Changes in chromosomal polymorphism and global warming: The case of Drosophila subobscura from Apatin (Serbia)

In this study, chromosomal inversion polymorphism data for a natural population of Drosophila subobscura from a swampy region near the town of Apatin (Serbia) were compared with data for the same population collected approximately 15 years earlier. The pattern of chromosomal inversion polymorphism changed over time. There were significant increases in the frequency of characteristic southern latitude (“warm” adapted) chromosomal arrangements and significant decreases in the frequency of characteristic northern latitude (“cold” adapted) chromosomal arrangements in the O and U chromosomes. The chromosomal arrangements O3+4 and O3+4+22 (derived from the O3+4 arrangement) showed significant increases in 2008 and 2009 with regard to the 1994 sample. There was also a significant increase (∼50%) in the U1+2 arrangement, while U1+8+2 (a typical southern arrangement) was detected for the first time. Since the Apatin swampy population of D. subobscura has existed for a long time in a stable habitat with high humidity that has not been changed by man our results indicate that natural selection has produced chromosomal changes in response to the increase in temperature that has occurred in the Balkan Peninsula of central southeastern European.


Introduction
The climate change of recent decades has led to heritable genetic changes in animal species as diverse as birds, squirrels and mosquitoes (Bradshaw and Holzapfel, 2006). Long-term changes in chromosomal inversion polymorphisms in Drosophila species suggest that the genetic constitution of populations is responding to climate change (Van Heerwaarden and Hoffmann, 2007). This suggestion is strongly supported by the fact that genes located within inversions are associated with a variety of traits including those involved in climate adaptation (Hoffmann and Rieseberg, 2008). This characteristic was observed in some species of Drosophila, such as D. melanogaster in Australia (Anderson et al., 2005;Umina et al., 2005), D. robusta in North America (Levitan and Etges, 2005) and D. subobscura in Europe (Rodríguez-Trelles and Rodríguez, 1998;Solé et al., 2002;Balanyà et al., 2004Balanyà et al., , 2006Balanyà et al., , 2009 and America (Balanyà et al., 2006(Balanyà et al., , 2009. In all D. subobscura populations from southwestern and central Europe studied to date an increase in the fre-quency of "warm"-adapted chromosomal arrangements typical of southern latitudes, and a decrease in the frequency of "cold"-adapted chromosomal inversions typical of northern latitudes, has been observed (Rodríguez-Trelles and Rodríguez, 1998;Solé et al., 2002;Balanyà et al., 2004Balanyà et al., , 2006Balanyà et al., , 2009. This species is an excellent model organism for such studies because of its extensive polymorphism for chromosomal inversions: 67 inversions distributed in 93 chromosomal arrangements have been identified, with the frequencies of most of these varying clinally with latitude (Krimbas and Loukas, 1980;Prevosti et al., 1988;Menozzi and Krimbas, 1992;Krimbas, 1993).
Of the five acrocentric chromosomes in D. subobscura, the O chromosome (homologous to the 3R chromosomal arm of D. melanogaster and second chromosome of D. pseudoobscura) is the longest and most polymorphic, with about 40 natural chromosomal arrangements (Krimbas, 1993). Since the 1960s more than 150 natural populations have been studied throughout the Palearctic distribution of this species. However, few long-term analyses of changes in inversion polymorphism have been done in southeastern Europe or the Balkan Peninsula. For this reason, the main aim of this work was to study a population of D. subobscura from Apatin (Serbia). This population is located in a swampy region on the left bank of the Danube river in the lowlands of the Pannonian plain. This area, which is midway along the Danube river about 1366-1433 km from the mouth of this river in the Black sea, is deliberately left to be flooded permanently or during spring across an area of at least 3500 ha. Consequently, the Apatin D. subobscura population has existed in a very humid habitat that has never been changed by man. This is probably the main reason for the large differences observed between the Apatin and other D. subobscura populations from the Balkan Peninsula (Zivanovic et al., 2002;Zivanovic and Marinkovic, 2003;Zivanovic, 2007;Zivanovic and Mestres, 2010a,b).
Specifically, we analyzed inversion polymorphisms in D. subobscura from Apatin 14 and 15 years after the initial studies, and used this information to assess variation in composition. Our intention was to compare samples from June 2008 and 2009 with the sample from June 1994 at exactly the same site, paying special attention to calendar and climatology data. The results may be especially interesting because during this period there were several heat waves, for example the first evidence of climate change in the region was recorded in summer 2003 (Schar and Jendritzky, 2004). The study should tell us whether long-term chromosomal inversion polymorphism changes, if they exist, are characteristic of all or only of specific chromosomes and their arrangements. Finally, the analysis of possible changes in frequency of some chromosomes in the O 3+4 group (O 3+4+1 , O 3+4+22 and so on) correlated to climate changes is of particular interest, because some genes involved in thermal adaptation are located in the O 3+4 region (Quintana and Prevosti, 1991;Moltó et al., 1992;Laayouni et al., 2007).

Material and Methods
Drosophila subobscura flies were collected from a poplar (Populus alba) wood in a swampy region located on the left bank of the river Danube, near the town of Apatin (45°40' N, 19°00' E, approximately 200 km NW of Belgrade). This region is included in the Ramsar List, which lists wetlands of international importance, and is known as the Upper Danube wetland area. Meteorological data for the site were obtained from the Serbian Republic Hydrometeorological Service. Samples were collected twice from exactly the same place in June 2008 and June 2009. To allow comparison with the June 1994 sample, the 2008 and 2009 samples were collected 2.5 days earlier per decade because spring/summer has advanced an average of 2.5 days per decade in Europe (Menzel et al., 2006). Unfortunately, the need to trap on precise days and Danube river flooding limited the sample size obtained. Only wild males (June 2008, June 2009 and the sons of wild females (June 2009) were used to assess chromosomal inversion polymorphism. The males were crossed individually with virgin females of the Kussnacht strain that were homokaryotypic for standard chromosomal arrangements in all five chromosomes.
Polytene chromosomes were stained and squashed in aceto-orcein solution. At least eight larvae from the progeny of each cross were examined. The chromosomal map of Kunze-Mühl and Müller (1958) was used for cytological analysis of the chromosomal arrangements and the nomenclature for these arrangements followed that of Kunze-Mühl and Sperlich (1955). Departure of the observed frequencies of chromosomal arrangements and chromosomal karyotypes from expectations was tested using c 2 test contingency tables. The degree of chromosomal inversion polymorphism in the population analyzed was assessed using the index of free recombination (IFR) (Carson, 1955). Chromosomal inversion polymorphism patterns were obtained by computing Nei's genetic distances (Nei, 1972) between Apatin 2008+2009 and Apatin 1994 and another five previously analyzed Balkan populations (Kamariste, June 1996; Djerdap, June 2001; Petnica, June 1995; Jastrebac, June 1993; Zanjic, June 1997) (Zivanovic et al., 2002;Zivanovic and Marinkovic, 2003;Zivanovic, 2007).

Results
The frequencies of chromosomal arrangements in the Apatin population from June 2008 and June 2009, together with those previously obtained for this population in June 1994 (Zivanovic et al., 2002), are shown in Table 1. A total of 18 chromosomal arrangements were found in the old and new samples. Some of the chromosomal arrangements found earlier (U 1 , E 1+2 , O 6 , O 3+4+2 , in 1994) were not observed in 2008 and 2009, whereas some chromosomal arrangements, such as U 1+8+2 , E 1+2+9+12 , O 22 and O 3+4+22 that were not found 15 years earlier were observed in the 2009 samples (Table 1). There were significant differences in the frequencies of the U chromosomal arrangements between the June 1994 and June 2009 samples (c 2 = 21.2, d.f. = 4, p < 0.001). Similarly, the 1994 and 2008+2009 samples also differed significantly in the frequencies of five U chromosomal arrangements (c 2 = 17.4, d.f. = 4, p < 0.01). There was a significant decrease in the frequency of the U st arrangement in the two comparisons (c 2 = 9.5, d.f. = 1, p < 0.01 and c 2 = 6.9, d.f. = 1, p < 0.01, for the first and second comparison, respectively). Over the same period of time, U 1+2 showed a significant increase in frequency (c 2 = 8.0, d.f. = 1, p < 0.01, for the first case and c 2 = 8. The variations in chromosomal polymorphisms were compared with the meteorological data shown in Table 3. The combined average for the maximum, minimum and mean temperatures (26.4°C, 14.4°C and 20.6°C, respectively) for 2008 and 2009 were higher than in 1994. A year-by-year analysis from 1994 onwards revealed several heat waves such as the well documented heatwave of 2003 492 Zivanovic and Mestres  (Schar and Jendritzky, 2004) and less drastic ones in 2000, 2002 and 2007. These data suggest a gradual increase in temperature at the locality studied. However, desiccation does not seem to affect the Apatin habitat: although the rainfall pattern is irregular, a general decrease has not been observed. Indeed, data from 2006 onwards suggest that there has been a tendency for rainfall to increase, in agreement with global warming predictions (Houghton, 2005;Wentz et al., 2007).
The IFR values obtained for the Apatin population were: 75.60 ± 1.88 (June 2008) and 77.13 ± 2.11 (2009), with a mean of 76.75 ± 1.63. These values were only slightly lower than that for 1994 (84.62 ± 1.55). Table 4 shows Nei's genetic distances between the Apatin and other Balkan populations. This table also provides older genetic distances between Apatin (1994) and these Balkan populations (Zivanovic et al., 2002;Zivanovic, 2007). The new genetic distances were lower than those obtained previously.

Discussion
The world's climate is changing rapidly, with a global temperature increase of 0.6°C in the past three decades, and 0.8°C in the past century (Hansen et al., 2006;Van Heerwaarden and Hoffmann, 2007). Many animals and plants adapt rapidly to variations in environmental conditions (Endler, 1986). In particular, widespread generalist species such as D. melanogaster and D. subobscura, which have short generation times and rapidly form phenotypic and genotypic clines, are excellent sensitive indicators of such global warming (Van Heerwaarden and Hoffmann, 2007;Balanyà et al., 2009). The results of long-term studies indicate that there have been changes in the frequency of chromosomal arrangements in D. subobscura. In recent decades, a significant decrease in the frequency of standard chromosomes ("cold" adapted and abundant in northern latitudes) and an increase in some non-standard chromosomal arrangements ("warm" adapted and characteristic of southern latitudes) have been described in natural D. subobscura populations from southwestern and central Europe (de Frutos and Prevosti, 1984;Gosteli, 1990;Orengo and Prevosti, 1996;Rodriguez-Trelles et al., 1996;Solé et al., 2002;Balanyà et al., 2004Balanyà et al., , 2006Balanyà et al., , 2009 and in American colonizing populations (Balanyà et al., 2006(Balanyà et al., , 2009). These changes have been attributed to the adaptation of chromosomal inversion polymorphism to a warmer climate in Europe and America.
The effect of global warming in the Apatin population is reflected as variation in chromosomal polymorphism. Specifically, there was a significant increase in the frequency of some chromosomal arrangements characteristic of southern latitudes ("warm" adapted) and a significant decrease in the frequency of some standard inversions ("cold" adapted) in the O and U chromosomes. Thus, O 3+4 showed a significant increase when considered together with O 3+4+x arrangements (~40% on average) in 2008 and 2009 compared to 1994. Similar results have been reported for other southwestern European populations (Orengo and Prevosti, 1996;Rodríguez-Trelles and Rodríguez, 1998). Interestingly, associations between heat-tolerance and the O 3+4 arrangement have been described previously for D.
Chromosomal polymorphism changes and global warming 493 Max. T, Min. T and Mean T stand for maximum, minimum and mean temperatures, respectively. subobscura (Quintana and Prevosti, 1991). Furthermore, several genes that code for heat stress proteins (Moltó et al., 1992) and others involved in thermal adaptation (Laayouni et al., 2007) vanovic et al., 2002;Zivanovic and Marinkovic, 2003;Zivanovic, 2007;Zivanovic and Mestres, 2010a,b). In contrast, the frequencies of the O 3+4 and O 3+4+x arrangements in the new Apatin population (2008 + 2009 sample) were no much lower (by 19% and 28%, respectively) than those previously reported for the southern Balkan populations. Data for the O-chromosome suggest that the Apatin population is evolving towards a more "southern" population. Similarly, for the U chromosome, a significant increase in the U 1+2 frequency (~50%) during the period studied, as well as the presence of the U 1+8+2 arrangement which was not previously observed, indicate evolution towards a "southern" population.
Our results agree with those from southwestern and central European populations (Rodríguez-Trelles and Rodríguez, 1998;Solé et al., 2002;Balanyà et al., 2004Balanyà et al., , 2006Balanyà et al., , 2009. The relatively low IFR values obtained in this study (not previously observed in 1994) are characteristic of ecologically central European populations of D. subobscura (Krimbas, 1993). This result, and the smaller genetic distances between the new Apatin population (2008+2009) and all other Balkan populations (Table 4) compared to the distances previously found, provide further evidence that the Apatin population is becoming more "southern". A similar decrease in genetic distances over time has been observed in Mediterranean, Atlantic and Central European populations of D. subobscura (Balanyà et al., 2004).
The absence of frequency changes for chromosomal polymorphism in the remaining chromosomes (A, J and E) suggests that our results were most probably not a consequence of gene flow from the south, where the frequencies of all the arrangements in these three chromosomes are completely different in all southern D. subobscura populations analyzed (Zivanovic et al., 2002;Zivanovic and Marinkovic, 2003;Zivanovic, 2007;Zivanovic and Mestres, 2010a,b). Together, these observations suggest that we have detected the effect of selection on chromosomal polymorphism composition (see also the discussion of this topic in Hoffman and Rieseberg, 2008). Some form of natural selection appears to be acting as a reaction to the increase in temperature associated with many heat waves that have markedly affected the Balkan Peninsula; random processes and genetic drift do not appear to be involved. The latter conclusion is supported by the high migration rate among D. subobscura populations from the Balkan Peninsula (Zi-vanovic et al., 2007) since migration among populations efficiently purges most effects of genetic drift.