Download Endosymbiont metacommunities, mtDNA diversity and the

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Molecular Ecology (2010)
doi: 10.1111/j.1365-294X.2010.04775.x
Endosymbiont metacommunities, mtDNA diversity and
the evolution of the Bemisia tabaci (Hemiptera:
Aleyrodidae) species complex
G W É N A E L L E G U E G U E N , * 1 F A B R I C E V A V R E , * O L I V I E R G N A N K I N E , † M I C H E L
PETERSCHMITT,‡ DELPHINE CHARIF,* ELAD CHIEL,§– YUVAL GOTTLIEB,** MURAD
G H A N I M , * * E I N A T Z C H O R I - F E I N § and F R É D É R I C F L E U R Y *
*Université de Lyon; Université Lyon 1; CNRS; UMR 5558, Laboratoire de Biométrie et Biologie Evolutive, 43 Boulevard du 11
Novembre 1918, Villeurbanne F-69622, France, †Laboratoire d’Entomologie Fondamentale et Appliquée, Département de Biologie
et Physiologie Animales, Université de Ouagadougou, 03 BP 7021 Ouagadougou 03, ‡CIRAD, UMR BGPI, TA A54 ⁄ K,
Campus International de Baillarguet, Montpellier Cedex, France, §The Agricultural Research Organization (ARO), Department
of Entomology, Newe Ya’ar Research Center, PO Box 1021, Ramat Yishay 30095, Israel, –Department of Evolutionary and
Environmental Biology, University of Haifa, Haifa, Israel, **The Agricultural Research Organization (ARO), Department of
Entomology, Volcani Center, PO Box 6, Bet-Dagan 50250, Israel
Abstract
Bemisia tabaci, an invasive pest that causes crop damage worldwide, is a highly
differentiated species complex, divided into biotypes that have mainly been defined
based on mitochondrial DNA sequences. Although endosymbionts can potentially
induce population differentiation, specialization and indirect selection on mtDNA,
studies have largely ignored these influential passengers in B. tabaci, despite as many as
seven bacterial endosymbionts have been identified. Here, we investigate the composition of the whole bacterial community in worldwide populations of B. tabaci, together
with host genetic differentiation, focusing on the invasive B and Q biotypes. Among 653
individuals studied, more than 95% of them harbour at least one secondary endosymbiont, and multiple infections are very common. In addition, sequence analyses reveal a
very high diversity of facultative endosymbionts in B. tabaci, with some bacterial genus
being represented by more than one strain. In the B and Q biotypes, nine different strains
of bacteria have been identified. The mtDNA-based phylogeny of B. tabaci also reveals a
very high nucleotide diversity that partitions the two ITS clades (B and Q) into six CO1
genetic groups. Each genetic group is in linkage disequilibrium with a specific
combination of endosymbionts. All together, our results demonstrate the rapid dynamics
of the bacterial endosymbiont–host associations at a small evolutionary scale, questioning the role of endosymbiotic communities in the evolution of the Bemisia tabaci species
complex and strengthening the need to develop a metacommunity theory of inherited
endosymbionts.
Keywords: Arsenophonus, community genetics, Hamiltonella, Rickettsia, Wolbachia
Received 1 March 2010; revision revised 3 May 2010; accepted 4 May 2010
Introduction
Correspondence: Frédéric Fleury, Fax: (33) 4 72 43 113 88;
E-mail: [email protected]
1
Present address: City College of New York, Biology department-Marshak building, 160 convent avenue, New York, 10031
NY, USA.
2010 Blackwell Publishing Ltd
The sweet potato whitefly, Bemisia tabaci (Hemiptera:
Aleyrodidae), is a worldwide invasive agricultural pest
recognized as a species complex consisting of morphologically indistinguishable entities, among which the
limits of reproductive barriers remain unclear (Frohlich
2 G. GUEGUEN ET AL.
et al. 1999; Perring 2001). This species diversity merge
with the fact that B. tabaci is divided into several biotypes or host-races differing by a number of ecologically and economically important traits such as host
plant range, induced plant physiological disorders,
insecticide resistance and virus transmission (Perring
2001). Allozymes analysis led to recognize more than
20 biotypes named A–S. Some of these biotypes were
thereafter recognized as true species based on genetic
isolation, as for example the new world A and the old
world B biotypes (Perring 2001; De Barro et al. 2005),
while others probably belong to the same species. This
confusing taxonomic status of B. tabaci populations
and frequent invasions on all continents with important associated economic losses has stimulated lively
debates about phylogeny, biotype identification and
geographical distribution of B. tabaci (De Barro et al.
2000; Boykin et al. 2007). Molecular markers (nuclear
and predominantly mtDNA) have been used extensively to define biotypes without considering any other
biological attributes, thus contributing to blur the transition from geographic populations to species. Based
on mitochondrial CO1 gene sequences, four major
phylogenetic clades are recognized (Boykin et al. 2007):
the Africa nonsilverleafing clade, the Uganda clade
specialized on sweetpotato, the Asia ⁄ New World ⁄ Australia clade and the Mediterranean ⁄ Africa silverleafing
clade. Using Bayesian phylogenetic methods on
numerous sequences of CO1 gene, Boykin et al. (2007)
have split these four clades into 12 well-resolved
genetic groups that probably represent different cryptic
species. The Mediterranean ⁄ Africa silverleafing clade
evidences four groups of populations that separate the
ubiquitous and invasive B and Q biotypes among
which reproductive barriers were observed (Moya
et al. 2001), the MS biotype localized in Indian ocean
(Delatte et al. 2005) and the Africa silverleafing populations (ASL group) (Boykin et al. 2007). However,
CO1 diversity is still observed within these groups,
and species or biotypes delineation and identification
using CO1 markers are strenuous without any other
biological data about interbreeding or specific ecological specialization. For example, the Q biotype displays
differentiated genetic groups with an apparently accelerated rate of divergence among mitochondrial lineages leading to define several haplotypes, whose
phenotypic traits remain unknown (Boykin et al. 2007;
Chu et al. 2008; McKenzie et al. 2009). This high level
of divergence in the mitochondrial lineages of the
B. tabaci species complex calls into question the phylogenetic and biological relevance of these genetic
groups, the extent to which they are reproductively
isolated, and the mechanisms responsible for mtDNA
evolution. Indeed, it is now well recognized that
mtDNA is not the ideal marker for molecular taxonomy because of the presence of recombination events,
variation in mutation rate and departure from neutrality that can result from the presence of selfish intracytoplasmic elements such as endosymbiotic bacteria
inducing selective sweeps (Shoemaker et al. 2004;
Hurst & Jiggins 2005).
Despite the ability of inherited bacterial endosymbionts both to induce indirect selection on insect
mtDNA and to contribute to population differentiation
up to species formation (Hurst & Jiggins 2005; Werren
et al. 2008), very few studies have taken into consideration the presence of several vertically transmitted
bacteria in the B. tabaci species complex. Maternal transmission of these bacteria, together with their linkage disequilibrium with mitochondria, may indeed induce
selective sweeps that indirectly impact the polymorphism of mtDNA when endosymbionts spread within a
host population. As a consequence, mtDNA-based
phylogenies may reflect endosymbiont rather than host
history. Moreover, recent studies have revealed that
arthropods can host more than one endosymbiont with
an unexpected wide diversity of host alterations (Haine
2008; Werren et al. 2008). Taking the whole endosymbionts community into account is especially important to understand B. tabaci complex, which has one of
the highest number of endosymbiotic elements, with
seven different vertically transmitted bacteria reported
so far (Zchori-Fein & Brown 2002). Portiera aleyrodidarum,
the obligatory primary endosymbiont of whiteflies,
provides its host with essential nutrients and has a long
co-evolutionary history with all members of the Aleyrodinae subfamily (Thao & Baumann 2004b). Six additional
facultative secondary endosymbionts, the effects of
which have yet to be determined, have also been
detected (Zchori-Fein & Brown 2002; Nirgianaki et al.
2003). Four of them (Wolbachia, Cardinium, Rickettsia and
Arsenophonus) are known to manipulate host reproduction in a wide range of insect species (Duron et al. 2008;
Werren et al. 2008). One (Hamiltonella defensa) induces
parasitoid resistance in the pea aphid (Oliver et al. 2002),
and one (Fritschea bemisiae) has unknown effect and has
so far only been reported in B. tabaci (Thao et al. 2003).
Although B. tabaci hosts all these bacteria with indications for nonrandom distribution among biotypes
(Chiel et al. 2007), the diversity of the whole secondary
endosymbiotic community and its variation at different
geographic and phylogenetic scales remains unknown.
Here, we analyse the diversity, prevalence and distribution of all known endosymbionts in different B. tabaci
populations at the worldwide scale, focusing on the
Mediterranean ⁄ Africa silverleafing clade, and we correlate the composition of the symbiotic community with
CO1 mtDNA diversity.
2010 Blackwell Publishing Ltd
ENDOSYMBIONT ASSEMBLAGE IN A SPECIES COMPLEX 3
Material and methods
Detection and molecular identification of
endosymbionts
Six hundred and fifty-three B. tabaci individuals (including 355 individuals from Israel, Chiel et al. 2007) were
collected from greenhouses or field worldwide, except
for two lines from La Reunion, which were laboratory
strains (Table S1, Supporting information). Individuals
were collected from different continents and host plants
(including a large sample from Africa). They all belong to
the Mediterranean ⁄ Africa silverleafing clade except few
individuals from Cambodia (Asian clade). Adults were
placed alive in ethanol (96%) and conserved at )20 C.
For most samples, DNA was extracted from individual whiteflies in 75 lL of chelex and stored at )20 C
(Walsh et al. 1991). For the last collected samples from
Africa, individuals were ground in 60 lL of Nonidet
extraction buffer (Delatte et al. 2005) and stored at
)20 C. DNA extraction methods were compared and
gave similar results. Individuals were screened for
endosymbiont infection using specific PCR primers targeting the 16S rRNA gene for P. aleyrodidarum, Hamiltonella, Cardinium and Rickettsia, the 23S rRNA gene for
Arsenophonus and Fritschea, and the wsp gene for Wolbachia (Table S2, Supporting information). All individuals
were tested for the presence of P. aleyrodidarum, the
obligate endosymbiont, to check for DNA extraction
quality. PCR analyses targeting symbiotic bacteria were
conducted as described in Table S2. Two multiplex
PCRs were developed, the first targeting Hamiltonella ⁄ Rickettsia and the second Cardinium ⁄ Wolbachia. DNA
was amplified in a final volume of 25 lL, containing 1·
of taq buffer, 1.5 mM MgCl2, 200 lM dNTPs, 200 nM of
each primer, 1.2 U taq polymerase and 2 lL of DNA.
For Hamiltonella ⁄ Rickettsia, the PCR program consisted
of an initial phase at 95 C lasting 3 min, followed by
35 cycles of 30 s at 95 C, 30 s at 58 C and 45 s at
72 C, with a final extension phase of 5 min at 72 C.
For Cardinium ⁄ Wolbachia, the PCR program was 92 C
for 1:30 min followed by 34 cycles of 30 s at 92 C, 40 s
at 56 C and 1:15 min at 72 C and a final extension of
10 min at 72 C (Table S2). Suitable positive controls
have been used and successfully amplified for each
endosymbiont and each PCR.
To display the prevalence of each endosymbiont
together with individual co-infections on the same
graph, a new graphical drawing was developed. Each
graph represents one population with different symbiotic bacterial strains on the x-axis. On the y-axis, host
individuals are ranked and grouped together according
to their infection status (noninfected, single-infection or
the composition of their bacterial assemblage). For each
2010 Blackwell Publishing Ltd
category of bacterial infection (x-axis), the coloured bars
indicate the prevalence of a given endosymbiont within
the population. When the graph is read horizontally,
according to y-axis, combinations of the different colours represent individuals sharing the same endosymbiotic community. When several mitochondrial genetic
groups of B. tabaci were found within a population, a
new x-axis category was added representing the host
genetic group; the percentage of each genetic group is
indicated using the same representation. All graphs
were constructed using R software (R Development
Core Team 2003).
To reconstruct endosymbiont phylogenies, at least
one sequence was obtained for each endosymbiont in
each genetic group where it had been detected. 23S
rRNA was sequenced for Arsenophonus, fbpA, ftsZ and
gatB for Wolbachia, sca1 and gltA for Rickettsia (Roux &
Raoult 1995; Ngwamidiba et al. 2006), gyrB for Hamiltonella and 16S rRNA for Cardinium (Table S2). These
choices were based on the availability of sequences for
phylogenetic reconstruction. PCR conditions were those
described in Table S2. PCR products were purified on
columns (NucloSpin Extract II; Macherey-Nagel) and
directly sequenced by Genoscreen (Lille, France).
Sequence analyses revealed that more than one bacterial
strain of Wolbachia, Cardinium and Arsenophonus infects
the B. tabaci complex. This bacterial strain diversity
should impose their identification by sequencing in all
host individuals but this falls beyond the scope of this
study. Strains identification was only performed for
Arsenophonus by sequencing the PCR products of all
individuals originating from African populations. For
all other bacteria and because the same sequence was
observed among populations belonging to the same
CO1 genetic group, we assumed that only one strain
infects one B. tabaci cytotype in one locality. This probably underestimates the whole bacterial diversity but
do not change our conclusions about the link between
bacterial assemblage and host genetic group.
All sequences were deposited in Genbank under
accession numbers from FJ766366 to FJ766372 for Arsenophonus, FJ766335 to FJ766342 for Cardinium, FJ766343 to
FJ766351 for Hamiltonella, FJ766352 to FJ766354 and
FJ766355 for Rickettsia gltA and sca1, respectively, and
FJ766356 to FJ766360, FJ766361 to FJ766365 and FJ766373
to FJ766375 for Wolbachia fbpA, ftsZ and gatB respectively.
Molecular identification of B. tabaci individuals and
sequencing
At least one individual per population and per infection
status was sequenced for mitochondrial CO1 and
nuclear ITS1. Sequences were used to reconstruct
4 G. GUEGUEN ET AL.
phylogenies and identify B. tabaci genetic groups. PCR
conditions and primers used for ITS1 and CO1 were
those described in (De Barro et al. 2000; Khasdan et al.
2005), respectively. PCR products were purified on
columns (NucloSpin Extract II; Macherey-Nagel) and
directly sequenced by Genoscreen (Lille, France). All
sequences were deposited in Genbank under accession
numbers from EU760702 to EU760718 and FJ766376 to
FJ766380 for ITS1 and EU760719 to EU760761 and
FJ766381 to FJ766437 for CO1. Sequences were also used
to estimate level of mtDNA and ITS1 polymorphism by
calculating the number of segregating sites, nucleotide
diversity, p (Tajima 1983; Nei 1987) and hw (Watterson
1975; Nei 1987). All these parameters were calculated
using DNASP software (Rozas et al. 2003).
Phylogenetic analyses
Mitochondrial CO1, nuclear ITS1 and endosymbionts
sequences obtained during this study and previously
reported sequences from Genbank were aligned with
Clustalw (Thompson et al. 1994). All phylogenetic trees
were constructed by the maximum-likelihood method
using PhymL software (Guindon & Gascuel 2003). Substitution models were determined by Modeltest, using
the AIC criterion for each alignment (Posada & Crandall 1998). ML trees were constructed using a HKY85
substitution model for CO1 and Arsenophonus, a JC69
substitution model for ITS1, Hamiltonella and Wolbachia
(Jukes & Cantor 1969) and a GTR+G model for Cardinium (Tavaré 1986). For Rickettsia, owing to a too high
divergence between Rickettsia from B. tabaci and the
Rickettsia group used as outgroup for the phylogeny,
sequences were translated using Transeq (Rice et al.
2000). Phylogeny was constructed on the protein
sequences using a jtt model of substitution (Jones et al.
1992). These models incorporate gamma-distributed rate
heterogeneity between sites. The robustness of clades
was assessed by 500 bootstrap replicates using Seqboot
(Felsenstein 1993).
Statistical analyses of association between B. tabaci
genetic groups and endosymbionts
Three correspondence analyses (CA) were performed to
test whether there was any correlation between host
genetic groups (based either on CO1 or on ITS1), and
composition of endosymbiotic communities (Benzecri
1973; Lobry & Chessel 2003). All computations were
carried out using ADE4 package (dudi.coa function)
(Thioulouse et al. 1997) in R software (R development
Core Team 2003). We first constructed a table in which
rows represent different populations and columns correspond to all infection statuses and symbiotic assemblages observed. When several genetic groups were
found in a population (Burkina Faso and Cameroon),
individuals were grouped according to them. This rowblock structure was used to perform the first CA (global
analysis), which made possible to represent on a factorial map the projection of rows and columns that maximizes the variation observed among populations. To
find out whether ITS1 and CO1 groupings are associated with a particular symbiotic community, two additional CA were performed to separate within- and
between-group variability. Populations were assigned
to a genetic group using either CO1 or ITS1. To obtain
the between-genetic group analysis, all rows corresponding to the same genetic group were summed, and
the correspondence analysis performed on the resulting
table retained only the between-genetic group variability. Because the sum of the variability explained by the
between- and the within-genetic groups analyses is
equal to the variability of the global analysis, the
within-genetic group variability was obtained by subtraction. Significance of the between-genetic group analysis was tested by Monte Carlo test with 1000
repetitions. If most of the symbiotic infection variability
is found between genetic groups, genetic assignation
and symbiotic corteges are strongly correlated, whereas
if most of the variability is found within genetic groups,
symbiotic infection and genetic groups are independent.
Results
Bemisia tabaci phylogenies based on nuclear and
mtDNA
The mitochondrial (CO1)- and nuclear (ITS1)-based
phylogenies (Figs 1 and 2) did not reveal major discrepancies and are consistent with previous studies (De
Barro et al. 2000; Boykin et al. 2007). However, CO1
mtDNA, which is the most common marker used to differentiate B. tabaci biotypes, is highly variable within
the Mediterranean ⁄ Africa silverleafing clade (86 segregating sites in 744 bp of sequence, estimated nucleotide
diversity: p = 0.02776 and hw = 0.02542) compared with
Fig. 1 CO1 gene–based phylogeny of B. tabaci is based on 850 bp. Only bootstraps higher than 60 are indicated. Tree was constructed by maximum-likelihood analysis using an HKY85 substitution model. Black type indicates sequences from Genbank. Various
colours indicate the different sequences obtained in this study. Red-B biotype, blue to violet-Q biotype and green-MS biotype.
Sequences are designated according to the geographical origin of B. tabaci individuals.
2010 Blackwell Publishing Ltd
ENDOSYMBIONT ASSEMBLAGE IN A SPECIES COMPLEX 5
2010 Blackwell Publishing Ltd
6 G. GUEGUEN ET AL.
the variation observed for ITS1 (14 segregating sites in
488 bp of sequence; p = 0.00547 and hw = 0.00885). Our
data set reveals, as expected, higher diversity for CO1
than for ITS1 because mtDNA in arthropods is known
to be evolving about twice as fast as ITS1 (Schlötterer
et al. 1994). However, in B. tabaci, overall CO1 nucleotide diversity is 5 times higher than nucleotide diversity
observed for ITS1. This ratio even increases to 10 within
Q biotype (p = 0.01576 for CO1 compared to
p = 0.00151 for ITS1), suggesting an increased rate of
divergence among mtDNA lineages. This higher rate of
mtDNA evolution is expected when different endosymbionts infect different populations, or because of
geographic and ⁄ or historical events (Hurst & Jiggins
2005; Galtier et al. 2009). However, without any further
information on species status of the different genetic
groups and absence of a close uninfected outgroup, we
were unable to disentangle these hypotheses. Overall,
the CO1 phylogeny differentiates six distinct genetic
groups (Fig. 1), while ITS1 phylogeny (Fig. 2) reveals
only two groups: Q biotype that cluster with Africa silverleafing populations (ASL) on the one hand and a
cluster composed of B and Ms biotypes which are
Fig. 2 ITS1 gene–based phylogeny of B. tabaci based on
490 bp. Only bootstraps higher than 60 are indicated. Tree was
constructed by maximum-likelihood analysis using a JC69 substitution model. Black type indicates sequences from Genbank.
Various colours indicate the different sequences obtained in
this study. Red-B biotype, blue-Q biotype, green-MS biotype
and grey-Asia and other biotypes. Sequences are designated
according to the geographical origin of B. tabaci individuals.
grouped together. It is noteworthy that B and MS,
which were previously defined as distinct biotypes
(Delatte et al. 2005), have identical ITS1 sequences. They
only differ by their CO1 sequences with Ms sequences
being more closely related to Q than to B. The high
diversity of CO1 gene also reveals important differentiation within Q biotype that does not appear when ITS1
is used, with four groups emerging from the phylogeny. One group was previously designated as ASL
genetic group (Boykin et al. 2007), two were recently
named Q1 (western Mediterranean populations) and Q2
(Middle Eastern populations) (Chu et al. 2008), and we
propose to designate the fourth one, consisting of populations from Burkina Faso only, as Q3 (Fig. 1). All these
genetic groups were then screened for endosymbiont
infection.
Prevalence, diversity and association of secondary
endosymbionts
The presence of all known endosymbionts and their
variants was determined by diagnostic-specific PCR (or
sequencing when necessary) for 18 populations belonging to the B–Q clade, one Asian population and 18
Israeli populations previously described in Chiel et al.
(2007) (Five of these Israeli populations are represented
on Fig. 3, two from B biotype and three from Q biotype). We estimated bacterial diversity at both individual and population levels and prevalence of
endosymbionts in each population. As expected, the
primary endosymbiont, essential for B. tabaci survival,
was detected in all tested individuals, providing proof
that all DNA extracts were of good quality. Fritschea
bemisiae was never observed in our samples, suggesting
that this bacterium is scarce or absent in the Mediterranean ⁄ Africa silverleafing clade, putting forward the
hypothesis that it could be restricted to the A biotype
(Thao et al. 2003). All other endosymbionts reported in
B. tabaci were observed, and sequence analysis revealed
that in addition to species diversity, several bacterial
strains may exist within the different bacterial genera.
A total of 10 endosymbiotic bacteria were found among
which two strains of Wolbachia (W1 and W2), two
strains of Cardinium (C1 and C2) and four strains of
Arsenophonus (A1, A2, A3 and A4) meaning that the
diversity of endosymbionts in B. tabaci is wider than
previously thought (Figs 4–6). One Cardinium strain
(C2) observed in the MS biotype and one Arsenophonus
strain (A2 strains) found in some African populations
are reported for the first time in B. tabaci. A2 Arsenophonus strain clusters with A. nasoniae endosymbiont of
the wasp Nasonia vitripennis and is characterized by an
intervening sequence insertion in the 23S ribosomal
DNA (Fig. 4) (Thao & Baumann 2004a). For Hamiltonella
2010 Blackwell Publishing Ltd
A
H
C1
W
R
A
Israel FP1 (n = 18)
Q2
0
0
Ivory Coast (n = 25)
ASL
H
C
W
R
A2
H
C
W2
R
A1
C
W
R
Cameroon Ma (n = 6)
Q1 & ASL
A
Israel FP4 (n = 20)
Q2
H
C
W2
R
A1
C1
W
R
A
Cameroon Me (n = 7)
Q1 & ASL
Q1 H C1 W1 R
100%
0
W
R
A
H
C2
W
R
A
Morocco Ag (n = 9)
Q1
0
100%
100%
0
H
0
Q1 H C1 W1 R
C
0
100%
A
100%
100%
0
H
H
France NP (n = 38)
Q1
H
C1 W1
R
A
Burkina B1 (n = 17)
Q1 & ASL
H
100%
Sudan (n = 8)
Q1
A
C1 W1
R
A
Burkina B2 (n = 18)
Q1 & ASL
0
A
R
A
Q1 H C W1 R A1 A2 A4
Q1 H C W R A1 A2 A4
Israel FP10 (n = 20)
Q2
Burkina O1 (n = 13)
Q3
Burkina O2 (n = 13)
Q3
H
C
W2
R
A1
100%
R
W
100%
W
C
France R (n = 33)
Q1
0
C1
H
Reunion (n = 8)
MS
0
R
A
100%
W
R
0
C1
100%
H
W
Israel FP6 (n = 20)
B
0
H
0
0
Morocco Ca (n = 11)
Q1
C
0
A3
100%
R
H
Tunisia Ke (n = 11)
Q1
0
W1
Israel FP17 (n = 20)
B
0
A
100%
R
100%
W
0
C
100%
0
C
Uruguay (n = 32)
Q1
100%
H
100%
H
100%
100%
A
Reunion (n = 19)
B
0
R
100%
W
100%
C
Cambodia (n = 3)
Thai
0
100%
H
100%
Antilles (n = 18)
B
0
Tunisia EH (n = 9)
B
0
100%
ENDOSYMBIONT ASSEMBLAGE IN A SPECIES COMPLEX 7
H
C
W
R
A2
H
C
W
R
A2
Fig. 3 Individual infection status observed in B. tabaci populations. For the Israeli populations, only two examples of B out of 16,
and three examples of Q2 out of 12 from Chiel et al. (2007) are shown. Red-Hamiltonella, dark violet and violet–Cardinium strains 1
and 2, yellow and orange–Wolbachia strains 1 and 2, blue—Rickettsia and green to dark green–Arsenophonus strains 1, 2 and 3. In
mixed populations, proportion of Q1 individuals is indicated in grey. The geographical origin of the population, the number of individuals tested and the main genetic group in the populations (defined using a mitochondrial marker) are indicated at the top of the
graphs.
and Rickettsia, only a single strain was detected (Figs S1
and S2, Supporting information) (Gueguen et al. 2009).
Secondary endosymbionts are very common, as over
95% (624 on 653) of tested individuals carried at least
one of them, and all B. tabaci genetic groups harbour at
least one and up to four bacteria when considering both
species and strains diversity (Fig. 3). However, presence and prevalence of each bacterium vary considerably in the different genetic groups, without
correspondence with the Bemisia phylogeny. For example, the frequency of Hamiltonella was close to 100% in
B and Q1 groups and close to 50% in some ASL populations, whereas this bacterium was never observed in
MS, Q2 or Q3 groups. In contrast, Arsenophonus was
almost fixed in Q3 (A2 strain) and shows high prevalence in Q2 (A1 strain) and ASL groups (different
strains) but was not found in B, Q1 or MS groups.
When considering the whole bacterial community, as
many as 20 different endosymbiotic combinations are
observed in our samples (Fig. 3). Most individuals
(65%; 407 ⁄ 625 infected individuals) harbour two or
more secondary endosymbionts, and double infections
are particularly common (59%; 369 ⁄ 625 infected individuals). For example, Rickettsia and Hamiltonella in B
biotype or Hamiltonella and Cardinium strain C1 in Q
2010 Blackwell Publishing Ltd
biotype are frequently found together within the same
host individual (Fig. 3). In contrast, some bacteria have
never been found co-infecting the same host individual
(Cardinium and Rickettsia, Hamiltonella and Arsenophonus
A1; Fig. 3).
Linkage disequilibrium between B. tabaci mtDNA
and endosymbiont communities
Given the high diversity of bacterial communities and
the different levels of genetic differentiation revealed by
CO1 and ITS1 sequences, we further analysed the association between infection status and genetic groups,
focusing on the Mediterranean ⁄ Africa silverleafing
clade (Cambodian population was excluded from this
analysis). To detect and quantify these associations, global and between-group correspondence analyses (CA)
were performed. CA analysis was used to plot the variability of endosymbiont communities among different
populations or genetic groups on a two-dimensional
factorial map. Infection status and populations were
both projected on the same map (Fig. 7). The first two
axes account for 39% of the total variability (the first
accounts for 20.5% and the second for 18.4%), revealing
a clear differentiation of B. tabaci populations according
8 G. GUEGUEN ET AL.
Fig. 4 Arsenophonus 23S gene–based phylogeny based on
530 bp. Only bootstraps higher than 60 are indicated. Tree was
constructed by maximum-likelihood analysis using an HKY85
substitution model. Black type indicates sequences from Genbank. Various colours indicate the sequences obtained in this
study. Blue to violet-Q biotype and grey-other biotypes.
Sequences are designated according to the geographical origin
of the B. tabaci individuals.
to their infection status. Global CA analysis separates
populations infected by Hamiltonella from those that are
not (first axis), and the second axis separates populations depending on whether they are infected by one or
the other strains of Arsenophonus and Wolbachia (Fig. 71). To study the association between genetic groups and
bacterial communities, decomposition of the variance
was analysed using different groupings based either on
ITS1 (Two groups that separate B ⁄ MS and Q ⁄ ASL) or
on CO1 (Six genetic groups B, Ms, Q1, Q2, Q3 and
ASL). When ITS1 grouping is considered, only 16.75%
of the total variability was explained by differences in
bacterial communities between groups (Monte Carlo
test, n = 1000, P > 0.05), showing that infection status
among individuals is highly diverse within these
groups. In contrast, variation in infection status
accounted for 64.61% of the total variability when considering CO1 grouping (Monte Carlo test, n = 1000,
P < 0.01). This correlation is illustrated in Fig. 7-2,
which shows that the different mitochondrial lineages
occupy separate positions on the factorial map. Indeed,
Fig. 5 Cardinium 16S gene–based phylogeny based on 1300 bp.
Only bootstraps higher than 60 are indicated. Tree was constructed by maximum-likelihood analysis using a GTR + G
substitution model. Black type indicates sequences from Genbank. Various colours indicate the sequences obtained in this
study. Blue-Q biotype, green-MS group and grey-other biotypes. Sequences are designated according to the geographical
origin of the B. tabaci individuals.
all B biotype populations harbour the same Hamiltonella
and Rickettsia combination with almost all individuals
co-infected by these two bacteria, while MS genetic
group is infected at low frequency with Cardinium C2
and Rickettsia. The situation is more pronounced in the
Q related genetic groups, where the composition of
endosymbiont communities differs between Q1, Q2, Q3
and ASL groups despite infection status of individuals
being highly polymorphic (Fig. 3). While Q1 is infected
with Hamiltonella at very high frequency, Cardinium C1
and Wolbachia W1, Q2 and Q3 never harbour these
same bacteria. Q2 is infected with Arsenophonus A1,
Rickettsia and Wolbachia W2, whereas Q3 is infected
with the same Rickettsia but the Arsenophonus A2 at high
prevalence. A more complicated situation is observed in
ASL group with the presence of 3 Arsenophonus strains
(A1, A2 and A4), which were never found in co-infection (Fig. 3). Overall, differentiation observed using the
CO1 gene is strongly associated with a particular symbiotic complement, revealing high level of linkage disequilibrium between mitochondrial lineages and
bacterial composition of endosymbiotic community. It is
however noticeable that this association is weaker or
even disappears when only part of the symbiotic community or one bacterium is considered.
2010 Blackwell Publishing Ltd
ENDOSYMBIONT ASSEMBLAGE IN A SPECIES COMPLEX 9
Endosymbiont diversity and dynamics of endosymbiont
communities
Fig. 6 Wolbachia phylogeny based on concatenation of gatB,
ftsZ and fbpA genes (1220 bp). Only bootstraps higher than 60
are indicated. Tree was constructed by a maximum-likelihood
analysis using a JC69 substitution model. Black type indicates
sequences from Genbank. Various colours indicate the
sequences obtained in this study. Blue to violet-Q biotype and
grey-other biotypes. Sequences are designated according to the
geographical origin of the B. tabaci individuals.
Discussion
Despite the high number of endosymbionts reported
in B. tabaci and their potential role in biotype differentiation and genetic groups biological properties, global
analyses of endosymbiont infection were lacking in
this species complex. Our results show that endosymbiont diversity is wider than previously described.
This high diversity of bacterial endosymbionts associated with their high prevalence results in multiple
infections that can deeply modify the host phenotype
(according to so far unknown effects of the bacteria)
which could, in turn, considerably influence the
evolution of this species complex. The rapid dynamics
of bacterial assemblages, together with the strong linkage disequilibrium with genetic groups defined on
mtDNA, advocates for an increased attention to these
potentially influential passengers in the study of
B. tabaci species complex.
2010 Blackwell Publishing Ltd
Previous studies have described six facultative bacterial
species in B. tabaci (Zchori-Fein & Brown 2002; Nirgianaki et al. 2003) with some reports indicating possible
infections by different strains of the same bacterium
(Nirgianaki et al. 2003; Li et al. 2007; Novakova et al.
2009). Here, we observed as many as nine different bacterial strains within the Mediterranean ⁄ Africa silverleafing clade. Occurrence of multiple infection in B. tabaci
is not an isolated case, and other insects, such as aphids
or tsetse flies, have also been shown to harbour several
bacterial endosymbiotic species (Aksoy 2003; Haynes
et al. 2003). All these insects harbour a nutritional endosymbiont. Even if some bias might exist because endosymbionts may have been more extensively screened in
these species, it is possible that these insects are more
susceptible to endosymbiont infection. Possible explanations for this higher susceptibility include manipulation
of host defences by the nutritional endosymbiont or
specific immune profile exhibited by bacteriocytes (Anselme et al. 2008; Wang et al. 2009), where all B. tabaci
endosymbionts are found (Gottlieb et al. 2008).
We also highlighted important variations of endosymbiotic communities between closely related genetic
groups. This distribution suggests that the composition
of these communities is a highly dynamic process resulting from horizontal transfers, spreading and loss events,
making it difficult to reconstruct a scenario of endosymbiont infection history. Horizontal transmission seems
probable for Rickettsia (present in B, Q2 and Q3 groups
but absent from Q1 and ASL), and Hamiltonella
(observed in B and Q1 but not in Q2). Even though
infection could be the ancestral state and subsequent
losses in some lineages could explain this pattern,
absence of sequence variation within these two symbionts makes shared ancestry unlikely and suggests horizontal transmission and subsequent invasion of the
bacteria. The possible different routes for horizontal
transmission include host plants, parasitoids and mating
(Vavre et al. 1999; Moran & Dunbar 2006; Sintupachee
et al. 2006). Rickettsia of B. tabaci is easily acquired by its
parasitoid Eretmocerus sp (Chiel et al. 2009), which then
could be a candidate responsible for horizontal transmission of this bacterium between B and Q2 genetic groups.
The high frequency of horizontal transfers certainly
explains why B. tabaci individuals often host several
bacteria. These multiple infections raise the question of
the interactions among bacteria that take place within a
B. tabaci individual. These interactions may occur
between nutritional and facultative endosymbionts or
between facultative bacteria. They may involve (i) competition between bacteria for space and resources, as
10 G . G U E G U E N E T A L .
d = 0.5
HR
R
RW2
A1
C2
N1
H
A2
HW1
1
d = 0.5
B
Q2
ASL
Q1
MS
Q3
2
Fig. 7 Correspondence analysis performed on the symbiotic community variability according to host biotype. Forty-one populations
are represented by dots. 7-1 Projection of populations and infection status on a factorial map. 7-2 Populations grouped according to
their CO1 phylogeny (6 groups).
suggested by their colocalization within the same host
tissues (Gottlieb et al. 2008) or (ii) cooperative processes
facilitating their co-occurrence (Vautrin et al. 2008; Jaenike et al. 2010). An interesting point is that some combinations of bacteria are seldom or never found in
B. tabaci. It is of course possible that these strains have
never been found together in the same host because of
the history of horizontal transfers, but direct or indirect
interactions between bacteria within or among host
individuals may also have shaped endosymbiont
assemblage.
Prevalence of infection by secondary symbionts
Prevalence of secondary endosymbionts in B. tabaci is
globally very high. However, infection frequencies vary
between bacteria and between genetic groups. Analysis
of the prevalence of each endosymbiont is complex
because it relies on multiple factors (history of horizontal transfers, interactions between endosymbionts and
phenotypic effects). For example, Hamiltonella is strikingly found at very high frequency in some genetic
groups, while it is absent in others. Under the hypothe-
sis of a recent horizontal transfer of this bacterium
between groups, this suggests that invasion of Hamiltonella has repeatedly occurred and that Hamiltonella
has a major impact on its host phenotype. In the aphid
Acyrtosiphon pisum, Hamiltonella induces parasitoid resistance and reaches a prevalence of 70% in some populations (Oliver et al. 2002; Darby et al. 2003). If the same
effect applies to B. tabaci, this could contribute to its
high prevalence in this species.
Endosymbionts known to alter insect reproduction
were not fixed in any of the B. tabaci populations
screened and sometimes reached only very low frequencies. This low prevalence may result from (i) very low
effects of these endosymbionts on host phenotype such
that they behave as nearly neutral; (ii) direct or indirect
interactions between bacteria within the same host or
among host individuals; (iii) inefficient vertical transmission; (iv) induction of a high physiological cost. To
date, all these parameters remain unknown in B. tabaci,
except a probable cost of Rickettsia in presence of insecticides (Kontsedalov et al. 2008; Ghanim & Kontsedalov
2009). One main issue of future studies would be to
determine the phenotypic effects of all these bacteria.
2010 Blackwell Publishing Ltd
E N D O S Y M B I O N T A S S E M B L A G E I N A S P E C I E S C O M P L E X 11
Radiation of mitochondrial lineages and their
association with endosymbiotic community
Clearly, endosymbionts are not randomly distributed
among the six CO1 genetic groups composing the Mediterranean ⁄ Africa silverleafing clade. This pattern of
association between mitochondrial variants and composition of the endosymbiotic community stresses the linkage disequilibrium that exists between the bacterial
community and mtDNA. This pattern of association
could have two major explanations. (i) Demographic
processes. B and Q biotypes of B. tabaci are known to be
invasive (McKenzie et al. 2009; Wan et al. 2009). Invasion is often associated with founder effects that can fix
one particular mtDNA variant together with the particular endosymbiotic community it is associated with. (ii)
Bacterial endosymbionts-induced selective sweeps.
Recurrent bacterial-induced selective sweeps induce a
decrease in mtDNA diversity within each group but an
increase in divergence between groups (Jiggins & Tinsley 2005), while nuclear diversity should not change. In
B. tabaci, (i) the complex history of endosymbiont infection with frequent spread and loss of bacteria, (ii) the
low diversity of ITS1 nuclear marker compared to CO1
and (iii) the absence of clear geographical pattern of bacterial distribution all make the recurrent selective
sweeps hypothesis more likely than the other one. However, the only way to properly disentangle these scenarios (bottlenecks vs. selective sweeps) would be to
compare the molecular evolution of mitochondrial
markers with suitable nuclear markers.
Whatever the hypothesis, the close association
between mitochondrial diversity and bacterial endosymbiont community raises the question of the relevance of
mtDNA markers for B. tabaci biotype identification and
phylogeny. With regard to the phylogeny and evolutionary history of B. tabaci complex, mtDNA is clearly
unsatisfactory, because endosymbionts can modify the
branching pattern of the mtDNA phylogeny by inducing selective sweeps (Hurst & Jiggins 2005). With regard
to the identification of B. tabaci biotypes, the usefulness
of mtDNA markers will greatly depend on the bacterial
phenotypic effects. If the mitochondrial genetic groups
are not associated with different biological specificities,
mtDNA markers will tend to overestimate the number
of biological entities and lead to an unnecessary multiplication of hypothetical biotypes. However, if endosymbionts do play a role in reproductive isolation
and ⁄ or biological attributes of their hosts, mtDNA
might provide a good marker for identifying the biotypes, sometimes even better than nuclear genes when
the endosymbiont-induced host phenotype modifications are not associated with genetic isolation. Determination of endosymbionts effects in B. tabaci is therefore
2010 Blackwell Publishing Ltd
an essential step in attempting to unravel these different
hypotheses, and understanding the biology and evolution of this pest species. Moreover, as endosymbiotic
bacteria may induce reproductive isolation between
individuals with different symbiotic complement and
given their high prevalence in B. tabaci species complex,
it is now essential that all studies aiming at defining
biotype or species boundaries take into consideration
these influential microbial partners.
Towards the metacommunity concept of secondary
endosymbionts
Altogether, our results shed light on the importance of
taking these microbial partners into account when
studying this species complex. The phenotypic variations observed in B. tabaci might result from both
nuclear and cytoplasmic (endosymbiont) genes and their
interactions. Individuals should be seen as a community
of genes from different origins, which, as a whole, constitute a selective unit. This means that it is necessary to
include the study of the B. tabaci complex in a community genetics perspective (Whitham et al. 2006) where
different levels of selection are taken into account.
The high incidence of endosymbionts also calls into
question the ecological processes that govern the distribution and abundance of these intracellular bacteria.
Interactions among bacterial endosymbionts occur at
three different levels: within the host individual (competition or cooperation), locally among host individuals
in the same population (horizontal transfers, competition mediated by the provision of ecologically important traits by symbionts, indirect interactions mediated
by endosymbiont-induced incompatibility) and regionally among geographic populations linked by host
migration. This clearly falls into the framework of the
metacommunity concept defined as a set of local communities linked by the dispersal of multiple, potentially
interacting species (Leibold et al. 2004). This approach
was recently applied to understanding the biogeography of free-living bacteria (Van der Gucht et al. 2007),
and it could be helpful for elucidating how the local
and regional processes interact to shape the spatiotemporal variation of inherited intracellular endosymbiont communities and their effects on host local
adaptation.
Acknowledgements
This research was supported by the High Council for Scientific
and Technological Cooperation between France and Israel (program on sustainable agriculture CNOUS 05F14) and by the
French National Research Agency (project ANR-06-PADD-04,
BemisiaRisk). We are grateful to Netta Mozes-Daube for technical help and to Samuel Nibouche, Sana Khalifa Mukhtar, Pedro
12 G . G U E G U E N E T A L .
Benoit and César Basso for samples from Cameroon, Sudan,
Dominican Republic and Uruguay and to Sylvain Mousset for
his valuable advice on molecular diversity analysis.
References
Aksoy S (2003) Symbiosis in Tsetse. In: Insect Symbiosis (eds
Bourtzis K, Miller TA), pp. 53–65. CRC Press, Boca Raton.
Anselme C, Perez-Brocal V, Vallier A et al. (2008) Identification
of the weevil immune genes and their expression in the
bacteriome tissue. BMC Biology, 6, 43.
Benzecri JP (1973) Correspondence Analysis (In French). Bordas,
Paris.
Boykin LM, Shatters Jr RG, Rosell RC et al. (2007) Global
relationships of Bemisia tabaci (Hemiptera: Aleyrodidae)
revealed using Bayesian analysis of mitochondrial CO1 DNA
sequences. Molecular Phylogenetics and Evolution, 44, 1306–
1319.
Chiel E, Gottlieb Y, Zchori-Fein E et al. (2007) Biotype-dependent
secondary symbiont communities in sympatric populations
of Bemisia tabaci. Bulletin of Entomological Research, 97, 407–
413.
Chiel E, Inbar M, Mozes-Daube N et al. (2009) Assessments of
Fitness Effects by the Facultative Symbiont Rickettsia in the
Sweetpotato Whitefly (Hemiptera: Aleyrodidae). Annals of the
Entomological Society of America, 102, 413–418.
Chu D, Wan FH, Tao YL et al. (2008) Genetic differentiation of
Bemisia tabaci (Hemiptera: Aleyrodidae) biotype Q based on
mitochondrial DNA markers. Insect Sciences, 117, 117–125.
Darby AC, Tosh CR, Walters KFA, Douglas AE (2003) The
significance of a facultative bacterium to natural populations
of the pea aphid Acyrtosiphon pisum. Ecological Entomology,
28, 145–150.
De Barro PJ, Driver F, Trueman JWH, Curran J (2000) Phylogenetic relationships of world populations of Bemisia tabaci
(Gennadius) using ribosomal ITS1. Molecular Phylogenetics and
Evolution, 16, 29–36.
De Barro PJ, Trueman JWH, Frohlich DR (2005) Bemisia
argentifolii is a race of B. tabaci (Hemiptera: Aleyrodidae): the
molecular genetic differentiation of B. tabaci populations
around the world. Bulletin of Entomological Research, 95, 193–
203.
Delatte H, Reynaud B, Granier M et al. (2005) A new
silverleaf-inducing biotype Ms of Bemisia tabaci (Hemiptera:
Aleyrodidarum) indigenous to the islands of the south-west
Indian ocean. Bulletin of Entomological Research, 95, 29–35.
Duron O, Bouchon D, Boutin S et al. (2008) The diversity of
reproductive parasites among arthropods: Wolbachia do not
walk alone. BMC Biology, 24, 6–27.
Felsenstein J (1993) Phylip package.
Frohlich DR, Torres-Jerez I, Bedford ID, Markham PG, Brown
JK (1999) A phylogeographical analysis of the Bemisia tabaci
species complex based on mitochondrial DNA markers.
Molecular Ecology, 8, 1683–1691.
Galtier N, Nabholz B, Glemin S, Hurst GDD (2009)
Mitochondrial DNA as a marker of molecular diversity: a
reappraisal. Molecular Ecology, 18, 4541–4550.
Ghanim M, Kontsedalov S (2009) Susceptibility to insecticides
in the Q biotype of Bemisia tabaci is correlated with bacterial
symbiont densities. Pest Management Science, 65, 939–942.
Gottlieb Y, Ghanim M, Gueguen G et al. (2008) Inherited
intracellular ecosystem: symbiotic bacteria share the
bacteriocytes of whiteflies. FASEB Journal, 22, 2591–2599.
Gueguen G, Rolain J, Zchori-Fein E et al. (2009) Molecular
detection and identification of Rickettsia endosymbiont in
different biotypes of Bemisia tabaci. Clinical Microbiol and
Infection, 15, 271–272.
Guindon S, Gascuel O (2003) PhyML—A simple, fast, and
accurate algorithm to estimate large phylogenies by
maximum likelihood. Systematic Biology, 52, 696–704.
Haine ER (2008) Symbiont-mediated protection. Proceedings of
the Royal Society of London. Series B, 275, 353–361.
Haynes S, Darby AC, Daniell TJ et al. (2003) Diversity of
bacteria associated with natural aphid populations. Applied
and Environmental Microbiology, 69, 7216–7223.
Hurst GDD, Jiggins FM (2005) Problems with mitochondrial
DNA as a marker in population, phylogeographic and
phylogenetic studies: the effects of inherited symbionts.
Proceedings of the Royal Society of London Series B, 272, 1525–
1534.
Jaenike J, Stahlhut JK, Boelio LM, Unckless RL (2010)
Association between Wolbachia and Spiroplasma within
Drosophila neotestacea: an emerging symbiotic mutualism?
Molecular Ecology, 19, 414–425.
Jiggins FM, Tinsley MC (2005) An ancient mitochondrial
polymorphism in Adalia bipunctata linked to a sex-ratiodistorting bacterium. Genetics, 171, 1115–1124.
Jones D, Taylor W, Thornton J (1992) The rapid generation of
mutation data matrices from protein sequences. Computer
Applications in the Biosciences, 8, 275–282.
Jukes T, Cantor C (1969) Evolution of Protein Molecules, pp. 21–
132. Academic Press, New York.
Khasdan V, Levin I, Rosner A et al. (2005) DNA markers for
identifying biotypes B and Q of Bemisia tabaci (Hemiptera:
Aleyrodidae) and studying population dynamics. Bulletin of
Entomological Research, 95, 605–613.
Kontsedalov S, Zchori-Fein E, Chiel E et al. (2008) The
presence of Rickettsia is associated with increased
susceptibility of Bemisia tabaci (Homoptera: Aleyrodidae) to
insecticides. Pest Management Science, 64, 789–792.
Leibold MA, Holyoak M, Mouquet N et al. (2004) The
metacommunity concept: a framework for multi-scale
community analysis. Ecology letters, 7, 601–613.
Li ZX, Lin HZ, Guo XP (2007) Prevalence of Wolbachia
Infection in Bemisia tabaci. Current Microbiology, 54, 467–471.
Lobry JR, Chessel D (2003) Internal correspondence analysis of
codon and amino-acid usage in thermophilic bacteria. Journal
of Applied Genetics, 44, 235–261.
McKenzie CL, Hodges G, Osborne LS, Byrne FJ, Shatters RG Jr
(2009)
Distribution
of
Bemisia
tabaci
(Hemiptera:
Aleyrodidae) biotypes in Florida-investigating the Q
invasion. Journal of Economic Entomology, 102, 670–676.
Moran NA, Dunbar HE (2006) Sexual acquisition of beneficial
symbionts in aphids. Proceedings of the National Academy of
Sciences of the United States of America, 103, 12803–12806.
Moya A, Guirao P, Cifuentes D, Beitia F, Cenis JL (2001)
Genetic diversity of Iberian populations of Bemisia tabaci
(Hemiptera: Aleyrodidae) based on random amplified
polymorphic DNA–polymerase chain reaction. Molecular
Ecology, 10, 891–897.
2010 Blackwell Publishing Ltd
E N D O S Y M B I O N T A S S E M B L A G E I N A S P E C I E S C O M P L E X 13
Nei M (1987) Molecular Evolutionary Genetics. Columbia university press, New York.
Ngwamidiba M, Blanc G, Raoult D, Fournier PE (2006) ScaI, a
previously undescribed paralog from autotransporter proteinencoding genes in Rickettsia species. BMC Microbiology, 6, 12–
23.
Nirgianaki A, Banks GK, Frohlich DR et al. (2003) Wolbachia
infections of the whitefly Bemisia tabaci. Current microbiology.
Current Microbiology, 47, 93–101.
Novakova E, Hypsa V, Moran NA (2009) Arsenophonus, an
emerging clade of intracellular symbionts with a broad host
distribution. BMC Microbiology, 9, 143–157.
Oliver KM, Russell JA, Moran AN, Hunter MS (2002)
Facultative bacterial symbionts in aphids confer resistance to
parasitic wasps. Proceedings of the National Academy of Sciences
of the United States of America, 100, 1803–1807.
Perring TM (2001) The Bemisia tabaci species complex. Crop
Protection, 20, 725–737.
Posada D, Crandall KA (1998) Modeltest: testing the model of
DNA substitution. Bioinformatics, 14, 817–818.
R development Core Team (2003) R software.
Rice P, Longden I, Bleasby A (2000) EMBOSS: the European
molecular biology open software suite. Trends in Genetics, 16,
276–277.
Roux V, Raoult D (1995) Phylogenetic analysis of the genus
Rickettsia by 16S rDNA sequencing. Research in Microbiology,
146, 385–396.
Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003)
DnaSP, DNA polymorphism analyses by the coalescent and
other methods. Bioinformatics, 19, 2496–2497.
Schlötterer C, Hauser M, Von Haeseler A, Tautz D (1994)
Comparative evolutionary analysis of rDNA ITS regions in
Drosophila. Molecular Biology and Evolution, 11, 513–522.
Shoemaker DD, Dyer KA, Ahrens M, McAbee K, Jaenike J
(2004) Decreased diversity but increased substitution rate in
host mtDNA as a consequence of Wolbachia endosymbiont
infection. Genetics, 168, 2049–2058.
Sintupachee S, Milne JR, Poonchaisri S, Baimai V, Kittayapong
P (2006) Closely related Wolbachia strains within the pumpkin
arthropod community and the potential for horizontal
transmission via the plant. Microbial ecology, 51, 294–301.
Tajima F (1983) Evolutionary relationship of DNA sequences in
finite populations. Genetics, 105, 437–460.
Tavaré S (1986) Some probabilistic and statistical problems in
the analysis of DNA sequences. In:Some Mathematical
Questions in Biology—DNA Sequence Analysis (ed. Miura RM),
pp. 57–86, American Mathematical Society, Providence.
Thao M, Baumann P (2004a) Evidence for multiple acquisition
of Arsenophonus by whiteflies species (Sternorrhyncha:
Aleyrodidae). Current Microbiology, 48, 140–144.
Thao M, Baumann P (2004b) Evolutionary relationships of
primary prokaryotic endosymbionts of whiteflies and their
hosts. Applied and Environmental Microbiology, 70, 3401–3406.
Thao ML, Baumann L, Hess JM et al. (2003) Phylogenetic
evidence for two new insect associated Chlamydia of the
family Simkaniaceae. Current Microbiology, 47, 46–50.
Thioulouse J, Chessel D, Doledec S, Olivier JM (1997) ADE-4: a
multivariate analysis and graphical display software.
Statistics and Computing, 7, 75–83.
Thompson JD, Higgins DG, Gibson TJ (1994) ClustalW:
improving the sensitivity of progressive multiple sequence
2010 Blackwell Publishing Ltd
alignment through sequence weighting, position-specific gap
penalties and weight matrix choice. Nucleic Acids Research,
22, 4673–4680.
Van der Gucht K, Cottenie K, Muylaert K et al. (2007) The
power of species sorting: local factors drive bacterial
community composition over a wide range of spatial scales.
Proceedings of the National Academy of Sciences of the United
States of America, 104, 20404–20409.
Vautrin E, Genieys S, Charles S, Vavre F (2008) Do vertically
transmitted symbionts co-existing in a single host compete
or cooperate? A modelling approach. Journal of Evolutionary
Biology, 21, 145–161.
Vavre F, Fleury F, Lepetit D, Fouillet P, Bouléteau M (1999)
Phylogenetic evidence for horizontal transmission of
Wolbachia in host-parasitoid associations. Molecular Biology
and Evolution, 16, 1711–1723.
Walsh PS, Metzger DA, Higuchi R (1991) Chelex-100 as a
medium for simple extraction of DNA for PCR based typing
from forensic material. BioTechniques, 10, 506–513.
Wan F, Zhang G, Liu S et al. (2009) Invasive mechanism and
management strategy of Bemisia tabaci (Gennadius) biotype
B: progress report of 973 Program on invasive alien
species in China. Science in China Series C, Life Sciences, 52,
88–95.
Wang JW, Wu YN, Yang GX, Aksoy S (2009) Interactions
between mutualist Wigglesworthia and tsetse peptidoglycan
recognition protein (PGRP-LB) influence trypanosome
transmission. Proceedings of the National Academy of Sciences of
the United States of America, 106, 12133–12138.
Watterson GA (1975) On the number of segregating sites in
genetical models without recombination. Theoretical Population
Biology, 7, 256–276.
Werren JH, Baldo L, Clark ME (2008) Wolbachia: master
manipulators of invertebrate biology. Nature Reviews Microbiology, 6, 741–751.
Whitham TG, Bailey JK, Schweitzer JA et al. (2006) A
framework for community and ecosystem genetics: from
genes to ecosystems. Nature Reviews Genetics, 7, 510–523.
Zchori-Fein E, Brown JK (2002) Diversity of prokaryotes
associated with Bemisia tabaci (Gennadius) (Hemiptera:
Aleyrodidae). Annals of Entomological Society of America, 95,
711–718.
This work is part of GG’s PhD on the endosymbiotic multiple
infections in insect vectoring viruses. FV, DC, EC, YG, MG,
EZF and FF are interested in host-parasite interactions. OG is
interested in Bemisia tabaci biology and damage in Africa.
MP is a virologist interested in viruses that are transmitted by
Bemisia tabaci
Supporting information
Additional supporting information may be found in the online
version of this article.
Fig. S1 Hamiltonella GyrB gene–based phylogeny based on
815 bp. Only bootstraps higher than 60 are indicated. Tree was
constructed by maximum-likelihood analysis using a JC69 substitution model. Black type indicates sequences from Genbank.
14 G . G U E G U E N E T A L .
Various colours indicate the different sequences obtained in
this study. Red-B biotype, blue-Q biotype and grey-other biotypes. Sequences are designated according to the geographical
origin of the B. tabaci individuals.
Fig. S2 Rickettsia phylogeny based on concatenation of sca1
and gltA genes (380 aa). Only bootstraps higher than 60 are
indicated. Tree was constructed by maximum-likelihood analysis using a JTT protein evolutionary model. Black type indicates sequences from Genbank. Various colours indicate the
different sequences obtained in this study. Blue to violet-Q biotype, green-MS group and red-B group. Sequences are desig-
nated according to the geographical origin of the B. tabaci
individuals.
Table S1 Date, geographical origin and number of individuals
for B. tabaci samples
Table S2 Primers and PCR conditions
Please note: Wiley-Blackwell are not responsible for the content
or functionality of any supporting information supplied by the
authors. Any queries (other than missing material) should be
directed to the corresponding author for the article.
2010 Blackwell Publishing Ltd