low genetic variability in a natural alpine marmot population (marmota marmota, sciuridae) revealed...

7
Molecular Ecology (1994) 3,347-353 Low genetic variability in a natural alpine marmot population (Mamota marmota, Sciuridae) revealed by DNA fingerprinting K. RASSMANN, W. ARNOLD* and D. TAUTZ Zoologisches lnstitut der Universitat Munchen, Luisenstr. 14, 80333 Miinchen and *Max Plnnck lristitut fiir Verhaltensphysiologie, Abt. Wickler, 82319 Seewiesen, Germany Abstract Genetic heterogeneity is usually considered an important factor for the viability of a population, yet there are cases in which populations sustain themselves despite virtual homozygosity. A prior step to studying the effects of such low levels of genetic variability can be the analysis of its causes. We analysed a population of the highly social alpine marmot (Mumatu mumota, Sciuridae) by multilocus DNA fingerprinting. The finger- print patterns revealed avery low degree of polymorphism in our main study population. We show that this lack of hypervariability is caused by a low effective population size, rather than by an unusual low mutation rate of the fingerprint loci studied. However, the current number of breeding pairs was found to be about an order of magnitude larger than the one that would be expected to lead to such a low degree of heterozygosity. We conclude that there must have been bottlenecks in the history of the Berchtesgaden marmot population that have severely affected its genetic heterozygosity. Keywards: Marmota, DNA-fingerprinting, effective population size, bottleneck Received 19 August 2993; revision received 1 December 1993; accepted 23 December 1993 Introduction Genetic diversity is generally assumed to be an important factor contributing to the viability and long term persist- ence of a natural population (Soul4 1987). The study of hypervariable DNA loci, such as minisatellites or micro- satellites, has indeed shown a high degree of genetic het- erogeneity in many natural populations (Burke etal. 1991). On the other hand, there have also been reports on populations that show much lower degrees of variability for such DNA fingerprint markers, as for example in some populations of the Channel Island fox (Urocyon littoralis) (Gilbert et al. 1990) and in the naked mole rat (Heterocephalus glaber) (Faulkes et al. 1% Reeve et al. 1990). In this study we analysed a population of the alpine marmot (Mannota mannotu L.), in respect to its genetic variability. Today’s populations of alpine marmots are most likely the remainder of a population that occupied a large area in central Europe during the late Pleistocene. Alpine marmots are well adapted to open cold habitats, typically found in zones marginal to glacial ice. With the change from ‘periglacial’towards modem climatic condi- tions at the late Pleistocene and at the beginning of Holocene (c. 9000-10 000 years ago), alpine marmots were forced back into their current narrow range in the Alps and Carpathians (Zimina & Gerasimov 1973; Forter 1975). Alpine marmots are highly social. They live in family groups with up to eleven adult members, since dispersal of offspring is delayed beyond sexual maturity (Arnold 1990a,b). While only one female reproduces in a given group, males are apparently not able to monopolize the reproduction in a similar way. Even though a clearly dominant male exists, all adult males within one family group may copulate multiply with a receptive female (Arnold 1990a). The original aim of our study was to ana- lyse the adaptive value of the mating system found in the alpine marmots by measuring the individual reproduc- tive success through DNA fingerprinting. A previous analysis of blood allozymes had already shown a very low level of polymorphism within the study population. Only two diallelic systems were found among 53 enzyme loci tested (Arnold 1990a). However, since it has been proven that fingerprinting can reveal polymorphism where enzyme electrophoresis fails (Turner et 01. 1990),

Upload: k-rassmann

Post on 01-Oct-2016

213 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Low genetic variability in a natural alpine marmot population (Marmota marmota, Sciuridae) revealed by DNA fingerprinting

Molecular Ecology (1994) 3,347-353

Low genetic variability in a natural alpine marmot population (Mamota marmota, Sciuridae) revealed b y DNA fingerprinting

K. RASSMANN, W. ARNOLD* and D. TAUTZ Zoologisches lnstitut der Universitat Munchen, Luisenstr. 14, 80333 Miinchen and *Max Plnnck lristitut fiir Verhaltensphysiologie, Abt. Wickler, 8231 9 Seewiesen, Germany

Abstract

Genetic heterogeneity is usually considered an important factor for the viability of a population, yet there are cases in which populations sustain themselves despite virtual homozygosity. A prior step to studying the effects of such low levels of genetic variability can be the analysis of its causes. We analysed a population of the highly social alpine marmot (Mumatu mumota, Sciuridae) by multilocus DNA fingerprinting. The finger- print patterns revealed avery low degree of polymorphism in our main study population. We show that this lack of hypervariability is caused by a low effective population size, rather than by an unusual low mutation rate of the fingerprint loci studied. However, the current number of breeding pairs was found to be about an order of magnitude larger than the one that would be expected to lead to such a low degree of heterozygosity. We conclude that there must have been bottlenecks in the history of the Berchtesgaden marmot population that have severely affected its genetic heterozygosity.

Keywards: Marmota, DNA-fingerprinting, effective population size, bottleneck

Received 19 August 2993; revision received 1 December 1993; accepted 23 December 1993

Introduction

Genetic diversity is generally assumed to be an important factor contributing to the viability and long term persist- ence of a natural population (Soul4 1987). The study of hypervariable DNA loci, such as minisatellites or micro- satellites, has indeed shown a high degree of genetic het- erogeneity in many natural populations (Burke etal. 1991). On the other hand, there have also been reports on populations that show much lower degrees of variability for such DNA fingerprint markers, as for example in some populations of the Channel Island fox (Urocyon littoralis) (Gilbert et al. 1990) and in the naked mole rat (Heterocephalus glaber) (Faulkes et al. 1% Reeve et al. 1990).

In this study we analysed a population of the alpine marmot (Mannota mannotu L.), in respect to its genetic variability. Today’s populations of alpine marmots are most likely the remainder of a population that occupied a large area in central Europe during the late Pleistocene. Alpine marmots are well adapted to open cold habitats, typically found in zones marginal to glacial ice. With the change from ‘periglacial’ towards modem climatic condi-

tions at the late Pleistocene and at the beginning of Holocene (c. 9000-10 000 years ago), alpine marmots were forced back into their current narrow range in the Alps and Carpathians (Zimina & Gerasimov 1973; Forter 1975).

Alpine marmots are highly social. They live in family groups with up to eleven adult members, since dispersal of offspring is delayed beyond sexual maturity (Arnold 1990a,b). While only one female reproduces in a given group, males are apparently not able to monopolize the reproduction in a similar way. Even though a clearly dominant male exists, all adult males within one family group may copulate multiply with a receptive female (Arnold 1990a). The original aim of our study was to ana- lyse the adaptive value of the mating system found in the alpine marmots by measuring the individual reproduc- tive success through DNA fingerprinting. A previous analysis of blood allozymes had already shown a very low level of polymorphism within the study population. Only two diallelic systems were found among 53 enzyme loci tested (Arnold 1990a). However, since it has been proven that fingerprinting can reveal polymorphism where enzyme electrophoresis fails (Turner et 01. 1990),

Page 2: Low genetic variability in a natural alpine marmot population (Marmota marmota, Sciuridae) revealed by DNA fingerprinting

348 K. RASSMA" ef a/.

we tested several DNA fingerprinting techniques on this population, mainly focusing on the simple sequence oli- gonucleotide hybridization approach (Ali et al. 1987). However, none of the genetic markers that we detected with our probes showed a high degree of polymorphism. Ln this paper we discuss the different parameters that may account for this low level of variability.

Methods

Study population and sampling

The animals studied were from a native marmot popula- tion (Jenner population) in the National Park of Berchtesgaden, located in the most eastern region of the Bavarian Alps. Part of this population (about 130 indi- viduals per year), which is found in an altitude between 1100 and 1500 m a.s.l., has been studied intensively dur- ing the past ten years (Arnold 1990a,b). About 500 ani- mals have been marked individually since the beginning of the field studies in 1982. Blood samples were taken from almost all of the animals of this population, initially only for hormone analysis and enzyme electrophoresis (Arnold 1990a). Two hundred and thirty of these Jenner population samples could be used for DNA fingerprint- ing. The sample covers two overlapping generations of the Jenner population, including nearly all individuals of the current population in the main study area. In 1990 we sampled 23 individuals from seven family groups of a marmot population in the Berchtesgaden Alps 13 km away from the Jenner population (Funtensee popula- tion). We also processed tissue samples from marmot populations in the Swiss and Austrian parts of the Alps (Switzerland: Bern, Graubuenden; Austria: Arlberg, Karwendel, Zillertal, Grossglockner, Ankogel, Kreuzeck, Eisenerz Alps).

Sample processing and DNA extraction

Many blood clots (0.3-1 mL) from those samples that were originally taken for hormone analysis and stored at -70 "C, could still be used for extracting DNA for finger- printing. Blood samples that were taken specifically for use in fingerprinting, were collected directly into lysis buffer (2-3 mL of blood in 1 volume of lysis buffer: 2% SDS, 100 m~ Tris, 100 mM EDTA) and could be stored for at least one year at ambient temperature. This buffer has been specifically designed for field studies as its compo- nents can be combined as powders and then dissolved in the field with clean water. It is not necessary to adjust the pH since the EDTA and the Tris will neutralize each other.

DNA extraction began with extensive Proteinase K di- gestion (500 pg/mL Proteinase K (Sigma), incubation

overnight at 45 "C), preceded by mechanical maceration of the blood clots if necessary. The solution was then ex- tracted twice with phenol and the DNA precipitated with ethanol. On average, the yield of DNA was about 50 pg per mL blood, but was very variable. Some DNA samples were also extracted from kidney or liver by homogeniza- tion in a buffer containing 50 mM Tris (pH 7.5) and 50 mM EDTA using a tissue grinder. SDS was then added to 1% w/v and DNA extraction was done as described above.

DNA fingerprinting

A number of restriction enzymes were tested in combi- nation with the subsequent hybridization of the finger- print blots with several simple sequence oligonucleotide probes, as well as the human minisatellite probes 33.15 and 33.6 (Jeffreys 1985). Scorable band patterns were ob- tained with both minisatellite probes and the oligonucle- otide probes (ATCC), and (CCA), when using the restric- tion enzyme HinfI. The DNA was digested with this enzyme in a buffer as recommended by the suppliers, supplemented with spermidine (Sigma) to a final concen- tration of 2 m. About 3 units of enzyme/pg DNA were used and the digestion was incubated at 37 "C overnight. These extensive digestion conditions were necessary to achieve complete digestion. We have observed at least two bands with the (ATCC), probe which disappeared only after prolonged digestion and which could easily be mistaken as true polymorphic alleles. The DNA digest (2-4 pg per individual) was then separated on a 0.8% agarose gel (20 cm long, run at 40 V for 36 h at 8 "C). The gels were dried and hybridized with the simple sequence oligonucleotide (ATCC), at 43 "C or the oligonucleotide (CCA), at 45 "C and then washed in 5 x SSPE at hybridi- zation temperature. Alternatively, the gels were blotted, hybridized with the minisatellite probes at 60 "C and washed in 1 x SSPE/O.l%SDS. Since hybridization with these probes resulted in a very simple band pattern, in- volving only very few and mostly invariable bands with each probe, identification of the same bands in different lanes of each gel/blot was unequivocal. Comparison be- tween gels was possible by running a standard marmot DNA digest on either side of the gel.

Results

Oligonucleotide fingerprinting in the main study population

The two simple sequence oligonucleotide probes (ATCC), and (CCA), together yielded 20 different restric- tion fragments in the resolvable size fraction between 3 kb and 23 kb (Fig. 1). Fourteen of these bands were shared between all individuals in the population, whereas only

Page 3: Low genetic variability in a natural alpine marmot population (Marmota marmota, Sciuridae) revealed by DNA fingerprinting

LOW GENETIC VARIABILITY IN MARMOTS 349

Fig. 1 Example for the hybridization pattern of the oligonucle- otide probes (CCA),and (ATCC), tohlinfl-digested DNA from one marmot family. Eight bands were normally scored for the (CCA), probe, twelve for the (ATCC), probe (marked with bars). The bands which were found to be variable in the population are marked with arrowheads. The upper two polymorphic bands with the (CCA),probe behave as alleles as judged from their Men- delian segregation behaviour. The second upper band in the (ATCC), fingerprint is monomorphic in the family shown, but is variable in the population. Note that the lower portion of the gel was frequently not very well resolved and was therefore omitted from the population analysis. A marker scale is shown to the right. Lanes 1 and 2 dominant male and female during 1984 until 1987; lanes 3-5: offspring in these years; lanes 6-8: offspring after the female from lane 4 became the new dominant female in 1988.

six showed polyrnorphisms and could be scored as present or absent in the blots. Four of these variable bands were apparently length alleles of different loci while the remaining two polymorphic bands represented a single diallelic locus ((CCA), probe; upper two bands in Fig. 1). This could be shown by the segregation analysis of these alleles in the 31 marmot families in the Berchtesgaden population which included 181 analysed parent/off- spring groups. The genotypes found for this diallelic lo- cus are within the range of Hardy-Weinberg expectation (Table 1).

Other multilocus fingerprint probes yielded a similar low degree of polymorphism. Figure 2 shows an example of the patterns obtained for the human minisatellite probes 33.15 and 33.6 on blots of unrelated animals from the Berchtesgaden population. These revealed an even lower degree of polymorphism. Only two out of 27 differ- ent restriction fragments in the resolvable size fraction between 3 and 23 kb were variable (Fig. 2).

A higher degree of polymorphism was found when comparing individuals from different regions of the Alps. Using the oligonucleotide probes, these fingerprints showed a number of bands different from those found in

Table 1 Frequencies of the variable bands detected with the (ATCC), and (CCA), oligonucleotide probes in the Jenner and the Funtensee population. The (ATCC), bands ATCCl to ATCC3 cor- respond to the three marked in Fig. 1 and are considered to be in- dependent loci (non-allelic behaviour was evident from the family analysis). The band frequency for the lower (CCA), band is not listed, since it was not clearly resolvable in all individuals tested. The upper two (CCA),bands are allelic (compare Fig. 1). The fre- quencies of the (CCA), genotypes do not deviate from those ex- pected for a population in Hardy-Weinberg equilibrium (Jenner population: G, = 2.54, P > 0.1; Funtensee population: G, = 3.06, P > 0.05). The statistical test for the homogeneity between both sub- populations is not included because of the small sample size of the Funtensee population, though the values found are roughly in line with the values one would expect under the assumption of homo- geneity between the two populations. Abbreviations: N = sample size, b = band frequency, a = allele frequency, g = genotype fre- quency

Jenner Fun tensee

(ATCC), b N bandl 0.68 230 band2 0.97 230 band3 0.78 197

lower(1) 0.53 231 upper(u) 0.47 231

11 0.31 72 pL 0.45 104 uu 0.24 55

(CCA), n N

g N

b N 0.65 23 0.74 23 0.87 23 a N 0.41 23 0.59 -23

0.09 2 0.65 15 0.26 6

g N

Fig. 2 Example for the hybridization pattern of the minisatellite probes 33.15 and 33.6 to HinfI-digested DNA from three unrelated individuals of the Jenner population. This pattern is representa- tive for the one found in further seven unrelated individuals. In combination, the two minisatellite probes detected 27 resolvable bands (marked with bars), only two of which were variable for the 33.15 probe (labelled with arrowheads) and none for the 33.6 probe.

Page 4: Low genetic variability in a natural alpine marmot population (Marmota marmota, Sciuridae) revealed by DNA fingerprinting

350 K. RASSh4ANN et al.

Fig. 3 Example for the hybridization pattern of the oligonucle- otide probes (CCA),and (ATCC), to Hinfl-digested DNA of indi- viduals from different regions of the Alps. Lanes 1-5 show indivi- duals from three different areas in Austria, lanes 6-11 individuals from two different areas in Switzerland, lane 12 an individual from the study area in Berchtesgaden for comparison. The 20 re- solvable bands for this individual are marked (cf. Fig. 1). Arrow- heads denote those of these 20 bands which are shared among all Berchtesgaden and all other alpine marmots tested.

the Berchtesgaden population (Fig. 3). In addition, all of the 20 bands that were revealed in the fingerprints of the Berchtesgaden population could be seen in this blot as well (Fig. 3). Twelve of these 20 bands were now seen to be variable, including the six that were polymorphic in the Berchtesgaden population. Thus, we conclude that at least the major fraction of the bands revealed by the oligo- nucleotide fingerprints of the Berchtesgaden marmot population seems to represent potentially hypervariable loci.

These results suggest that the low degree of polymor- phism found in these animals is not due to molecular pe- culiarities of the loci analysed. Instead, it appears evident that population genetic parameters must be analysed to explain this low degree of polymorphism.

Population size

Field observations, as well as our molecular data allow to judge on the migration behaviour of the marmots in the study area. This in turn gives us a rough estimate of the current size of the interbreeding population in the Berchtesgaden Alps. To get an idea on the extent of gene flow within this area, we have analysed animals from another area in the Berchtesgaden Alps (Funtensee). These animals live about 13 km away, but are separated from the Jenner animals by several uninhabitable rocky

stretches that lie between these two areas. Nonetheless, the fingerprints of this population revealed the same bands as seen in the Jenner population, even showing similar frequencies for the variable bands (Table 2). This suggests that there is a continuous gene flow between these areas which is sufficiently high to maintain genetic similarity between these two subpopulations. Recent studies using radiotelemetry have provided further evi- dence that dispersing individuals migrate much further than it was previously assumed (Arnold 1993). Therefore, we have to expect that marmot groups throughout the Berchtesgaden area are in frequent genetic exchange, whereas they are probably more isolated from the other regions of the Alps. A systematic census of all populated marmot burrows has been performed for approximately one-fifth of the potential marmot habitat in the Berchtes- gaden Alps. On the basis of these data, we estimate the interbreeding population to be in the order of at least 500 groups, i.e. breeding pairs, which is equivalent to a mini- mum effective population size of 1000. A potential par- ticipation of subadult males in the reproduction would increase this value, yet the actual contribution is difficult to assess and shall therefore be omitted in the following calculation in favour of a more conservative estimate of the effective population size.

Observed heterozygosity

The observed degree of average heterozygosity (H) can be estimated from multilocus fingerprint data in the follow- ing way: allele frequencies (a) can be calculated from the band frequencies @) of variable bands by a = 1 - d(1- b).

These can then be used to determine the approximate heterozygosity values for each locus by calculating

The sum of all heterozygosity values divided by the number of loci (L) analysed then renders an estimate of the average heterozygosity

H = Xb - Z[1- d(1- b)?] /L .

H, = b - a2 .

An estimate of the number of loci is crucial for this type of analysis. However, it can easily be obtained from highly invariant multilocus fingerprints such as ours, since allelic bands are obvious and all other resolvable bands can be assumed to belong to different loci. This es- timate is more critical when the fingerprint patterns are much more variable. Since allelic relationships between the bands are difficult to determine in these cases, L must be estimated from the frequencies of the bands (Stephens ef al. 1992). However, this can lead to an underestimate of L (and therefore an overestimate of H ) , since alleles that are running below the resolvable size region are ne- glected. In our case, we prefer therefore the direct count- ing method.

Page 5: Low genetic variability in a natural alpine marmot population (Marmota marmota, Sciuridae) revealed by DNA fingerprinting

LOW GENETIC VARIABILITY IN MARMOTS 351

The following calculaticrns are based on the data from the oligonucleotide probes. These showed 19 resolvable restriction fragments which we can treat as independent loci. Fourteen of these were monomorphic and were therefore scored with an allele frequency of 1 for each. The allele frequencies for the three variable loci detected with the (ATCC), probe (ATCC,, ATCC,, ATCC,; see Ta- ble 1) were calculated from their band frequencies. The respective heterozygosity values were found to be 0.49, 0.28 and 0.5. The small variable fragment in the (CCA), fingerprint was not clearly resolvable in all animals tested and the band frequency could not be determined un- equivocally. We assume therefore the maximum possible heterozygosity of 0.5 for this locus. The heterozygosity of the diallelic (CCA), locus (0.47) was directly calculated from the observed genotype frequency (Table 1). The av- erage heterozygosity for the oligonucleotide fingerprint loci in the Berchtesgaden marmot population was thus found to be H = 2.24/19 = 0.12.

Mutation rate

The average mutation rate (p) for a human minisatellite locus is approximately 0.004 per gamete, though the val- ues vary considerably between different loci (Jeffreys et al. 1985; Jeffreys et al. 1988). A slightly lower rate, about 0.001 per gamete, was calculated for human simple se- quence loci which are recognized by the (CCA), probe (Nurnberg rt nl. 1989). In our own data we have seen one apparent mutational event which may be used to estimate a mutation rate. Among all fingerprints of the Jenner population we find only one individual with a novel re- striction fragment. The (ATCC), fingerprint of this indi- vidual shows a band of approximately 4.2 kb which is not found in its parents nor in any other marmot of the Berchtesgaden area. Thus we presume that it is due to a mutational event and exclude migration as a possible source for this allele. All of the (ATCC), fingerprint bands are present in the respective individual, except of the polymorphic ATCC, band. To obtain a mutation rate for this locus we assume that the novel band is derived from this band. The allele frequency for this band is 0.43 (Ta- ble 1) and a total of 253 animals were analysed with re- spect to this band. This means that among all gametes. tested, 218 could have carried this allele. Since we found one mutation in this sample size, we arrive at an estimate of the mutation rate for the ATCC, locus of p = 0.005. On the other hand, it is clear that this cannot be taken as an average rate for all loci, since in this case we should have observed far more mutational events. A minimum aver- age mutation rate for the ATCC, loci can be calculated as follows: the band pattern of the whole population is so invariant that any new band found should be the result of a mutational event. About 250 animals i.e. 500 gametes

were analysed. Since each locus could have contributed a new allele independently, we have to multiply this by 12 and arrive at an average mutation rate of p = 1/6000 = 0.00017. This represents a minimum estimate, since it dis- regards all new alleles that would have run below the separation range or that would have comigrated with other bands in the gel.

Expected average heterozygosity and efective population size

We can now use the estimated values for the observed effective population size (N,) and the average mutation rate in a formula, which links these variables with the ex- pected average heterozygosity in Hardy-Weinberg populations (Kimura & Crow 1964). However, the results should be seen as rough estimates, since this formula is limited to populations under equilibrium, a condition that is generally difficult to determine in natural popula- tions. The formula is given as follows:

H = 4Nep/(l + 4Ne;t).

Accordingly, the average heterozygosities for the maximal and minimal values of p that are expected in a population with N , = 1000 is 0.95 for p = 0.005 and 0.41 fn: 11 = 0.00017. Thus, even the most conservative esti- mate is higher than the observed value of H = 0.12 (see above). We believe therefore that the genetically effective population size of the Berchtesgaden marmot population is much smaller than the currently observed number of breeding pairs. Keeping in mind the limitations of the formula, we can roughly estimate the actual effective population size by solving N, = H/[4p(1- H)]. Under the premise that H = 0.12 we find an N, of 7 for p = 0.005 and N, of 200 for 11 = 0.00017. Given that the two estimates of p are extremes, we can conclude that the genetically effec- tive population size is apparently an order of magnitude lower than the one currently observed.

Discussion

The low degree of genetic variability found in the Berch- tesgaden marmot population could be due to either a low mutation rate of the oligonucleotide fingerprint loci stud- ied, or to a small effective size of the Berchtesgaden mar- mot population. Yet another explanation for a lack of ge- netic heterogeneity has been suggested in another social animal, the naked mole rat (Reeve et 01. 1990). Inbreeding can be the result of a small population size, but it can also represent a reproductive strategy of a species, e.g. when matings occur preferably within the same family group. In marmots, short distance dispersal or replacement of a dominant animal by an offspring is known to occur (Arnold 1990a,b) and would favour inbreeding. Also, it

Page 6: Low genetic variability in a natural alpine marmot population (Marmota marmota, Sciuridae) revealed by DNA fingerprinting

352 K. RASSMANN et al.

has been found that the sons of territorial females fre- quently mate with their own mother. However, restricted gene flow within a population promotes the genetic dif- ferentiation between its subgroups and therefore should even increase the genetic variance of this population (Chesser & Ryman 1986). In this discussion we will there- fore focus on the two explanations mentioned first.

One of these potential causes, the low mutation rate, seems a rather unlikely explanation in our case. First, the analysis of individuals from more distant arias of the Alps has shown the expected higher degree of variability of the loci studied. Second, independent probes such as the minisatellite probes showed also a very low variabil- ity. In addition, we have meanwhile obtained data on the distribution of alleles of polymorphic simple sequence loci and found that they too are extremely monomorphic in the marmots of the Berchtesgaden area (Klinkicht 1993). It seems therefore much more likely that the low degree of genetic heterogeneity of the marmots in the Berchtesgaden Alps is due to a small effective population size.

On the other hand, our estimate of the current size of the interbreeding marmot population in this area is an order of magnitude higher then the one that we would have expected to explain the low degree of heterozygos- ity. We have to conclude therefore that our population has experienced a drastic loss of breeding pairs in the past (bottleneck), or represents a founder population. Our data do not allow us to distinguish between these two possibilities. Though the Berchtesgaden population is known to have existed continuously during recent centu- ries, one could imagine that a colonisation with a few founder individuals has occurred within the last millen- nium, which could explain the observed pattern. A more likely possibility is that the population experiences fre- quently bottlenecks. The most likely cause for severe bot- tlenecks are harsh winters. Death during hibernation is the major reason for mortality in alpine marmots. Harsh winters can cause the extinction of whole family groups, especially when pairs and infants hibernate without the presence of subadults (Arnold 1990b). The most severe winter during the field study in Berchtesgaden was ob- served in 1989/90, which resulted in a loss of 18% of all reproducing animals in the main study population as compared to an average winter mortality of 8%. Histori- cal records show that winter temperatures can be much lower than in 1989/90. Therefore, we believe that high losses due to winter mortality could be a reason for re- peated bottlenecks.

The low levels of genetic heterogeneity could poten- tially have adverse effects on the viability of the popula- tion for two reasons. First, it leads inevitably to inbreed- ing between genetically closely related individuals and thus to a high probability of creating offspring with

homozygous lethal or sublethal alleles. Interestingly, as mentioned above inbreeding occurs also as a mating strategy in these animals. This may already have led to a selection against homozygous lethal or sublethal alleles. Secondly, lack of genetic variability might render a popu- lation more susceptible to changes in the environment. This could affect the long-term viability of the popula- tion. Unfortunately, we are not in a position to judge whether this is indeed the case. Our data and observa- tions deal only with a small time window of the popula- tion history and thus allow no long-term predictions. A possible way to get more information OR the long term viability of the Berchtesgaden population would be com- parative studies on similar marmot populations in other regions of the Alps with a higher degree of genetic vari- ability. These could then be used for comparative studies on winter mortality, disease resistance or parasite load, which could be predictive parameters for long term vi- ability. Such a study might eventually lead to more con- clusive insights into the effects of low genetic variability on natural populations.

Acknowledgements

We thank Agnes Turk for expert field assistance, Gabi Buttner for assistance in the laboratory and the National Park administration of Berchtesgaden for permission to work and for accommodation. We are particularly grate- ful to Monika Preleuthner for providing tissue samples from marmots from other areas in the Alps, and to Jorg T. Epplen, Lutz Roewer and Hans Zischler for advising us on the use of simple sequence oligonucleotide probes. We wish to thank Josephine Pemberton and Hans Siegis- mund for the critical reading of some earlier versions of this manuscript and Pim Arntzen, Iris van Pijlen and Esther Signer for very fruitful and happy discussions. This study was financed by the Max-Planck-Gesellschaft and grant AR180/1-1 of the Deutsche Forschungsgemein- schaft to W. Arnold and D. Tautz.

References Ali S, Miiller CR, Epplen JT (1987) DNA-fingerprinting by oligo-

nucleotide probes specific for simple repeats. Human Genetics, 74,239-243.

Arnold W (1990a) The evolution of marmot sociality: I. Why dis- perse late? Behavioural Ecology and Sociobiology, 27,229-237.

Arnold W (1990b) The evolution of marmot sociality: 11. Costs and benefits of joint hibernation. Behaviournl Ecology and So- ciobiology, 27,239-246.

Arnold W (1993) Social evolution in marmots and the adaptive value of joint hibemation. Verhandungen der Deutschen Zoologischen Gesellschaft, 86,79-93.

Burke T, Dolf G, Jeffreys AJ, Wolff R (eds) (1991) D N A Fingerprint- ing: Approaches and Applications. Birkhauser Verlag, Base1

Chesser RK, Ryman N (1986) inbreeding as a strategy in subdi-

Page 7: Low genetic variability in a natural alpine marmot population (Marmota marmota, Sciuridae) revealed by DNA fingerprinting

LOW GENETIC VARIABILITY IN MARMOTS 353

vided populations. Evolirtioii, 40,616-624. Faulkes CG, Abbott DH, Mellor A t (1990) Investigation of genetic

diversity in wild colonies of naked mole-rats (Helt.rocrphnliis glabrr) by DNA-fingerprinting. /ourid of Zoology (London), 221,

Ziir Okologie und Vrrbreitungsgesclricllfe des Alpennrurnteltieres im Berner Oberland. Dissertation, Universita t Bern.

Gilbert DA, Lehmann N, O'Brien SJ, Wayne RK (1990) Genetic fin- gerprinting reflects population differentiation in the Califor- nia channel island fox. Nature, 344, 764-767.

Jeffreys AJ, Wilson V, Thein SL (1985) Individual-specific finger- prints of human DNA. Nature, 316,76-79.

Jeffreys AJ, Royle NJ, Wilson V, Wong Z (1988) Spontaneous mu- tation rates to new length alleles at tandem-repetitive hypervariable loci in human DNA. Nature, 332,278-281.

Kimura M, Crow JF (1964) The number of alleles that can be main- tained in a finite population. Genetics, 49,725-738.

Klinkicht M (1993) Untersuchiingen zcim Pnnrriiigssysfem des Alpenmurmeltiers, Marmota m. marmota (L ime , 1758) mittels DNA-Fingerprinting. PhD thesis, University of Munich.

Numberg P, Roewer L, Neitzel Het nl. (1989) DNA fingerprinting with the oligonucleotide probe (CAC),/(GTG),: somatic stabil- ity and germline mutations. Human Genetics, 84,7578.

Reeve HK, Westneat DF, Noon WA, Sherman PW, Aquadro CF (1990) DNA 'fingerprinting' reveals high levels of inbreeding in colonies of the eusocial naked mole-rat. Proceedings of the Nntional Academy of Sciences of the USA, 87,2496-2500.

87-97. Forter D (1975)

soul6 M (ed.) (1987) vhb/C P U \ 7 l / / ~ 7 t ~ l l l l S ~ J ~ ~ ~ J J l W r Z V ~ ~ i U J J . Cambridge University Press, Cambridge.

Stephens JC, Gilbert DA, Yukhi N, O'Brien SJ (1992) Estimation of heterozygosity for single probe multilocus DNA fingerprints. Molrcrilar Biology nrld E i ~ ~ l ~ i t i ~ r ~ , 9, 729-743.

Turner BJ, Elder JF, Laughlin TF, Davis WP (1990) Genetic varia- tion in clonal vertebrates detected by simple-sequence DNA- fingerprinting. Procrediifgs o f t h Nntiorlol Acniferiry of Sciuiccps of the USA, 67,5653-5657.

Zimina RP, Gerasimov IP (1973) The periglacial expansion of mar- mots (Marnrotn) in middle Europe during late Pleistocene. joro-. ~ i n l of Mnnrnialogj, 54,327-340.

The work presented in this paper is part of a collaboration between Walter Arnold (formerly Max-Planck Institut f u r Verhaltensphysiologie in Seewiesen, now Zoologisches Insitut der Universitat Marburg) and Diethard Tautz (Zoologisches Institut der Universitat Miinchen). The data were obtained by Kornelia Rassmann during the work for her Diploma thesis. Walter Arnold has studied the Berchtesgaden marmot population for ten years and provided most of the blood samples from the Jenner popula- tion. Kornelia Rassmann collected the blood samples from the Funtensee population with the help of AgnesTiirk. The workon the paternity analysis has meanwhile been continued by Markus Klinkicht in the laboratory of Diethard Tautz.