Conservation Genetics

, Volume 14, Issue 1, pp 231–236

Low genetic diversity of a high mountain burnet moth species in the Pyrenees

Authors

    • National Museum of Natural History Luxembourg
    • Department of Community EcologyCentre for Environmental Research UFZ
  • Claudia Drees
    • Biocentre Grindel and Zoological MuseumUniversity of Hamburg
  • Thomas Schmitt
    • Department of BiogeographyTrier University
  • Thorsten Assmann
    • Institute of EcologyLeuphana University Lüneburg
Short Communication

DOI: 10.1007/s10592-012-0424-0

Cite this article as:
Dieker, P., Drees, C., Schmitt, T. et al. Conserv Genet (2013) 14: 231. doi:10.1007/s10592-012-0424-0

Abstract

The burnet moth Zygaena anthyllidis, endemic to the high elevations of the Pyrenees, is vulnerable to land-use. In order to identify conservation priorities based on an assessment of genetic diversity within populations and gene flow among populations, we examined Z. anthyllidis’ genetic variability and differentiation based on allozyme electrophoresis from seven populations scattered across its entire range. In comparison to other mountain Lepidoptera, the populations studied exhibit a low level of genetic diversity. Remarkable between-population differentiation (FST = 0.053), the presence of private alleles, and the lack of significant isolation-by-distance pattern characterises the genetic make-up of the species. We interpreted the pattern of genetic differentiation as a consequence of low dispersal power in combination with insufficient landscape connectivity. Ongoing land-use change might reinforce genetic differentiation due to habitat fragmentation and additionally affect negatively allozyme variability at shifting range margins, i.e. the capacity to adapt to changing environments. We therefore suggest creating a network of suitable habitats at the landscape scale to facilitate genetic exchange and to conserve the species’ overall genetic variability.

Keywords

AllozymesConservation geneticsGenetic diversityGenetic differentiationPyreneesZygaena anthyllidis

Introduction

The recent, drastic decline in biodiversity is linked to profound and rapid changes in climate and land-use (Walther et al. 2002; Rosenzweig et al. 2008). Cold-adapted, mountain species are particularly sensitive to global warming, especially endemic species which are confined to local climatic conditions in small geographical areas (Dirnböck et al. 2011). Numerous mountain species react to climate change by shifting their ranges towards higher altitudes. Such shifts in the upper altitudinal range limits in particular are attributed to global warming (Holzinger et al. 2008). However, changes in the lower range margins are often caused by the combination of changes in climate and land-use with the impact of land-use changes sometimes outweighing that of climatic changes (Lenoir et al. 2010). In this case, there is a need to implement conservation measures to preserve alpine species by paying special attention to populations located at the species’ lower range margin which often function as source populations for higher ones (Forister et al. 2010).

Since the 1950′s, profound socio-economic changes have led to the collapse of traditional farming systems such as the transhumance, a rotating grazing system that was widely distributed in European mountain ranges. As a consequence, in many places pastures were abandoned or grazing became intensified, mainly at well accessible sites (MacDonald et al. 2000). This situation presents a challenge for species depending on the continuation of extensive grazing regimes. Habitat loss and fragmentation thus are accompanied by a reduction of landscape permeability which in turn can affect gene flow between populations. In the long-term this can alter the genetic structure of populations (Keyghobadi et al. 2005).

We studied Zygaena anthyllidis Boisduval 1828, an endemic burnet moth species (Lepidoptera, Zygaenidae) found at high altitudes along the main ridge of the Western and Central Pyrenees. As a characteristic species of subalpine and alpine grasslands, its occurrence is highly dependent on low-intensity grazing practices. Changes in the traditional grazing management during the last decades, however, have led to drastic, localised uphill shifts in the species’ lower altitudinal range limits (Dieker et al. 2011). In order to examine the impacts of these changes on the species’ genetic make-up, we analysed the pattern of current genetic diversity as well as intra- and interpopulation variation in seven Z. anthyllidis populations scattered across the species’ entire distribution. We discuss the threats posed by ongoing land-use change, and provide recommendations for effective management measures in order to maintain long-term genetic variability within and among populations.

Materials and methods

Collection of samples

We collected 249 specimens from seven populations of Z. anthyllidis across its entire distribution range (Fig. 1; Table 1). Geographical distances between the populations ranged from 14 to 140 km. Adults were netted during the end of the flight period of 2008 (late July), frozen in liquid nitrogen and stored at −80 °C until analysis.
https://static-content.springer.com/image/art%3A10.1007%2Fs10592-012-0424-0/MediaObjects/10592_2012_424_Fig1_HTML.gif
Fig. 1

Map of the Pyrenees, showing altitudinal zones as well as the sampling localities and the distribution range of Z. anthyllidis. Population numbers are as given in Table 1

Table 1

Overview of the seven studied Z. anthyllidis populations

Name of sampling locality

Population no.

Coordinates N; E

N

AR30

P (%)

P95 (%)

Mean heterozygosity

HWE

No. of private alleles

Frequency of private alleles

Ho

HE

E-Aísa

P1

42.77; −0.59

40

1.51

35.3

29.4

5.6

7.0

+

1

3.7

F-Artouste

P2

42.86; −0.34

40

1.51

41.2

35.3

7.6

9.2

1

3.7

E-Panticosa

P3

42.68; −0.28

40

1.42

17.6

35.3

4.6

4.9

+

1

1.3

F-Estaragne

P4

42.84; 0.15

40

1.32

23.5

23.5

4.0

4.8

+

2

7.5; 1.3

E-Balcón de Pineta

P5

42.68; 0.05

13

1.32

29.4

23.5

6.8

4.1

+

F-Pic du Midi de Bigorre

P6

42.94; 0.14

40

1.20

16.7

11.8

3.0

3.3

+

1

1.3

F-Port d’Aula

P7

42.77; 1.11

36

1.28

23.5

23.5

4.1

4.6

+

Mean (SD)

   

1.36 (±0.12)

29.4 (±8.3)

21.9 (±8.1)

5.1 (±1.6)

5.7 (±1.9)

 

1.2 (±0.45)

 

Given are the location, population number, the sampling coordinates, the number of specimens genetically analyzed (N), the allelic richness across all loci (AR30; extrapolated to 30 specimens), the percentage of polymorphic loci (P) and the percentage of loci where the most frequent allele did not exceed 95 %; (P95), the observed (HO) and expected (HE) heterozygosity, the presence of the Hardy–Weinberg–Equilibrium (HWE: + p > 0.05; − deviation from HWE with p < 0.05), the number of private alleles, and the frequency of private alleles

Allozyme electrophoresis

Half of each specimen’s abdomen was homogenized in Pgm buffer using ultrasound and centrifugation at 17,000 g for 5 minutes. The clear supernatants were stamped on cellulose acetate plates (Helena Laboratories, Beaumont, TX, USA), and electrophoresis was conducted following standard methods (Hebert and Beaton 1993). We analyzed 17 allozyme loci [Gpdh, Hbdh, Mdh 1+2 on TC buffer (pH: 7.0); Me, Idh, 6Pgdh, G6pdh, G3pdh, Aat 1+2, Fum, Acon 1+2, Mpi, Pgi, Pgm on TM buffer (pH: 8.0)]. Allozyme conditions followed the ones used for Z. exulans (Schmitt and Hewitt 2004).

Population genetic analyses

The following parameters of genetic diversity were calculated: (i) Observed (HO) and (ii) expected (HE) heterozygosity which were calculated over all loci for each population, (iii) allelic richness (AR) which was determined by extrapolating to a sample size of 30, and (iv) the proportion of those loci where the most frequent allele did not exceed 95 % (P95). All diversity parameters were calculated using the ARLEQUIN package, version 3.5.1.2 (Excoffier et al. 2005), except AR which was assessed using the ARES package (Van Loon et al. 2007) for R (R Development Core Team 2011). We tested populations for conformity to the Hardy–Weinberg–Equilibrium (HWE) and checked for genotypic linkage disequilibrium by using ARLEQUIN. This program was also used to assess the population structure to calculate the F-statistics FIS and FST and a hierarchical analysis of molecular variance (AMOVA). A Mantel test as implemented in AIS 1.0 (Miller 2005) was performed to test whether geographic and genetic distances were correlated. The software was also used to test for spatial autocorrelation. We used the software BOTTLENECK (Cornuet and Luikart 1996) to identify populations that might have gone through recent bottlenecks.

Results

In total, 33 different alleles were detected at 17 allozyme loci, nine of which were polymorphic (6Pgdh, G6pdh, Hbdh, Idh2, Aat2, Me, Pgi, Mpi, Pgm). The highest number of alleles (four) was detected in Hbdh and Mpi. Four of the nine polymorphic loci (G6pdh, Hbdh, Aat2, Mpi) had private alleles scattered across five populations (Table 1).

Linkage disequilibrium was observed only in the populations P1 (locus pairs 6Pgdh and Gpdh) and P3 (locus pair Mpi and Pgm). Population 2 showed significant deviation from the HWE (Table 1) due to heterozygote deficiency in the loci 6Pgdh and Pgm. The mean allelic richness per locus calculated over all populations was 1.36 (AR30: 1.20–1.51). The average P95 is 21.9 % (P95: 11.7–35.3 %). The genetic diversity (HE) between the populations ranged from 3.3 to 9.2 % (mean: 5.7 %). The lowest values of the different diversity parameters were found in the northernmost sample site P6, the highest values were detected in the western populations P1 and P2.

The studied populations were significantly differentiated (FST = 0.053, p < 0.0001). The variance analysis revealed 5.3 % of the total variance among populations, 12.3 % of the variation among individuals within populations (FIS = 0.129, p < 0.0001), and about 80 % of the genetic variation within individuals. The populations did not show a significant isolation-by-distance pattern (Mantel test correlation coefficient: −0.032; p > 0.05). No significant pattern of spatial autocorrelation was found. No recent bottleneck (heterozygote excess) is suspected in any of the populations.

Discussion

Genetic diversity and differentiation

The overall genetic diversity of the Z. anthyllidis populations is low in comparison to other mountain burnet moth and butterfly species, even if the comparison focuses only on Pyrenean populations (Table 2). This might be due to the geographic restriction of Z. anthyllidis (distribution range < 2,000 km2; see Fig. 1) which is concomitantly accompanied by an overall lower population size (e.g. Gitzendanner and Soltis 2000; Frankham 2003). The detected level of genetic diversity of Z. anthyllidis populations is comparable to that of other endangered butterfly species such as the Violet Copper Lycaena helle (Habel et al. 2011), the Alcon Blue Maculinea alcon (Gadeberg and Boomsma, 1997) or the burnet moth Z. carniolica (Habel et al. 2012).
Table 2

Comparision of genetic diversity and structure of Z. anthyllidis and other mountain butterfly species

Species

Mountain range

Max. distance studied (km)

No. of populations studied

No. of loci studied

Over all populations

Author

Polymorphic loci Ptot/P95 [%] mean (min–max)

Max. no. of alleles per locus A mean (min–max)

Ho mean (min–max)

HE mean (min–max)

FST

Oeneis iva

Sierra Nevada

65–240

20

16

37.5

1.6

12.8

13.3

Porter and Shapiro (1989)

Oeneis chryxus stanislaus

Sierra Nevada

 

19

16

37.5

1.6

14.6

13.2

Coenonympha gardetta

Alps

30–430

2

18

61.2/41.7

2.2

14.3

13.7

Schmitt and Besold (2010)

Erebia epiphron silesiana

Jesenik Mountains, Krkonose Mountains

3–8

6 a

17

48.5/27.9

1.6

8.9

Schmitt et al. (2005)

Erebia epiphron

Alps, Pyrenees, Jesenik Mountains

15–1,500

16

18

69.7/43.2

2.1

15.4

0.283***

Schmitt et al. (2006)

 

Eastern Pyrenees

15–100

4

 

69.5/44.4

2.0

13.5

 

Central Pyrenees/Western Alps

20–500

5

 

77.8/46.3

2.3

16.4

Erebia melampus

Alps

40–600

23

18

47.7/33.1

(22.2–77.8)/(16.7–38.9)

1.8

(1.44–2.1)

9.8

(6.4–11.8)

10.8

(6.4–15.7)

Haubrich and Schmitt (2007)

Erebia sudetica inalpina

Alps

 

1

18

42.8/25.0

1.6

9.7

8.9

Erebia euryale

Pyrenees, Alps, Carpathians, Rila Mountains

20–1,900

11

17

68.5/41.2

(47.1–88.2)/(17.6–64.7)

2.3

(1.7–3.1)

15.2

(7.9–18.2)

15.6

(8.2–19)

0.199***

Schmitt and Haubrich (2008)

 

Pyrenees

 

1

 

47.1/41.2

1.7

13.7

16.1

Zygaena exulans

Pyrenees, Alps

15–600

7

17

55.2/37.1

(26.7–86.7)/(26.7–60)

1.8

(1.3–2.3)

10.5

(6.6–17.2)

0.054***

Schmitt and Hewitt (2004)

 

(Eastern) Pyrenees

15

2

 

50/26.7

(46.7–53.3)/(26.7)

1.7

(1.6–1.9)

4.2

(6.6–7.7)

0.0015

Zygaena exulans

Pyrenees, Alps, Apennines, Carpathians

7–2,050

24

18

57.6/34.7

(22.2–77.8)/(5.6–50)

1.8

(1.3–2.1)

11.3

(1.4–19)

10

(1.4–16)

0.189***

Dieker et al. (unpublished data)

 

(Western and Central) Pyrenees

7–100

5

 

38.9/24.5

(33.50–50)/(16.7–27.8)

1.5

(1.4–1.8)

7.4

(6.7–8.2)

6.9

(6.3–7.6)

0.036***

Mean (±SD)

    

56.1/35.4

(±12.5)/(±8.0)

1.8

(±0.3)

11.2

(±2.9)

11.4

(±4.0)

  

Zygaena anthyllidis

(Western and Central) Pyrenees

14–140

7

17

28.7/21.9

(17.7–1.2)/(11.8–35.3)

1.4

(1.4–1.5)

5.1

(3.0–7.6)

5.7

(3.3–9.2)

0.053***

This study

Only studies using allozyme electrophoresis were considered. For explanations of genetic diversity measures see Table 1. In addition, FST-values are given (*** p < 0.001)

aThe population in the Krkonoske Mountains was recently established with individuals from Jesenik Mountains

Z. anthyllidis is a characteristic species of extended subalpine and alpine pastures in the Western and Central Pyrenees. Therefore, it is surprising that the studied populations are well differentiated, especially in comparison to the congeneric, arctic-alpine species Z. exulans (FST = 0.0015) (Schmitt and Hewitt 2004) which mainly colonizes passes, plateaus and exposed habitats, and therefore, has a scattered distribution pattern in the Pyrenees (Dieker et al. 2011, own observations).

The remarkable genetic differentiation between Z. anthyllidis populations is highlighted by the significant FST-value and the presence of private alleles which have been detected in two-thirds of the studied populations. The limitation of genetic exchange among populations is in turn reflected in the positive FIS-value. The lack of evidence for population-specific causes of strong differentiation, such as recent bottlenecks (although there might be methodological limitations due to the low polymorphisms), and the lack of a significant isolation-by-distance relationship can be interpreted as an old genetic pattern.

In general, strong genetic differentiation can be explained by low dispersal power and/or low habitat connectivity. Both drivers affect the spatial mobility of genotypes, especially as the effects of genetic drift and mutational load are reduced by dispersal (Ronce 2007). Long-distance dispersers are of significance as they link local populations as well as colonizing empty habitat patches (Hughes et al. 2007). However, no evidence for such dispersers is available in this species. In addition, the found differentiation pattern may be due to a strong influence of landscape connectivity and structural composition on the species’ dispersal (Baguette and van Dyck 2007). Landscape elements such as mountain ridges and scree slopes may function as barriers which lead to naturally isolated habitats. The detected heterozygosity deficit in population P2 (deviation from the HWE) may be due to differentiated subpopulations (Wahlund effect) reflecting small-scale differentiation within the studied sample site (e.g. Castric et al. 2001). This assumption fits to both low power of dispersal and/or low habitat connectivity and may support our suggestions of the main genetic drivers in Z. anthyllidis.

Recent land-use change in the Pyrenees may reinforce the effects of natural habitat isolation by habitat fragmentation and loss due to either grazing intensification or mainly due to abandonment of pastures. Thus, forest encroachment into subalpine pastures (Batllori et al. 2010) with the consequence of increasing fragmentation might nowadays have led to a decrease in successful exchange of individuals among populations and thus might also have affected the genetic population structure, as studies in other butterfly species have shown (e.g. Matter et al. 2011).

In conclusion, a combination of the low Z. anthyllidis dispersal capacity together with the reduction in landscape connectivity might have produced the found genetic differentiation which is reinforced by recent land-use changes. The low level of genetic diversity might pose a problem for the endemic species to adapt to changing environmental conditions. In particular, populations at shifting range margins are expected to suffer from a decline in allozyme variability (for an example see Hill et al. 2006), and therefore may be susceptible to long-term genetic erosion processes.

Implications for conservation

Due to the supposed low gene flow found among Z. anthyllidis populations, we recommend that habitat management strategies should take into account not only conventional species conservation goals, but also ensure viable rates of genetic exchange between populations. This should be achieved by (1) maintaining or re-establishing grazing rates favourable for Z. anthyllidis and (2) conserving and restoring a functional network of suitable habitats across the species range. Such a network would facilitate the dispersal of Z. anthyllidis at the landscape scale and concomitantly foster genetic exchange among populations.

After realising a proper habitat management (Dieker et al. 2011), a re-establishment of formerly existing populations should be attempted in areas where local extinctions of Z. anthyllidis have occurred. The reintroduced individuals should originate from neighbouring populations due to the remarkable genetic differentiation within the entire range.

Acknowledgments

We are grateful to the respective authorities for the necessary permissions. PD was funded by the Scholarship Programme AFR of the National Research Fund (FNR), Luxembourg. Furthermore, we thank the National Museum of Natural History Luxembourg for financial support.

Copyright information

© Springer Science+Business Media Dordrecht 2012