Coevolution: Evolutionary Interactions Between Herbivorous Insects and Plants

1
Coevolution Evolutionary Interactions Between
Herbivorous Insects and Plants

• Peter B. McEvoy
• Ent 420/520 Insect Ecology

2
Questions (Futyma and Slatkin 1983)

• Adaptive Radiation. How often has the adaptive
  radiation of a group depended on radiation of
  other groups with which they interact?
• Speciation. Does speciation of hosts and parasite
  often occur in parallel?
• Defense. Do defense systems of prey become more
  complex over evolutionary time with addition of
  new defenses to the armory, or are old defenses
  traded in for new ones?
• Specialization, Virulence-resistance. Do
  parasites tend toward specialization or toward
  benign or even mutualistic relationship with
  their hosts?
• Nature and power of historical explanation. In
  general, how much of the history of evolution
  must be explained in terms of the evolutionary
  effects of interspecific interactions?

3
Evolution of Insect-plant Relationships

• Plant effects on insects. Do plants exert
  selection pressure on the insects?
• Insect effects on plants. Do insects exert
  selection pressure on the plants?
• Symmetry. Are the selective interactions between
  insects and plant reciprocal?
• Do related insects feed on related hosts?
• Is specialization an evolutionary dead end?

4
Alternative Theories on Herbivore-Plant Evolution

• Coevolution (Ehrlich and Raven 1964)
• Diffuse Coevolution, Community Coevolution (Fox
  1988)
• Geographic Mosaic Theory of Coevolution (Thompson
  1994, 2005)
• Sequential Evolution (Jermy 1976, 1984, 1991)

5
Escape and Radiation Coevolution Hypothesis
(Ehrlich and Raven)

• Many factors influence the evolution of
  herbivorous diets, but plant chemistry is central
• Secondary chemicals produced by chance and
  modified under selective pressure from herbivores
  
• Plant escape and radiation chemically-defended
  plants escape herbivores and radiate
• Insect escape and radiation counter-adaptations
  by insects allow insects to overcome host
  defenses, exploit previously defended hosts,
  escape competition from other insects, and
  radiate
• Parallel diversification in insects and plants
  Insects and plants augment one anothers
  diversity through coevolution, reciprocal
  evolutionary change in interacting species

6
Critique of Coevolution (Schoonhoven et al. 1998)

• Allelochemicals. Same secondary plant compound
  may have different functions in different insects
  toxic, deterrent, attractant negative,
  neutral, positive
• Plant defense and radiation. Little evidence that
  plants have evolved defenses under herbivore
  selective pressure or that such plants could
  diversify more effectively because they were
  defended
• Insect offense and radiation. Processes of
  speciation currently known do not support
  hypothesized radiation since interspecific
  competition generally weak
• Reciprocal speciation rarely found except in
  pollinators and plants, microbial symbionts and
  their hosts. Related insects often feed on
  unrelated hosts.

7
No single factor drives insects to specialize

• Escape from interspecific competition
• Reduced exposure to predators (enemy escape)
• Increased efficiency at detoxifying plant
  allelochemicals
• Genetically based trade-offs in offspring
  performance
• Increased efficiency in host finding

8
No clear link between chemistry and speciation

• Spatial barriers
• Behavioral barriers
• Asynchrony in life history
• Hybrid incompatibility

Speciation
Reproductive isolation
Chemistry
9
Asynchrony in Emergence of Three Apple Maggot
RacesChanges in preference and performance
conserved by temporal separation of host races
and assortative mating
10
Oviposition Preferences of Euphydryas editha
Changed as ancestral host decreased and novel
host increased
Novel Host increased
Ancestral Host decreased
11
Geographic Mosaic Theory (Thompson) recognizes
the importance of spatial heterogeneity and
movement Why retain the concept of
coevolution?

• Local variation in environment and population and
  community structure
• Local variation in outcome of interactions and
  specialization
• Balance between gene flow and selection
• Results in a shifting, geographic mosaic of
  interaction that need not result in a simple
  escalation of adaptations and counter-adaptations

12
Sequential or Asymmetrical Evolution Theory
(Jermy)

• Sequence of events. Evolution of insects follows
  the evolution of plants without significantly
  affecting plant evolution
• Interaction strength. Plants exert strong
  selective pressure on insects, whereas reciprocal
  negative effect of insects on plants is rare and
  weak
• Speciation in insects may be mediated by plants,
  but speciation in plants unrelated to insects
• Origin of plant traits. Switches in host occur,
  but no evidence that acquisition of new host has
  changed the traits of a plant

13
Tree of Plants
ucjeps.berkeley.edu/TreeofLife/hyperbolic.php
14
Chronology of Plant and Insect Evolution (Smart
and Hughes 1972)

• Potential resources
• Increased stature up-standing plants ?
  overtopping ? trees
• Appearance of Vasculature (Cordaites)
• Appearance of seeds (spores ?terminal sporangia
  ?terminal spike ?heterosporous condition ?seeds)
• Leaves microphylls ? megaphylls ? laminate
  leaves ? insect damaged leaves
• Flowers and fruits
• Species - diversification of Angiosperms toward
  end of early Cretaceous
• Modes of exploitation - Sucking probably earliest
  mode of feeding, followed by chewing, while
  mining and galling appear much later

15
More Plant Species ? More Insects Species
Higher Plant Species Diversity ? More Insects
Species
16
More Plant Architectural Diversity ? More Insect
Species
Number of phytophage species S increases with
area of host For a given host area, the number
of phytophage species increases with the
architectural diversity of the host plant (e.g.
we find progressively fewer phytophages on trees,
shrubs, herbs)
17
Diversities of Fossil Families Within Insect
Orders of Phanerozoic (Labandeira et al 1993)
Angiosperms appear 2/3 of the way up the band of
Mesozoic, radiate extensively in Cenozoic
(Tertiary) Possibly accelerated Hym, Lep, Dip,
Col No such effect on Orthop, Homop,
Heterop Insect diversification depends on
intrinsic trends rather than environmental
(ecological) conditions
18
Insect Familial Diversity From Triassic to Present
Overall, appearance and ascendancy of the
angiosperms associated with a slow-down rather
than an acceleration of insect familial
diversification Caveat family diversification
not necessarily identical with species
diversification
19
Summary From Fossil Evidence

• Timing of events. No precise coincidence in time
  between evolution of higher plants and insect
  taxa. Hymenoptera, Lepidoptera, Diptera,
  Coleoptera radiations may have accelerated with
  appearance and radiation of angiosperms
• Expanding resources. An increase in plant
  structural and architectural diversity, plant
  taxonomic diversity, opens up new possibilities
  for insect diversification.

20
Phylogenetic Evidence

• Host records - field and literature surveys
• Phylogenies (plant and insect) are constructed
  based on morphological, secondary chemical,
  molecular, and other characters
• Patterns - inferences about the predictability
  and stability of the host range
• Pattern 1 Taxonomically related insects feed on
  taxonomically related hosts
• Pattern 2 Parallel evolution of insects and
  plants (Coevolution)
• Pattern 3 Non-Parallel evolution of insects and
  plants

21
Phylogenetic Patterns 2Parallel evolution of
insects and plants (Coevolution)
22
Phylogenetic patterns 3 Non-Parallel evolution
of insects and plants
23
Phylogenetic Analysis Sample of 25 insect-plant
associations (Mitter and Farrell 1991, Farrell
et al. 1992, Farrell and Mitter 1993, Futuyma and
Mitter 1997)

• Taxonomic similarity. Shifts among plant
  families are relatively rare, but shifts within
  plant families are relatively common
• Ancient origin. Conservative plant-insect
  associations are probably very old, from about 70
  to 100 million years old
• Seldom evolve by strict coevolution. The
  exception, rather than the rule, is finding close
  concordance in insect and plant phylogenies
  matching different insect species with different
  plant species in a tight coevolutionary relation,
  but broad concordance is found higher in the
  taxonomic hierarchy
• Similar hosts in different regions. Similar host
  ranges are revealed by comparing related insects
  in different biogeographic regions, another
  indication that that diets are phylogenetically
  conservative
• Chemical similarity. Some host ranges are
  conservative with respect to phytochemistry

24
Challenges to Conventional Wisdom

• Phylogenetically conserved mechanisms. Host use
  is not necessarily phylogenetically
  conservative taxonomically related insects may
  feed on taxonomically unrelated plants
• Origins and Potentials of specialists.
  Specialization is not necessarily derived
  both primitive and derived taxa may specialize,
  and specialization need not be an evolutionary
  dead end

25
Plant coumarins, Swallowtails, and Other Insects
An Example of a Coevolved System (Berenbaum)

• Diet phylogenies of several herbivore clades
  using taxonomically disparate hosts may be
  interpretable as conserved associations with
  coumarin-bearing plants
• Escalation of defense
• Radiation of plants
• Radiation of insects

26
Parsnip Moth Depressaria pastinacella
Parsnip Moth Depressaria pastinacella
Heracleum sphondylium
27
Coumarins Vary From Simple to Complexin
Escalating Plant Defense
30 plant families
8 plant families
2 plant families Fabaceae (2) Apiaceae (11)
28
PLANT DIVERSITY More Extensive Radiation by
Plants With More Complex furanocoumarins
29
INSECT DIVERSITY More Extensive Radiation in
Insects Associated With Plants Having More
Complex Coumarins
30
Greater Degree of Specialization Among Insects
Feeding on Plants With More Complex Coumarins
31
Genetics

• Genetic basis for chemical variation among
  individual wild parsnip plants Pastinaca sativa
  attacked by the parsnip webworm Depressaria
  pastinacella (Oecophoridae)
• Nature of inheritance for plant resistance.
  Quantitative and polygenic resistance
• Negative correlations in resistance traits of the
  plant. Negative correlation between two traits
  (bergapten and sphondin) conveying resistance
  limits resistance in the plant population when
  herbivore present
• Costs of defense for plant. Negative correlation
  between several resistance factors and total seed
  production in greenhouse where herbivore is
  absent.
• Lingering Questions Are there similar genetic
  constraints on response of insect to selection by
  plant chemistry? Polygenic control of
  resistance-breaking ability? Negative
  correlations in traits conferring resistance
  breaking ability?

32
Overall Summary

• Adapative hurdles. Feeding on plants presents
  formidable hurdles to insects, but once hurdles
  are cleared, radiation may be dramatic
• Coevolution. Evolutionary interactions between
  insects and plants have been described as
  coevolution, but strict reciprocity has not been
  demonstrated for any complex of plants and
  herbivores. This has spawned alternative
  theories.
• Coevolution is more likely to occur when an
  insect has few host species and its hosts harbor
  few natural enemies.

33
Parsnip Webworm interaction
Pastinaca sativa or parsnip introduced to NA
almost 400 yrs ago Heracleum lanatum or cow
parsnip is an alternate host
Parsnip Pastinaca sativa
Heracleum sphondylium
Parsnip Moth Depressaria pastinacella
reassociated with parsnip in mid 19th
century Larvae attack seeds
CalPhotos
34
Geographic Mosaic Hypothesis(Thompson 1994,
1999, 2005)

• Outcomes of interspecific interactions can vary
  among populations due to structural differences
  in the communities in which interactions are
  embedded
• Where selection intensity is great, reciprocal
  responses are likely in so-called hotspots
• In contrast, where selection pressures are
  relaxed, reciprocal responses in coldspots are
  far less likely to occur

35
Geographic Variation in Outcomes has been
Reported for

• Plants and Pathogens
• Hosts and Parasites
• Hosts and Parasitoids
• Competitors
• Pollinators or floral parasites and plants
• Herbivores and plants
• Coevolutionary changes in chemical defense have
  rarely been examined in this context

36
Department of EntomologyUniversity of Illinois
May R. Berenbaum
Arthur R. Zangerl
37
(No Transcript)
38
A Model System

• Plate 1. Clockwise from left wild parsnip
  damaged by parsnip webworm in The Netherlands,
  parsnip webworm in the United States, and
  prepupal parsnip webworm parasitized by
  Copidosoma sosares in The Netherlands. Photo
  credits A. Zangerl Copidosoma Sosares (Walker)
  (Hymenoptera Encyrtidae)

39
Evidence of hot spots and cold spotsPhenotype
matching between plant furanocoumarin profile and
insect detoxification profile

• Credentials of the sample. Four populations of
  webworms and wild parsnips in the midwestern
  United States
• Operational definitions of resistance and
  virulence. Evaluated correspondence between
  resistance in the wild parsnip populations
  (capacity to produce furanocoumarins) with
  virulence (capacity to detoxify
  furanocoumarins) in each webworm population
• Resistance factors. Four furanocoumarins known
  to influence the interaction between webworms and
  parsnips - bergapten, isopimpinellin,
  xanthotoxin, and sphondinwere quantified in the
  seeds of wild parsnips and, as well, rates of
  metabolism of these four furanocoumarins were
  quantified in the associated insect populations

40
How does interaction strength vary with community
structure?

• Interaction with natural enemies and alternative
  host plants can reduce the intensity of
  coevolutionary interaction between webworms and
  parsnips
• Many of the models of coevolutionary dynamics are
  based on two-species interactions and relatively
  few empirical studies involve multispecies
  interactions
• Restated diagrammatically as follows

41
Expected Effects of Varying Community Structure
Parasitoid
Herbivore
Plant
Plant
Coevolutionary Temperature Expectation
--- Cold Spots ---
Hot Spot
H-P Well-Matched
H-P Mismatched
42
Hypotheses

• Outcomes of interaction between the wild parsnip
  and the parsnip webworm, compared at home (area
  of indigeneity) and abroad (area of introduction)
• If webworms act as selective agents on
  furanocoumarin content of hosts, they will likely
• Avoid or fail to utilize plants with
  furanocoumarin profiles that confer resistance
• Thrive on plants with furanocoumarin profiles
  that are insufficient to confer resistance.
• The selective impact of the parsnip webworm on
  the wild parsnip is likely reduced by
• alternate, chemically different host plants
• a specialist parasitoid natural enemy that
  inflicts significant mortality
• Leading to an absence of a predictive
  relationship between webworm presence and host
  plant chemical composition

43
Methods

• Pastinaca sativa and Heracleum spp. surveyed in
  Europe during only one year 2003 ignores
  temporal variation (within and between years) at
  each location i.e. sample long on space, short
  on time
• Counted parasitized and unparasitized webworms
  with what probability of detection?
• Strength of interactions interaction
  temperature classified as categorical variable,
  cold or hot, depending on whether webworms were
  prevalent or rare a cursory measure of
  interaction strength
• Webworms collected to test detoxification
  capacity
• Five furanocoumarins measured in seeds of P.
  sativa collected in USA and EU in 2004 (damage
  also recorded). Compared furanocoumarin seed
  concentrations as a function of continent and
  interaction temperature.
• In Europe 4 of 14 (30)populations were cold and
  in the United States, 3 of 9 (33) populations
  were cold so frequency of hot and cold sites
  similar between continents

44
Results to be examined

• Surveys of webworms and parasitoids in European
  populations of P. sativa and H. sphondylium
• Furanocoumarin detoxification capacity of
  European and midwestern U.S. populations of
  parsnip webworms
• Differences in seed furanocoumarin content
  between European host plants
• Evidence for selection by webworms on
  furanocoumarin chemistry of European host plants
• Differences between hot and cold regions within
  continents and not between continents in seed
  furanocoumarin content

45
Surveys of webworms in EU
hot

• Parsnip webworm infestation levels in European
  populations generally higher for Heracleum
  sphondylium (black circles) than Pastinaca sativa
  (gray circles) in 2003

cold
Surveys of parasitoids less conclusive
46
Detoxification ability of webworms lower at home
(Neth) than abroad (Midw)
indicate magnitudes of differences

• Fig. 4. Least-square means and standard errors of
  furanocoumarin detoxification rates for
  midwestern U.S. and Netherlands populations of
  parsnip webworms. Percentages above bars indicate
  magnitudes of significant differences between
  continents based on a mixed/nested ANOVA
  (continent was a fixed effect, and population was
  a random effect nested within continent). For
  imperatorin, bergapten, isopimpinellin,
  xanthotoxin, and sphondin, P values for the main
  effect of continent were 0.004, 0.031, 0.049,
  0.021, and 0.033, respectively

47
Chemical profiles differ on two hosts in Europe
indicate magnitudes of differences

• Fig. 5. Least-square means and standard errors of
  seed furanocoumarin content of Heracleum
  sphondylium and Pastinaca sativa in Europe.
  Percentages above bars indicate magnitudes of
  significant differences between species based on
  a mixed/nested ANOVA (species was a fixed effect,
  and population was a random effect nested within
  species). For imperatorin, bergapten,
  isopimpinellin, and xanthotoxin, P values for the
  main effect of species were lt0.001, 0.012, 0.008,
  0.021, and lt0.0013, respectively. Tests were not
  conducted for sphondin, which was not detected
  (n.d.) in H. sphondylium, or for angelicin, which
  was not detected in P. sativa

48
Evidence of Selection by Webworms on
furanocoumarin chemistry in EU hosts is spotty

• Three under selection 3 of 11 populations show
  evidence of selection
• Two hot, one cold. 2 of these 3 had high webworm
  infestations (54 and 56 of plants attacked),
  the other low (8.9)
• Match within one. In 1 of the 2, plants free of
  webworms had higher xanthotoxin
• Match within other. In the other, plants without
  webworm damage had higher imperatorin,
  isopimpinellin, and sphondin content

49
Difference in furanocouarin content between hot
and cold regions within continents, but not
between continents
indicate magnitudes of differences

• Fig. 6. Least-square means and standard errors of
  seed furanocoumarin concentration for wild
  parsnips as a function of prevalence of webworms
  (interaction temperature webworms are rare in
  cold regions and common in hot regions).
  Percentages above bars indicate magnitudes of
  significant differences between hot and cold
  regions based on a mixed/nested ANOVA (continent
  and interaction temperature were fixed effects,
  and population was a random effect nested within
  the main effects)

Continent and Hot/Cold interact in case of
sphondin
50
Conclusions

• Home and Away Populations Similar. Patterns are
  more reflective of interaction temperature than
  of continental origin
• Introduced System is Simplified. The ubiquitous
  two-species interaction in North America is in
  fact exceptional in Europe webworms could more
  reliably be found infesting H. sphondylium even
  where P. sativa was available as well.
• Interactions among three trophic levels. A
  preference for H. sphondylium exists despite the
  comparatively high probability of parasitism
  associated with this host plant and may reflect
  the overall lower furanocoumarin content of H.
  sphondylium

51
Assumptions

• Phenotypic variation in plants and insects has a
  genetic basis good evidence
• Webworms influence parsnip fitness and population
  dynamics evidence?
• Parasitoids influence webworm population dynamics
  - evidence?

52
A Novel Use of Herbarium RecordsIncrease in
toxicity of an invasive plant following
reassociation with its coevolved herbivorePNAS
2005 102 (43)
53
Weevil-Camellia Geographic Mosaic

• Figure 1. A camellia weevil (Curculio camelliae)
  boring into a Japanese camellia (Camellia
  japonica) fruit, and a cross-section of a
  camellia fruit showing two tunnels excavated by
  the rostrum of adult females. (Photographs
  courtesy of Hirokazu Toju.)

54
Boring success varies in relation to host
pericarp thickness and weevil rostrum length
A Case study involving physical rather than
chemical resistance Toju, H., and T. Sota. 2005.
Imbalance of predator and prey armament
geographic clines in phenotypic interface and
natural selection. American Naturalist 167.
55
Implications for Biological Control

• Introduced species often escape their coevolved
  competitors and natural enemies
• Simplification of biological control systems may
  lead to stronger, pairwise interactions between
  herbivore and host plant
• Simplification may increase the frequency of
  coevolutionary host spots within the geographical
  mosaic
• Parsnip-Webworm interaction in NA is older than
  the oldest biocontrol interaction in NAwith time
  we may expect to see coevolution in control
  organisms and their hosts, with changes in
  virulence and resistance