Population
Genetics of Plant Pathogens
APSnet
Education Center Advanced Topics
Gene and Genotype Flow
Gene flow is a unifying force that prevents populations from diverging. Gene flow breaks down the geographical or other boundaries that could otherwise isolate populations. As a result of isolation between populations, and the consequent limitations in exchange of genes, we expect that populations will diverge by genetic drift or as a result of selection for alleles that adapt each population to its local niche. But if gene flow occurs at a sufficiently high level, then otherwise isolated populations will not diverge genetically. Instead, they become united and evolve as a single evolutionary unit. Gene flow is especially important for plant pathogens in agroecosystems because it is the process that introduces new genes into agricultural fields distant from the site of the original mutation. This process is probably important in natural ecosystems as well. Gene flow moves virulence alleles into new populations. Gene flow thus introduces new alleles that can displace old alleles, if they are better adapted to the current host.
In populations which are made up of one or a few clonal lineages, a special case of gene flow can occur in which each clone (i.e. a genotype) has several mutations that differentiate it from the dominant, pre-existing clone. Given the fact that many genes move together as a block in asexual clones, it is better to think of "genotype flow." Genotype flow then refers to the movement of entire genotypes (usually clones or clonal lineages) between distinct populations. Genotype flow occurs only for organisms that have a significant asexual component to their life cycle. As an example, genotype flow occurs when a genotype (clone) of Fusarium oxysporum f. sp. melonis (cause of Fusarium wilt on melons) moves from North America to Israel on the muddy boots of an agricultural scientist. In this case, F. oxysporum does not have a sexual cycle, so the entire set of alleles in the clone is introduced into a new population. If this clone has a high degree of fitness, it can become established in the new location. Though recombination is possible for bacteria and viruses, it is reasonable to consider these pathogens as exhibiting mainly genotype flow, while fungi can exhibit a mixture of gene and genotype flow.
Population Subdivision and Gene Flow
The population subdivision that results from genetic drift can be overcome by gene flow. The easiest model to consider how this process works is the Continent-Island model proposed by the population geneticist Sewall Wright. The following example will illustrate this model.
Assume that a is the virulent mutant allele that occurs at an avirulence locus (and A is the corresponding avirulence allele). Assume further that the frequency of the mutant a allele is q. Represent the frequency of the a allele as f(a) and the frequency of the A allele is f(A).
Figure 6. The continent-island model assumes that gene flow occurs in only one direction, from a donor population (continent) to a recipient population (island).
Let:
m = the proportion of the island population that consists of migrants
1-m = the proportion of the island population that consists of natives
Q = the frequency of the virulence allele a in the "donor" (continent) population
qo = the frequency of the virulence allele a in the "recipient" (island) population
After one cycle of gene flow, we find that:
q1 = (1-m)qo + mQ
q = -m(qo - Q)
where
This formula can be used to calculate how fast allele frequencies will change through gene flow. As an example, let’s consider the hypothetical movement of a virulence allele for leaf rust from the UK to France.
f(a) = 0.50, the UK population has a high frequency of the virulence allele because most UK wheats have Lr13 R-gene.
qo = 0.00,
m = 0.05, the migration rate is high because a large number of spores were deposited by a migration event caused by a wind storm moving spores across the English Channel.
q
= -0.05(0.00-0.50) = 0.025
q1 = 0.025 ~3% of the French population now contains avrLr13
At equilibrium (after many cycles of gene flow driven by many storms sweeping across the English Channel), allele frequencies of the donor and recipient populations become the same, qo = Q. So the frequency of avrLr13 will go to 0.50 in France even if French wheat breeders never use Lr13 in their resistant wheat cultivars. This is one possible explanation for the unexpected high frequency of virulence alleles in populations of some pathogens when the host population lacks the corresponding resistance gene (Bousset et al. 2002; Caffier et al. 1996; Hovmoller 2001).
Other Models for Gene Flow
Many other models of gene flow have been described in addition to the island model. Figure 7 shows examples of one- and two-dimensional stepping-stone models and more complex multidimensional models of gene flow. Each of these models represents a permutation of the same scheme and can be adapted to the reality of the agricultural or natural ecosystem under study.
 |
| Figure 7. Illustration of different models of gene flow. A) Continent-island model; B) Full island model; C) One-dimensional stepping stone model; and D) Two-dimensional stepping stone model. Click here to see an enlargement of this figure. |
The end result of gene flow is to make populations become genetically similar. This is illustrated in Figure 8, which shows how quickly geographically separated populations converge on the same allele frequency when 10% of each population is made up of immigrants from the other populations.
 |
| Figure 8. Change of allele frequency against time in five subpopulations exchanging migrants at the rate m=0.1 per generation. Note the rapid convergence to a common equilibrium frequency. |
Examples of Gene Flow in Plant Pathology
Several good examples of long-distance regional and global gene flow exist for fungal pathogens in agricultural ecosystems.
 |
| Figure 9. Shared RFLP alleles at the pSTL10 RFLP locus in Mycosphaerella graminicola populations from 3 continents. |
Example 1: Evidence for global gene flow among populations of the
wheat leaf blotch pathogen Mycosphaerella graminicola (anamorph Septoria
tritici). RFLP (restriction fragment length polymorphism) alleles are
shared between populations around the world (Figure 9) and allele
frequencies are remarkably similar among populations on different
continents (Table 2). But no isolates with shared DNA fingerprints were
found in different populations (Zhan et al. 2003). This shows that the
individual genotype that moved to a new population did not persist, but
its genes were passed into the recipient population through its sexual
offspring. The greatest gene diversity was found in the population from
Israel, which is the center of origin of the wheat host (Zhan et al.
2003). The center of diversity for the pathogen suggests that this also is
the center of origin of M. graminicola. This fits the standard
model for gene flow. Zhan et al. (2003) hypothesized that ascospores
disseminate genes over distances of 100s of km, while infected seeds
disseminate genes between continents.
Example 2: Evidence for regional genotype flow for the banana
wilt pathogen Fusarium oxysporum f. sp. cubense, which
causes Panama disease DNA fingerprints detected the same genotypes in different countries
(Koenig et al. 1997; Bentley et al. 1998). This fungus probably moves
regionally and between plantations on infected banana cuttings that are
used to start new plantations. The greatest genotypic diversity in the
pathogen population was found at the center of origin of bananas which is
in Southeast Asia. Example 3: Evidence for global movement of a single clone of the
potato late blight pathogen Phytophthora infestans. DNA
fingerprints (Figure 11, Goodwin et al. 1992) were used to show that the
global pandemic in the 1840s was most likely due to movement of a single
clone out of Mexico, which is the center of diversity and the likely
center of origin of this fungus. After moving
into North America, the fungus migrated on infected potatoes to Europe,
and then migrated globally via trade (Figure 12, Goodwin 1997; Goodwin et
al. 1994). This fungus requires two mating types for sexual reproduction.
Since only one mating type escaped originally, all P. infestans
populations were asexual until recently. Beginning in the late 1970s, new
clones "escaped" from Mexico, including the opposite mating type
and now there is increasing genotypic diversity in P. infestans
populations worldwide. The metalaxyl fungicides do not work as well
against the "new" populations and new populations are beginning
to show signs of sexual reproduction. The first confirmation of the A2
mating type outside of Mexico was in Switzerland in 1980. The effects of genetic drift can be overcome by gene flow. If enough
individuals are exchanged between two populations that are experiencing
independent genetic drift, then the drifting populations become
genetically linked and population subdivision will not occur. Sewall
Wright explained this best with his population genetic parameter Nem.
As before, Ne is the effective population size (a measure of genetic
drift), and m is
the percentage of the recipient population made up of immigrants (a
measure of gene flow). The product of these two parameters, Nem,
is a measure of the average number of migrants exchanged among populations
each generation. A value for Nem can be estimated using a
measure of population subdivision called FST, or by using private alleles,
alleles found only in one population. If Nem = 0, no migrants are exchanged between populations.
The result is that different alleles can be fixed in different populations
through genetic drift. Populations diverge and population subdivision
occurs. If Nem >1, meaning that on average one or more
individuals are exchanged between populations each generation, then
populations will not diverge by genetic drift and they will gradually
become similar. Very little gene flow is needed to counteract genetic
drift. If Nem = 1, the effects of drift are exactly counterbalanced
by the effects of gene flow, and the populations do not diverge or
converge. This principle is best illustrated with an example. Assume p = q = 0.5, in other words, the two alleles at a locus are
present at equal frequencies. With Ne = 10, the effects of drift are expected to be large:
Var(p) = 0.0125 (s.e. 0.11). In this population, 1 immigrant (Nem
= 1) corresponds to m = 0.10; thus, 10% of the population is made up of
migrants. To counteract a small Ne, m must be relatively large. With Ne = 10,000, the effects of drift are expected to be small: Var(p)
= 0.0000125 (s.e. 0.0035). In this population, 1 immigrant (Nem
= 1) corresponds to m = 0.0001; thus, one-hundredth of one percent of the
population is made up of migrants. To counteract a large Ne, m
can be very small. A metapopulation is a set of local populations connected by migrating
individuals. The local populations may undergo repeating cycles of
extinction and recolonization, while the metapopulation can remain
relatively constant. A metapopulation is a population of populations. To understand metapopulations, it helps to realize that populations are
never really at equilibrium (except in mathematical models), so you can
consider a species as a collection of small populations that are not at
equilibrium. Metapopulation models may offer a good representation of how pathogens
evolve in agroecosystems, especially if the pathogen is a biotroph that
cannot exist without a living host. In agricultural ecosystems, a new
niche for a pathogen opens when a field is planted to a susceptible crop.
Colonization (in this case possibly representing a founder
effect) occurs when the pathogen encounters the crop. The pathogen niche is
removed when the crop is harvested. After the plant host is removed, the
pathogen population goes extinct or experiences a bottleneck. If the pathogen produces
long-lived overseasoning survival structures, then the metapopulation
model is not such a good representation. As an example of plant pathogens that fit the metapopulation model
quite well, consider the case of the cereal rusts that colonize wheat,
barley, and oats each year in the "Puccinia pathway" in North
America, illustrated in Figure 13. As a result of removal of the alternate barberry host, the
overwintering stage of the wheat stem rust fungus Puccinia graminis f.
sp. tritici is practically non-existent in this area. Many rust fungi
overwinter in the southern-most state of Texas, or in Mexico. Spores move
north on the prevailing wind, following the developing cereal crops, and
arrive in Canada in time to infect spring-planted cereals in the summer.
The specific pathotypes that colonize each cereal field will be determined
by the specific resistance genes present in the cereal cultivars grown in
each field. This process will be explained further in the section on selection. During the fall when
cereals are harvested in Canada and the northern USA, the prevailing wind
shifts to a southerly direction and some rust spores are able to move
south and infect volunteer cereal plants, thus reversing the direction of
migration. The cold winters in Canada and the northern USA ensure that no
spores survive the winter to begin an epidemic cycle in the following
year, so the local rust population goes extinct during the winter if no
alternate hosts are available. Cereal crops are recolonized by migrants
from the south during the summer of the following year. The wheat leaf rust pathogen Puccinia triticina (Puccinia
recondita f. sp.
tritici) reproduces only
asexually in North America, so the pathogen population is composed of a
series of clones and clonal lineages
that move north,
following the wheat crop each year. Imagine a series of farmer's fields distributed along the Puccinia
pathway. These fields are colonized by urediniospores that come from distant
fields (from Southern USA or Mexico) and from neighboring fields. In each
farmer's field, the local fungal population can go to extinction by: 1) harvesting the crop, After extinction has occurred, these fields can be recolonized when the
farmer plants a new wheat crop. The primary inoculum that initiates the
epidemic in each field can come from distant populations or from
neighboring fields. The dispersal distance and the amount of primary inoculum introduced
into uninfected fields play a large role in determining the neighborhood
size for the pathogen. The genetic neighborhood for a pathogen is the
geographical area over which populations exchange enough migrants to
evolve as a single unit. Go
to Knowledge Test for Gene/Genotype Flow
Table 2. Similarity in
RFLP allele frequencies
at three RFLP loci among Mycosphaerella graminicola populations from
Oregon, Israel, Denmark, United Kingdom, and Uruguay. N refers to
the number of isolates assayed in each population
Figure 10. The distribution of clonal
lineages of Fusarium oxysporum f. sp. cubense on
banana plantations around the world. Each symbol represents a
different clonal lineage. The most diverse concentration of
lineages corresponds with the presumed center of origin in
Southeast Asia (circled). Click
here to see an enlargement of this figure.
Figure 11. DNA fingerprints based on
hybridization with the probe RG57 were used to identify clones of Phytophthora
infestans.
Figure 12. The
series of migration events that likely led to the panglobal spread
of a single clone of Phytophthora infestans (Goodwin 1997).
Click here
to see an enlargement of this figure.
Nem: The Connection Between Genetic
Drift and Gene Flow
The Metapopulation Concept and Plant Pathogens
Figure 13. The Puccinia pathway in North America offers a
good example of a pathogen metapopulation. Rust spores move north
with prevailing winds during the spring and summer, and return to
the south when prevailing winds shift direction in the fall.
2) applying a fungicide,
3) rotating to a non-host crop,
4) planting a resistant cultivar, and/or
5) a cold winter.
by The American Phytopathological Society
American Phytopathological Society
3340 Pilot Knob Road
St. Paul, MN 55121
e-mail: