Fragmented Habitats Accelerate Butterfly Evolution - And Extinction

Genetic analysis of more than 90 butterfly museum specimens from Finland revealed that populations from highly fragmented habitats can undergo rapid evolutionary changes as well as extraordinary losses in genetic variation. By comparing extinct populations to those that still live, researchers found natural selection for populations that were more likely to disperse and colonise new habitats. But they also found that even accelerated evolution may not be fast enough to save shrinking populations from extinction. This study emphasises the value of old museum collections for demonstrating how populations and species respond to a rapidly changing world.

The Glanville fritillary, Melitaea cinxia, is a medium-sized orange, black and white "checkerspot" butterfly that lives in dry meadows and grasslands throughout temperate regions of Asia and parts of Europe. A dimorphic species, adult females are slightly larger than the more intensely coloured males. Glanville fritillaries spend most of their lives as very hungry caterpillars: adults live for only a few weeks and produce just one generation each year.

Unfortunately for this handsome butterfly, humans really like their habitats. Since agriculture and other land uses have transformed or destroyed many dry meadows and grasslands throughout western and northern Europe, the Glanville fritillary has experienced correspondingly severe population declines in those parts of its range. For example, in southwest (SW) Finland, which is the northernmost edge of this species' range, all of the mainland and southwestern archipelago populations went extinct in the 1970s. Currently, this species is restricted to the Åland Islands (Figure 1A), an archipelago that lies between Finland and Sweden in the Baltic Sea where a large "metapopulation" occurs in a highly fragmented landscape consisting of some 4,000 dry meadows located across many islands.

A metapopulation is comprised of a loose network of many smaller geographically separate populations that are nevertheless genetically interconnected. Whilst the metapopulation itself remains stable, each subpopulation is yoyoing along its own demographic trajectory: some are approaching extinction whilst others are rapidly increasing in number. Over time, subpopulations tend to go extinct due to inbreeding and other localised events but are re-established through colonisation by individuals that disperse away from nearby thriving populations. These metapopulation extinction-colonisation dynamics are one example of rapid evolutionary change.

"I originally come from the UK where we are unfortunately no stranger to extinctions", said the study's lead author, evolutionary biologist Toby Fountain, a postdoctoral researcher at the University of Helsinki in Finland.

"In fact our study organism, the Glanville fritillary, gets its name from Lady Eleanor Glanville who collected the butterfly in my home county of Lincolnshire in central England in the 17th century", said Dr Fountain in email. According to Dr Fountain, changing land use and subsequent habitat fragmentation in the UK pushed this species into extinction, so the only surviving UK populations occur on the South Coast and on the Isle of Wight.

Despite these widespread extinctions, this butterfly species has the capacity to quickly adapt to highly fragmented habitats, too. For example, previous studies found that Glanville fritillaries living in the naturally fragmented landscapes of Åland have significantly higher flight metabolic rates than do those living in continuous areas (doi:10.1038/33136). An increased flight metabolism enhances mobility -- exactly what you would expect in a butterfly metapopulation that shows high dispersal and colonisation rates.

Predictably, these physiologic adaptations are correlated with molecular changes. Baseline gene expression profiles are different in butterflies living in fragmented landscapes when compared to profiles from butterflies found in continuous landscapes (doi:10.1371/journal.pone.0101467) -- although neither the identities nor the functions of these up-regulated genes are known at this time.

Although preliminary, these observations highlight some interesting questions regarding the ongoing evolutionary dynamics in the extant Åland metapopulation and how they might contrast to those in the extinct SW Finland metapopulations. Did the extinct SW Finland metapopulations have less genetic variability (that may have contributed to their ultimate extinction) when compared to the thriving Åland metapopulation? Did the extinct SW Finland metapopulations show similar genotype shifts (presumably indicating natural selection specifically for genes related to flight) when compared to genetic changes seen in the Åland metapopulation? And how might researchers learn more about extinct populations and the evolutionary processes that preceded their extinctions?

This is where Natural History Museum collections come to the rescue: because museums house collections of animals and plants -- many of them quite old -- it is possible to use modern technologies to capture snapshots through time of past events (doi:10.1111/mec.13529).

Extinct populations became increasingly distinct from extant populations over time

Dr Fountain and his team of researchers collected samples from more than 90 museum specimens representing extinct and extant populations of the Glanville fritillary in Finland, some of which were more than 100 years old, and analysed single nucleotide polymorphisms (SNPs -- pronounced “snips”). SNPs are a common type of genetic variation, or allele. Each SNP is composed of a tiny difference in just one DNA building block, or nucleotide, at the same location in two otherwise-identical genes.

Genetic analysis of 222 SNPs distributed throughout the genome revealed that historic Åland samples (n = 74; green dots in Figure 2) clustered neatly together with samples from Saltvik (n = 39; blue dots in Figure 2), a small region in Åland that serves as the modern reference population for this study, having been sampled between 2007 and 2011. This clustering indicates there was no loss of genetic diversity from the Åland metapopulation during the previous 100 years. But when these metapopulations were compared to the extinct SW Finland metapopulations, we see that the oldest samples are closest to the Åland/Saltvik samples, whereas the more recently collected SW Finland samples show increasing genetic divergence (Figure 2):

Different populations show similar allelic shifts to habitat fragmentation

Although it is significantly smaller than the extinct SW Finland metapopulations, the Sottunga metapopulation (purple dots; Figure 1) provides another example of how the Glanville fritillary is responding to fragmented landscapes. The Sottunga metapopulation was introduced only 24 years ago into a small and highly fragmented landscape consisting of just 51 suitable habitat patches. Thus, the team was able to study three distinct metapopulation types: an old living population (Åland/Saltvik), several old dead populations (SW Finland) and a new living population (Sottunga).

To learn whether all of these populations responded similarly to fragmented habitats, Dr Fountain and his team compared the magnitude and direction of allele frequency changes away from the Saltvik frequencies (solid middle line; Figure 3) for the Sottunga metapopulation (white circles; Figure 3) and for the extinct SW Finnish populations (black circles; figure 3):

Although allele frequencies shifted in mostly the same direction away from the Saltvik frequencies, there is one important difference: allele frequency shifts were significantly smaller in the newly-established Sottunga metapopulation than in the extinct SW Finland metapopulations, which had far longer -- 80 years, on average -- to diverge. Clearly these shifts did not result from genetic drift, which would have been scattered: the consistency and directionality of these changes indicate that they were influenced by nonrandom processes.

Alleles associated with flight showed strong selective pressure, even in extinct populations

Dr Fountain and his team then analysed 12 of the 15 "outlier" alleles that had been excluded from the previous analysis. These "outlier" alleles were analysed separately because they consist of SNPs that were correlated in some way with flight and thus, the team already knew they were under strong selective pressure in the metapopulations that were living in highly fragmented habitats. But did these alleles show similar changes in old and new metapopulations? Did living and extinct populations show similar allele shifts?

To answer these questions, the team compared outlier allele shifts for the extant Sottunga and extinct SW Finland metapopulations with four additional regional populations in the Baltic region (see Figure 1 for their locations). Two of these regional metapopulations, Åland in Finland and Uppland in Sweden, live in highly fragmented landscapes, whereas the other two, Saaremaa in Estonia and Öland in Sweden, occur in more continuous (less fragmented) habitats.

The data reveal that outlier allele frequencies for these four regional populations were strongly positively correlated with allele frequencies in fragmented populations and strongly negatively correlated with allele frequencies in the continuous populations (Figure 4). Basically, outlier allele frequencies in new living populations (Sottunga, see Figure 4A) and old extinct populations (SW Finland, see Figure 4B) shifted in the same direction as did allele frequencies for old established populations (Åland) living in fragmented landscapes.

"We compared the genetic composition of the extinct population with several independent datasets and found evidence of selection for enhanced dispersal and colonisation capacity, an adaptive response to habitat fragmentation, in the now extinct population", explained Dr Fountain.

Unfortunately, these studies are handicapped by unavoidably small sample sizes.

"The inference of allele frequency change is based on very small sample sizes for SW Finland, but the study gets around this problem by making multiple comparisons with different contemporary and historical population samples from fragmented and continuous habitat landscapes", said evolutionary ecologist Ilik Saccheri, a senior lecturer at the University of Liverpool, who was not part of this study.

"The correlational approach is necessary and appropriate here but they could have done a bit more to support the claim that the allele frequency changes reflect natural selection for better colonising capacity", explained Dr Saccheri in email.

Yet, these preliminary findings are attractive because they are consistent and repeatable.

"The finding that surprised us most in the study was how repeatable some of the genetic changes the extinct population experienced were", said Dr Fountain.

Despite having gone extinct in the 1970s, the team found evidence that the SW Finland metapopulations were rapidly adapting to life in highly fragmented landscapes -- albeit, not fast enough.

"This must be what is happening to many many declining populations -- rapid adaptation that ultimately falls short of the growth rate required for long term persistence", said Dr Saccheri.

Based on what is currently known, is it possible to predict whether the very small, very young, Sottunga metapopulation will go extinct in the future?

"Sottunga is definitely vulnerable", said Dr Fountain. "There have been a number of years where it has almost gone extinct, but the population has bounced back the next year. Isolated populations with low genetic diversity are often the most vulnerable, particularly to environmental change."

Considering the many, often subtle, effects of increasing habitat fragmentation, environmental change and global climate change, it is important to learn how to identify populations that are vulnerable to extinction, although they may hang on, barely, for a very long time before eventually winking out -- thereby confounding predictions.

"For example, a completely isolated population of the Glanville fritillary has survived on the island of Pikku Tytärsaari in the Baltic Sea, for at least 75 years despite high genetic load", said Dr Fountain. "Thus it can be challenging to predict whether and when a population will go extinct."

Museums can shed light on extinct populations

"Fountain et. al. confirm the importance of museum sampling. They are able to examine population genetics before and after an extinction event, something that would be otherwise impossible without such repositories,” said biologist Michael Holmes, an assistant professor at Coastal Carolina University, who was not part of this study.

"This article is a great example of the power of museum samples to study evolutionary processes in extant, and in this case extirpated, populations and species, especially as it pertains to human-induced environmental change,” said Professor Holmes in email.

"Museum samples are an excellent resource as they can give us a snap shot of evolution in action", said Dr Fountain. "These samples allow us to track genetic changes over time, giving us information on the population's demographic history as well as how they may respond to changing environments."

"Using museum samples allows us to go back in time to see how rapid environmental change affects the evolution of populations", explained Dr Fountain. "This is particularly important for populations that have now gone extinct".

Museum collections can provide valuable glimpses into the ongoing evolutionary process -- but only if they are regularly updated.

"Continuous sampling of populations gives us a time-capsule record of that biological unit,” said Professor Holmes.

Modern technologies allow scientists to unveil mysteries contained in museum collections that would otherwise remain invisible.

"Until relatively recently scientists were really only able to study phenotypic changes in museum specimens,” said Professor Holmes in email. "We saw this with Kettewell’s observations of peppered moth populations before, during and after the Industrial Revolution.”

"However, with PCR and now next-gen sequencing and genomics we can see these evolutionary changes at the molecular level and in cases such as Fountain et al’s paper, we can attribute some of these genomic changes to phenotypic changes”, explained Professor Holmes.

"The implications [are] that species may have the potential to adapt to rapidly changing environments, but that these evolutionary changes may not be strong enough to rescue the population from extinction", said Dr Fountain. "This is important for any species which is experiencing rapid environmental change, as well as other members of the ecological community that rely on those species (eg for pollination)."

Museum Specimens Reveal Habitat Fragmentation Increases Butterfly Evolution & Extinction | @GrrlScientist

Sources:

Toby Fountain, Marko Nieminen, Jukka Sirén, Swee Chong Wong, and Ilkka Hanski (2016). Predictable allele frequency changes due to habitat fragmentation in the Glanville fritillary butterfly, Proceedings of the National Academy of Sciences | doi:10.1073/pnas.1600951113

Also cited:

Michael W. Holmes, Talisin T. Hammond, Guinevere O. U. Wogan, Rachel E. Walsh, Katie LaBarbera, Elizabeth A. Wommack, Felipe M. Martins, Jeremy C. Crawford, Katya L. Mack, Luke M. Bloch, and Michael W. Nachman (2016). Natural history collections as windows on evolutionary processes, Molecular Ecology 25, 864–881 | doi:10.1111/mec.13529

Ilik Saccheri, Mikko Kuussaari, Maaria Kankare, Pia Vikman, Wilhelm Fortelius & Ilkka Hanski (1998). Inbreeding and extinction in a butterfly metapopulation, Nature, 392 (6675) 491-494 | doi:10.1038/33136

Panu Somervuo, Jouni Kvist, Suvi Ikonen, Petri Auvinen, Lars Paulin, Patrik Koskinen, Liisa Holm, Minna Taipale, Anne Duplouy, Annukka Ruokolainen, Suvi Saarnio, Jukka Sirén, Jukka Kohonen, Jukka Corander, Mikko J. Frilander, Virpi Ahola, and Ilkka Hanski (2014). Transcriptome analysis reveals signature of adaptation to landscape fragmentation, PLoS ONE 9(7): e101467 | doi:10.1371/journal.pone.0101467