(Open Access) Hybridization may facilitate in situ survival of endemic species through periods of climate change (2013) | Matthias Becker

Q: What are the contributions in "Hybridization may facilitate in situ survival of endemic species through periods of climate change" ?

Here the authors report statistical evidence that the transfer of genetic information through hybridization is a feature of species from the plant genusPachycladon that survived the Last Glacial Maximum in geographically separated alpine refugia in New Zealand ’ s South Island. The authors show that transferred glucosinolate hydrolysis genes also exhibit evidence of intralocus recombination. The authors use a mathematical model to show that when hybridization increases the adaptive potential of species, future biodiversitywill be best protected by preserving closely related species that hybridize rather than by conserving distantly related species that are genetically isolated. Species affected by climate change in the past—the consequences of which are manifested in extant species ’ ranges and patterns of genetic diversity—provide models to test for signatures of hybridization. Here the authors studied Pachycladon ( Brassicaceae ), an allopolyploid genus of 11 species9 that have radiated in the New Zealand Alps during the Pleistocene period9,10. Such gene exchange and recombination has the potential to alter the chemical defence in the offspring of hybridizing species. Although ecological modelling has an important place in understanding this impact2, an accurate prediction of range shift and extinction of species also requires determining their adaptive potential, and in particular the frequency with which hybridization facilitates adaptation4.

LETTERS

PUBLISHED ONLINE: 3 NOVEMBER 2013 | DOI: 10.1038/NCLIMATE2027

Hybridization may facilitate in situ survival of

endemic species through periods of climate change

Matthias Becker

, Nicole Gruenheit

1,2

, Mike Steel

, Claudia Voelckel

, Oliver Deusch

Peter B. Heenan

, Patricia A. McLenachan

, Olga Kardailsky

, Jessica W. Leigh

and Peter J. Lockhart

Predicting survival and extinction scenarios for climate change

requires an understanding of the present day ecological

characteristics of species and future available habitats, but also

the adaptive potential of species to cope with environmental

change. Hybridization is one mechanism that could facilitate

this. Here we report statistical evidence that the transfer

of genetic information through hybridization is a feature of

species from the plant genus Pachycladon that survived the Last

Glacial Maximum in geographically separated alpine refugia

in New Zealand’s South Island. We show that transferred

glucosinolate hydrolysis genes also exhibit evidence of intra-

locus recombination. Such gene exchange and recombination

has the potential to alter the chemical defence in the offspring

of hybridizing species. We use a mathematical model to show

that when hybridization increases the adaptive potential of

species, future biodiversity will be best protected by preserving

closely related species that hybridize rather than by conserving

distantly related species that are genetically isolated.

Predicting the response of organisms and estimating loss of

genetic diversity are important challenges for evaluating the impact

of global climate change on biodiversity

1–3

. Although ecological

modelling has an important place in understanding this impact

an accurate prediction of range shift and extinction of species also

requires determining their adaptive potential, and in particular

the frequency with which hybridization facilitates adaptation

Determining this is important, because although hybridization

can be a maladaptive phenomenon

, it might also help species

to acquire adaptive traits, respond successfully to environmental

change and invade new habitats

1,3,4,6–8

. A better understanding of

its positive and negative contributions is essential for evaluating

biodiversity impacts.

Species affected by climate change in the past—the consequences

of which are manifested in extant species’ ranges and patterns

of genetic diversity—provide models to test for signatures of

hybridization. Here we studied Pachycladon (Brassicaceae), an

allopolyploid genus of 11 species

that have radiated in the New

Zealand Alps during the Pleistocene period

9,10

. Figure 1 shows ice

cover at the height of the Last Glacial Maximum (LGM; 21,000–

18,000 years ago), the present day distribution of three Pachycladon

species, and a chloroplast TCS (statistical parsimony) haplotype

network indicating relationships among accessions of the three

species. At present, all three species are restricted to greywacke rock

in the central and northern regions of the South Island of New

Institute of Fundamental Sciences, Massey University, Palmerston North 4442, New Zealand,

Faculty of Life Sciences, University of Manchester, Oxford

Road, Manchester M13 9PP, UK,

Allan Wilson Centre, University of Canterbury, Christchurch 8140, New Zealand,

Landcare Research, Lincoln 7640,

New Zealand,

Department of Anatomy, University of Otago, Dunedin 9054, New Zealand,

Department of Mathematics and Statistics, University of

Otago, Dunedin 9054, New Zealand. *e-mail: m.becker@massey.ac.nz

Zealand

. On the basis of ecological characteristics and prediction

of available habitats, it has been proposed that there was nunatak

survival in the central South Island at the height of the LGM for

Pachycladon enysii, and survival in more northern and peripheral

South Island refugia for P. fastigiatum and P. stellatum

However, hybridization and the introgression of adaptively

significant traits between species could challenge our understanding

of this scenario and other similar survival scenarios for biodiversity

that are based on present day species’ attributes and present

habitats. To investigate this with Pachycladon, we tested for evidence

of past introgression with chloroplast markers from five loci

(Supplementary Fig. 1) in accessions sampled across the geographic

ranges of P. enysii, P. fastigiatum and P. stellatum using the JML

(Joly, McLenachan, Lockhart) test of ref. 11 and a species tree

reconstructed from seven nuclear gene loci (Supplementary Fig. 2).

We found that incomplete lineage sorting was unable to

explain identical and highly similar chloroplast haplotypes shared

by P. enysii, P. fastigiatum and P. stellatum accessions (Fig. 1

and Supplementary Fig. 3). At probability value p < 0.1 there

was evidence for introgression of chloroplast genomes between

adjacent central populations of P. enysii and P. fastigiatum

and between adjacent northern populations of P. enysii and

P. stellatum (Supplementary Fig. 3 and Table 4). This suggests

that there has been regional chloroplast capture through sympatric

events of hybridization.

Relaxed molecular clock estimates (Supplementary Fig. 4) in-

dicate that the chloroplast haplotypes of the central popula-

tions of P. enysii and P. fastigiatum diverged from the haplo-

types of northern populations at least 190,000 years ago (95%

highest posterior density lower bound: 0.19 million years ago

(Ma)—95% highest posterior density upper bound: 1 Ma). In

contrast, analyses of nuclear molecular markers identify iden-

tical or highly similar alleles common to northern and cen-

tral populations of Pachycladon species. Thus, although re-

stricted seed dispersal and in situ survival of central pop-

ulations of Pachycladon through successive glacial–interglacial

cycles over the Pleistocene seems likely, central and north-

ern populations also seem to have been connected by pollen-

mediated gene flow.

In general, species boundaries seem to be maintained between

P. enysii, P. fastigiatum, and P. stellatum. This is suggested from

Bayesian delimitation analysis for seven nuclear genes (S. Joly,

P. B. Heenan and P. J. Lockhart, manuscript in preparation).

NATURE CLIMATE CHANGE | VOL 3 | DECEMBER 2013 | www.nature.com/natureclimatechange 1039

LETTERS

NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2027

41°26'54.070"S¬

43°29'27.163"S

43°30'13.419"S¬

44°33'45.497"S

Pf8

Pf9

Pf15

Pf14

Pf3

Pf21

Pf1

Pf16

Pf17

Pf22

Pf18

Pf13, Pf19

Ps2

Pf4

Pf7

Pf20

Pf10

Pf5, Pf6

Pf11, Pf12

Pe18

Pe8

Pe7

Pe12

Pe11

Pe17

Pe9

Pe5

Pe3

Pe2

Pe6

Pe15

Pe10

Pe13

Pe4

Pe1

Pe14, Pe16

100 km

P. enysii

P. fastigiatum

Ps1

Pf22.1

Pf18.3

Pe5.1

Pe12.1

Pe1.1

Ps1.4

Ps1.3

Ps1.2

Ps1.1

Pf1

Pe2.1

Ps2.2

Ps2.3

Pf21.1

Pe10.1

Pe10.2

Pe10.3

Pe10.4

Pe13.1

Pe3.1

Pe6.1

Pe15.1

Pe15.2

Pf15.1

Pf13.1

Pf16.1

Pf16.2

Pf16.3

Pf17.1

Pf18.1

Pf18.2

Pf19.1

Pf3.1

Pf3.2

Pf14.1

Pf14.2

Pf8.1

Pf9.2

Pf9.3

Pf9.1

Pe5.2

Pf4.1

Pf5.1

Pf5.2

Pf5.3

Pf5.4

Pf6.1

Pf6.2

Pf6.3

Pf7.1

Pf10.1

Pf10.2

Pf10.3

Pf11.1

Pf20.1

Pe7.1

Pe7.2

Pe7.3

Pe8.1

Pe9.1

Pe9.2

Pe9.3

Pe9.4

Pe11.1

Pe14.1

Pe18.1

Figure 1 | Phylogeographic distribution of P. enysii (Pe), P. fastigiatum (Pf) and P. stellatum (Ps) populations. a, Ice cover (dark grey) and peripheral

refugia (black) at the height of the LGM. Coordinates for range limits of haplotype clusters are given. Haplotypes are colour coded; multi-coloured circles

identify populations (for example, Pf18) with multiple genotypes (for example, Pf 18.1–4). b, Sampling performed across distribution areas of P. enysii and

P. fastigiatum: ﬁlled shapes represent sampled populations and open shapes represent unsampled populations (P. stellatum not shown as it occurs only in

two adjacent locations, Fig. 1a and Supplementary Table 1). c, TCS network of relationships among P. enysii, P. fastigiatum, and P. stellatum

chloroplast haplotypes.

However, despite this finding there is also evidence for introgression

of nuclear genes of potential adaptive significance between these

species (Fig. 2). These are the epithiospecifier modifier (ESM1;

ref. 12), epithiospecifier protein (ESP; ref. 13), and modified

vacuole phenotype1 (MVP1; ref. 14) genes whose differential

expression is associated with species-specific glucosinolate hydrol-

ysis phenotypes in Pachycladon

and which, in Arabidopsis thaliana,

have been demonstrated to differentially affect herbivore fitness

The expressed alleles of ESP and ESM1 showed strong signatures

of positive and purifying selection in Pachycladon (Fig. 3). JML

1040 NATURE CLIMATE CHANGE | VOL 3 | DECEMBER 2013 | www.nature.com/natureclimatechange

NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2027

LETTERS

Most common Pf allele

Most common Pe allele

Pe7.1, Pe8.1

Pe10.1

Pe4.1

Pf8.1

Ps1.1

Ps2.1

Pf allele and minor variants

Upregulated allele

Gene expression ratio of Pf:Pe

Recombination break point

Stop codon

Stop

# In Arabidopsis orthologue

Pe allele and minor variants

Ps allele and minor variants

Pe4.1-specific allele

Ps2.1-specific allele

Pe1.1, Pf9.1

Allelic

variation

Stop

Stop Stop

Stop

Stop Stop

Stop

Stop Stop

Stop

MVP1

homeologue 1

MVP1

homeologue 2

MVP1

paralogue

ESM1

ESP

homeologue 1

ESP

homeologue 2

1:127Δ 22:1

24:1

Chromosome #1 Chromosome #1 Chromosome #3

Figure 2 | Recombination and allelic variants of glucosinolate hydrolysis genes in Pachycladon. Alleles shared between P. enysii, P. fastigiatum and

P. stellatum were found for six loci implicated in glucosinolate hydrolysis: two homeologues and one paralogue of MVP1, two homeologues of ESP and a

single ESM1 gene. Allele sharing due to recent introgression was conﬁrmed for MVP1 homeologue 2, ESP homeologue 2 and ESM1. Recombination within

ESM1 of P. enysii and P. fastigiatum led to introgression of the 5

exon region of the P. fastigiatum ESM1 allele and the neighbouring MVP1 paralogue allele into

accessions of P. enysii (for a detailed discussion of these results refer to Supplementary Information 1).

(ref. 11) identified alleles of glucosinolate hydrolysis genes—MVP1

homeologue 2, ESP homeologue 2 and ESM1—to be introgressed

between northern populations of P. enysii, P. fastigiatum and P.

stellatum (Supplementary Table 5). Allele sharing between species

is also observed with the other glucosinolate hydrolysis loci under

investigation (MVP1 homeologue 1, ESP homeologue 1 and MVP1

paralogue; Fig. 2). However, this observation is more difficult

to assess because of low sequence variation. In the instances of

statistically inferred introgression, allele sharing has altered the

functional allelic diversity of populations in all three species of

Pachycladon (Supplementary Tables 6 and 7). Furthermore, a

likelihood ratio test indicates that intra-genic recombination has

occurred within the ESM1 locus in P. enysii (Supplementary Fig. 5;

p = 4.7 × 10

−11

). This hybridization-mediated recombination has

replaced a neighbouring MVP1 paralogue allele, characterized by

a frame shift and premature stop codon, with an allele that has

an open reading frame and is expressed

in P. fastigiatum. This

modification alters the functional allelic diversity of the MVP1

paralogue locus in P. enysii (Fig. 2).

These above instances of introgression and recombination,

are likely to impact on glucosinolate hydrolysis chemotypes

(Supplementary Table 6), herbivore fitness and the adaptive

potential of Pachycladon species. In contrast to the chloroplast

results, introgression of glucosinolate hydrolysis genes is not

restricted to geographically adjacent/sympatric populations: The

same recombination within ESM1 was found in northern (Pe10)

and central populations of P. enysii (Pe7, Pe8). Thus, it seems

that pollen-mediated nuclear gene flow has dispersed glucosinolate

hydrolysis genotypes more widely and facilitated genetic exchange

between disjunct populations.

Hybridization is increasingly speculated as an important mecha-

nism for rapid adaptation to environmental change

1,3,4,7

. However,

presumably because of uncertainty in its possible outcomes

hybridization is yet to be incorporated into models that predict

the survival and extinction scenarios of biota. Previously, it was

suggested

that P. enysii survived Pleistocene glaciation in central

South Island nunataks, whereas P. fastigiatum was exterminated

from such areas. However, here we show that introgression between

these species is a feature of populations that have persisted in the

central part of the South Island of New Zealand over periods of

Pleistocene climate change when Pachycladon species ranges would

have expanded and contracted with available habitat and changes

in biotic interactions. Furthermore, our results suggest that intro-

gression and recombination are mechanisms that have the potential

in Pachycladon to rapidly generate adaptive chemotypes important

in herbivory and pathogen defence

15,16

. Thus, if introgression and

recombination have similarly acted on other adaptive gene sets,

they are likely to have aided survival of Pachycladon species in the

past and will impact on their adaptive potential in the future. This

hypothesis is important to test given recent observations suggesting

the adaptive significance of hybridization in animal

8,18,19

and plant

species

, as well as interest in determining what processes are most

important in shaping ecological interactions

The suggestion that hybridization and introgression might act

as a general mechanism for increasing the adaptive potential of

species has implications for predicting and maximizing future

NATURE CLIMATE CHANGE | VOL 3 | DECEMBER 2013 | www.nature.com/natureclimatechange 1041

LETTERS

NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2027

0.0

0.5

1.0

1.5

2.0

2.5

ESP homeologue 2

0.0

50 140 230 326

GDSL-like lipase/acylhydrolase

416 506 596 686 782 881 971 1,061

0.5

1.0

1.5

2.0

2.5

3.0

ESM1

83 173 263 353 443 533 623 713 803 893

Kelch 1 Kelch 1 Kelch 6 Kelch 1 Kelch 4

Figure 3 | Selection signatures in open reading frames for Pachycladon

ESP (homeologue 2) and ESM1. a, Comparison of ESP homeologue 2 from

P. enysii and P. stellatum. b, Comparison of ESM1 from P. enysii and

P. fastigiatum. Domains of ESP homeologue 2 (Kelch 1, 4, and 6,

respectively) and ESM1 (GDSL-like lipase/acylhydrolase) conﬁdently

predicted by Pfam (http://pfam.sanger.ac.uk/) contain signatures for

positive and purifying selection (K

ratios 6= 1, where K

is the ratio

of the number of non-synonymous substitutions per non-synonymous site

) to the number of synonymous substitutions per synonymous site

)). These observations are consistent with a hypothesis that

glucosinolate hydrolysis genes comprise adaptive gene sets for plant

defence in the Brassicaceae

phylogenetic diversity through conservation efforts

. Although

hybrids and hybridization have a long and uncomfortable history

in conservation biology

22,23

, the preservation and protection

of hybrids and/or individuals with the potential to hybridize

in some cases could lead to greater preservation of future

phylogenetic and biological diversity. We illustrate this conjecture

with a simple example.

Suppose there are two groups of closely related species (A, B)

and (C, D) that are phylogenetically distant from each other (Fig. 4).

The species within each pair occupy overlapping geographic areas.

Assume that the probability p of any of these species surviving

environmental change in situ is close to zero, but that with

conservation intervention we are able to increase the probability

of survival of any one species. We will assume that our resources

(budget) allow only two species to be conserved. The Noah’s Ark

problem

in this simple setting is to decide which two species

should be preserved, where the aim is to increase expected future

phylogenetic diversity.

Ignoring for the moment the influence of hybridization, it is

well known that the pair of species having maximal phylogenetic

diversity (PD) would have maximal separation in the tree (for

example species A and C in Fig. 4, Scenario 1). In this case the

expected future PD is precisely 2pT, when the time from the

ancestor to the present is T and the length of the four short pendant

edges is ignored. In contrast, conserving A and B (Scenario 2) leads

to a future expected PD value of (2p − p

)T , where p

is the

probability that both A and B survive (p

would be p

if these events

are independent, but as A and B are closely related it will probably

be closer to p). Scenario 1 therefore always leads to higher PD than

Scenario 2 in this simple setting.

∗

Time Time

Scenario 2Scenario 1

∗

Figure 4 | Scenarios for conserving phylogenetic diversity. Scenario 1,

conservation of two distantly related species (indicated by ∗). Scenario 2,

conservation of two closely related species (indicated by ∗).

However, the situation can change if hybridization occurs within

the (A, B) clade. For this case we assume that: because the species

are highly specialized for their respective habitats, there is limited or

no adaptive potential in each individual species, and thus without

conservation of their habitat their probability for survival is zero;

under conditions of environmental change hybrids will survive and

reproduce, in which case there is a further probability that some

hybrids will have increased adaptive potential, which will allow

them to occupy altered as well as colonize new habitats.

Formally, if A (and not B) is subject to conservation efforts,

let x

be the probability that a hybrid from (A, B) will become

established and survive owing to an adaptive potential greater than

that of its parents. Similarly, let x

be the corresponding probability

if both A and B receive conservation efforts (for example, by

protecting the habitats they both occur in). For Scenario 1, the

expected future PD is at most (2p + x

)T , whereas for Scenario 2

it is at least x

T (Methods). Thus, it is now entirely possible for

Scenario 2 to have higher expected future PD than Scenario 1, as

this will hold whenever

> 2p +x

(1)

Inequality (1) will be satisfied if p and x

are sufficiently small

compared with x

If, in addition, hybrids can also arise within the (C, D) clade, then

the corresponding condition sufficient for Scenario 2 to have higher

expected PD than Scenario 1 is

+ x

> 2p +x

+ x

(2)

where x

is the probability that a hybrid from (C, D) will become

established if neither C nor D is conserved, and where x

is the

corresponding probability if C (but not D) is conserved (Methods).

Although the example here is simple, in principle the calculation

can be applied to any phylogenetic tree with arbitrary branch

lengths, extinction probabilities, and groups of taxa capable

of forming hybrids. This result complements earlier studies

21,25

that have investigated how biodiversity optimization decisions

that treat species independently can be modified in respect

of interactions such as predator–prey relationships. Extinction

risk is also unevenly distributed amongst taxa

26,27

. However,

hybridization probabilities have the property that, in contrast

to the other interaction effects previously studied, they correlate

negatively with phylogenetic distance and therefore can provide

contrasting PD predictions.

It is important to stress that only in instances where hybridiza-

tion increases the adaptive potential does the above model apply.

What is not modelled here are the conditions under which this

occurs, and the contrasting situations and scenarios under which

hybridization might reduce fitness

. A greater chance of adaptation

is expected where a large number of hybridization events can occur

between closely related species

. However, even when adaptive

potential is increased, hybridization will be a double-edged sword.

It has the potential to promote both long-term survival of endemic

1042 NATURE CLIMATE CHANGE | VOL 3 | DECEMBER 2013 | www.nature.com/natureclimatechange

NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2027

LETTERS

lineages as well as facilitate biological invasions and the replacement

of endemics by exotic species

19,28

. As genome science delivers

more detail on the nature of species, the evolutionary impact of

hybridization will become clearer. A likely dilemma for future

conservation efforts will be in deciding whether to work with or

against hybridization and whether or not recognized species are the

best units for conservation

2,22

Methods

Empirical data analyses. A map showing ice cover and potential peripheral refugia

in the South Island of New Zealand during the LGM was constructed as described

in Supplementary Information 1 and plotted against extant species’ distributions

(Fig. 1a). Accessions were genotyped using an ABI3730 sequencing protocol

(Massey Genome Service) after first characterizing hot spot regions of chloroplast

genome variation (Supplementary Fig. 1). These hypervariable regions were

identified by Illumina sequencing and comparative analysis of chloroplast genomes

for multiple accessions of P. fastigiatum/P. stellatum and P. enysii as also described

in Supplementary Information 1. A haplotype network (Fig. 1c) was constructed

for 1,463 base pairs (bp) of gene ycf1 using the TCS software (see Supplementary

Information 1 for details). For this analysis, indels were each recoded as single

point mutations. Poly-A and poly-T sequences longer than ten bp and one

site with low sequencing quality due to flanking poly-A and poly-T stretches

were removed. For relaxed clock estimates of chloroplast haplotype divergence

times, indels were excluded from the analyses. The alignment of concatenated

sequences is given in Supplementary Information 1. A normal distribution prior

of 10 Ma ± 1 Ma was placed on the divergence time of A. thaliana from A. lyrata.

An exponential distribution prior (mean 1.0 Ma, offset = 0.8 Ma) was used to

bound the diversification age of alpine Pachycladon. Further details have been

provided in Supplementary Information 1. The program DnaSPv5 (ref. 29) was

used to calculate K

ratios in comparisons of allelic variation for glucosinolate

hydrolysis genes. A 99 bp window and step size of 9 bp was used. Gene-wide K

values were used to avoid the problem of K

= 0 within individual windows.

For statistical tests of hybridization we used: JML (ref. 11); the species tree from

Joly et al. (manuscript in preparation); and concatenated sequences of chloroplast

genes or glucosinolate hydrolysis genes. Details of computational analyses have also

been included in Supplementary Information 1.

Phylogenetic diversity. For our model of phylogenetic diversity loss accompanying

possible hybridization within the AB clade, the expression (2p + x

)T in Scenario

1 arises from writing the random variable PD as PD = X

T + X

T , where X

and

are 0,1 random variables with X

= 1 if either A survives or an (A, B) hybrid

becomes established and survives, and X

= 1 if C survives. Then the expected PD is

Pr(X

= 1)T + Pr(X

= 1)T and because Pr(X

= 1) ≤ p + x

by Boole’s inequality,

and Pr(X

= 1) = p, the expected PD is no more than (2p + x

)T . For Scenario 2

we have PD = XT , where X is the 0,1 random variable that equals 1 precisely when

at least one of the following three events occur: A survives, B survives, or the event

(H) that a hybrid of A and B becomes established and survives. Thus, the expected

PD in this case is Pr(X

= 1)T ≥ Pr(H)T = x

T . Thus, if Inequality (1) holds,

Scenario 2 will have higher expected PD than Scenario 1.

This analysis extends to the setting where hybrids can also arise within the

(C, D) clade by noting that the expected PD in Scenario 2 is at least x

T + x

whereas for Scenario 1 it is at most (p + x

)T + (p + x

)T so when Inequality (2)

holds, Scenario 2 will again have a higher expected PD than Scenario 1.

Received 3 May 2013; accepted 16 September 2013;

published online 3 November 2013

References

1. Hoffmann, A. A. & Sgrò, C. M. Climate change and evolutionary adaptation.

Nature 470, 479–485 (2011).

2. Bálint, M. et al. Cryptic biodiversity loss linked to global climate change.

Nature Clim. Change 1, 313–318 (2011).

3. Anderson, J. T., Panetta, A. M. & Mitchell-Olds, T. Evolutionary and ecological

responses to anthropogenic climate change: Update on anthropogenic climate

change. Plant Physiol. 160, 1728–1740 (2012).

4. Abbott, R. et al. Hybridization and speciation. J. Evol. Biol. 160, 229–246 (2013).

5. Muhlfeld, C. C. et al. Hybridization rapidly reduces fitness of a native trout in

the wild. Biol. Lett. 5, 328–331 (2009).

6. Rieseberg, L. H. Evolution: Replacing genes and traits through hybridization.

Curr. Biol. 19, R119–R122 (2009).

7. Arnold, M. L. Transfer and origin of adaptations through natural hybridization:

Were Anderson and Stebbins right? Plant Cell 16, 562–570 (2004).

8. The Heliconius Consortium, Butterfly genome reveals promiscuous exchange

of mimicry adaptations among species. Nature 487, 94–98 (2012).

9. Heenan, P. B. & Mitchell, A. D. Phylogeny, biogeography and adaptive

radiation of Pachycladon (Brassicaceae) in the mountains of South Island, New

Zealand. J. Biogeograph. 30, 1737–1749 (2003).

10. Joly, S., Heenan, P. B. & Lockhart, P. J. A Pleistocene inter-tribal

allopolyploidization event precedes the species radiation of Pachycladon

(Brassicaceae) in New Zealand. Mol. Phylogenet. Evolut. 51, 365–372 (2009).

11. Joly, S. JML: Testing hybridization from species trees. Mol. Ecol. Res. 12,

179–184 (2012).

12. Zhang, Z., Ober, J. A. & Kliebenstein, D. J. The gene controlling the quantitative

trait locus EPITHIOSPECIFIER MODIFIER1 alters glucosinolate hydrolysis

and insect resistance in Arabidopsis. Plant Cell 18, 1524–1536 (2006).

13. Lambrix, V., Reichelt, M., Mitchell-Olds, T., Kliebenstein, D. J. &

Gershenzon, J. The Arabidopsis epithiospecifier protein promotes the

hydrolysis of glucosinolates to nitriles and influences Trichoplusia ni herbivory.

Plant Cell 13, 2793–2807 (2001).

14. Agee, A. E. et al. Modified vacuole phenotype1 is an Arabidopsis

myrosinase-associated protein involved in endomembrane protein trafficking.

Plant Physiol. 152, 120–132 (2010).

15. Voelckel, C. et al. Transcriptional and biochemical signatures of divergence

in natural populations of two species of New Zealand alpine Pachycladon.

Mol. Ecol. 17, 4740–4753 (2008).

16. Wittstock, U. & Burow, M. Glucosinolate breakdown in Arabidopsis:

Mechanism, regulation and biological significance. Arabidopsis Book 8,

e0134 (2010).

17. Voelckel, C., Gruenheit, N., Biggs, P., Deusch, O. & Lockhart, P. J. Chips and

tags suggest plant-environment interactions differ for two alpine Pachycladon

species. BMC Genom. 13, e322 (2012).

18. Nadachowska-Brzyska, K., Zieliński, P., Radwan, J. & Babik, W. Interspecific

hybridization increases MHC class II diversity in two sister species of newts.

Mol. Ecol. 21, 887–906 (2012).

19. Abi-Rached, L. et al. The shaping of modern human immune systems by

multiregional admixture with archaic humans. Science 334, 89–94 (2011).

20. Benton, M. J. The red queen and the court jester: Species diversity and the role

of biotic and abiotic factors through time. Science 323, 728–732 (2009).

21. Moulton, V., Semple, C. & Steel, M. Optimizing phylogenetic diversity under

constraints. J. Theoret. Biol. 246, 186–194 (2007).

22. O’Brien, S. J. & Mayr, E. Bureaucratic mischief: Recognizing endangered

species and subspecies. Science 251, 1187–1188 (1991).

23. Allendorf, F. W., Leary, R. F., Spruell, P. & Wenburg, J. K. The problems with

hybrids: Setting conservation guidelines. Trends Ecol. Evol. 16, 613–622 (2001).

24. Weitzman, M. L. The Noah’s Ark problem. Econometrica 66, 1279–1298 (1998).

25. Witting, L., Tomiuk, J. & Loeschecke, V. Modelling the optimal conservation

of interacting species. Ecol. Model. 125, 123–144 (2000).

26. Purvis, A., Agapow, P. M., Gittleman, J. L. & Mace, G. M. Nonrandom

extinction and the loss of evolutionary history. Science 288, 328–330 (2000).

27. Thuiller, W. et al. Consequences of climate change on the tree of life in Europe.

Nature 470, 531–534 (2011).

28. Ryan, M. E., Johnson, J. R. & Fitzpatrick, B. M. Invasive hybrid tiger

salamander genotypes impact native amphibians. Proc. Natl Acad. Sci. USA

106, 11166–71 (2009).

29. Librado, P. & Rozas, J. DnaSP v5: A software for comprehensive analysis of

DNA polymorphism data. Bioinformatics 25, 1451–1452 (2009).

Acknowledgements

N.G. and O.D. were supported by Postdoctoral Fellowships from the German Academic

Exchange Service (DAAD). C.V. was a recipient of an Alexander von Humboldt Feodor

Lynen Research Fellowship. This work was initiated with financial support from the New

Zealand Marsden Fund and received additional project funding from Massey University.

P.J.L. and M.S. contributed to this work while New Zealand Royal Society James Cook

Fellows. We thank B. Martin, S. Joly and K. Sluis (Illumina) for their support and

encouragement, and V. Symonds for constructive comments.

Author contributions

M.B., N.G., O.D., C.V., P.B.H. and P.J.L. designed the experiments and conducted

most analyses. P.A.M. and O.K. provided technical support. The authorship order

reflects relative contributions. J.W.L. designed and conducted the recombination

breakpoint test. P.J.L. developed the conjecture, and M.S. the mathematical model

described in the manuscript.

Additional information

Supplementary information is available in the online version of the paper. Reprints and

permissions information is available online at www.nature.com/reprints. Correspondence

and requests for materials should be addressed to M.B.

Competing ﬁnancial interests

The authors declare no competing financial interests.

NATURE CLIMATE CHANGE | VOL 3 | DECEMBER 2013 | www.nature.com/natureclimatechange 1043

Hybridization may facilitate in situ survival of endemic species through periods of climate change

Figures

Citations

Plant DNA barcoding: From gene to genome

Adaptive introgression as a resource for management and genetic conservation in a changing climate

Invasive hybridization in a threatened species is accelerated by climate change

Biological invasions, climate change and genomics.

Whole-genome sequence analysis shows that two endemic species of North American wolf are admixtures of the coyote and gray wolf

References

DnaSP v5

Climate change and evolutionary adaptation

Hybridization and speciation

The problems with hybrids: setting conservation guidelines

Butterfly genome reveals promiscuous exchange of mimicry adaptations among species

Related Papers (5)

Extinction by hybridization and introgression

Climate change and evolutionary adaptation

Inference of population structure using multilocus genotype data

Hybridization as an invasion of the genome

Hybridization and speciation

Frequently Asked Questions (1)

Q1. What are the contributions in "Hybridization may facilitate in situ survival of endemic species through periods of climate change" ?