Content uploaded by Maria Navajas
Author content
All content in this area was uploaded by Maria Navajas
Content may be subject to copyright.
REVIEWS
NATURE REVIEWS | GENETICS VOLUME 4 | NOVEMBER 2003 | 889
The goal of biological control is to suppress the popula-
tion density of a pest organism through the use of preda-
tors, parasites, pathogens or herbivores. These organisms
might be indigenous, in which case natural enemies are
augmented or their effects on pests are facilitated, or they
might be non-indigenous, in which case the organisms
are purposefully introduced
(BOX 1).The latter case is
referred to as classical biological control
1,2
and involves
the introduction of whole genomes into a new environ-
ment. Some species that have been introduced for bio-
logical control purposes have also been artificially
selected in the laboratory or manipulated genetically to
express traits that are desirable for pest control.
Classical biological control is more than a century
old. Despite its age, researchers have questioned whether
the field has reached the status of a predictive science
3,4
.
In contrast to other modern practical sciences, such as
medicine, engineering and meteorology, in which we
customarily expect high rates of experimental success,
the probability that any given attempt at classical biolog-
ical control will lead first to establishment and subse-
quently to the control of a pest is relatively low
5
.Recent
estimates indicate that in the biological control of
arthropods only 34% of all introductions have resulted
in establishment, only 47% of which provided control
of the targeted pest, giving an overall success rate of
16%
6,7
.The biological control of weeds has a better rate
of establishment (60–63%) but a similar rate of overall
success in control (10–18%)
8,9
.
There are large gaps in our understanding of classical
biological control and, in particular, of what determines
the success of colonization and establishment. To com-
pound matters, unexpected consequences have dogged
the field, with some notable outcomes. Of particular
concern is whether species that are introduced to control
one pest will feed on other
HOSTS in the same environ-
ment. These so-called non-target effects can vary in
severity
10–12
,but in extreme cases they have included the
extinction of non-target hosts
4,13
.Clearly, we need to
understand more about the factors that determine estab-
lishment and success to avoid what Peter Kareiva, a
noted ecologist, described as “hit-or-miss pest control,
with the misses possibly disrupting natural communities
far more than they reduce pest populations”
3
.
Here,we outline the present possibilities for genetics
as a tool in biological control and note areas in which
progress is likely to accelerate in the future. We discuss
the probable roles of evolution and
ADAPTATION in bio-
logical control, and contrast findings in this area that
focus on microorganisms with those that focus on
GENES IN NEW ENVIRONMENTS:
GENETICS AND EVOLUTION IN
BIOLOGICAL CONTROL
George K. Roderick* and Maria Navajas
‡
The availability of new genetic technologies has positioned the field of biological control as a
test bed for theories in evolutionary biology and for understanding practical aspects of the release
of genetically manipulated material. Purposeful introductions of pathogens, parasites, predators
and herbivores, when considered as replicated semi-natural field experiments, show the
unpredictable nature of biological colonization. The characteristics of organisms and their
environments that determine this variation in the establishment and success of biological control
can now be explored using genetic tools. Lessons from studies of classical biological control can
help inform researchers and policy makers about the risks that are associated with the release of
genetically modified organisms, particularly with respect to long-term evolutionary changes.
HOSTS
Prey for organisms that are
introduced for biological control.
*Environmental Science,
Policy and Management,
201 Wellman Hall MC 3112,
University of California,
Berkeley, California
94720-3112, USA.
‡
Institut National de la
Recherche Agronomique,
Centre de Biologie
et Gestion des Populations,
Campus International
de Baillarguet, CS 30 016,
34980 Montferrier sur Lez,
France.
Correspondence to G.K.R.
e-mail: roderick@nature.
berkeley.edu
doi:10.1038/nrg1201
ADAPTATION
Evolution as a result of selection.
SOURCE POPULATION
Ancestral population; the pest
might have descended from this
population recently or many
generations in the past.
890 | NOVEMBER 2003 | VOLUME 4 www.nature.com/reviews/genetics
REVIEWS
Genetic markers for population origins and spread.
Because of the genealogical nature of genetic material,
so-called neutral genetic markers are ideal for tracing
the movement patterns of organisms. This information
can be used to determine the population origins and
spread of target pests, as well as the patterns of establish-
ment and spread of organisms that have been intro-
duced for biological control
18,19
.
The usual strategy for finding species that are suit-
able for classic biological control involves searching in
regions where the pest is indigenous, according to the
hypothesis that the most specific and efficient natural
enemies are to be found in the localities from which the
pest originates
20
.Every attempt is usually made to focus
the search for natural enemies on the
SOURCE POPULATION
of the pest, which can be inferred from genetic data.
Unfortunately, recently founded populations of pests
macroorganisms. Finally, we examine the risks that are
associated with biological control and ask whether
research into this subject can be used to inform
researchers and policy makers as to the risks of using
genetically modified organisms (GMOs)
14
.
Genetics in biological control
Advances in genetics have already contributed much to
our understanding of biological control and with
genomic tools becoming available for more and more
species, progress is likely to increase exponentially
15–17
.
Three areas of genetic research have been especially influ-
ential: the development of genetic markers to examine
population origins and spread; the isolation of genes that
are involved in development, reproduction and behav-
iour, with much progress from new genomic informa-
tion; and the development of gene-transfer technology.
Box 1 | Ty pes of biological control
The biological control of pests can
involve several different types of
procedure
1,2
.In classical biological
control (
a), non-indigenous predators,
parasites, pathogens or herbivores are
typically, although not always,
collected in the native habitat of the
pest
20
. The organisms to be released
are quarantined to address biosafety
concerns while non-target-host testing
and other tests take place
41
.
Introductions can involve large
numbers of individuals in one
environment or smaller numbers in
several different habitats or localities.
After establishment, the introduced
species and the pest are both expected
to persist indefinitely at low numbers,
which provides long-term control.
This method is suitable if long-term
persistence is desired and if non-target
risks are low.
In inoculation biological control (
b),
the organisms are intentionally
released with the expectation that they
will then multiply and control the pest
species for an extended period, but not
permanently. The organisms
reproduce after release, but
establishment is neither expected nor desired. For this reason, this
method is well suited to ecosystems that are periodical or contained,
such as glass houses
105
.Because establishment and long-term
persistence do not occur, the procedure must be repeated regularly.
In inundative biological control (
c), the organisms to be released are
reared, often commercially, to increase their numbers before release
60
and control is then achieved exclusively by the released organisms themselves. In contrast to inoculation biological control, reproduction is not expected
after release — this distinction is often important for regulatory agencies. Of course, unexpected inoculative effects might occur if reproduction does
take place after release.
In conservation biological control (
d), aspects of the habitat or cultural practices are altered to increase the number or efficiency of predators,
parasites, pathogens or herbivores. Such alternations might include providing increased spatial complexity of the habitat, for example through
providing further crops, or changing the timing of various aspects of cultivation or other means of pest control, such as pesticide applications.
When indigenous natural enemies are involved, this method presents limited risks to non-target hosts.
a Classical
b Inoculation
c Inundative
d Conservation
Establishment of
control agent
Control agent
reproduces but
does not persist
Habitat altered to
favour natural
control agent
Control agent
does not reproduce
Biological control agent
Control agents propagated
before release
Pest species
NATURE REVIEWS | GENETICS VOLUME 4 | NOVEMBER 2003 | 891
REVIEWS
present EFFECTIVE POPULATION SIZE
39
,post-colonization
growth rate
27
and pathways of dispersal
35
.Manipulated
releases of species for biological control provide an
opportunity to test these approaches. For example,
known numbers of the cane toad Bufo marinus were
serially introduced into the Pacific islands and Australia
to control a beetle pest of sugar cane, which allowed
replicated assessments of genetic bottlenecks that were
associated with the introduction and of the subsequent
population growth
27
.
Genetic markers are also powerful tools for moni-
toring biological control releases and for assessing rele-
vant biological traits. In quarantine situations, for
example, genetic markers can be used to examine the
effect that rearing populations of a biological control
agent can have on genetic diversity, or simply for diag-
nostics and quality control
40,41
.In the field, the impor-
tance of particular traits in the success of establishment
and control can be assessed. Introductions of individu-
als from different sources or from laboratory colonies
that were selected for different characteristics, such as
insecticide resistance
34
or host detection
42
, can be exam-
ined after introduction to determine which phenotypes
or genotypes have prevailed
43
.As a way of identifying
individuals in the field, marker genes have recently
been used to genetically modify organisms before their
release. These genes can be subsequently monitored to
track the organisms and their offspring in nature, as
well as to test hypotheses about the risks that are associ-
ated with the release of GMOs. Green fluorescent pro-
tein (GFP), which was originally isolated from jellyfish,
is the best known, but similar markers in other colours,
such as DsRed, are also available
44
.The validity of marker
genes to trace the fate of genotypes in the environment
rests on the assumption that these genes are selectively
neutral — this has yet to be tested in the field.
and most organisms that are introduced for biological
control do not satisfy the assumptions of traditional pop-
ulation-genetic approaches that assess dispersal and gene
flow — in particular, recently founded populations are
unlikely to be at
MIGRATION–DRIFT GENETIC EQUILIBRIUM
21,22
(BOX 2).When potential source populations have diverged
genetically,
PHYLOGEOGRAPHIC APPROACHES
23
that trace the his-
tory of single genes can be used for determining origins
24
;
studies of the West Nile virus
25
and the invasive weed
Ta m a r i x
26
are examples in which this approach was used.
Unfortunately, these approaches are typically not applica-
ble to macroorganisms, particularly those that have been
artificially introduced, often serially, worldwide
27
.For
such organisms, potential source populations often have
not been isolated from other populations for long
enough to diverge genetically. Consequently, markers are
rarely unique to particular source populations and no
single genetic locus can provide a clear genetic signature
of history for these species
28
.
The lack of divergence in recently established popu-
lations is compounded by a reduced level of genetic
variation as a consequence of genetic bottlenecks in
the founding populations
27–30
.The recent application
of
MULTI-LOCUS GENETIC APPROACHES coupled with the use of
statistical
ASSIGNMENT TESTS now provides a means to
determine the origins of populations with some statisti-
cal confidence
28,31–33
(BOX 3).These approaches use
CO-DOMINANT MARKERS,such as MICROSATELLITES, single
nucleotide polymorphisms (SNPs) and
ALLOZYMES.As
well as being used to determine population origins, such
markers can be used to trace the spread of new geno-
types after their introduction
34
.
New analytical approaches that involve
MARKOV-CHAIN
MONTE CARLO
simulations
35
and BAYESIAN APPROACHES
36,37
now make it possible to estimate population parameters,
such as the effective number of colonists
38
, ancestral and
GENETIC DRIFT
The random change in allele
frequencies.
MIGRATION–DRIFT GENETIC
EQUILIBRIUM
The balance between the loss of
alleles through genetic drift and
the gain of alleles through
migration.
PHYLOGEOGRAPHIC APPROACH
The use of estimated gene
genealogies to study the
geographical history and
structure of populations or
species.
MULTI-LOCUS GENETIC
APPROACHES
Genetic methods that make use
of information from many loci;
such approaches use nuclear loci
because mitochondrial genes are
typically inherited as one locus.
ASSIGNMENT TESTS
Statistical procedures in which
individuals can be assigned to
probable source populations.
CO-DOMINANT MARKERS
Genetic markers that allow the
determination of both alleles at a
diploid locus (for example,
microsatellites, allozymes and
single nucleotide
polymorphisms); these differ
from dominant markers in
which the determination of
heterozygotes is not always
possible (or example, RAPDs
and AFLPs).
MICROSATELLITES
Co-dominant nuclear DNA
markers that consist of sets of
repeated short nucleotide
sequences.
ALLOZYMES
Co-dominant nuclear DNA
markers that consist of enzymes
that differ in their mobility on a
charged gel.
MARKOV-CHAIN MONTE CARLO
A computational technique for
the efficient numerical
calculation of likelihoods.
BAYESIAN APPROACH
A statistical perspective that
focuses on the probability
distribution of parameters before
and after observing the data.
Box 2 | Migration–drift equilibrium
Populations are often modelled with the assumption of
migration–drift genetic equilibrium, in which groups of
populations exchange migrants and also undergo
GENETIC DRIFT (a).
Under these conditions, populations exchange migrants (Nm),
where N is the effective size of each population and m is the
proportion of individuals that migrate. Nm — shown in the figure
as 3 red individuals — is a useful parameter. For example, for a
large population (population 1), in which genetic drift is low, if
Nm = 3, the proportion that migrates is relatively small; whereas,
with the same number of migrants in a small population
(population 2), the proportion that migrate is relatively large, as is
the effect of genetic drift. So, regardless of population size, new
alleles that arrive in a population by migration are balanced by the
loss of alleles through genetic drift. Genetic equilibrium of this
type might take in the order of N or more generations to be
established, but the exact time that is necessary depends on many
factors, including the nature of the population structure, the
constancy of migration rates through time and the extent to
which new populations arise or go extinct
106
.Non-equilibrium
conditions occur when migration and drift are not in balance; for example, when one population has been recently
founded by colonists from another population (
b).Accordingly, genetic approaches that rely on the assumption of
migration–drift equilibrium are not appropriate for the study of recently founded populations.
a Equilibrium
1
b Non-equilibrium
1
N
1
m
1
N
1
m
1
N
2
m
2
2
2
892 | NOVEMBER 2003 | VOLUME 4 www.nature.com/reviews/genetics
REVIEWS
corn and cotton
47,48
.Surprisingly, the factors that under-
lie the original evolution of the Bt toxin in Bacillus are
unknown. Insecticidal toxins have also been identified
from scorpions and mites and have been tested in insect
control
49
.
The rate of identification of new genes that are
involved in the development, reproduction and behav-
iour of organisms that are used for biological control is
likely to accelerate as the availability of genomic infor-
mation increases. For example, rapid progress has been
made in identifying mosquito genes using a comparative
approach with the genomes of Drosophila and the mos-
quito Anopheles gambiae
44
.Also, microarray technology
50
is already contributing to studies of species that are used
for biological control, especially to topics such as virus
transmission, host adaptation and other general biologi-
cal questions, for example, predator–prey interactions
51
.
Gene-transfer technology. Gene transfer has been impor-
tant in crops such as maize, cotton, potatoes and rice,
which, as mentioned above, have been genetically
manipulated to express the Bt protein to control insect
pests, primarily Lepidoptera. Such crops are now in
Identification and isolation of genetic elements.
Molecular genetic techniques have been used to identify
traits of interest for potential use in biological control.
These traits originate both from the organism that is to
be used for biological control and from other species,
the genes of which can be used for the genetic manipu-
lation of other organisms
44
.
Tr aditional breeding is one way to modify natural
enemies before their release for biological control, with
the best-studied cases involving selection for insecticide
resistance
45
.Other traits have also been proposed,
including
DIAPAUSE
characteristics, fecundity and tem-
perature tolerance. In a transgenic example, the insect
gene that encodes chitinase, which was originally iso-
lated from the tobacco hornworm Manduca sexta, was
inserted into tobacco. It had a notable influence on
infestations of both the hornworm and the tobacco
budworm Heliothis virescens when combined with
non-lethal doses of the bacterial Bt insectical protein-
containing insecticide
46
. Bt has been isolated from the
bacterium Bacillus thuringiensis and, in its various
forms, has been used as a live inoculant, insecticide and
most recently as a transgene in crops such as potato,
EFFECTIVE POPULATION SIZE
The population size that
responds identically to that
modelled genetically; that is, the
size of the population that
matters for genetic concerns.
DIAPAUSE
A resting stage for insects,
typically during winter or dry
periods.
MAXIMUM LIKELIHOOD
A procedure in phylogenetic
reconstruction in which a tree is
chosen that maximizes the
probability of the data given the
model and the tree hypothesis.
PARSIMONY
A procedure in phylogenetic
reconstruction in which a tree is
chosen because it requires the
fewest possible mutations to
explain the data.
NESTED CLADE ANALYSIS
A statistical parsimony
procedure that constructs sets of
nested clades. With knowledge
of geographic distribution, the
clades can be examined for
evidence of processes that are
associated with geographic
structure, such as isolation by
distance, allopatric
fragmentation and long-distance
colonization.
Box 3 | Genetic approaches for determining population origins
Tw o general approaches for determining the
population origins of individuals
31
are shown. In the
phylogeographic approach (
a), haplotypes are typically
sampled from potential source populations as well as
from the population under study (denoted by an
asterisk). A genealogy of haplotypes is inferred using
one of a range of methods, which include
MAXIMUM
LIKELIHOOD, PARSIMONY and NESTED CLADE ANALYSIS
19,23,24,26
.
The locality from which each haplotype was collected is
noted on the resulting tree or network, and the origin of
the haplotype under study can be determined by its
inferred genealogical history. In the figure, each
haplotype is connected to others and the lines show one
mutational step. The white circles (denoted by a plus
sign) are the inferred haplotypes that were either not
sampled or have since become extinct. The origin of the
haplotype under study is inferred from the genealogy.
For example, the haplotype under study is one
mutational step from that found in the second potential
source population (red), and is therefore more closely
related to those individuals than individuals from any
other population. In this case, the haplotype under
study was not actually sampled in the presumed source
population. This approach is possible when the rate of
mutation of the haplotype involved is high relative to
the time since the potential source populations
diverged genetically.
In the frequency-assignment approach (
b),
a multilocus genotype is determined not only for
the individual in question but also for a large sample
of individuals from the potential source populations.
Assignment tests are used to statistically assign the
genotype in question to one of the potential sources. There are several variations of this approach, each with its own
advantages and disadvantages
28,31
.Because this approach uses the frequency of alleles at several loci, it can be used even
when potential source populations do not have unique population-specific haplotypes.
Potential source populations
*Haplotype
a Phylogeographic approach
b Frequency-assignment approach
Potential source populations
Locus 1
Locus 3
Locus 2
*Genotype
+
+ +
+
NATURE REVIEWS | GENETICS VOLUME 4 | NOVEMBER 2003 | 893
REVIEWS
immediate effects on rabbit populations in each loca-
tion, including high infectiousness and lethality
(FIG. 1).
However, within a year, some rabbits evolved resistance
in Australia. Re-sampling viruses over time and com-
paring their virulence against susceptible rabbits made
it clear that the virus became less virulent, at least ini-
tially, in each location. Recent studies have provided
insights into the genetics of this interaction. The myx-
oma virus genome encodes several gene products that
interfere with and/or modulate apoptosis in specific cell
types, immune responses and the recruitment of
inflammatory cells. Viral immunomodulatory mole-
cules that have been characterized include: M-T1,
which is a chemokine-binding protein; M-T2, which is
a soluble homologue of the tumour necrosis factor
(TNF)-receptor; M-T7,which is a soluble interferon-γ
(IFN-γ)-receptor homologue that also binds chemokine
receptors; myxoma growth factor (MGF), which is a
member of the epidermal growth factor (EGF) family;
M-T4, M-T5 and M11L,all of which modulate the
apoptosis of host cells; and three
SERPINS (SERP1, SERP2
and SERP3)
58
.Finally, in culture, myxoma virus also
down-modulates major histocompatability complex
(MHC) class 1 by inducing the loss of cell-surface MHC
class I and by preventing the movement of newly synthe-
sized MHC class I from the endoplasmic reticulum (ER).
Evolutionary change associated with organisms that
move into new habitats is not uncommon
59
and the
rapid co-evolution of the rabbit and the myxoma virus
might not be an isolated event — other such interac-
tions might go unnoticed, particularly if the players are
not economically important
4,60
.
A similar but less well-documented case involves the
introduction of the fungus Entomophaga maimaiga
from Japan into Massachusetts in the United States for
gypsy moth control in 1910. The fungus was not
detected again until 1989, when it caused massive epi-
zootics in gypsy moth populations in New England. The
evolution of a strain with increased virulence was put
forward as one hypothesis to explain both the reappear-
ance and the increased effect of the fungus after the long
period of latency; it might also be that the original strain
died out and was replaced by another
61
.In some situa-
tions, the evolutionary response of microorganisms to
new environments might be accelerated by the acquisi-
tion of new genetic material through mechanisms such
as recombination or transposable elements. For exam-
ple, co-infection studies have shown that the host ranges
of baculoviruses can be expanded, which indicates that
genetic exchange between strains has occurred
49
.Also, in
bacteria that have been genetically marked, gene transfer
from a resident population to an introduced bacteria
has been detected
62
.
Macroorganisms. In contrast to microorganisms, in
introduced macroorganisms there is little or no evidence
that genetic adaptation is important for the success of
biological control, or even that evolution occurs in such
macroorganisms within the time period over which bio-
logical control is usually monitored
54
.This result seems
to hold even in cases for which there is evidence of
widespread production
48
.Gene transfer has also been
important in the study and use of baculoviruses for bio-
logical control. Several toxin genes have been inserted
into baculoviruses to increase their effectiveness in killing
their insect hosts; for example, Buthus eupeus (lesser
Asian scorpion) insect toxin 1 (BeIT), Androctonus
australis (North African scorpion) insect toxin (AaHIT)
and female mite Pyemotes tritici insect toxin (TxPI),
which have been used to genetically manipulate the
Autographa californica multiple nucleopolyhedrovirus
(AcMNPV)
49
.Gene-transfer technology also provides a
way of creating transgenic pest strains or unwanted vec-
tors of disease that eventually die out on their own. This
strategy, termed
AUTOCIDAL CONTROL, is not new, but with
transgenic strains the lethal transgenes can be designed
to activate in response to changes in environmental con-
ditions (such as temperature or diet) or in the presence
of another new gene
44
.Sex-ratio distortion, perhaps
mediated by symbiotic organisms such as Wolbachia,has
also been proposed as a strategy of control
52,53
.However,
for the last two tactics, issues of risk and ecological safety
have yet to be determined.
Evolution in biological control
A common assumption that underlies the effectiveness
of classical biological control is that predators, parasites,
pathogens and herbivores will adapt to exploit the new
habitat and the target pest
4,41
.Ifadaptation is important
in biological control, then preserving high levels of
genetic diversity in introduced populations is also
important. Although there is abundant evidence to
indicate that introduced microorganisms, such as
viruses and bacteria, do adapt quickly to new condi-
tions, there is little evidence for the importance of adap-
tation in effective biological control that involves
macroorganisms, such as insect parasites,
PAR ASITOIDS and
herbivores
54
.Whether this difference reflects a disparity
in generation time between microorganisms and macro-
organisms or whether there is something fundamentally
different about the role of evolution in biological con-
trol that involves microorganisms and macroorganisms
is not known, but the issue is crucial for long-term risk
assessment.
Microorganisms. Studies of microorganisms clearly
show that adaptive evolution can take place as part of
purposeful introductions and that such evolution
affects the dynamics of biological control. The biologi-
cal control of domestic European rabbits (Oryctolagus
cuniculus) with the myxoma pox virus
55,56
in Europe
and in Australia, where they were introduced in 1950,
best illustrate the point. The myxoma virus is indige-
nous to South America, where its natural reservoir is
the Brazilian rabbit Sylvilagus brasiliensis and its vectors
are biting mosquitoes. In S. brasiliensis,myxoma pro-
duces localized skin lesions that heal but are a source of
infection for the virus. In European rabbits, soon after
infection, the virus replicates in the lymphocytes and
destroys them
57
.Myxoma virus was first introduced to
Australia in 1950, and subsequently to France and
Britain in 1952 and 1953, respectively, with almost
AUTOCIDAL CONTROL
The introduction of an
organism that causes its own
population to decline without
interaction with other species.
PARA SITOID
An insect that kills only one host
individual in its lifetime and has
a free-living adult stage; this
differs from a predator, which
kills many host individuals in its
lifetime, and from a parasite,
which typically does not kill the
host and can persist for several
generations in one host.
SERPINS
Irreversible inhibitors of serine
proteases that regulate a diverse
array of physiological processes,
including apoptosis,
inflammation, angiogenesis,
complement activation,
fibrinolysis and coagulation.
894 | NOVEMBER 2003 | VOLUME 4 www.nature.com/reviews/genetics
REVIEWS
For example, in a study of plants that grew taller when
they were invasive,
COMMON GARDEN EXPERIMENTS showed
that this was a plastic response to the new environment
64
;
there was no evidence of post-invasion evolution despite
the increased size in colonizing populations.
Only one study, to our knowledge, has been designed
specifically to test whether adaptation has occurred in
macroorganisms after introduction. In a study of an
aphid parasitoid that was introduced from France to
New York, Hufbauer
5
compared presumed source and
introduced populations and found no evidence of adap-
tation. In fact, the introduced populations lost abilities
that would have made them more successful as para-
sites. These results, and others from continuing studies
65
,
indicate that genetically based adaptation has not
occurred in macroorganisms that have been introduced
for biological control over the timescales in which
they have been studied. It should be noted that genetic
variation might be crucial to initial establishment, but
expansion of host range and other non-target effects. In
a review of 10 biological control projects that involved
non-target effects, Louda et al.
11
argued that adaptation
was neither observed nor seemed important. The
authors noted, however, that these studies were not
specifically designed to test for adaptation. Several other
studies have failed to document changes associated with
species introduced for biological control that could not
be accounted for by pre-adaptation; that is, traits that
were present before introduction
41
.In an examination
of 352 introductions of insects for the biological control
of weeds, the cited examples of so-called ‘host shifts’
from the targeted weed species did not seem to result
from adaptive genetic change
63
;in other words, the use of
non-target hosts involved host species that would be
considered part of the
FUNDAMENTAL HOST RANGE of the
introduced herbivores. Even when introduced organisms
show marked morphological change after colonization,
there need not be a genetic basis for such modifications.
EPIZOOTICS
Outbreaks of organisms that
feed on other organisms.
VIRULENCE GRADES
Categories of virus virulence
that are based on host (rabbit)
survival (measured in days) and
case mortality (expressed as a
percentage).
FUNDAMENTAL HOST RANGE
The actual host range of a
species before any evolutionary
change.
COMMON GARDEN
EXPERIMENTS
Ecological transplant studies in
which organisms are reared
under identical conditons.
Epizootic
0 1234
Britain
Australia
France
Australia
1950 1960
1950 198019701960
a Rabbit case mortality
b Myxoma virulence
Figure 1 | Co-evolution and biological control. An illustration of the co-evolution of European rabbits and the myxoma pox virus,
which is associated with biological control in Australia, Britain and France
55,107
. a | Rabbit mortality is shown as the percentage
mortality of a group of infected rabbits (green) that have been challenged with the grade III strain of the virus (mean rabbit survival
time 17–28 days), as a function of the number of
EPIZOOTICS that the source population had experienced. b | Myxoma virulence is
plotted as the proportion of all field samples that fall into the following classes of
VIRULENCE GRADES sampled over time: grades I and II
combined (red; mean rabbit survival time <16 days and case mortality >95 %), and grades III–V combined (blue; mean rabbit
survival time >16 days and case mortality <95 %). Virulence might be higher in Britain than in France because one species of rabbit
flea, Spilopsyllus cuniculi, is probably the only important vector.
NATURE REVIEWS | GENETICS VOLUME 4 | NOVEMBER 2003 | 895
REVIEWS
Co-evolution. Relatively little is known about the
genetics of the interactions between hosts and natural
enemies that are associated with biological con-
trol
4,78–80
.As is the case with rabbits and the myxoma
virus, several pest insects have evolved resistance to
the microorganisms that were used to control them
54
,
and even to particular components of control, such as
the Bt protein
81,82
.Less is known about the genetic
interactions between pests and macroparasites or
predators. The few studies that have been carried out
indicate that genetic variation in both pests and ene-
mies is not limiting. Controlled quantitative genetic
experiments have found genetic differentiation
among host-associated populations for both para-
sitoids and pests
83,84
.These results are consistent with
molecular-marker data that document statistical dif-
ferences in gene frequencies between parasitoid pop-
ulations. For example, Vaughn and Antolin
85
found
genetic differences at a local scale between parasitoid
populations that were associated with two host aphid
species.
Why co-evolutionary responses have not been
observed more often in biological control manipula-
tions is not clear. As argued above, there might not have
been sufficient time to see these responses in macroor-
ganisms — studies of older introductions can test this
hypothesis. Alternatively, it has been argued that co-
evolution that involves reciprocal changes in virulence
and resistance is unlikely because of a lack of genotype-
specific virulence and defence
86
.Another hypothesis is
that natural selection in such systems might be relatively
weak as a result of subdivision among populations, with
each population being under slightly different selective
pressures
54
.Yet another model indicates that if the fre-
quency of encounters fluctuates between generations,
generalist host resistance and partial specialist parasitoid
virulence are favoured, which is consistent with the idea
of the host ‘hedging its bets’ in an unpredictable envi-
ronment
87
.This work indicates that asymmetries might
develop in host–parasitoid co-evolution, an idea that
has gained support from recent field studies
88
.In each of
these scenarios, genetic population structure and geo-
graphic variation of both pests and natural enemies are
clearly important
4,59
.
Neither the roles of hybridization and horizontal
gene exchange nor the interactions that are associated
with symbiosis have been well studied in the context of
biological control
53,89
.Insight into these issues will not
only add to our understanding of how evolution oper-
ates in managed systems, but will also bear directly on
the effectiveness of biological control. For example, it has
been argued that the lack of an evolutionary response by
pests allows biological control to be more evolutionarily
stable than chemical forms of pest control
54
.It has also
been proposed that hosts that are not adapted to their
parasites are more likely to be controlled, in what has
been referred to as the ‘new-association hypothesis’
90
,
and many cases of successful biological control involve
new associations. Whether biological control efforts that
involve new associations are successful more often than
those that do not has yet to be rigorously tested
4
.
that we only subsequently observe the successful geno-
types (that is, there is a non-reporting bias). Despite the
lack of evidence for any genetic basis for adaptation in
the above studies, ecological changes are common,
which shows that the introduced species already pos-
sessed the ability to exploit expanded ecological ranges.
If genetic adaptation is not important in the success
of biological control, then preserving the genetic diver-
sity of biological control organisms for this reason is not
crucial. Indeed, many invasive species thrive while expe-
riencing a genetic bottleneck that is associated with
the colonization of a new habitat
27,28,66
. One example is
the varroa mite, which is a parasite that switched from
the eastern honeybee Apis cerana to the European hon-
eybee Apis mellifera when the latter was introduced into
Asia for
APICULTURAL reasons
67
.The parasite has spread
almost worldwide in the past 30 years. Interestingly,
only 2 of the 18 Asian varroa
HAPLOTYPES that parasitize
A. cerana in Asia have switched to the new host
68
. One
of these two haplotypes has been successful in coloniz-
ing new geographical regions, despite the fact that its
genetic variability is low (M. Solignac et al., unpublished
observations). This haplotype has also subsequently
developed strong pesticide resistance
69
,which shows
that widespread ecological success does not depend
on initial genetic variability. One possible reason for
this is that diversity might recover quickly after a bot-
tleneck
70
,long before new alleles can arise through
mutation
71
(BOX 4).Alternatively, many insects might
be well adapted to withstand low genetic diversity; for
example, many successful insect parasitoids have low
levels of genetic diversity because of brother–sister
mating
22
or because of the effects of MICROBE-ASSOCIATED
PARTHENOGENESIS
54,72
.
That genetic adaptation is not important in the suc-
cess of biological control introductions is a bold supposi-
tion in light of the abundant evidence that indicates that
insects might switch plant hosts in evolutionary time,
often leading to the formation of new species
73,74
.Indeed,
studies that have examined adaptation over several hun-
dreds of years, such as those of Rhagoletis fruitflies on
hawthorn and apple
75
, show that plant host shifts can
lead to genetic divergence and even
SYMPATRIC SPECIATION.
In Hawaii, the moth genus Hedylepta contains five
species that feed on banana, which must have evolved
since the Polynesian introduction of banana to Hawaii
within the past 1,000 years
76
.So,why has adaptation not
been seen in systems that are the subject of biological
control? One explanation is that adaptation in macroor-
ganisms does occur, but there have not been enough
controlled studies to provide evidence
4,54,59
.Alternatively,
in the ecological timescale under which biological con-
trol takes place, there might not have been sufficient
time to see the evolutionary responses associated with
colonization that are typically observed over longer
periods
77
.For example, adaptation in Rhagoletis fruitflies
spans several hundred years; similar genetic changes in
organisms that are used for biological control might be
expected over the same time period. We might therefore
be observing only the first stages of more complex and
unpredictable evolutionary events in biological control.
APICULTURE
The practice of bee
domestication.
HAPLOTYPE
The allelic configuration of
multiple genetic markers that is
present on a single chromosome
of a given individual.
MICROBE-ASSOCIATED
PARTHENOGENESIS
The occurrence of reproduction
without males, which is caused
by the presence of a microbe.
SYMPATRIC SPECIATION
Genetic divergence that leads to
species formation in the same
habitat.
896 | NOVEMBER 2003 | VOLUME 4 www.nature.com/reviews/genetics
REVIEWS
and Stiling
4
note that “forecasting the impacts of non-
indigenous species is an imprecise science”, yet we know
that the effects of non-indigenous species can be devas-
tating both environmentally and economically
94,95
.In
comparison with the use of GMOs, classical biological
control can involve inserting entire non-indigenous
genomes into ecosystems. Adaptation and genetic
change is often assumed to be important in the success
of species that are used for biological control. By con-
trast, the manipulation of GMOs involves inserting
genes into genomes with the goal of creating an organ-
ism with superior properties for a trait of interest. As
yet, the risks that are associated with genetic modifica-
tion are not fully known, although there is a recognized
risk of their spread or hybridization with wild-type
organisms. Interestingly, scientific and public concern
over the unintended effects of classical biological control
is mild compared with concern over the release of
GMOs, despite the fact that the insertion of genes into an
organism that is already present in a given environment
might have considerably less effect than the introduc-
tion of a fully functional genome into a new environ-
ment. So, why is there limited protest about releasing
entire new genomes when so much concern is raised
over releasing organisms with a few new genes?
One reason for the greater concern about GMOs is
that single genes with relatively simple effects are usu-
ally inserted. So, evolutionary responses to a change in a
single gene might be more easily ascertained than
responses to an entire organism. Selection pressure is
also increased when the trait of interest is expressed
Community genetics. Recently, the topic of ‘community
genetics’ has emerged in community biology
91
.It recog-
nizes that non-equilibrium conditions are common, if
not the norm, in ecological communities and that evolu-
tion is important in community studies particularly in
species interactions. Although the need for a separate
name for this topic has been questioned — it can be con-
sidered as a combination of population and ecological
genetics
92
— its discussion has helped to re-emphasize
two important, although perhaps often forgotten, points:
that all ecological interactions have evolutionary histo-
ries and that interactions can evolve, even over short
periods of time
77,93
.Observations of species that are
involved in biological control can be interpreted in this
light — agents of biological control have the elements
that are necessary for adaptive change, namely heritable
genetic variation, and this variation is associated with
different hosts, at least in some cases. What is not under-
stood about biological control in this context is whether
the ability to feed on new hosts is the result of adaptation
or merely a consequence of a pre-existing wide funda-
mental host range. A community genetics viewpoint also
makes it obvious that there is no real distinction between
ecological and evolutionary timescales, but rather that
both ecology and evolution are important features of
biological communities.
GMOs versus biological control
The assessment of the risks that are associated with bio-
logical control is difficult and might even be impossible
— in a well-publicized review of the topic, Simberloff
Box 4 | What determines genetic diversity after a bottleneck?
Genetic bottlenecks occur when
samples of individuals are drawn from
larger populations, which is common
for populations of invasive species or
organisms that are introduced for
biological control
19,27,29
.In this process,
heterozygosity initially decreases, but
eventually is restored depending on the
size of the bottleneck and the growth
rate of the population. Alleles are also
lost, particularly rare ones.
Importantly, this loss of alleles is faster
than the loss of heterozygosity.
Accordingly, after the bottleneck there
is a transient excess in heterozygosity
relative to that predicted on the basis of
the number of alleles. The important
elements of this process are the period
of time since the beginning of the
bottleneck, the ratio between the
effective population size (N
e
) before
and after the bottleneck, the per-locus mutation rate and the sample size of genes. Cornuet and Luikart
29
suggested
several ways in which this excess in heterozygosity can be used to estimate the size of a bottleneck. Their analysis shows
that the time of the maximum excess and the magnitude of the excess will differ depending on the number of alleles —
see the comparison between two alleles (red) and five alleles (blue) in the example depicted. They also showed that loci
that follow the infinite-allele model, in which each unique allele is assumed to evolve only once, are the best suited for
detecting bottlenecks.
0.00
0.02
0.04
0.06
0.08
0.10
0.12
Heterozygosity excess
Time after bottleneck (in 2N
e
generations)
5 alleles
2 alleles
102345678
NATURE REVIEWS | GENETICS VOLUME 4 | NOVEMBER 2003 | 897
REVIEWS
detailed scientific guidelines for the biological-safety
testing of transgenic crops. The project involves rep-
utable scientists in developed and developing countries
who coordinate and implement the guidelines as a
dynamic process that will build local scientific and
technical expertise and communication among scien-
tists and policy makers. The aim of this project is to
rapidly publish sections of the guidelines as they are
completed and to revise previously published sections
in a timely manner. The underlying principle is that the
development and release of transgenic products should
only take place when a firm understanding of the
manipulated system and its potential environmental
implications has been achieved and maximum
biosafety measures are in place
48
.Whether similar
efforts are possible for assessing the risk associated with
biological control is not clear, but efforts are underway
as a result of a European Union-financed project to
evaluate the environmental risk of inundative biologi-
cal control
(BOX 1) introductions into Europe
60
.In this
project, risk evaluation centres around host range.
Particularly difficult to address in this regard will be the
potential role of evolution in biological control systems,
given their recent history and non-equilibrium nature.
The future of biological control
Given the present public concern about the use of
GMOs
14
, there is a question as to whether classical bio-
logical control has a future. The answer, in large part,
rests on the level of risk that policy makers and the pub-
lic are willing to accept, as well as on economic factors.
This risk can be assessed — pre-release host testing is
the norm for weed biological control programmes and
is becoming more common for the biological control
of other organisms, particularly insects
104
.The extent
of pre-testing that is possible is usually determined by
economics. Of course, acceptable levels of risk are con-
text dependent and the effectiveness and risks that are
associated with other methods of control (such as
chemical insecticides and GMO crops), as well as the
perception of the seriousness of the pest problem, are of
crucial importance. With the public endorsing ‘food
safety’ and agriculture with a low environmental
impact, carefully practiced biological control is still a
strong alternative to chemical pesticides and GMOs.
The assessment of risk would be easier if there was
a better understanding of what causes some organisms
to extend their host range or even to switch hosts.
Many biological control releases represent a series of
semi-natural experiments, each of which offers infor-
mation about factors that influence rates of establish-
ment and success. Understanding these experiments
requires post-release monitoring, which is typically
difficult to fund. The application of biological control
would gain in both precision and predictive power if
genetic and ecological approaches were combined in
experimental work.
One possibility would be to improve biological con-
trol organisms before their release — either through
gene transfer or traditional breeding — as has been done
for mites and insecticide resistance. Host specificity is
continually. For example, insects are likely to develop
resistance to transgenic crops with one new gene that
codes for a simple toxin that is constantly expressed,
and this has, indeed, occurred — Bt is a noted exam-
ple
47
.Specific risks, such as the likelihood that gene flow
will move inserted genes into new populations or new
species
96,97
,might be more easily identified for GMOs
than for introduced organisms
60
.For example, the
potential non-target species that are at risk might not
be well characterized, or even known, and many indi-
rect non-target effects are possible. Nevertheless, both
modelling and empirical studies have shed light on the
factors that are important in predicting the risks associ-
ated with classical biological control; these include the
potential for establishment, dispersal, host range, direct
and indirect non-target effects, competitiveness relative
to wild types and features of the environment
54,60,98–100
.
Risk assessment in biological control might be best
developed in studies of GM baculoviruses for which
there is a conceptual framework to organize relevant
factors in relation to one another
98
.This framework has
three main components: the identification of effects, the
identification of exposure and the evaluation of effects.
The first assesses whether genetic modification has pro-
duced changes in the virus beyond those that were
intended, such as changes to host range, sublethal
effects, genetic identity and toxicology of the foreign
gene product. The identification of exposure examines
genetic and environmental stability and the routes by
which a baculovirus might be dispersed from the release
site. Impact evaluation combines these data to predict
the potential influence of the baculovirus on the envi-
ronment, particularly non-target effects that are usually
assessed by phylogenetic extrapolation. There might
also be indirect effects, for example, mediated through
KEYSTONE SPECIES or indirect competition. Many factors,
such as environmental stability and dispersal routes, can
only be assessed by studies of wild-type baculovirus. For
example, models have been constructed to predict
whether an engineered strain of baculovirus will outcom-
pete wild-type strains. Results indicate that dominance of
one strain over another is likely, with the outcome depen-
dent on the speed with which the baculovirus kills its
host and how infectious it is
101
.Most recombinant
strains kill their hosts faster but are less infectious. The
difference in infectiousness must be small for engineered
viruses to have competitive advantage. However, even if
engineered strains are at a competitive disadvantage, they
might remain in the environment for decades before
going extinct. Both environmental variables and the
densities of introduced material might also affect com-
petitive interactions. Similar to the studies of bac-
uloviruses, studies of bacteria have documented few dif-
ferences in survival, spread, persistence in the field and
ecological effects between GM bacteria and the corre-
sponding unmodified parent strains
62
. Likewise, studies
of the predatory mite Metaseiulus occidentalis have com-
pared biological traits between transgenic and wild
colonies under laboratory conditions
102
.
Capalbo et al.
103
have recently documented progress
in the development of scientific principles and have
KEYSTONE SPECIES
Species in ecological
communities that have
disproportional direct and
indirect effects on other species,
which are usually regulated
through top-down processes,
such as predation.
898 | NOVEMBER 2003 | VOLUME 4 www.nature.com/reviews/genetics
REVIEWS
planned, the organism will multiply faster in the field
than in the laboratory and disperse freely, thereby obvi-
ating the need for a commercial rearing facility.
For GMOs (and synthetic pesticides), patent rights and
the need to continually manufacture the product to sell
it, allow organizations to continue to profit from their
research investment. So, although the transgenetic alter-
ation of natural enemies has much merit, in reality it is
still not feasible.
As a test bed, biological control has much to offer —
efforts in classical biological control have resulted in
numerous semi-natural field experiments that can be
used to test hypotheses about the factors that influence
establishment in new habitats and successful pest con-
trol. In particular, much is to be learned at the interface
of genetics and community ecology. Whether biological
control can be a predictive science is less obvious, but
clearly the answer will depend on the timescale over
which studies are possible.
crucial in the selection and approval of new biological
control agents, and so would be the highest priority
attribute for genetic manipulation. Yet, identifying and
transferring the appropriate genes would pose consider-
able challenges. Diverse organisms are now used for bio-
logical control and great effort would be required for
each new agent. Also, the reasons why a particular bio-
logical control agent is effective or ineffective depend on
a complex interaction of factors: adaptation to local
environment, susceptibility to predators or parasites,
adaptation to the target and so on. Unfortunately, solv-
ing one of these problems by genetic manipulation does
not guarantee that the agent will not be limited by
another. As well as insufficient funding for pre-release
testing and post-release monitoring, there is little eco-
nomic incentive to develop more efficient biological
control agents. Once an agent is released it becomes
public property and no one company can monopolize
the profits. Also, if biological control proceeds as
1. Eilenberg, J., Hajek, A. & Lomer, C. Suggestions for unifying
the terminology in biological control. BioControl 46, 387–400
(2001).
A guide to the different types of biological control
procedures and their uses.
2. Garcia, R., Caltagirone, L. E. & Gutierrez, A. P. Comments on
a redefinition of biological control. Roundtable. BioScience 38,
692–694 (1988).
3. Kareiva, P. Contributions of ecology to biological control.
Ecology 77, 1963–1964 (1996).
4. Simberloff, D. & Stiling, P. How risky is biological control?
Ecology 77, 1965–1974 (1996).
A rigorous assessment of the risks that are
associated with biological control, including non-
target effects.
5. Hufbauer, R. A. Evidence for nonadaptive evolution in
parasitoid virulence following a biological control introduction.
Ecol. Appl. 12, 66–78 (2002).
A review of the ecological and evolutionary changes
that are associated with biological control.
6. Hall, R. W. & Ehler, L. E. Rate of establishment of natural
enemies in classical biological control. Bull. Entomol. Soc.
Am. 25, 280–282 (1979).
7. Van Lenteren, J. C. The potential of entomophagous
parasites for pest control. Agri. Ecosyst. Environ. 10, 143–158
(1983).
8. Crawley, M. J. The successes and failures of weed biocontrol
using insects. Biocontrol News Info. 10, 213–223 (1989).
9. Lawton, J. H. in Alternative to the Chemical Control of Weeds:
Proceedings of an International Conference (eds Bassett, C.,
Whitehouse, L. J. & Zabkiewicz, J. A.) (Ministry of Forestry,
Rotorua, New Zealand, 1990).
A food-web approach to assess the trophic
relationships that are associated with biological
control introductions in Hawaii.
10. Henneman, M. L. & Memmott, J. Infiltration of a Hawaiian
community by introduced biological control agents. Science
293, 1314–1316 (2001).
11. Louda, S. M., Pemberton, R. W., Johnson, M. T. &
Follett, P. A. Nontarget effects — the Achilles’ heel of
biological control? Retrospective analyses to reduce risk
associated with biocontrol introductions. Annu. Rev. Entomol.
48, 365–396 (2003).
An examination of non-target effects in 10 case
studies of biological control.
12. Boettner, G. H., Elkinton, J. S. & Boettner, C. J. Effects of a
biological control introduction on three nontarget native
species of saturniid moths. Conserv. Biol. 14, 1798–1806
(2000).
13. Howarth, F. G. Environmental impacts of classical biological
control. Annu. Rev. Entomol. 36, 485–511 (1991).
A highly publicized synthesis of the environmental
effects that are associated with classical biological
control.
14. Hoddle, M. S. Is the anti-GMO stance warrented in Australia
and New Zealand? Myrmecia May 10–11 (2003).
15. Unruh, T. R. & Woolley, J. B. in Handbook of Biological Control
(eds Bellows, T. S. & Fisher, T. W.) 57–85 (Academic Press,
San Diego, California, 1999).
16. Ehler, L. E., Sforza, R. & Mateille, T. (eds.) Genetics,
Evolution, and Biological Control 288 (CABI, New York, in
the press).
An up-to-date review of genetics and evolutionary
biology as applied to biological control.
17. Ashburner, M., Hoy, M. A. & Peloquin, J. J. Prospects for the
genetic transformation of arthropods. Insect Mol. Biol. 7,
201–213 (1998).
18. Navajas, M., Lagnel, J., Gutierrez, J. & Boursot, P. Species
wide homogeneity of nuclear ribosomal ITS2 sequences in
the spider mite Tetranychus urticae contrasts with extensive
mitochondrial COI polymorphism. Heredity 80, 742–752
(1998).
19. Roderick, G. K. in Genetics, Evolution, and Biological Control
(ed. Mateille, T.) (CABI, New York, in the press).
20. Waage, J. in Critical Issues in Biological Control (eds
Mackauer, M., Ehler, L. E. & Roland, J.) 135–157 (Intercept,
Andover, United Kingdom, 1990).
21. Slatkin, M. Gene flow and the geographic structure of natural
populations. Science 236, 787–792 (1987).
22. Roderick, G. K. Geographic structure of insect populations:
gene flow, phylogeography, and their uses. Annu. Rev.
Entomol. 41, 263–290 (1996).
23. Holder, M. & Lewis, P. O. Phylogeny estimation: traditional
and bayesian approaches. Nature Rev. Genetics 4, 275–284
(2003).
24. Rosenberg, N. A. & Nordborg, M. Genealogical trees,
coalescent theory and the analysis of genetic polymorphisms.
Nature Rev. Genetics 3, 380–390 (2002).
25. Lanciotti, R. S. et al. Origin of the West Nile virus responsible
for an outbreak of encephalitis in the northeastern United
States. Science 286, 2333–2337 (1999).
26. Gaskin, J. F. & Schaal, B. A. Hybrid Tamarix widespread in
U.S. invasion and undetected in native Asian range. Proc. Natl
Acad. Sci. USA 99, 11256–11259 (2002).
27. Estoup, A., Wilson, I. J., Sullivan, C., Cornuet, J. M. & Moritz,
C. Inferring population history from microsatellite and enzyme
data in serially introduced cane toads, Bufo marinus. Genetics
159, 1671–1687 (2001).
28. Davies, N., Villablanca, F. X. & Roderick, G. K. Determining the
sources of individuals in recently founded populations:
multilocus genotyping in non-equilibrium genetics. Trends
Ecol. Evol. 14, 17–21 (1999).
29. Cornuet, J. M. & Luikart, G. Description and power analysis of
two tests for detecting recent population bottlenecks from
allele frequency data. Genetics 144, 2001–2014 (1996).
30. Hufbauer, R. A., Bogdanowicz, S. M. & Harrison, R. G. The
population genetics of a biological control introduction:
mtDNA and microsatellite variation in native and introduced
populations of Aphidius ervi, a parasitoid wasp. Mol. Ecol. (in
the press).
31. Cornuet, J. M., Piry, S., Luikart, G., Estoup, A. & Solignac, M.
New methods employing multilocus genotypes to select or
exclude populations as origins of individuals. Genetics 153,
1989–2000 (1999).
32. Pritchard, J. K., Stephens, M. & Donnelly, P. Inference of
population structure using multilocus genotype data. Genetics
155, 945–959 (2000).
33. Rannala, B. & Mountain, J. L. Detecting immigration using
multilocus genotypes. Proc. Natl Acad. Sci. USA 94,
9197–9201 (1997).
34. Navajas, M. et al. Field releases of the predatory mite
Neoseiulus fallacis (Acari: Phytoseiidae) in Canada, monitored
by pyrethroid resistance and allozyme markers. Biol. Control
20, 191–198 (2001).
35. Bohonak, A. J., Davies, N., Villablanca, F. X. & Roderick, G. K.
Invasion genetics of New World medflies: testing alternative
colonization scenarios. Biol. Invasions 3, 103–111 (2001).
36. Beerli, P. & Felsenstein, J. Maximum likelihood estimation
of a migration matrix and effective population sizes in
n subpopulations by using a coalescent approach. Proc. Natl.
Acad. Sci. USA 98, 4563–4568 (2001).
37. Nielsen, R. & Wakeley, J. Distinguishing migration from
isolation: a Markov chain Monte Carlo approach. Genetics
158, 885–896 (2001).
38. Navajas, M. et al. Genetic structure of the greenhouse
population of the spider mite Tetranychus urticae:
spatio-temporal analysis with microsatellite markers.
Insect Mol. Biol. 11, 157–165 (2002).
39. Storz, J. F. & Beaumont, M. A. Testing for genetic evidence of
population expansion and contraction: an empirical analysis
of microsatellite DNA variation using a hierarchical Bayesian
model. Evolution 56, 154–166 (2002).
40. Roush, R. T. & Hopper, K. R. Use of single-family lines to
preserve genetic-variation in laboratory colonies. Ann.
Entomol. Soc. Am. 88, 713–717 (1995).
41. Hopper, K. R., Roush, R. T. & Powell, W. Management of
genetics of biological control introductions. Annu. Rev.
Entomol. 38, 27–51 (1993).
42. Perez-Maluf, R., Kaiser, L. & Wajnberg, E. Genetic variability of
conditioned probing responses to a fruit odor in Leptopilina
boulardi (Hymenoptera: Eucoilidae), a Drosophila parasitoid.
Behav. Genet. 28, 67–73 (1998).
43. Bidochka, M. J. et al. Fate of biological control
introductions: monitoring an Australian fungal pathogen of
grasshoppers in North America. Proc. Natl Acad. Sci. 93,
918–921(1996).
44. Handler, A. M. & Beeman, R. W. United States Department
of Agriculture Agricultural Research Service: advances in
the molecular genetic analysis of insects and their
application to pest management. Pest Manag. Sci. 59,
728–735 (2003).
A summary of the ways in which molecular genetic
approaches are used in pest management.
45. Whitten, M. J. & Hoy, M. in Handbook of Biological Control
(eds Bellows, T. S. & Fisher, T. W.) 271–296 (Academic Press,
San Diego, California, 1999).
46. Ding, X. et al. Insect resistance of transgenic tobacco
expressing an insect chitinase gene. Transgenic Res. 7,
77–84 (1998).
47. Federici, B. A. in Handbook of Biological Control (eds Bellows,
T. S. & Fisher, T. W.) 575–593 (Academic Press, San Diego,
California, 1999).
48. Hilbeck, A. Implications of transgenic, insecticidal plants for
insect and plant biodiversity. Perspect. Plant Ecol. Evol. Syst.
4, 43–61 (2001).
NATURE REVIEWS | GENETICS VOLUME 4 | NOVEMBER 2003 | 899
REVIEWS
49. Possee, R. D., Barnett, A. L., Hawtin, R. E. & King, L. A.
Engineered baculoviruses for pest control. Pestic. Sci. 51,
462–470 (1997).
50. Holloway, A. J., Van Laar, R. K., Tothill, R. W. &
Bowtell, D. D. L. Options available — from start to finish —
for obtaining data from DNA microarrays. Nature Genet. 32
(Suppl.), 481–489 (2002).
51. Hui, D. et al. Molecular interactions between the specialist
herbivore Manduca sexta (Lepidoptera, Sphingidae) and its
natural host Nicotiana attenuata: V. Microarray analysis and
further characterization of large-scale changes in herbivore-
induced mRNAs. Plant Physiol. 131, 1877–1893 (2003).
52. Sinkins, S. P., Curtis, C. F. & O’Neill, S. L. in Influential
Passengers: Inherited Microorganisms and Arthropod
Reproduction (eds O’Neill, S. L., Hoffmann, A. A. &
Werren, J. H.) 155–208 (Oxford Univ. Press, Oxford, 1997).
53. Stouthamer, R. in Genetics, Evolution, and Biological
Control (eds. Ehler, L. E., Sforza, R. & Mateille, T.) 235–252
(CABI, New York, in the press).
54. Holt, R. D. & Hochberg, M. E. When is biological control
evolutionarily stable (or is it)? Ecology 78, 1673–1683 (1997).
55. Fenner, F. Biological control as exemplified by small pox
eradication and myxomatosis. Proc. R. Soc. Lond. B 218,
259–285 (1983).
56. Fenner, F. Adventures with poxviruses of vertebrates. FEMS
Microbiol. Rev. 24, 123–133 (2000).
57. Best, S. M., Collins, S. V. & Kerr, P. J. Coevolution of host and
virus: cellular localization of virus in myxoma virus infection of
resistant and susceptible European rabbits. Virology 277,
76–91 (2000).
58. Zúñiga, M. C. A pox on thee! Manipulation of the host immune
system by myxoma virus and implications for viral–host co-
adaptation. Virus Res. 88, 17–33 (2002).
An update outlining the present understanding of the
genetic mechanisms that are associated with
myxoma/rabbit co-evolution.
59. Thompson, J. N. Evolutionary ecology and the conservation
of biodiversity. Trends Ecol. Evol. 11, 300–303 (1996).
60. van Lenteren, J. C. et al. Environmental risk assessment of
exotic natural enemies used in inundative biological control.
Biocontrol 48, 3–38 (2003).
61. Hajek, A. E., Humber, R. A. & Elkinton, J. S. Mysterious origins
of Entomophaga maimaiga in North America. Am.
Entomolog. 41, 31–42 (1995).
62. Amarger, N. Genetically modified bacteria in agriculture.
Biochimie 84, 1061–1072 (2002).
63. van Klinken, R. D. & Edwards, O. R. Is host-specificity of weed
biological control agents likely to evolve rapidly following
establishment? Ecol. Lett. 5, 590–596 (2002).
64. Willis, A. J., Memmott, J. & Forrester, R. I. Is there evidence for
the post-invasion evolution of increased size among invasive
plant species? Ecol. Lett. 3, 275–283 (2000).
65. Cromartie, W. J. Is adaptation after release necessary for
successful classical biological control? Entomol. News 108,
252, 258, 304 (1997).
66. Tsutsui, N. D., Suarez, A. V., Holway, D. A. & Case, T. J.
Reduced genetic variation and the success of an invasive
species. Proc. Natl Acad. Sci. USA 97, 5948–5953 (2000).
An elegant study of the effects of reduced genetic
diversity, which led to changes in behaviour that were
associated with the invasion biology of Argentine ants.
67. De Jong, D. D., Morse, R. A. & Eickwort, G. C. Mite pests of
honey bees. Annu. Rev. Entomol. 27, 229–252 (1982).
68. Anderson, D. L. & Trueman, J. W. H. Varroa jacobsoni (Acari:
Varroidae) is more than one species. Exp. App. Acarol. 24,
165–189 (2000).
69. Elzen, P. J., Eischen, F. A., Baxter, J. R., Elzen, G. W. & Wilson,
W. T. Detection of resistance in US Varroa jacobsoni Oud.
(Mesostigmata: Varroidae) to the acaricide fluvalinate.
Apidologie 30, 13–17 (1999).
70. Nei, M., Maruyama, T. & Chakraborty, R. The bottleneck effect
and genetic variability in populations. Evolution 29, 1–10
(1975).
71. Luikart, G. & Cornuet, J. M. Empirical evaluation of a test for
identifying recently bottlenecked populations from allele
frequency data. Conserv. Biol. 12, 228–237 (1998).
72. Stouthamer, R. & Kazmer, D. J. Cytogenetics of microbe-
associated parthenogenesis and its consequences for gene
flow in Trichogramma wasps. Heredity 73, 317–327 (1994).
73. Dres, M. & Mallet, J. Host races in plant-feeding insects and
their importance in sympatric speciation. Phil. Trans. R. Soc.
Lond. B 357, 471–492 (2002).
74. Farrell, B. D. “Inordinate fondness” explained: why are there
so many beetles? Science 281, 555–559 (1998).
75. Berlocher, S. H. & Feder, J. L. Sympatric speciation in
phytophagous insects: moving beyond controversy? Annu.
Rev. Entomol. 29, 403–433 (2002).
76. Zimmerman, E. C. Possible evidence of rapid evolution in
Hawaiian moths. Evolution 14, 137–138 (1960).
77. Gillespie, R. G. & Roderick, G. K. Arthropods on islands:
colonization, speciation, and conservation. Annu. Rev.
Entomol. 47, 595–632 (2002).
78. Sorci, G., Moller, A. P. & Boulinier, T. Genetics of host–parasite
interactions. Trends Ecol. Evol. 12, 196–200 (1997).
79. Thompson, J. N. The Coevolutionary Process (Univ. of
Chicago Press, Chicago, 1994).
80. Tuda, M. & Bonsall, M. B. Evolutionary and population
dynamics of host–parasitoid interactions. Res. Popul. Ecol.
41, 81–91 (1999).
81. Morin, S. et al. Three cadherin alleles associated with
resistance to Bacillus thuringiensis in pink bollworm. Proc.
Natl Acad. Sci. USA 100, 5004–5009 (2003).
82. Genissel, A. et al. Initial frequency of alleles conferring
resistance to Bacillus thuringiensis poplar in a field population
of Chrysomela tremulae. Proc. R. Soc. Lond. B 270, 791–797
(2003).
83. Hufbauer, R. A. Pea aphid–parasitoid interactions: have
parasitoids adapted to differential resistance? Ecology 82,
717–725 (2001).
84. Hufbauer, R. A. & Via, S. Evolution of an aphid–parasitoid
interaction: variation in resistance to parasitism among aphid
populations specialized on different plants. Evolution 53,
1435–1445 (1999).
85. Vaughn, T. T. & Antolin, M. F. Population genetics of an
opportunistic parasitoid in an agricultural landscape. Heredity
80, 152–162 (1998).
86. Kraaijeveld, A. R. & Godfray, H. C. J. Geographic patterns in
the evolution of resistance and virulence in Drosophila and its
parasitoids. Am. Nat. 153, 61–74 (1999).
87. Lapchin, L. Host–parasitoid association and diffuse
coevolution: when to be a generalist? Am. Nat. 160, 245–254
(2002).
88. Baker, D. A. The significance of host specialization in the
parasitoid wasp, `Diaeretiella rapae’ (M’Intosh). Thesis, Univ.
of Western Australia (2002).
89. O’Neill, S. L., Hoffmann, A. A. & Werren, J. H. (eds). Influential
Passengers: Inherited Microorganisms and Arthropod
Reproduction (Oxford Univ. Press, Oxford, 1997).
90. Hokkanen, H. M. T. & Pimentel, D. New associations in
biological control: theory and practice. Can. Entomol. 121,
829–840 (1989).
91. Neuhauser, C. et al. Community genetics: expanding the
synthesis of ecology and genetics. Ecology 84, 545–558
(2003).
92. Ricklefs, R. E. Genetics, evolution, and ecological
communities. Ecology 84, 588–591 (2003).
93. Palumbi, S. R. Evolution — humans as the world’s greatest
evolutionary force. Science 293, 1786–1790 (2001).
A review that documents the role of evolution in
human civilization.
94. Wilcove, D. S., Rothstein, D., Dubow, J., Phillips, A. & Losos,
E. Quantifying threats to imperiled species in the United
States. BioScience 48, 607–615 (1998).
95. Pimentel, D., Lack, L., Suniga, R. & Morrison, D.
Environmental and economic costs of nonindigenous species
in the United States. Bioscience 50, 53–65 (2000).
96. Hoy, M. A. Transgenic arthropods for pest management
programs: risks and realities. Exp. App. Acarol. 24, 463–495
(2000).
97. Soboleva, T. K., Shorten, P. R., Pleasants, A. B. & Rae, A. L.
Qualitative theory of the spread of a new gene into a resident
population. Ecol. Modell. 163, 33–44 (2003).
98. Richards, A., Matthews, M. & Christian, P. Ecological
considerations for the environmental impact evaluation of
recombinant baculovirus insecticides. Annu. Rev. Entomol.
43, 493–517 (1998).
99. Jung, C. & Croft, B. A. Aerial dispersal of phytoseiid mites
(Acari: Phytoseiidae): estimating falling speed and
dispersal distance of adult females. Oikos 94, 182–190
(2001).
100. Pearson, D. E. & Callaway, R. M. Indirect effects of
host-specific biological control agents. Trends Ecol. Evol.
18, 456–461 (2003).
101. Dushoff, J. & Dwyer, G. Evaluating the risks of engineered
viruses: modeling pathogen competition. Ecol. Appl. 11,
1602–1609 (2001).
102. Li, J. & Hoy, M. A. Adaptability and efficacy of transgenic and
wild-type Metaseiulus occidentalis (Acari: Phytoseiidae)
compared as part of a risk assessment. Exp. App. Acarol. 20,
563–573 (1996).
103. Capalbo, D. M. F. et al. Brazil and the development of
international scientific biosafety testing guidelines for
transgenic crops. J. Invertebr. Pathol. 83, 104–106 (2003).
104. Hoddle, M. in Encyclopedia of Insects (eds Resh, V. H. &
Cardé, R.) (Academic Press, San Diego, California, 2003).
105. Ravesberg, W. J. in Advances in Insect Rearing for
Research and Pest Management (eds Anderson, T. E. &
Leppla, N. C.) 465–487 (Westview, Boulder, Colorado,
1992).
106. Whitlock, M. C. & McCauley, D. E. Indirect measures of gene
flow and migration: FST not equal to 1/(4Nm+1). Heredity 82,
117–125 (1999).
107. Dwyer, G., Levin, S. A. & Buttel, L. A simulation-model of the
population-dynamics and evolution of myxomatosis. Ecol.
Monogr. 60, 423–447 (1990).
Acknowledgements
We thank R. Gillespie, R. Hufbauer, O. Edwards, B. Croft, M. Hoddle,
L. Smith and three anonymous reviewers for valuable insights and
suggestions. This work is supported by grants from the National
Science Foundation, the United States Department of Agriculture, the
California Department of Food and Agriculture, the University of
California and the French Institut National de la Recherche
Agronomique.
Online links
DATABASES
The following terms in this article are linked online to:
SwissProt: http://us.expasy.org/sprot
M11L | M-T1 | M-T4 | M-T5 | M-T7 | MGF | SERP1 | SERP2 |
SERP3 | TxPI
FURTHER INFORMATION
Berkeley Natural History Museums:
http://bnhm.berkeley.museum
Biological control: a guide to natural enemies in North
America: http://www.nysaes.cornell.edu/ent/biocontrol
CABI-Bioscience: http://www.cabi-bioscience.org/html/
Biocontrol.htm
French Institut National de la Recherche Agronomique in
Montpellier: http://www.montpellier.inra.fr/CBGP
International Organisation for Biological Control (IOBC):
http://www.oilb.agropolis.fr
University of California Berkeley’s Gump South Pacific
Research Station in Moorea, French Polynesia:
http://moorea.berkeley.edu
Access to this interactive links box is free online.
