(i) A genetically unique species

D. ananassae occupies a unique status among Drosophila species due to certain peculiarities in its genetic behaviour (Singh, Reference Singh1985a, Reference Singh2000). It exists in highly structured populations in Asia and South Pacific (Johnson, Reference Johnson1971; Stephan, Reference Stephan1989; Stephan & Langely, Reference Stephan and Langley1989; Tomimura et al., Reference Tomimura, Matsuda, Tobari, Cariou, Da Lage, Stephan, Langley and Tobari1993; Stephan et al., Reference Stephan, Xing, Kirby and Braverman1998; Vogl et al., Reference Vogl, Das, Beaumont, Mohanty and Stephan2003; Das et al., Reference Das, Mohanty and Stephan2004; Schug et al., Reference Schug, Smith, Tozier-Pearce and McEvey2007, Reference Schug, Baines, Amanda, Mohanty, Das, Grath, Smith, Zhargam, McEvey and Stephan2008) and its biogeographical history is well characterized. Other peculiar characteristics are the existence of spontaneous crossing-over in males, which is meiotic in origin (Kikkawa, Reference Kikkawa1937; Moriwaki, Reference Moriwaki1937, Reference Moriwaki1940; Moriwaki et al., Reference Moriwaki, Tobari and Oguma1970; Moriwaki & Tobari, Reference Moriwaki and Tobari1973; Matsuda et al., Reference Matsuda, Imai and Tobari1983; Kale, Reference Kale1969; Hinton, Reference Hinton1970; Singh & Singh, Reference Singh and Singh1988); presence of chromosome rearrangements, such as, pericentric inversions, translocations, transpositions, deficiencies and extrabands, reflecting high mutability in D. ananassae (Kikkawa, Reference Kikkawa1938). D. ananassae harbours a large number of chromosome rearrangements, 78 paracentric inversions, 21 pericentric inversions and 48 translocations in its natural populations (Singh & Singh, Reference Singh and Singh2007b). Most of these paracentric inversions have restricted distribution, while three cosmopolitan inversions (Futch, Reference Futch1966), namely, alpha (AL) in 2L, delta (DE) in 3L and eta (ET) in 3R show worldwide distribution (Singh, Reference Singh1996). The same inversions were given different names by other investigators. In the present paper, nomenclature by Ray-Chaudhuri & Jha (Reference Ray-Chaudhuri and Jha1966) as, alpha (AL), delta (DE) and eta (ET) will be followed. The optic morphology (Om), hyper-mutability system (Hinton, Reference Hinton1984; Matsubayashi et al., Reference Matsubayashi, Tobari and Hori1991; Awasaki et al., Reference Awasaki, Juni and Yoshida1996); ZAM, a retrovirus-like element has also been reported in D. ananassae (Baldrich et al., Reference Baldrich, Dimitri, Desset, Leblanc, Codipietro and Vaury1997); spontaneous bilateral genetic mosaic, which was characterized by three mutant characters (cu, e, se) on the left side and all normal characters on the right side was detected in D. ananassae; parthenogenesis has been reported in the light and dark forms of D. ananassae by Futch (Reference Futch1972) ; segregation distortion; Y-4 linkage of nucleolus organizer (contrary to X–Y nucleolus organizer in Drosophila) has also been reported (Hinton & Downs, Reference Hinton and Downs1975).

(i) Inversion polymorphism in natural populations

Since the establishment of the modern synthesis, inversions have been a privileged system to study such diverse subjects as phylogenies, geographical clines, temporal cycles and meiotic drive, and, of course, to look for evidence of natural selection (Krimbas & Powell, Reference Krimbas and Powell1992). Study of chromosomal polymorphism in populations show the interplay of evolutionary factors in the maintenance and improvement of their adaptation to the environment. The development of polymorphism through natural selection is one of the ways through which a population may improve its capacity to utilize the environment and survive through temporal changes of it. In natural populations of Drosophila, chromosomal polymorphism due to inversions is common and is an adaptive trait (Da Cunha, Reference Da Cunha1960; Dobzhansky, Reference Dobzhansky1970; Sperlich & Pfriem, Reference Sperlich, Pfriem, Ashburner, Carson and Thompson1986).

Studies on chromosomal polymorphism in Indian populations of D. ananassae were initiated by Ray-Chaudhuri & Jha (Reference Ray-Chaudhuri and Jha1966, Reference Ray-Chaudhuri and Jha1967); since then, a number of investigations on chromosomal polymorphism in Indian populations of D. ananassae have been carried out. Quantitative data on the frequencies of three cosmopolitan inversions in Indian natural populations of D. ananassae show that there are significant variations in the frequencies of these inversions (also showing north–south trends) and the level of inversion heterozygosity among the populations, and that the natural populations are geographically differentiated at the level of inversion polymorphism (see review by Singh, Reference Singh1996; Singh & Singh, Reference Singh and Singh2007a). Populations from the similar eco-geographic regions show similar trends in inversion frequencies and level of inversion heterozygosity. There is no strong positive relation between genetic differentiation and geographic distance although many pair-wise comparisons show that populations separated by small geographic distances show higher genetic identity (see review by Singh, Reference Singh1996; Singh & Singh, Reference Singh and Singh2007a). However, in the study by Das et al. (Reference Das, Mohanty and Stephan2004), genetic differentiation was found to correlate significantly with geographic distance.

These three cosmopolitan inversions differ in their distribution and prevalence and do not show variation with geographical and other parameters as revealed by correlation and regression analysis. No temporal divergence was found between spatially similar but temporally different populations (sampled at different times), i.e. none of the populations showed long-term directional changes. This reinforces the concept of rigid polymorphism in natural populations of D. ananassae, as such a system does not show temporal variation or variation with geographical parameters (Singh & Singh, Reference Singh and Singh2007a).

Nei's (Reference Nei1973) gene diversity estimates were calculated using quantitative data on the frequencies of three cosmopolitan inversions in 45 Indian natural populations of D. ananassae to deduce the distribution of genetic differentiation when populations were grouped according to the time of collection (years and months), regions (coastal and mainland regions) and seasons. Major proportion of this diversity is distributed among populations of different groups rather than within-populations of the same group. The association of genetic variation with environmental and geographical heterogeneity could be due to natural selection operating on chromosomal variability in D. ananassae (Singh & Singh, unpublished).

Singh (Reference Singh2001) reviewed the work done on inversion polymorphism in Indian natural populations of three species, viz. Drosophila melanogaster, Drosophila bipectinata and D. ananassae, which clearly demonstrates that these three species vary in their patterns of inversion polymorphism and have evolved different mechanisms for adjustment to their environments, although they belong to the same species group.

Thus, there is geographic variation of chromosomal polymorphism in D. ananassae populations due to their adaptation to varying environment and natural selection operates to maintain inversions. Since, the three cosmopolitan inversions in D. ananassae are widely distributed and occur in high frequencies, they may be regarded as very old in the evolutionary history of the fly and adaptively important for the species.

(ii) Inversion polymorphism in laboratory populations

Chromosomal polymorphism due to three cosmopolitan inversions often persists in laboratory populations of D. ananassae established from females collected from nature (Singh, Reference Singh1982a, Reference Singh1983b, Reference Singhc, Reference Singh1987). These laboratory populations when compared with the corresponding natural populations show both increasing and decreasing trends in inversion frequencies and level of inversion heterozygosity though most of the populations have maintained more or less similar trends (Singh & Singh, Reference Singh and Singh2008). This demonstrates that heterotic buffering is associated with these inversions and chromosomal polymorphism is balanced due to adaptive superiority of inversion heterozygotes (Moriwaki et al., Reference Moriwaki, Ohnishi and Nakajima1956; Singh & Ray-Chaudhuri, Reference Singh and Ray-Chaudhuri1972; Singh, Reference Singh1982a; Tobari & Moriwaki, Reference Tobari, Moriwaki and Tobari1993). However, the degree of heterosis may vary depending on the allelic contents of the chromosome variants (Singh, Reference Singh1983b). Genetic identity (I) and genetic distance (D) values calculated following the formula of Nei (Reference Nei1972) to determine the degree of genetic divergence between natural and laboratory populations indicate that there is variation in the degree of genetic divergence in D. ananassae populations transferred and maintained for several generations under laboratory conditions. Populations collected from similar environmental conditions that initially show high degree of similarity have diverged to different degrees. This randomness could be due to genetic drift, though inversions in this species are subject to selection (Singh & Singh, Reference Singh and Singh2008).

(iii) Genetic co-adaptation

The results obtained in D. ananassae with respect to the phenomenon of genetic co-adaptation (Singh, Reference Singh1972, Reference Singh1974b, Reference Singh1981, Reference Singh1985b) conflicts with what has been found in other species of Drosophila. In D. ananassae, the inversion heterozygotes produced by chromosomes derived from distant localities exhibit heterosis. Evidence for persistence of heterosis associated with cosmopolitan inversions in interracial hybridization experiments has been presented, involving chromosomally polymorphic and monomorphic strains of D. ananassae (Singh, Reference Singh1972, Reference Singh1974b, Reference Singh1981, Reference Singh1985b). Based on these findings, it has been suggested that heterosis associated with cosmopolitan inversions in D. ananassae appears to be simple luxuriance rather than population heterosis (co-adaptation), and thus luxuriance can function in the adjustment of organisms to their environment (Singh, Reference Singh1985b). This provides evidence against selectional co-adaptation hypothesis.

(iv) Lack of evidence for intra- and interchromosomal interactions

Inversion polymorphism found in different species of Drosophila offers a good material for testing epistatic interactions. The phenomenon of non-random associations between linked inversions is documented in D. ananassae (Singh, Reference Singh1983a, Reference Singh1984). Two linked inversions, namely, delta (3L) and eta (3R) of the third chromosome are associated randomly in natural populations (Singh, Reference Singh1974a, Reference Singh1984; Singh & Singh, unpublished). However, the same two inversions show non-random association in laboratory stocks, which could be due to suppression of crossing-over and random genetic drift (Singh, Reference Singh1983a, Reference Singh1984; Singh & Singh, Reference Singh and Singh1988, Reference Singh and Singh1990, Reference Singh and Singh1991, unpublished). For unlinked inversions, no evidence of interchromosomal interactions has been found in D. ananassae in both natural and laboratory populations (Singh, Reference Singh1982b, Reference Singh1983a; Singh & Singh, Reference Singh and Singh1989, unpublished).

Tobari & Kojima (Reference Tobari and Kojima1967, Reference Tobari and Kojima1968) and Kojima & Tobari (Reference Kojima and Tobari1969) studied the selective modes of inversion polymorphism of a single pair of arrangements of either II or III chromosomes singly and of two pairs of the arrangements of both the chromosomes jointly. Their results indicate that interaction between arrangements of 2L and 3L can be responsible for the differences in the succession of frequencies approaching equilibrium between populations containing different genetic conditions, however balanced polymorphisms were established in all populations. The fitness of the karyotypes gradually changes, depending upon the frequencies of the karyotypes, which are successively changing in the population. Thus, the fitness of the karyotypes is a function of their frequencies in the population (Tomimura et al., Reference Tomimura, Matsuda, Tobari, Cariou, Da Lage, Stephan, Langley and Tobari1993).

(v) Allozyme polymorphism

Enzyme polymorphism has been used mainly to detect selection acting on specific loci, to understand genetic structure of populations, and to analyse the patterns of geographic differentiation. Results of amylase electrophoresis in D. ananassae (Doane, Reference Doane and Hanly1969) revealed some polymorphism and striking geographic pattern throughout the world (Da Lage et al., Reference Da Lage, Cariou and David1989). African populations were much more polymorphic than those from far East and showed multibanded phenotypes, suggesting multiplication of Amy structural gene, with at least 4 copies per haploid genome in certain populations. Nine other species of D. ananassae subgroup exhibited weak amylase activity (Da Lage et al., Reference Da Lage, Cariou and David1989). Unlike the case in D. melanogaster, the isozymes in this species show considerable temporal variation in expression (Da Lage & Cariou, Reference Da Lage, Cariou and Tobari1993). Though, D. ananassae is in the melanogaster group, it has evolved a very different set of regulatory patterns for amylase than D. melanogaster, though both are, ancestrally, tropical fruit breeders. Number of copies and allozymic variations are higher in D. ananassae (Da Lage et al., Reference Da Lage, Lemeunier, Cariou and David1992; Cariou & Da Lage, Reference Cariou, Da Lage and Tobari1993). An analysis of D. ananassae subgroup including D. ananassae itself has shown that a maximum of 30% of the loci are polymorphic and that even the most polymorphic (Estc, Acph, Ca, Pgm) loci show similar variability in all species (see Tobari, Reference Tobari1993 for references).

Similar studies on Adh isozymes between different geographic strains show biochemical genetic differentiation (Jha et al., Reference Jha, Mishra and Pandey1978; Parkash & Jyoutsna, Reference Parkash and Jyoutsna1988; Sharma et al., Reference Sharma, Sharma and Prakash1993).

Considering enzyme polymorphism, the main conclusion emerging is that populations of D. ananassae have a moderate level of genetic variability in spite of their worldwide distribution (Tobari, Reference Tobari1993). Compared to allozymes, the picture of geographic differentiation appears to be different for chromosomes, which are more variable and more differentiated even over short distances. This could be due to the fact that allozymes in general are more neutral than chromosome arrangements (Tobari, Reference Tobari1993).

In numerous studies of allozyme variation in D. ananassae (Gillespie & Kojima, Reference Gillespie and Kojima1968; Johnson, Reference Johnson1971), authors have often attempted to detect linkage disequilibrium between loci, reasoning that if selection affects these polymorphisms, one might expect such disequilibrium at least in some loci. The conclusion from all these studies is that virtually no linkage disequilibrium exists among allozyme loci.

(vi) DNA polymorphism

Using data of DNA sequence variation, theories of population genetics and evolution can be tested more rigorously than with previously available methods (Tobari, Reference Tobari1993). Das et al. (Reference Das, Mohanty and Stephan2004) had inferred the population structure and demography of D. ananassae using multilocus DNA sequence (10 netural loci) and 16 populations covering entire species range (Asia, Australia and America). Using putatively neutral nuclear DNA sequence polymorphisms from 10 independent loci, central populations were discerned from the peripheral populations. The levels of nucleotide diversity, the number and frequency of haplotypes, and the amount of linkage disequilibrium vary among the populations. In comparison with the previous studies of two neutral loci [Om (1D) and forked] with samples from Asia and South America (Stephan, Reference Stephan1989; Stephan & Langley, Reference Stephan and Langley1989; Stephan et al., Reference Stephan, Xing, Kirby and Braverman1998), their (Das et al., Reference Das, Mohanty and Stephan2004) analysis finds lower estimates of nucleotide diversity. In D. ananassae, reduced recombination is associated with low levels of DNA polymorphism. In other studies (Stephan, Reference Stephan1989; Stephan & Langley, Reference Stephan and Langley1989; Stephan & Mitchell, Reference Stephan and Mitchell1992) of DNA polymorphism in D. ananassae at four loci of the X-chromosome: vermilion (v), furrowed (fw), forked (f) and Om (ID), genes experiencing normal amounts of polymorphism, Om (ID) and forked (f), were found to be 10 times more variable than genes located in regions of very low recombination, furrowed (fw) and vermilion (v). These effects are due to restricted migration between populations and differences in recombination rates of the chromosome regions in which the various loci lie. Recombination is the main factor determining nucleotide variability in different regions of the genome. Chromosomal inversions are known to reduce and redistribute recombination, and thus their specific effect on nucleotide variation may be of major importance as an explanatory factor for levels of DNA variation (Navarro et al., Reference Navarro, Barbadilla and Ruiz2000). Reduction in average heterozygosity in the v and fw regions can be explained based on the models of directional selection and genetic hitchhiking. Recurrent fixation of few alleles will wipe out standing variation in a population by this process and thus reduce the level of heterozygosity, if the recombination rate is low. The hitchhiking effect is less strong in regions with intermediate or high recombination rates, such as f and Om (1D). Effect of population subdivision on variation among populations with different distances from the species centre in Southeast Asia, i.e. Myanmar, India and Brazil, was also examined. The between-population differences in average heterozygosity may be explained via neutral theory of molecular evolution (Kimura, Reference Kimura1983), which predicts that average nucleotide heterozygosity is proportional to effective population size, so that average heterozygosity follows the order Myanmar>India>Brazil. Since, D. ananassae spreads from its zoogeographical centre in Southeast Asia (Myanmar) to India and then to Central America and South America via restricted migration, hence reduced population size and average heterozygosity (Stephan, Reference Stephan1989; Stephan & Langley, Reference Stephan and Langley1989; Stephan & Mitchell, Reference Stephan and Mitchell1992). Natural selection may have a strong influence on the broad expanses of genome in populations from Northern versus Southern Asia (Stephan et al., Reference Stephan, Xing, Kirby and Braverman1998; Chen et al., Reference Chen, Marsh and Stephan2000; Kim & Stephan, Reference Kim and Stephan2000; Baines et al., Reference Baines, Das, Mousset and Stephan2004) and because of the obvious genetic drift that may accompany South Pacific Islands and potentially some of the ancestral populations from Southeast Asia that surround the ancestral geographic range in Indonesia. However, the young age of population makes it extremely unlikely that natural selection have a role in DNA sequence variation (Das et al., Reference Das, Mohanty and Stephan2004), but evidence of adaptive evolution was inferred from the pattern of DNA sequence variation in northern versus southern populations of Asia (Stephan et al., Reference Stephan, Xing, Kirby and Braverman1998; Chen et al., Reference Chen, Marsh and Stephan2000; Kim & Stephan Reference Kim and Stephan2000; Baines et al., Reference Baines, Das, Mousset and Stephan2004). These studies (Stephan et al., Reference Stephan, Xing, Kirby and Braverman1998; Chen et al., Reference Chen, Marsh and Stephan2000; Baines et al., Reference Baines, Das, Mousset and Stephan2004) suggest that extensive physical and genetic maps based on molecular markers and detailed studies of population structure may provide insights into the degree to which natural selection affects DNA sequence polymorphism across broad regions of chromosome. In other studies (Vogl et al., Reference Vogl, Das, Beaumont, Mohanty and Stephan2003; Das et al., Reference Das, Mohanty and Stephan2004; Schug et al., Reference Schug, Regulski, Pearce and Smith2004, Reference Schug, Smith, Tozier-Pearce and McEvey2007), we found that the level of molecular variation is quite variable among the populations. Populations in ancestral range in Indonesia and peripheral range in Asia and Australia show lower genetic differentiation than populations from the Pacific Island (Schug et al., Reference Schug, Baines, Amanda, Mohanty, Das, Grath, Smith, Zhargam, McEvey and Stephan2008). Molecular variation varies considerably among the ancestral, peripheral and South Pacific populations consistent with the previous studies of intron polymorphism (Vogl et al., Reference Vogl, Das, Beaumont, Mohanty and Stephan2003; Das et al., Reference Das, Mohanty and Stephan2004; Schug et al., Reference Schug, Smith, Tozier-Pearce and McEvey2007) and microsatellites (Schug et al., Reference Schug, Smith, Tozier-Pearce and McEvey2007). In contrast to polymorphism, divergence between D. ananassae populations and its sibling species D. pallidosa is constant across loci.

Genome size differences are usually attributed to the amplification and deletion of various repeated DNA sequences, including transposable elements (TEs), when species encounter a new environment. Nardon et al. (Reference Nardon, Deceliere, Levenbruck, Weiss, Vieira and Biemont2005) conducted a study to find out whether genome size is influenced by colonization of new environments in Dipteran species, including D. ananassae. Results show that D. ananassae does not display obviously smaller average genomes in their probable region of origin, and variability in genome size of Indian populations of D. ananassae have been found. This could be due to different colonization routes followed by this species and different environmental conditions encountered by the populations.

Using genomic data from five closely related species of Drosophila (D. melanogaster, Drosophila simulans, Drosophila yakuba, Drosophila erecta and D. ananassae), a maximum likelihood framework was applied to calculate rates of protein evolution and to test the evidence of positive selection. In all comparisons, weak positive correlation between expression divergence and protein evolution was found (Good et al., Reference Good, Hayden and Wheeler2006).