of
functional groups from gene expression data by decorrelating GO
graph structure. Bioinformatics, 22(13), 1600–1607. https://doi.
org/10.1093/bioin forma tics/btl140
Ali, M. R., Lim, J., & Kim, Y. (2015). Transcriptome of a specialized ex-
tra-embryonic cell, teratocyte, and its host immunosuppressive role
reveale d by ex vivo RNA interference: Teratocyte transcriptome. Insect
Molecular Biology, 24(1), 13–28. https://doi.org/10.1111/imb.12132
Ar vids on, R., Kais er, M., Le e, S. S., Ure nda , J.-P., Dai l, C. , Mo hamm ed, H., …
Adams, M. E. (2019). Par asit oi d jewel wasp mou nt s mu lt ipronged neu-
rochemical attack to hijack a host brain. Molecular & Cellular Proteomics,
18(1), 99–114. https://doi.org/10.1074/mcp.RA118.000908
As gar i, S. (201 2) . Venom s fro m end opara sitoi ds. In N. E. Beckage & J.-M .
Drezen (Eds.), Parasitoid viruses (p p . 217–2 31). Lo n d on, UK : Ac a d emic
Press. https://doi.org/10.1016/B978-0-12-38485 8-1.00018 -7
Asgari, S., & Rivers, D. B. (2011). Venom proteins from endoparasitoid
wasps and their role in host-parasite interactions. Annual Review
of Entomology, 56(1), 313–335. https://doi.org/10.1146/annur ev-
en to-1 207 0 9 -14 48 49
Beckage, N. E., & Drezen, J.-M. (Eds.) (2012). Parasitoid viruses: Symbionts
and pathogens (1st ed.). London, UK and New York, NY: Elsevier/
Academic Press.
Beckage, N. E., & Gelman, D. B. (2004). Wasp parasitoid disruption of
host development: Implications for new biologically based strategies
for insect control. Annual Review of Entomolog y, 49, 299–330. https://
doi.org/10.1146/annur ev.ento.49.061802.123324
Belle, E., Beckage, N. E., Rousselet, J., Poirie, M., Lemeunier, F., & Drezen,
J.-M. (2002). Visualization of polydnavirus sequences in a parasitoid
wasp chromosome. Journal of Virology, 76(11), 5793–5796. https://
doi.org/10.1128/JVI.76.11.5793-5796.2002
Benoist, R., Chantre, C., Capdevielle-Dulac, C., Bodet, M., Mougel, F.,
Calatayud, P. A., … Kaiser, L. (2017). Relationship between ovipo-
sition, virulence gene expression and parasitism success in Cotesia
typhae nov. sp. parasitoid strains. Genetica, 145 (6), 469–479. https://
doi.org/10.10 07/s1070 9-017-9987-5
Benoist, R., Paquet, S., Decourcelle, F., Guez, J., Jeannette, R., Calatayud,
P.-A., … Kaiser, L. (2020). Role of egg-laying behavior, virulence
and local adaptation in a parasitoid’s chances of reproducing in
a new host. Journal of Insect Physiology, 120, 103987. https://doi.
org/10.1016/j.jinsp hys.2019.103987
Bézier, A., Annaheim, M., Herbiniere, J., Wetterwald, C., Gyapay, G.,
Bernard-Samain, S., … Drezen, J.-M. (2009). Polydnaviruses of brac-
onid wasps derive from an ancestral nudivirus. Science, 323(5916),
926–930. https://doi.org/10.1126/scien ce.1166788
Bézier, A ., Herbinière, J., Lanzrein, B., & Drezen, J.-M. (2009).
Polydnavirus hidden face: The genes producing virus particles of
parasitic wasps. Journal of Invertebrate Pathology, 101(3), 194–203.
https://doi.org/10.1016/j.jip.2009.04.006
Bézier, A., Louis, F., Jancek, S., Periquet, G., Theze, J., Gyapay, G., …
Drezen, J.-M. (2013). Functional endogenous viral elements in the
genome of the parasitoid wasp Cotesia congregata: Insights into the
evolutionary dynamics of bracoviruses. Philosophical Transactions of
the Royal Society B: Biological Sciences, 368(1626), 20130 047. https://
doi.org/10.1098/rstb.2013.0047
Blumberg, D., & Luck, R. F. (1990). Differences in the rates of super-
parasitism between two strains of Comperiella bifasciata (Howard)
(Hymenoptera: Encyrtidae) parasitizing California red scale
(Homoptera: Diaspididae): An adaptation to circumvent encapsula-
tion? Annals of the Entomological Society of America, 83(3), 591–597.
https://doi.org/10.1093/aesa/8 3.3.591
Boratyn, G. M., Schäffer, A. A., Agarwala, R., Altschul, S. F.,
Lipman, D. J., & Madden, T. L. (2012). Domain enhanced lookup
time accelerated BLAST. Biology Direct, 7(1), 12. https://doi.
org /10.1186/1745- 6150-7-12
Branca, A., Le ru, B. P., Vavre, F., Silvain, J.-F., & Dupas, S. (2011).
Intraspecific specialization of the generalist parasitoid Cotesia sesa-
miae revealed by polyDNAvirus polymorphism and associated with
different Wolbachia infection. Molecular Ecology, 20(5), 959–971.
https://doi.org/10.1111/j.1365-294X.2010.04977.x
Broman, K. W. (20 01). Review of statistical methods for QTL mapping in
experimental crosses. Lab Animal, 30(7), 44–52.
Broman, K. W., & Sen, S. (2009). A guide to QTL mapping with R/qtl.
Dordrecht, The Netherlands: Springer.
Burke, G. R., & Strand, M. R . (2014). Systematic analysis of a wasp par-
asitism arsenal. Molecular Ecology, 23(4), 890–901. https://doi.
org /10.1111/me c.12648
Burke, G. R., Thomas, S. A., Eum, J. H., & Strand, M. R. (2013). Mutualistic
polydnaviruses share essential replication gene functions with patho-
genic ancestors. PLoS Pa th, 9(5), e10033 48. https://doi.org/10.1371/
journ al.ppat.1003348
Burke, G. R., Walden, K. K. O., Whitfield, J. B., Robertson, H. M., & Strand,
M. R. (2014). Widespread genome reorganization of an obligate virus
mutualist. PLOS Genetics, 10 (9), e1004660. https://doi.org/10.1371/
journ al.pgen.1004660
Catchen, J., Hohenlohe, P. A., Bassham, S., Amores, A., & Cresko, W.
A. (2013). Stacks: An analysis tool set for population genomics.
Molecular Ecology, 22(11), 3124–3140. https://doi.org/10.1111/
mec.12354
Cavigliasso, F., Mathé-Hubert, H., Kremmer, L., Rebuf, C., Gatti, J. L.,
Malausa, T., … Poirié, M. (2019). Rapid and differential evolution of
the venom composition of a parasitoid wasp depending on the host
strain. Toxi n s, 11(11), 629. https://doi.org/10. 3390/toxin s1111 0629
Charnov, E. L., & Skinner, S. W. (1985). Complementary approaches to
the understanding of parasitoid oviposition decisions. Environmental
Entomology, 14(4), 383–391. https://doi.org/10.1093/ee/14.4.383
Chassain, C., & Bouletreau, M. (1987). Genetic variability in the egg-lay-
ing behaviour of Trichogramma maidis. Entomophaga, 32(2), 149–157.
https://doi.org/10.10 07/BF023 73125
16
|
BENOIST ET al.
Chevignon, G., Cambier, S., Da Silva, C., Poulain, J., Drezen, J.-M., Huguet,
E., & Moreau, S. J. M. (2015). Transcriptomic response of Manduca
sexta immune tissues to parasitization by the bracovirus associated
wasp Cotesia congregata. Insect Biochemistry and Molecular Biology,
62, 86–99. https://doi.org/10.1016/j.ibmb.2014.12.008
Chevignon, G., Periquet, G., Gyapay, G., Vega-Czarny, N., Musset, K.,
Drezen, J.-M., & Huguet, E. (2018). Cotesia congregata bracovirus
circles encoding PTP and Ankyrin genes integrate into the DNA of
parasitized Manduca sexta hemocytes. Journal of Virology, 92(15),
e00438-18. https://doi.org/10.1128/JVI.00438 -18
Chevignon, G., Thézé, J., Cambier, S., Poulain, J., Silva, C. D., Bézier, A .,
… Huguet, E. (2014). Functional annotation of Cotesia congregata
bracovirus: Identification of the viral genes expressed in parasitized
host immune tissues. Journal of V irolog y, 88, 8795–8812. https://doi.
org/10.1128/JVI.00209 -14
Churchill, G. A., & Doerge, R. W. (1994). Empirical threshold values for
quantitative trait mapping. Genetics, 138 (3), 963–971.
Colinet, D., Kremmer, L., Lemauf, S., Rebuf, C., Gatti, J.-L., & Poirié,
M. (2014). Development of RNAi in a Drosophila endoparasitoid
wasp and demonstration of its efficiency in impairing venom pro-
tein production. Journal of Insect Physiology, 63, 56–61. https://doi.
org/10.1016/j.jinsp hys.2014.02.011
Colinet, D., Mathé-Huber t, H., Allemand, R., Gatti, J.-L., & Poirié, M.
(2013). Variability of venom components in immune suppressive par-
asitoid wasps: From a phylogenetic to a population approach. Journal
of Insect Physiology, 59 (2), 205–212. https://doi.org/10.1016/j.jinsp
hys.2012.10.013
Crawley, M. J. (2012). The R bo ok (2nd ed.). Chichester, UK:
Wiley-Blackwell.
Danneels, E. L., Rivers, D. B., & de Graaf, D. C. (2010). Venom pro-
teins of the parasitoid wasp Nasonia vitripennis: Recent discovery
of an untapped pharmacopee. Toxi n s, 2(4), 494–516. https://doi.
org/10.3390/toxin s2040494
Darvasi, A. (1998). Experimental strategies for the genetic dissection
of complex traits in animal models. Nature Genetics, 18(1), 19–24.
https://doi.org/10.1038/ng 019 8-19
de Givry, S., Bouchez, M., Chabrier, P., Milan, D., & Schiex, T. (2005).
carhta gene : Multipopulation integrated genetic and radiation hybrid
mapping. Bioinformatics, 21(8), 1703–1704. https://doi.org/10.1093/
bioin forma tics/bti222
Drezen, J.-M., Leobold, M., Bézier, A., Huguet, E., Volkoff, A .-N., &
Herniou, E. (2017). Endogenous viruses of parasitic wasps: Variations
on a common theme. Current Opinion in Virology, 25, 41–48. https://
doi.org/10.1016/j.coviro.2017.07.002
Dubuffet, A ., Colinet, D., Anselme, C., Dupas, S., Carton, Y., & Poirié, M.
(2009). Variation of Leptopilina boulardi success in Drosophila hosts:
What is inside the black box? In Advances in parasitology (vol. 70, pp.
147–188). Elsevier. https://doi.org/10.1016/S0065 -308X(09)70006
-5
Dupas, S., & Car ton, Y. (1999). Two non-linked genes for specific vir-
ulence of Leptopilina boulardi against Drosophila melanogaster
and D. yakuba. Evolutiona ry Ecology, 13 (2), 211–220. https://doi.
org/10.1023/A:10066 91431658
Dupas, S., Frey, F., & Carton, Y. (1998). A single parasitoid segregating
factor controls immune suppression in Drosophila. The Journal of
Heredity, 89(4), 306–311. https://doi.org/10.1093/jhere d/89.4.306
Edson, K. M., Vinson, S. B., Stoltz, D. B., & Summers, M. D. (1981). Virus
in a parasitoid wasp: Suppression of the cellular immune response
in the parasitoid’s host. Science, 211(4482), 582–583. https://doi.
org/10.1126/scien ce.7455695
Falabella, P., Varricchio, P., Provost, B., Espagne, E., Ferrarese, R.,
Grimaldi, A., … Pennacchio, F. (2007). Characterization of the I B-like
gene family in polydnaviruses associated with wasps belonging to
different Braconid subfamilies. Journal of General V irolog y, 88(1),
92–104. https://doi.org/10.1099/vir.0.82306 -0
Flint, J., Greenspan, R. J., & Kendler, K. S. (2010). How genes influence
behavior (1st ed.). Oxford, UK and New York, NY: Oxford University
Press.
Friedrich, P., & Bozóky, Z. (2005). Digestive versus regulatory proteases:
On calpain action in vivo. Biological Chemistry, 386(7), 609–612.
https://doi.org/10.1515/BC.2005.071
Gauthier, J., Boulain, H., van Vugt, J. J. F. A., Baudry, L., Persyn, E., Aury,
J.-M., & Drezen, J.-M. (2020). Chromosomal resolution reveals sym-
biotic virus colonization of parasitic wasp genomes. BioRxiv. h tt ps: //
doi.org/10.1101/2020.03.19.994459
Gauthier, J., Gayral, P., LeRu, B. P., Jancek, S., Dupas, S., Kaiser, L., …
Herniou, E. (2018). Genetic footprints of adaptive divergence in the
bracovirus of Cotesia sesamiae identified by targeted re-sequenc-
ing. Molecular Ecology, 27, 2109–2123. https://doi.org/10.1111/
mec.145 74
Gitau, C. W., Gundersen-Rindal, D., Pedroni, M., Mbugi, P. J., & Dupas,
S. (2007). Differential expression of the CrV1 haemocyte inactiva-
tion-associated polydnavirus gene in the African maize stem borer
Busseola fusca (Fuller) parasitized by two biotypes of the endopara-
sitoid Cotesia sesamiae (Cameron). Journal of Insec t Physiology, 53(7),
676–684. https://doi.org/10.1016/j.jinsp hys.2007.04.008
Godfray, H. C. J. (1994). Parasitoids: Behavioral and evolutionary ecology.
Princeton, NJ: Princeton University Press.
Gueguen, G., Kalamarz, M. E., Ramroop, J., Uribe, J., & Govind, S. (2013).
Polydnaviral ankyrin proteins aid parasitic wasp survival by coordi-
nate and selective inhibition of hematopoietic and immune NF-kappa
B signaling in insect hosts. PLoS Path, 9(8), e1003580. https://doi.
org/10.1371/journ al.ppat.1003580
Hansson, B., Sigeman, H., Stervander, M., Tarka, M., Ponnikas, S.,
Strandh, M., … Hasselquist, D. (2018). Contrasting results from
GWAS and QTL mapping on wing length in great reed war-
blers. Molecular Ecology Resources, 18 (4), 867–876. https://doi.
org /10.1111/1755-0998.12785
Heimpel, G. E., & de Boer, J. G. (2008). Sex determination in the
Hymenoptera. Annual Review of Entomology, 53(1), 209–230. https://
doi.org/10.1146/annur ev.ento.53.103106.093441
Henter, H. J. (1995). The potential for coevolution in a host-parasitoid
system. ii. Genetic variation within a population of wasps in the abil-
ity to parasitize an aphid host. Evolution, 49(3), 439–445. https://doi.
org /10.1111/j.1558-5646.1995.t b022 76.x
Herniou, E., Huguet, E., Theze, J., Bézier, A., Periquet, G., & Drezen,
J.-M. (2013). When parasitic wasps hijacked viruses: Genomic and
functional evolution of polydnaviruses. Philosophical Transactions
of the Royal Society B: Biological Sciences, 368(16 26) , h tt p s://d oi.
org/10.1098/rstb.2013.0051
Herniou, E., Olszewski, J. A., Cory, J. S., & O’Reilly, D. R. (2003). The
genome sequence and evolution of baculoviruses. Annual Review
of Entomology, 48(1), 211–234. https://doi.org/10.1146/annur
ev.ento.48.091801.112756
Jancek, S., Bézier, A., Gayral, P., Paillusson, C., Kaiser, L., Dupas, S., …
Herniou, E. A. (2013). Adaptive selection on bracovirus genomes
drives the specialization of Cotesia parasitoid wasps. PLoS One, 8(5),
https://doi.org/10.1371/journ al.pone.0064432
Kaiser, L., Couty, A., & Perez-Maluf, R. (2009). Dynamic use of fruit
odours to locate host larvae: Individual learning, physiological
state and genetic variability as adaptive mechanisms. Advances
in Parasitolog y, 70, 67–95. https://doi.org/10.1016/S0065
-308X(09)70003 -X
Kaiser, L., Dupas, S., Branca, A., Herniou, E. A., Clarke, C. W., Capdevielle
Dulac, C., … Le Ru, B. P. (2017). The Cotesia sesamiae story: Insight
into host-range evolution in a Hymenoptera parasitoid and implica-
tion for its use in biological control programs. Genetica, 145(6), 455–
468. https://doi.org/10.1007/s1070 9-017-9989-3
Kaiser, L., Fernandez-Triana, J., Capdevielle-Dulac, C., Chantre, C.,
Bodet, M., Kaoula, F., … Le Ru, B. (2017). Systematics and biology of
|
17
BENOIS T ET al.
Cotesia typhae sp. n. (Hymenoptera, Braconidae, Microgastrinae), a
potential biological control agent against the noctuid Mediterranean
corn borer, Sesamia nonagrioides. ZooKeys, 682, 105–136. https://doi.
org/10.3897/zooke ys.682.13016
Kaiser, L., Le Ru, B. P., Kaoula, F., Paillusson, C., Capdevielle-Dulac, C.,
Obonyo, J. O., … Dupas, S. (2015). Ongoing ecological speciation in
Cotesia sesamiae, a biological control agent of cereal stem borers.
Evolutionary Applications, 8(8), 807–820. https://doi.org/10.1111/
eva.12260
Kapranas, A., Tena, A ., & Luck, R. F. (2012). Dynamic virulence in a par-
asitoid wasp: The influence of clutch size and sequential oviposition
on egg encapsulation. Animal Behaviour, 83(3), 833–838. https://doi.
org/10.1016/j.anbeh av.2012.01.004
Kitano, H. (1986). The role of Apanteles glomeratus venom in the defensive
response of its host, Pieris rapae crucivora. Jour nal of Insect Physio logy,
32(4), 369–375. https://doi.org/10.1016/0022-1910(86)90050 -8
Kwon, B., & Kim, Y. (2008). Transient expression of an EP1-like gene en-
coded in Cotesia plutellae bracovirus suppresses the hemocyte pop-
ulation in the diamondback moth, Plutella xylostella. Developmental &
Comparative Immunology, 32(8), 932–942. https://doi.org/10.1016/j.
dci.2008.01.005
Lander, E. S., & Botstein, D. (1989). Mapping mendelian factors underlying
quantitative traits using RFLP linkage maps. Genetics, 121(1), 185–199.
Larièpe, A., Mangin, B., Jasson, S., Combes, V., Dumas, F., Jamin, P., …
Moreau, L. (2012). The genetic basis of heterosis: Multiparental
quantitative trait loci mapping reveals contrasted levels of apparent
overdominance among traits of agronomical interest in maize (Zea
mays L.). Genetics, 190(2), 795–811. https://doi.org/10.1534/genet
ics.111.133447
Le Masurier, A. D. (1987). A comparative study of the relationship be-
tween host size and brood size in Apanteles spp. (Hymenoptera:
Braconidae). Ecological Entomology, 12(4), 383–393. https://doi.
org /10.1111/j.1365-2311 .1987.tb010 19.x
Le Masurier, A. D. (1991). Effect of host size on clutch size in Cotesia glomerata.
The Journal of Animal Ecology, 60(1), 107. https://doi.org/10.2307/5448
Legner, E. F. (1987). Inheritance of gregarious and solitary oviposi-
tion in Muscidifurax raptorellus kogan and legner (Hymenoptera:
Pteromalidae). The Canadian Entomologist, 119(9), 791–808. https://
doi.org/10.4039/Ent11 9791-9
Li, H., Handsaker, B., Wysoker, A ., Fennell, T., Ruan, J., & Homer, N., …
1000 Genome Project Data Processing Subgroup (2009). The se-
quence alignment/map format and SAMtools. Bioinformatics, 25(16),
2078–2079. https://doi.org/10.1093/bioin forma tics/btp352
Mack ay, T. F. C. , St one, E. A., & Ayrol es, J. F. (2 009). The gene tics of qua n-
titative traits: Challenges and prospects. Nature Reviews Genetics,
10(8), 565–577. https://doi.org/10.1038/nrg2612
Marco Antonio, D. S., Guidugli-Lazzarini, K. R ., do Nascimento, A. M.,
Simões, Z. L. P., & Hartfelder, K. (2008). RNAi-mediated silencing
of vitellogenin gene function turns honeybee (Apis mellifera) work-
ers into extremely precocious foragers. Naturwissenschaften, 95 (10),
953–961. https://doi.org/10.1007/s0011 4-008-0413-9
Marti, D., Grossniklaus-Bürgin, C., Wyder, S., Wyler, T., & Lanzrein, B.
(2003). Ovary development and polydnavirus morphogenesis in the
parasitic wasp Chelonus inanitus. I. Ovary morphogenesis, amplifica-
tion of viral DNA and ecdysteroid titres. Journal of General Virology,
84(5), 1141–1150. https://doi.org/10.1099/vir.0.18832 -0
Martin, M. (2011). Cutadapt removes adapter sequences from
high-throughput sequencing reads. EMBnet.Journal, 17 (1), 10.
https://doi.org/10.14806/ ej.17.1.200
Mochiah, M. B., Ngi-Song, A. J., Overholt, W. A., & Stouthamer, R. (2002).
Wolbachia infection in Cotesia sesamiae (Hymenoptera: Braconidae)
Causes cytoplasmic incompatibility: Implications for biological con-
trol. Biological Control, 25(1), 74–80. https://doi.org/10.1016/S1049
-9644(02)00045 -2
Moreau, S., & Asgari, S. (2015). Venom proteins from parasitoid wasps
and their biological functions. Toxins, 7(7), 2385–2412. https://doi.
org/10.3390/toxin s7072385
Mrinalini, A., & Werren, J. H. (2016). Parasitoid wasps and their venoms. In
P. Gopalakrishnakone, & A. Malhotra (Eds.), Evolution of venom ous ani-
mals and th eir toxins (pp. 1–26) . Dordrech t: Springer S cience +Business
Media. https://doi.org/10.1007/978-94-007-6727-0_2-2
Na chtergaele, S ., & He, C. (2017 ). The eme rging biolog y of RN A po s t- t r a n -
scriptional modifications. RNA Biology, 14(2), 156–163. https://doi.
org /10.1080/15476 286.2016.1267096
Nurk, S., Bankevich, A., Antipov, D., Gurevich, A., Korobeynikov, A.,
Lapidus, A., … Pevzner, P. A. (2013). Assembling genomes and
mini-metagenomes from highly chimeric reads. In M. Deng, R. Jiang,
F. Sun, & X. Zhang (Eds.), Research in computational molecular biology
(pp. 158–170). Berlin, Heidelberg: Springer.
Or zack, S. H., & Gla dstone, J. (1994). Qua nti t ati ve gene tic s of sex rati o tra its
in the parasitic wasp, Nasonia vitripennis. Genetics, 137(1), 211–220.
Pannebakker, B. A., Trivedi, U., Blaxter, M. A., Watt, R., & Shuker, D. M.
(2013). The transcriptomic basis of oviposition behaviour in the par-
asitoid wasp Nasonia vitripennis. PLoS One, 8(7), e68608. https://doi.
org/10.1371/journ al.pone.0068608
Pannebakker, B. A., Watt, R., Knott, S. A., West, S. A., & Shuker, D. M.
(2011). The quantitative genetic basis of sex ratio variation in Nasonia
vitripennis: A QTL study. Journal of Evolutionary Biology, 24(1), 12–22.
https://doi.org/10.1111/j.1420-9101.2010.02129.x
Pennacchio, F., & Strand, M. R. (2006). Evolution of develop-
mental strategies in parasitic Hymenoptera. Annual Review of
Entomology, 51(1), 233–258. https://doi.org/10.1146/annur
ev.ento.51.110104.151029
Pichon, A., Bézier, A., Urbach, S., Aury, J.-M., Jouan, V., Ravallec, M., …
Volkoff, A.-N. (2015). Recurrent DNA virus domestication leading
to different parasite virulence strategies. Science Advances, 1(10),
e1501150. https://doi.org/10.1126/sciadv.1501150
Pruijssers, A . J., & Strand, M. R. (2007). PTP-H2 and PTP-H3 from
Microplitis demolitor bracovirus localize to focal adhesions and are
antiphagocytic in insect immune cells. Journal of Virology, 81(3),
1209–1219. https://doi.org/10.1128/JVI.02189 -06
Quicke, D. L. J. (2015). The braconid and ichneumonid parasitoid wasps:
Biology, systematics, evolution and ecology. John Wiley & Sons.
https://doi.org/10.1002/97811 18907085
R Core Team (2018). R: A language and environment for statistical com-
puting. Vienna, Austria: R Foundation for Statistical Computing.
Retrieved from https://www.R-proje ct.org/
Rochette, N. C., & Catchen, J. M. (2017). Deriving genotypes from R AD-
seq short-read data using Stacks. Nature Protocols, 12(12), 264 0–
2659. https://doi.org/10.1038/nprot.2017.123
Rosenheim, J. A., & Hongkham, D. (1996). Clutch size in an obligately
siblicidal parasitoid wasp. Animal Behaviour, 51(4), 841–852. https://
doi.org/10.1006/anbe.1996.0088
Santure, A. W., & Garant, D. (2018). Wild GWAS-association mapping
in natural populations. Molecular Ecology Resources, 18(4), 729–738.
https://doi.or g/10.1111/1755-0998.12901
Sim, A. D., & Wheeler, D. (2016). The venom gland transcriptome of
the parasitoid wasp Nasonia vitripennis highlights the importance of
novel genes in venom function. BMC Genomics, 17(1), 571. https://
doi.org/10.1186/s1286 4-016-2924-7
Stapley, J., Feulner, P. G. D., Johnston, S. E., Santure, A. W., & Smadja,
C. M. (2017). Variation in recombination frequency and distribution
across eukaryotes: Patterns and processes. Philosophical Transactions
of the Royal Society of London. Series B, Biological Science s, 372(1736),
20160 455. https://doi.org/10.1098/rstb.2016.0 455
Strand, M. R. (2014). Teratocytes and their functions in parasitoids.
Current Opinion in Insect Science, 6, 68–73. https://doi.org/10.1016/j.
cois.2014.09.005
18
|
BENOIST ET al.
Strand, M. R., & Burke, G. R. (2015). Polydnaviruses: From discov-
ery to current insights. Virology, 479–48 0, 393–402. https://doi.
org/10.1016/j.virol.2015.01.018
Tanaka, T. (1987). Calyx and venom fluids of Apanteles kariyai
(Hymenoptera: Braconidae) as factors that prolong larval period of
the host, Pseudaletia separata (Lepidoptera: Noctuidae). Annals of
the Entomological Society of America, 80(4), 530–533. https://doi.
org/10.1093/aesa/80.4.530
Tanksley, S. D. (1993). Mapping polygenes. Annual Review of Genetics, 27,
205–233. https://doi.org/10.1146/annur ev.ge.27.120193.001225
Tellier, A., & Brown, J. K. M. (2007). Polymorphism in multilocus host–
parasite coevolutionary interactions. Genetics, 177(3), 1777–1790.
https://doi.org/10.1534/genet ics.107.074393
Thoetkiattikul, H., Beck, M. H., & Strand, M. R. (2005). Inhibitor kap-
paB-like proteins from a polydnavirus inhibit NF-kappaB activation
and suppress the insect immune response. Proceedings of the National
Academy of Sciences of the United States of America, 102(32), 11426–
11431. https://doi.org/10.1073/pnas.05052 40102
Thompson, J. N. (2009). The coevolutionary process. Chicago, IL:
University of Chicago Press.
Thompson, J. N., & Burdon, J. J. (1992). Gene-for-gene coevolution be-
tween plants and parasites. Nature, 360(6400), 121–125. https://doi.
org/10.1038/360121a0
Ueno, T., & Tanaka, T. (1997). Comparison between primary and second-
ary sex ratios in parasitoid wasps using a method for observing chro-
mosomes. Entomologia Experimentalis et Applicata, 82(1), 105–108.
https://doi.org/10.1046/j.1570-7458.1997.00119.x
Velth ui s, B.-J., Yang , W., Van Opijnen, T., & Werr en , J. H. (2005). Gen et ics
of female mate discrimination of heterospecific males in Nasonia
(Hymenoptera, Pteromalidae). Animal Behaviour, 69(5), 1107–1120.
https://doi.org/10.1016/j.anbeh av.2004.06.032
Vieira, V., Cardoso, M. A., & Araujo, H. (2017). Calpain A controls mi-
totic synchrony in the Drosophila blastoderm embryo. Mechanisms
of Development, 14 4, 141–149. https://doi.o rg /10.1016/j.
mod.2016.05.005
Volkoff, A.-N., Jouan, V., Urbach, S., Samain, S., Bergoin, M., Wincker,
P., … Drezen, J.-M. (2010). Analysis of virion structural components
reveals vestiges of the ancestral ichnovirus genome. PLoS Path, 6(5),
e1000923. https://doi.org /10.1371/journ al.ppat.1000923
Waage, J. K., & Greathead, D. J. (Eds.) (1986). Insect parasitoids. London
and Orlando: Academic Press.
Waage, J. K., & Ming, N. S. (1984). The reproductive strategy of a par-
asitic wasp: I. Optimal progeny and sex allocation in Trichogramma
evanescens. The Journal of A nimal Ecology, 53(2), 401. https://doi.
org/10.2307/4524
Wang, Z., Liu, Y., Shi, M., Huang, J., & Chen, X. (2019). Parasitoid wasps as
effective biological control agents. Journal of Integrative Agriculture,
18(4), 705–715. https://doi.org/10.1016/S2095 -3119(18)62078 -7
Webb, B. A. (1998). Polydnavirus biology, genome structure, and evolution.
In L. K. Miller, & L. A. Ball (Eds.), The insect viruses (pp. 105–139). Boston,
MA: Springer. https://doi.org/10.1007/978-1-4615-5341-0_5
Wetterwald, C., Roth, T., Kaeslin, M., Annaheim, M., Wespi, G., Heller,
M., … Lanzrein, B. (2010). Identification of bracovirus particle pro-
teins and analysis of their transcript levels at the stage of virion for-
mation. Journal of General Virology, 91(10) , 2610–2619. ht tps ://doi.
org/10.1099/vir.0.02269 9-0
Woolhouse, M. E. J., Webster, J. P., Domingo, E., Charlesworth, B., &
Levin, B. R. (2002). Biological and biomedical implications of the
co-evolution of pathogens and their hosts. Nature Genetics, 32(4),
569–577. https://doi.org/10.1038/ng120 2-569
Wyler, T., & Lanzrein, B. (2003). Ovary development and polydna-
virus morphogenesis in the parasitic wasp Chelonus inanitus. II.
Ultrastructural analysis of calyx cell development, virion formation
and release. Journal of Ge neral Virolog y, 84(5), 1151–1163. http s: //doi.
org/10.1099/vir.0.18830 -0
Xu, Y., & Crouch, J. H. (2008). Marker-assisted selection in plant breed-
ing: From publications to practice. Crop Science, 48(2), 391–407.
https://doi.org/10.2135/crops ci2007.04.0191
SUPPORTING INFORMATION
Additional supporting information may be found online in the
Supporting Information section.
How to cite this article: Benoist R, Capdevielle-Dulac C,
Chantre C, et al. Quantitative trait loci involved in the
reproductive success of a parasitoid wasp. Mol Ecol.
2020;00:1–18. https://doi.org/10.1111/mec.15567
