Trends Ecol Evol (Amst) 2015 Dall SR

8/9/2019 Trends Ecol Evol (Amst) 2015 Dall SR

http://slidepdf.com/reader/full/trends-ecol-evol-amst-2015-dall-sr 1/7

Genes as cues: phenotypic integrationof genetic and epigenetic information

from a Darwinian perspectiveSasha R.X. Dall1, John M. McNamara1,2, and Olof Leimar3

1Centre for Ecology and Conservation, Biosciences, College of Life and Environmental Sciences, University of Exeter,

Penryn Campus, Tremough, Penryn TR10 9EZ, UK2School

of

Mathematics,

University

of

Bristol,

University

Walk,

Bristol

BS8

1TW,

UK3Department of Zoology, Stockholm University, SE-106 91 Stockholm, Sweden

The

development

of

multicellular

organisms

involves

a

delicate interplay between genetic and environmental

influences. It is often useful to think of developmental

systems as integrating available sources of informationabout current conditions to produce organisms. Genes

and inherited physiology provide cues, as does the state

of the environment during development. The integration

systems

themselves

are

under

genetic

control

and

sub-

ject to Darwinian selection, so we expect them to evolve

to produce organisms that fit well with current ecologi-

cal

(including

social)

conditions.

We

argue

for

the

scien-

tific value of this explicitly informational perspective by

providing detailed examples of how it can elucidate

taxonomically

diverse

phenomena.

We

also

present

a

general framework for linking genetic and phenotypic

variation from an informational perspective. This appli-

cation

of

Darwinian

logic

at

the

organismal

level

canelucidate genetic influences on phenotypic variation in

novel and counterintuitive ways.

Development

and

cue

integration

The

textbook

depiction

of

genes

as

‘recipes’

that

determine

organismal phenotypes is increasingly difficult to sustain

in the

modern

biological

sciences

[1,2].

Phenotypic

plastic-

ity

–

the

ability

of

a

genotype

to

produce

distinct

pheno-

types as a result of nongenetic influences during

development

–

may

be

ubiquitous

in

multicellular

organ-

isms [1,3]. Such nongenetic influences (here referred to

collectively as epigenetic) can include environmental cues

and ‘parental effects’ (including vertical cultural inheri-

tance

and

DNA-methylation

patterns)

[1,3,4], which

often

reflect

parental

experience

of

the

local

selective

environ-

ment [5]. However, it is underappreciated that genes

themselves

can

also

act

as

cues

of

the

current

environment

[5–9]. If

there

is

restricted

gene

flow

among

local

environ-

ments with differential fitness effects of alleles in different

environments,

allele

frequencies

will

vary

spatially

[10]. Therefore,

an

individual’s

genotype

will

statistically

contain

information

about

(correlate

with)

local

environ-

mental conditions and can be regarded as a genetic cue. In

effect, genes at polymorphic loci provide information thatthe phenotypic states they trigger have had local success

[5–7,10]. Furthermore,

at

the

level

of

developmental

mech-

anisms, there is considerable overlap in how environmen-

tal

cues

and

genetic

variation

determine

phenotypes

[1,11].

Here

we

aim

to

provide

a

general

framework

for

think-

ing about how potential cues are integrated during devel-

opment

to

maximise

the

fit

of

the

phenotype

to

current

ecological

(including

social)

conditions.

Such

integration

will be favoured by Darwinian selection if the cue integra-

tion

system

is

itself

under

genetic

control

[7].

For

instance,

a polymorphic ‘genetic cue’ locus might influence the hor-

monal

regulation

of

phenotypes,

either

by

encoding

tran-

scription

factors

that

modulate

gene

expression

at

otherloci [12] and influence circulating hormone levels or by

influencing

hormone

receptors

[11,13]. Similarly,

develop-

ing phenotypes can be modulated by epigenetic parental

effects

(e.g.,

genomic

imprints,

maternal

effects)

and

the

state

of

the

local

environment.

By

influencing

transcription

in target tissues, hormone receptors can ‘integrate’ these

inputs

[14,15], thereby

guiding

phenotypic

development

[14]. In

summary,

evolutionary

change

of

developmental

systems

occurs

via

selection

on

modifier

loci

(e.g.,

that

influence

tissue-specific

hormone

receptor

densities),

fi-

ne-tuning responses to developmental cues. Modifiers

might

often

be

regulatory

genes.

Adaptive

cue-integration

systems

of

the

sort

outlinedabove will take any available sources of information (epi-

genetic

and

genetic)

into

account

in

determining

pheno-

types. Our framework predicts that the relative weights

put on

each

source

of

information

should

depend

on

envi-

ronmental

characteristics,

such

as

its

temporal

stability,

and on cue accuracies. If correlations among allele frequen-

cies

at

a

cue

locus

and

local

ecological

conditions

provide

greater predictive accuracy than environmental cues, ge-

netic

polymorphism

can

be

favoured,

along

with

genes

that

are

selected

to

modify

phenotypic

development

appropri-

ately for each genetic morph [7] (such ‘specific’ modifiers

are known;

e.g.,

from

studies

of

sex

determination

[16]).

Thus by taking such an explicitly informational approach

Opinion

0169-5347/

2015Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tree.2015.04.002

Corresponding author: Dall, S.R.X. ([email protected]).

Keywords: development; information; genetic cues; epigenetic effects; environmental

cues.

TREE-1935; No. of Pages 7

Trends in Ecology & Evolution xx (2015) 1–7 1

http://dx.doi.org/10.1016/j.tree.2015.04.002

mailto:[email protected]

mailto:[email protected]




we

can

predict

epistatic

effects

among

genes

determining

the cue

integration

(developmental)

system

and

other

genes

that

provide

cues.

This informational perspective offers a powerful way to

understand

how

genetic

and

epigenetic

influences

are

integrated

during

development.

Since

this

perspective

was introduced [6,7,10] it is increasingly being used to

interpret

a

wide

variety

of

biological

phenomena

[3,5,17–21]. Here we promote the scientific value of this approach

by

outlining

the

underlying

logic

and

detailing

how

it

can

elucidate

taxonomically

widespread

phenomena.

We

show

how rigorously applying Darwinian logic at an organismal

level

(i.e.,

assuming

selection

favours

developing

pheno-

types that process information to maximise their fit to their

local

environment)

can

help

to

explain

genetic

influences

on phenotypic

variation

in

novel

and

counterintuitive

ways.

Being alive can be informative

Central to the idea that genes can be used as developmen-

tal

cues

of

the

forthcoming

environment

is

the

existence

of

persistent genetic variation with alleles differing in fre-quency

in

different

environments

[7,22].

If

the

ability

to

cope with local environments is (to some degree) controlled

genetically

and

conditions

remain

stable

enough

across

generations

(e.g.,

individuals

are

likely

to

inherit

parental

environments and do not disperse too far away to repro-

duce),

the

fact

that

an

individual

exists

suggests

that

it

is

likely

to

be

in an

environment

to

which

it

is

well

adapted.

This is because most of the individuals of its genotype are

likely

to

be

born

into

such

‘good’

habitats

precisely

because

genotype

fitness

(the

rate

at

which

it

is

copied)

is

max-

imised where it is best matched to environmental condi-

tions.

Thus,

over

generations,

genotype

members

will

accumulate

disproportionately

in

environments

to

whichthey

are

well

suited

[10,21]. We

call

this

demographic

accumulation the ‘Multiplier Effect’ (ME) and have dem-

onstrated

how

it

can

explain

the

evolution

and

mainte-

nance of

ostensibly

paradoxical

unconditional

strategies

such as natal philopatry (attempting to breed in natal

habitats)

[10]. We

briefly

illustrate

the

main

insights

(modelling

details

can

be

found

in

[10]).

Case

study:

natal

philopatry

Consider an organism with discrete, non-overlapping gen-

erations with a 1-year generation time. The environment

contains many breeding sites, some of which are good (it is

well

suited

to),

others

poor.

Sites

change

quality

over

time

but

site

quality

(good

or

poor)

is

positively

correlated

between years. During development individuals receive

environmental

cues

of

current

site

quality.

Individuals

can either

attempt

to

breed

on

their

birth

site

or

choose

another breeding site at random. After breeding the indi-

vidual

dies

and

its

offspring

face

the

same

decision.

Consider

first

a

genotype

that

ignores

environmental

cues and always attempts to breed at its birth site (i.e., is

natally

philopatric).

Since

good

sites

produce

more

surviv-

ing offspring

than

poor

sites,

by being

natally

philopatric

individuals

of

the

genotype

increase

their

chances

of

breed-

ing on a good site. Over generations this effect amplifies

(the

ME

[10]). This

results

in

a

higher

proportion

of

individuals

breeding

on

good

sites

at

demographic

equilib-

rium

than

would

be

expected

by

chance

given

the

frequency

of

good

sites

in the

environment.

Therefore,

breeding

on

birth sites is the best thing for individuals to do unless

environmental

cues

provide

very

strong

evidence

that

sites

are

poor.

In

other

words,

natal

philopatry

maximises

fitness unless environmental cues are highly accurate

(Figure 1). In

informational

(Bayesian)

terms

[23,24],the probability that an offspring came from a good site

acts

as

the

Bayesian

prior

and

the

posterior

probability

is

the

probability

of a

good

site

given

this

prior

and

the

cue.

Since an individual that does not attempt to breed at its

natal

site

chooses

a

site

at

random,

natal

philopatry

is

optimal provided the posterior probability exceeds the

proportion

of

good

sites

in

the

environment.

As

Figure

1

shows,

the

likelihood

that

natally

philopa-

tric individuals came from good sites (the Bayesian prior)

depends

strongly

on

the

ratio

of

surviving

offspring

on

the

two

site

types.

This

ratio

might

depend

on

other

aspects

of

the genome. If variation in these other aspects were main-

tained,

it

would

be

adaptive

to

have

a

cue

integration

(e.g.,

developmental) system that based its ‘decision’ on whetherto disperse

on

both

the

cue

value

and

cues

provided

by

these other genetic elements. In other words, one might

expect

epistasis

between

the

genes

controlling

the

cue

integration

system

and

those

affecting

the

ratio

of

surviv-

ing offspring produced on different patches.

There

has

been

a

recent

effort

to

link

ideas

about

the

ME

[10]

to

a

branch

of

population

genetics

theory

that

has

its roots in Fisher’s thinking about mutation–selection

balance

[21]. According

to

the

Reduction

Principle

(RP),

0

0.2

0.4

0.6

0.8

1 1.5 2 2.5 3 3.5 4

P r o b a b i l i t y

p a t c h

i s g o o d

Gain rao (good/poor)

e = 0.2

e = 0.025

e = 0.05

e = 0.1

TRENDS in Ecology & Evolution

Figure 1. Effect of the ratio of the numbers of surviving offspring on good sites to

numbers on poor sites on natal philopatry and the ‘Multiplier Effect’. The broken

line shows the probability that a randomly chosen site is good. The curves show

the posterior probability that the natal site is good given that the developmental

cue indicates a bad site (i.e., is 0; see below). It is optimal to ignore this cue if the

posterior probability exceeds the probability that a randomly chosen site is good.

Good sites remaingoodfor 10 yearson averagebeforeturning poor and poor sites

persist for 40 years before becoming good; the proportion of good sites in the

environment is 0.2. During development each population member receives a cue

thatis either0 or1. Ifthe siteis poor, the cue is0 withprobability1 – e and is1 with

probability e (where 0 < e < 0.5). If the site is good, the cue value is 0 with

probability e and is 1 with probability 1 – e . If e is greater than thecriticalcue error,

probability fitness is maximised by ignoring cues and being natally philopatric.

Individuals that attempt natalphilopatryfail to do sowith probability0.04. Thereis

no density dependence acting, althoughMcNamara andDall [10] showthat similar

results can be obtained if it is included. Adapted from [10].

Opinion Trends in Ecology & Evolution xxx

xxxx,

Vol.

xxx,

No.

x


2



populations

near

an

equilibrium

between

selection

and

any

‘transformation

processes’

(e.g.,

mutation,

dispersal,

re-

combination)

will

evolve

less

transformation

because

at

such an equilibrium any change in phenotypic state (e.g.,

where

individuals

are,

what

alleles

they

express)

is

likely

to

be

a

change

away

from

matching

the

selective

environ-

ment (i.e., maladaptive) as most individuals will already be

in

conditions

to

which

they

are

suited

[21]. It

can

be

shownthat the conditions under which the ME operates are

equivalent

to

those

specified

by

the

RP

[21]; therefore,

the genetic

cue

approach

is

also

likely

to

provide

insight

into the evolution of a wide range of population genetic

phenomena

including

modifier

loci,

linkage

disequilibri-

um, recombination, evolvability, and dispersal or migra-

tion

(reviewed

in

[21]). Thus,

we

might

expect

lower

mutation

or

recombination

rates

to

evolve

when

fitness

landscapes are rugged and do not vary much between

generations

as

this

makes

it

likely

that

any

phenotypic

change

would

be

maladaptive

[10,21].

Applying the ‘genetic cue’ perspective

Phenomena involving cue integration are taxonomically widespread.

To

further

illustrate

the

value

of

the

genetic

cue perspective, we discuss two examples where predation

or

herbivory

intensity

varies

and

individuals

can

invest

in

costly

defences

based

on

information

from

various

sources.

For instance, defences in both crustaceans and plants can

be

induced

by

individual

experience

and

maternal

cues.

Here

we

show

that

genetic

cues

are

also

likely

to

play

a

role.

Many

species of Daphnia

develop distinct

defensive

morphologies

against predators,

usually in response

to

waterborne chemical cues of predator presence (e.g.,

predatory

phantom

midge larvae kairomones). For Daph-

nia pulex these

defences include

dorsal

‘neck

teeth’ andtheir induction has been

studied in considerable detail

[25–30]. Phantom midge larvae are gape-limited preda-

tors and it is mainly juvenile daphnids

(up

to

the

fourth

instar)

that are vulnerable

to

them.Correspondingly, the

neck teeth arepresent during instars one to four and tend

to

bemost

strongly

expressed

in the third instar

[27].

The

development

of

neck teeth

in response

to

kairomones

can

be viewed as cue integration. If mothers experience kair-

omones,

but not their offspring, offspring

express neck

teeth mainly in the second instar [27]. Furthermore,

offspring have a developmental window for induction of

neck teeth, during the late embryonic stage [29] while in

maternal brood

chambers. Without

exposure to

kairo-

mones

then,

neck teeth

are not expressed.

If

offspring

experience kairomones in time, the degree of neck tooth

expression

depends

on

the strength

and

duration

of

kai-

romone

exposure [27,29]. Thus,

from

a

developmental cue

integration perspective, there should be an early indica-

tion

that defence

is needed,

through maternal or

embry-

onic experience,

and,

given

this, the strength

of

defence

investment depends on continued evidence that it is

needed. We

can extend

this to

include

genetic cues

as

follows.

There

is

substantial

variation

in

how

readily

individua-

ls induce defences, even among clones collected from the

same

ponds

[26,28]. Given

that

daphnids

go

through

bouts

of asexual reproduction

interspersed

with

episodes

of

sex,

there is ample opportunity

for

locally adapted

ge-

netic variation

to

accumulate

at different times

and in

different microhabitats. Thus, prevailing genotypes are

likely to

cue (correlate with) local conditions (e.g., local

abundance of

predators)

in a

similar way as

maternal

cues [5]. For instance, D. pulex clones taken when phan-

tom midge larval densities were at their highest fre-quently induced defensive morphologies at very low

kairomone concentrations,

while clones

taken when

phantom

midge

larvae

were

rare and fish predators

predominated showed weaker responses (few individuals

induced

defences, and only at higher

kairomone concen-

trations) [26]. The informational interpretation is that

the former clones

had high

inherited (genetic)

expecta-

tions of

predation

risk

by

phantom midge

larvae

and

therefore did not need much corroboration from environ-

mental cues to

develop

defences,

while the latter clones

had very

low

phantom midge predation risk

expectations

and therefore required stronger environmental evidence

to

induce defences. To date there have been

no empirical

tests of this explanation for variation in the inducibility of defences

in

daphnids.

In

the field,

clones

could be

sampled across their spatiotemporal niches (different

places

and different

seasons)

to

ascertain whether they

show

reaction norms that

on

average are

favourable

for

the conditions encountered. Laboratory manipulations of

the requisite spatial

and temporal variation in predation

risk would also

be

feasible.

Plants often induce biochemical defences in response to

insect

herbivory

[31]. In

addition,

it

has

been

suggested

that

antiherbivore

defences

might

commonly

be

induced

transgenerationally [32,33]. One reason to expect adaptive

maternal

effects

to

evolve

in

plants

is

that

pollen

dispersal

often

occurs

over

much

greater

distances

than

seed

dis-persal,

possibly

inhibiting

local

genetic

adaptation

and

favouring epigenetic signalling from mother plants to

seeds

[9,34,35].

There

are

numerous

molecular

mecha-

nisms

of

the

induction

of

plant

defences

involving

the

jasmonate signalling pathway that are being actively ex-

plored.

These

include

direct

responses

to

plant

tissue

wounding

[31]

and

transgenerational

induction

involving

small-RNA signalling from mother plant to seeds [32,33] as

well

as

genetic

(ecotypic)

variation

in

defence

[36]. Box

1

illustrates how integration of these different effects can

evolve, if they are explicitly regarded as potentially infor-

mative cues during development. See [5,9] for other exam-

ples

of

similar

evolutionary

models.

The novel

interpretation

we

argue

for

in

both

cases

above – that an individual’s genotype can serve as a cue

of

the

risk

of

predation

or

herbivory

in

the

near

future

–

has

so

far

not

been

tested

empirically.

Such

tests

would

be

important, because they can increase our understanding of

the

circumstances

that

favour

genetic

versus

environmen-

tal

determination

of

phenotypes

and,

more

generally,

the

integration of information during development. In contrast

to

traditional

approaches

to

the

study

of

phenotypic

plas-

ticity

and

genetic

polymorphism,

our

approach

allows

these

phenomena

to

be

conceptualised

within

a

common

framework, thereby facilitating predictions of the use and

integration

of

cues.


xxxx,

Vol.

xxx,

No.

x


3



Box 1. Plant defence induction against herbivory

Figure I i llustrates a simulation model of the induction of plant

resistance against herbivory as motivated in the main text, which

follows the general scheme outlined by Holeski et al. [33]. Local

patches vary spatially and temporally in the expected intensity of

herbivore attack, which can be either low or high. The defence

phenotype expressed in the mature plant u protects the plant by

changing herbivore consumption by a factor 1 – u but has the cost of

changing the

reproductive output of the unconsumed parts of theplant by 1 – u /2. In the model, herbivore attack acts both as an

environmental cue that induces defence and as a selective factor

(Figure IA). A maternal cue can also beproducedby the plant, priming

the seeds for defence induction, and it is possible for herbivory

intensity-correlated genetic variation in u (‘genetic cues’) to accumu-

late under certain conditions (see below). Performing evolutionary

simulations under various conditions, we explore the

extent to

which u is influenced by genetic, maternal, and direct environmental

cues (see the supplementary material online for modelling details).

We show that defences can be induced as a response to a

combination of maternal and environmental cues (Figure IB), a

combination of genetic and environmental cues (Figure IC), or just

environmental cues (Figure ID). These outcomes depend on which

influences can convey information about the likely herbivory

intensity, given temporal and spatial variation in conditions andgeneflow, and on the availability of mechanisms of cue transmission.

‘Genetic cue use’ (genetically polymorphic antiherbivore defences) is

predicted when there is substantial genetic variation among patches

at evolutionary equilibrium (e.g., Figure IC: very high patch temporal

stability and very limited seed andpollendispersal) and therefore cue

locus allele frequencies better predict the risk of herbivory than noisy

maternal cues.

Pollen Ovule

Maternal cueSeed

Seed dispersal

Growing plant Herbivory

Jasmonate signalling

Mature plant Herbivory

(A)

–1.0

0.0

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Maternal cue

D e f e n s e

p h e n o t y

p e

(B)

–1.0

0.0

1.0

0.0

0.2

0.4

0.6

0.8

1.0

Genec cue

D e f e n s e

p h e n o t y p e

(C)

0.0 1.0

0.0

0.2

0.4

0.6

0.8

1.0

Herbivory cue

D e f e n s e

p h e n o t y p e

(D)


Figure I. Evolution of inducible defences against herbivory in diploid hermaphroditic plants. See the supplementary material online for further details. (A)

Sketchof the

life history and information flow to themature plant.Ovules arefertilisedby dispersingpollen and seeds receive genetic cues (alleles at a cuelocus) from each parentas

well as small RNA-mediated epigenetic cues from themother plant. After seed dispersal the growing plant might be fedupon,which influences the level of jasmonate

signalling and is a cue of the risk of herbivory for the mature plant. (B–D) Evolutionaryequilibria in defence levels. (B) High temporal stability of herbivory, very limited

seeddispersal, and freeamong-patchpollendispersal. Thegreycurve shows the average induceddefenceas a functionof thematernalepigenetic cue.The points show

induceddefencesof individual plants in low (blue) andhigh(red) expectedherbivory patches. Thedeviationof thepoints from thegreylineis a responseto theintensity

of herbivory on thegrowing plant. There is little genetic variation in defence. (C) Very high temporal stability and very limited seed and pollen dispersal, together with

very noisy transmission of thematernal cue, results in an equilibrium with substantial genetic variationat thecue locus,whichapproximately correspondsto the three

genotypes of an additive two-allele locus. Deviations from the grey line are responses to the intensity of herbivory on the growing plant. There is little maternal

influence on thelevel of defence.(D) High temporal instabilitywith lowand high herbivory intensity varying randomlyacross generations,in which case only thedirect

environmental cue (herbivory on the growing plant) influences induced defences.


xxxx,

Vol.

xxx,

No.

x


4



New perspectives on dominance, epistasis, and genetic

conflict

For genes

that encode developmental systems,

the geno-

type at a genetic cue locus can play a similar role as

environmental cues, in thesense

of providing

information

to

the organismabout selectively relevant circumstances.

Based on this view of genes as cues of the selective

environment,

we should expect systems that fine-tunedevelopment to genetic variation to evolve [7,22]. This

can occur

via selection

for

genetic modifiers that adjust

developmental

trajectories

[12,15]. For

instance, if the

different homozygotes at a two-allele locus are adapted to

markedly different

environmental conditions and

there is

substantial gene flow in the population, the phenotype of

the heterozygote

is likely to

evolve

to

match one of

the

homozygotes (e.g.,

the most commonly

adaptive one).

This will occur through the invasion of modifier loci that

change

the phenotype of

the heterozygote

[37].

Thus

dominance

can be viewed

as evolving via the fine-tuning

of the developmental influence of genetic cues of the

selective environment [22].

Another consequence of viewing developmental sys-tems as potentially

using genetic cues

is that genetic

conflict [38] is likely to play a significant role in the

evolution of

these

systems

[7].

To see why this is so,

consider

two

habitats

with

different phenotypes

being

optimal in each habitat and, for simplicity, haploid indi-

viduals that reproduce sexually. If

there is a

two-allele

locus with

alleles

producing

phenotypes that are suitable

for each habitat, depending on the intensities of selection

and gene flow allele frequencies

will differ

between habi-

tats.

Potential modifiers that

are fully linked

to

an allele

that is common in one habitat and rare in the other will

mainly experience

one of

the habitats, so

their evolution-

ary ‘interest’

will be to

produce

a

phenotype

that is

suc-cessful in that habitat.

For

unlinked

modifiers, however,

thesituation canbe radically different. Amodifier allele at

an unlinked

locusmight be present

at similar frequencies

in both

habitats

and its

evolutionary interest will be to

produce phenotypes that do well overall. Therefore, un-

linked

modifiers

that produce

less extreme

phenotypes

will be favoured.

It

is even

possible that

unlinked

modi-

fiers favour genetically monomorphic generalist pheno-

types whereas fully linked

modifiers favour genetic

polymorphisms. Such differences in the evolutionary

interests of modifier genes with contrasting linkage to a

genetic cue locus represent genetic conflict [38]; evolution-

ary change

favoured

by linked

modifiers (increased

phe-

notypic divergence

between morphs) selects

for unlinked

modifiers to reverse this change.Aswell asless phenotypic

differentiation, unlinked modifiers can

favour

develop-

ment that

is sensitive

to

environmental cues [7],

in effect

favouring phenotypic plasticity over genetic polymor-

phism. The evolutionary outcomes

of

such ecologically

driven

genetic conflicts will depend

on

the mechanisms

of regulation of development and gene expression. In

general,

the genetic cue

approach highlights the need

for

further work on

ecologically driven

genetic conflicts,

having the potential

to

predict

patterns

of

linkage

and

epistasis in phenotype determination from Darwinian

first principles.

A Darwinian framework for development:

ecoevolutionary

feedback

is

key

Since

the

informational

interpretation

of

genotypes

was

introduced in the literature [6,7], it has attracted the

interest

of

researchers

exploring

a

range

of phenomena,

including

phenotypic

polymorphisms

(polyphenisms

[39]),

transgenerational epigenetic influences on development

[9,40],

mechanisms

of

ageing

[17], sex-determination

sys-tems [20,41], early-life influences on health and disease

[18], genetic

variation

in

stress

sensitivity

[19], and

eco-

logical

responses

to

anthropogenic

environmental

change

[42]. To encourage wider application, we encapsulate its

basic

logic

in

a general

informational

(Bayesian)

frame-

work [23] for development in Figure 2. Such approaches

embody

fundamental

Darwinian

logic

since

fitness

is

max-

imised

by

phenotypes

that

maximise

their

fit

to

current

conditions by utilising any sources of information available

when

it

is

cost-effective

to

do

so

[43].

A

crucial

feature

of

an

ecologically

consistent

[44,45]

informational framework for development is that it must

incorporate

various

ecoevolutionary

feedbacks.

Two

are

illustrated in Figure 2. If

an

environmentvaries

spatially,

thedistribution

across

habitats of individuals can be regarded as a Bayesian

prior distribution

for

the developmental

system they

possess. Local cues

during

development

provide further

information on local conditions. However, the manner

in

which

such

cues

modify

development

affects

how the

survival

and

reproduction

of

individuals

depend

on

local conditions, determining the spatial distribution of

individuals and hence the prior

distribution

itself.

Developmental

system Phenotype

Distribuon of

individuals acrosslocal habitats

Correlaon between

cue genes and

environment

Local

habitat

Cue genes

Parental

states


Figure 2.

A Darwinian framework for integrating genetic and epigenetic

information during development. We envisage environments comprising local

habitats linked by dispersal. Conditions can vary among habitats and over time.

Thelocal habitat, parental states, and genes allprovide cues that areintegrated by

the developmental system to determine the mature phenotype. The

developmental system is under selection to match phenotypes to environments.

The distribution of individuals across habitat types is determined by the

demographic processes of survival, reproduction, and dispersal. It is thus

influenced by the developmental system itself, through the phenotypes that

develop in different conditions. At phenotypic determination (during

development), this distribution acts as a Bayesian prior distribution on habitat

type. Given a cueof local conditions during development, the posterior probability

that the local habitat is of a given type depends on the developmental system.

Thus, selection on the developmental system to change its responses to

environmental cues depends on its existing configuration.


xxxx,

Vol.

xxx,

No.

x


5



At evolutionary stability the developmental system will

integrate

prior

information

anddevelopmental

cues

in

an

optimal

way

given

the prior

distribution resulting from

its particular mechanism for phenotype determination.

The

correlation

between

cue

genes

and

the

local

selective environment imposes selection on the cue

genes

themselves

and

how

they

are

integrated

by

the

developmental

system.

The

correlation

depends

on

the

developmental process itself; it will typically be low

when developmental systems ignore genetic cues, sincecue

genes

are

then

under

weak

selection,

and

higher

when phenotypes

depend

more

on

such

cues.

Thus

the

information

content

of

genetic

cues,

and

hence

the

selection pressure on developmental systems to inte-

grate

these

cues,

depends

on

the

developmental

system

itself. Again, at evolutionary stability the developmen-

tal system should be optimal given the selection

pressures

it

generates.

Such

feedback

raises

the

possibility

that

there

might

be

more

than

one

evolutionarily

stable

developmental

system

in any given environment.

Concluding

remarks:

empirical

and

philosophical

prospects

Substantial empirical effort remains to fully understand

any developmental

process

underpinning

phenotypic

vari-

ation

from

an

adaptive

perspective.

Our

framework

(Figure 2) lends itself naturally to generating hypotheses

that

are

testable

using

the

‘modern

toolkit’

of

approaches

for

studying

adaptation

in

evolutionary

biology.

To

illus-

trate

how

it

could

be

applied

we

give

a

few

examples

of

promising areas of inquiry in Box 2.

A major

challenge

in

modern

biology

remains

to

fully

understand

how

genetic

variation

maps

to

phenotypic

variation. We offer our developmental cue integration

framework

(Figure 2) as

a

powerful

basis

for

theorising

about

the

links

between

genotype

and

phenotype

in

Dar-

winian systems. There are important philosophical impli-

cations

of

this

perspective

[8,46,47].

The

idea

that

both

genetic and environmental cues are ‘read’ by the mecha-

nisms

of

development

to

maximise

phenotypic

‘fit’

to

the

environment demonstrates that thinking of inheritance

systems such as the genome as representing (biased sub-

sets of) environments has considerable explanatory utility in biology [47]. We are confident that explicitly informa-

tional

approaches

to

understanding

biological

systems,

centred

on

phenotypes

as

evolved

information-processing

systems, can offer profound new insights.

Acknowledgements

The authors thank Angus Buckling, Brian Charlesworth, Tim Coulson,

Peter Hammerstein, Andrew Higginson, Dave Hosken, Dustin Ruben-

stein, Judy Stamps, Tom Tregenza, Tobias Uller, and Nina Wedell for

helpful feedback. This work was supported by a Leverhulme Trust

International Network Grant to S.R.X.D., Peter Hammerstein, O.L., and

J.M.M., as well as grants from the Swedish Research Council (621-2010-

5437)

and the Strategic Research Program

Ekoklim at Stockholm

University to O.L.

Appendix A. Supplementary dataSupplementary data associatedwith thisarticle canbe found,in theonline

version, at http://dx.doi.org/10.1016/j.tree.2015.04.002.

References1 West-Eberhard, M.J. (2003) Developmental Plasticity and Evolution,

Oxford University Press

2 Lynch, M. and Walsh, B. (1998) Genetics and Analysis of Quantitative

Traits, Sinauer Associates

3 Beldade, P. et al. (2011) Evolution and molecular mechanisms of

adaptive developmental plasticity. Mol. Ecol. 20, 1347–1363

4 Danchin, E. et al. (2011) Beyond DNA: integrating inclusive

inheritance into an extended theory of evolution. Nat. Rev. Genet.

12, 475–486

Box 2. Promising areas for future applications of the ‘genetic cue’ perspective

Bacteria often combine activities of genetic elements (e.g., plas-

mids) with social cues (e.g., quorum sensing) in their responses to

local environments [48]. Because these responses often involve

producing some sort of public good (e.g., iron-sequestering side-

rophores [49,50]), cues of the local social environment can be

crucial for controlling such investment [50,51]. It is likely that

frequencies of public good alleles will vary according to the

historical social structure of populations (particularly if they assortindependently of the bacterial genome; e.g., on plasmids); there-

fore, patterns of bacterial ‘cooperation’ might be best understood in

terms of adaptive integration of environmental and genetic cues.

Such logic will be readily testableusing experimental evolution. For

instance, bacterial cultures could be maintained that differ in the

degree to which social structure (e.g., clonal diversity within a

culture) varies across generations. Our ‘genetic cue’ perspective

would predict quorum-sensing strains to predominate in ‘socially

unstable’ cultures.

Males of many fish species express various mating tactics within

populations, from simple differences in mating behaviour to the

development of alternative morphs [52]. Such variation can be

underpinned by genetic differences and/or environmental (includ-

ing social) influences in different species [52–56], which have

traditionally been thought of as representing distinct evolutionary

outcomes (e.g., status/condition-dependent versus mixed evolutio-narily stable strategies

[52]).

By contrast, our cue integration

perspective facilitates explicit analysis of the ecological conditions

favouring different forms of cue use (or combinations thereof), as

well as likely evolutionary transitions among systems [56]. This

lends itself naturally to empirical comparative analysis. For

instance, species in the genus Xiphophorus (swordtails) show a

range of different alternative male morph-determining systems,

from genetic polymorphisms (e.g., X. nigrensis , X. maculates , X.

mulilineatus [54,55]) to social determination (e.g., X. variatus [53]).

Given the well-resolved phylogenies for this genus [57,58], we

would expect closely related species that determine male morphsdifferently to also differ in how stable their population densities

(and therefore operational sex ratios) are.

Understanding the evolution of cooperative breeding in vertebrates

has typically focussed on explaining patterns of dispersal within

and among species [59]. A key driver of delayed dispersal (and

therefore helping by non-breeders) has been suggested to be the

existence of substantial spatial variation in breeding habitat quality

that is likelyto persist acrossgenerations [59]. This is reminiscent of

theconditions underwhich an inherited state is likely to reliably cue

the best fit to local conditions (i.e., the ME operates [10]),

suggesting that the genetic cue perspective is likely to be useful

in elucidating the complex interplaybetween ecological andgenetic

factors underpinning vertebrate societies. Formal cue integration

models would identify key socioecological drivers of patterns of

helping in facultative cooperative breeders (e.g., stronger habitat

quality differences favouring delayed dispersal [10]). These wouldspecify a priori statistical models

for

existing long-term field

datasets, as well as suggest bespoke manipulations of the key

factors in established field-study systems.


xxxx,

Vol.

xxx,

No.

x


6


http://refhub.elsevier.com/S0169-5347(15)00092-0/sbref0300


































5 Leimar, O. and McNamara, J.M. (2015) The evolution of

transgenerational integration of information in heterogeneous

environments. Am. Nat. 185, E55–E69

6 Leimar, O. (2005) The evolution of phenotypic polymorphism:

randomized strategies versus evolutionary branching. Am. Nat. 165,

669–681

7 Leimar,

O. etal.(2006)

Anew

perspectiveondevelopmentalplasticityand

the principles of adaptive morph determination. Am. Nat. 167, 367–376

8 Rivoire, O. and Leibler, S. (2014) A model for the generation and

transmission of variations in evolution. Proc. Natl. Acad. Sci. U.S.A.111, E1940–E1949

9 English, S. et al. (2015) The information value of non-genetic

inheritance in plants and animals. PLoS ONE 10, e0116996

10 McNamara,J.M. andDall,S.R.X. (2011) Theevolutionof unconditional

strategies via the ‘‘multiplier effect’’. Ecol. Lett. 14, 237–243

11 Nijhout, H.F. (1999) Controlmechanismsof polyphenicdevelopment in

insects. Bioscience 49, 181–192

12 Lehner, B. (2011) Molecular mechanisms of epistasis within and

between genes. Trends Genet. 27, 323–331

13 Nijhout, H.F. (2003) Development and evolution of adaptive

polyphenisms. Evol. Dev. 5, 9–18

14 Flatt, T. et al. (2005) Hormonal pleiotropy and the juvenile hormone

regulation of Drosophila development and life history. Bioessays 27,

999–1010

15 Martin, L.B. et al. (2011) Integrator networks: illuminating the black

box linking genotype and phenotype. Integr. Comp. Biol. 51, 514–52716 Beukeboom, L. and Perrin, N. (2014) The Evolution of Sex

Determination, Oxford University Press

17 Flatt, T. et al. (2013) Life-history evolution and the polyphenic

regulation of somatic maintenance and survival. Q. Rev. Biol. 88,

185–218

18 Gluckman, P.D. et al. (2007) Early life events and their consequences

for later disease: a life history and evolutionary perspective. Am. J.

Hum. Biol. 19, 1–19

19 Del Giudice, M. et al. (2011) The Adaptive Calibration Model of stress

responsivity. Neurosci. Biobehav. Rev. 35, 1562–1592

20 Uller, T. and Helantera,

H. (2011) From the origin of sex-determining

factors to the evolution of sex-determining systems. Q. Rev. Biol. 86,

163–180

21 Altenberg, L. (2012) The evolution of dispersal in random

environments and the principle of partial control. Ecol. Monogr. 82,

297–333

22 Leimar, O. (2009) Environmental and genetic cues in the evolution of

phenotypic polymorphism. Evol. Ecol. 23, 125–135

23 Dall, S.R.X. et al. (2005) Information and its use by animals in

evolutionary ecology. Trends Ecol. Evol. 20, 187–193

24 Stamps, J.A. and Krishnan, V.V. (2014) Combining information from

ancestors andpersonal experiences to predict individual differences in

developmental trajectories. Am. Nat. 184, 647–657

25 Tollrain, R. (1995) Predator-inducedmorphological defenses: costs, life

history shifts, and maternal effects in Daphnia pulex. Ecology 76,

1691–1705

26 Hammill, E. et al. (2008)Costs, benefits and the evolution of inducible

defences: a case study with Daphnia pulex. J. Evol. Biol. 21, 705–715

27 Imai, M. et al. (2009) Elaborate regulations of the predator-induced

polyphenism in the water flea Daphnia pulex:

kairomone-sensitive

periods and life-history tradeoffs. J. Exp. Zool. A Ecol. Genet.

Physiol. 311, 788–795

28 Riessen, H.P. (2012) Costs of predator-inducedmorphological defences

in Daphnia.

Freshw. Biol. 57, 1422–1433

29 Naraki, Y. et al. (2013) Identification of theprecise kairomone-sensitive

period and histological characterization of necktooth formation in

predator-induced polyphenism in Daphnia pulex. Zoolog. Sci. 30,

619–625

30 Dennis, S.R. et al. (2014) Endocrine regulation of predator-induced

phenotypic plasticity. Oecologia 176, 625–635

31 Howe, G.A. andJander, G. (2008)Plant immunity to insect herbivores.

Annu. Rev. Plant Biol. 59, 41–66

32 Rasmann, S. et al. (2012) Herbivory in the previous generation

primes plants for

enhanced insect

resistance. Plant Physiol. 158,

854–863

33 Holeski, L.M. et al. (2012) Transgenerational defense induction and

epigenetic inheritance in plants. Trends Ecol. Evol. 27, 618–626

34 Galloway, L.F. (2005)Maternal effects provide phenotypic adaptation

to local environmental conditions. New Phytol. 166, 93–99

35 Galloway, L.F. and Etterson, J.R. (2007) Transgenerational plasticity

is adaptive in the wild. Science 318, 1134–1136

36 Kusnierczyk, A. et al. (2007) Transcriptional responses of Arabidopsisthaliana ecotypes with different glucosinolate profiles after attack by

polyphagous Myzus persicae and oligophagous Brevicoryne brassicae.

J. Exp. Bot. 58, 2537–2552

37 Otto, S.P.

and Bourguet, D. (1999) Balanced polymorphisms and the

evolution of dominance. Am. Nat. 153, 561–574

38 Hurst, L.D. et al. (1996) Genetic conflicts. Q. Rev. Biol. 71, 317

39 Fusco, G. andMinelli, A. (2010) Phenotypic plasticity in development

and evolution: facts and concepts. Introduction. Philos. Trans. R. Soc.

Lond. B: Biol. Sci. 365, 547–556

40 Uller, T. (2008) Developmental plasticity and the evolution of parental

effects. Trends Ecol. Evol. 23, 432–438

41 Cornwallis, C.K. andUller, T. (2010) Towards an evolutionary ecology

of sexual traits. Trends Ecol. Evol. 25, 145–152

42 Chown, S.L. et al. (2009) Population responses within a landscape

matrix: a macrophysiological approach to understanding climate

change impacts. Evol. Ecol. 24, 601–61643 McNamara, J.M. and Dall, S.R.X. (2010) Information is a fitness

enhancing resource. Oikos 119, 231–236

44 McNamara, J.M. (2013) Towards a richerevolutionarygame theory. J.

R. Soc. Interface 10, 20130544

45 Houston, A.I. and McNamara, J.M. (2006) John Maynard Smith and

the importance of consistency in evolutionary game theory. Biol.

Philos. 20, 933–950

46 Shea, N. (2010) What’s transmitted? Inherited information. Biol.

Philos. 26, 183–189

47 Shea,N. (2013) Inherited representationsare read in development. Br.

J. Philos. Sci. 64, 1–31

48 Lopez,

D. et al. (2010) Biofilms. Cold Spring Harb. Perspect. Biol. 2,

a000398

49 Hider,R.C. andKong,X. (2010)Chemistryand biology of siderophores.

Nat. Prod. Rep. 27, 637–657

50 Taylor, T.B. et al. (2013) The social evolution of dispersal with public

goods cooperation. J. Evol. Biol. 26, 2644–2653

51 Van Gestel, J. et al. (2012) The evolution of cell-to-cell communication

in a sporulating bacterium. PLoS Comput. Biol. 8, e1002818

52 Gross, M.R. (1996) Alternative reproductive strategies and tactics:

diversity within sexes. Trends Ecol. Evol. 11, 92

53 Borowsky, R. (1978) Social Inhibition of maturation in natural

populations of Xiphophorus variatus (Pisces: Poeciliidae). Science

201, 933–935

54 Lampert,K.P. et al. (2010)Determination of onsetof sexualmaturation

andmatingbehavior bymelanocortin receptor 4 polymorphisms.Curr.

Biol. 20, 1729–1734

55 Volff, J-N. et al. (2013)Geneamplificationand functional

diversification

ofmelanocortin4 receptorat anextremelypolymorphiclocuscontrolling

sexual maturation in the platyfish. Genetics 195,

1337–1352

56 Schwander, T. andLeimar, O. (2011)Genesas leaders and followers in

evolution. Trends Ecol. Evol. 26, 143–151

57 Kang, J.H. et al. (2013) Comprehensive phylogenetic analysis of all

species of swordtails andplaties (Pisces: Genus Xiphophorus) uncovers

a hybrid origin of a swordtail fish, Xiphophorus monticolus, and

demonstrates that the sexually selected sword originated in the

ancestral lineage of the genus, but was lost again secondarily. BMC

Evol. Biol. 13, 25

58 Cui, R. et al. (2013) Phylogenomics reveals extensive reticulate

evolution in Xiphophorus fishes. Evolution 67, 2166–2179

59 Koenig, W.D. et al. (1992) The evolution of delayed dispersal in

cooperative breeders. Q. Rev. Biol. 67, 111


xxxx,

Vol.

xxx,

No.

x


7

















































































































































































































































































































































































































































































































Documents

Trends Ecol Evol (Amst) 2015 Dall SR