Molecular Evolution of Duplicated Amylase Gene Regions in Drosophila melanogaster: Evidence of Positive Selection in the Coding Regions and Selective Constraints in the cis-Regulatory Regions

Corresponding author: Tsuneyuki Yamazaki, Department of Biology, Faculty of Science, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka City, Japan 812-8581., tyamascb{at}mbox.nc.kyushu-u.ac.jp (E-mail)

Communicating editor: N. TAKAHATA

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In this study, we randomly sampled Drosophila melanogaster from Japanese and Kenyan natural populations. We sequenced duplicated (proximal and distal) Amy gene regions to test whether the patterns of polymorphism were consistent with neutral molecular evolution. F_st between the two geographically distant populations, estimated from Amy gene regions, was 0.084, smaller than reported values for other loci, comparing African and Asian populations. Furthermore, little genetic differentiation was found at a microsatellite locus (DROYANETSB) in these samples (G^'_st = -0.018). The results of several tests (Tajima's, Fu and Li's, and Wall's tests) were not significantly different from neutrality. However, a significantly higher level of fixed replacement substitutions was detected by a modified McDonald and Kreitman test for both populations. This indicates that positive selection occurred during or immediately after the speciation of D. melanogaster. Sliding-window analysis showed that the proximal region 1, a part of the proximal 5' flanking region, was conserved between D. melanogaster and its sibling species, D. simulans. An HKA test was significant when the proximal region 1 was compared with the 5' flanking region of Alcohol dehydrogenase (Adh), indicating a severe selective constraint on the Amy proximal region 1. These results suggest that natural selection has played an important role in the molecular evolution of Amy gene regions in D. melanogaster.

IN Drosophila, levels of polymorphism have been examined frequently using allozymes, RFLPs, and DNA sequences to test evolutionary hypotheses regarding genetic variation in natural populations. The Amy gene encoding -amylase in Drosophila is one of the most extensively studied genes. The structural Amy gene in Drosophila melanogaster is composed of one exon of 1482 bp (BOER and HICKEY 1986 ) and is inversely duplicated on the second chromosome (BAHN 1967 ; LEVY et al. 1985 ). The duplicated genes are known as the proximal and distal genes, respectively. Because -amylase interacts directly with food, Amy is a potential target gene of adaptive evolution. Indeed, -amylase is highly polymorphic at the isozyme level (KIKKAWA 1964 ; DOANE 1969 ; HICKEY 1979 ; SINGH et al. 1982 ; YAMAZAKI et al. 1984 ; DAINOU et al. 1993 ) and there is a large amount of genetic variability with respect to the levels of -amylase activity (DOANE 1969 ; YAMAZAKI and MATSUO 1984 ; LANGLEY et al. 1988 ). In particular, the regulation of Amy has been the focus of a great deal of research. It has become clear that factors such as the response to dietary carbohydrates (YAMAZAKI and MATSUO 1984 ; BENKEL and HICKEY 1986 ; INOMATA et al. 1995A ; INOMATA and YAMAZAKI 2000 ), tissue specificity (ABRAHAM and DOANE 1978 ; POWELL et al. 1980 ; KLARENBERG et al. 1986 ), and stage-specific expression patterns (YAMAZAKI 1986 ; DA LAGE and CARIOU 1993 ; INOMATA and YAMAZAKI 2000 ) are all involved in the regulation of this gene. MATSUO and YAMAZAKI 1984 showed that amylase regulation (inducibility) is positively correlated with fitness in the productivity and life span of individuals.

The selective significance of the "structural" gene has also been investigated in several studies, which have involved two contrasting strategies. One approach is to measure fitness differences between isozymes directly, in laboratory cage populations (DE JONG and SCHARLOO 1976 ; HICKEY 1977 ; YARDLEY et al. 1977 ; POWELL and ANDJELKOVIC 1983 ; MATSUO and YAMAZAKI 1984 ). Unfortunately, most of the results obtained in this manner have been inconclusive, mainly due to linkage disequilibrium (see DOANE 1980 ). The other approach is to study molecular evolutionary genetics to assess the evolutionary history of the gene. Two Amy structural genes in D. melanogaster species subgroups have evolved in a concerted fashion, in contrast to the apparently divergent evolution seen in the 5' flanking regions (HICKEY et al. 1991 ; SHIBATA and YAMAZAKI 1995 ). Furthermore, heterogeneity has also been observed in the inferred relative numbers of synonymous and replacement substitutions that occurred in the speciation of the D. melanogaster species subgroups or subgenus Sophophora of Drosophila (SHIBATA and YAMAZAKI 1995 ; INOMATA et al. 1997 ). The level of polymorphism has already been examined for Amy coding regions in D. melanogaster, but the sequences have previously been sampled only with respect to different isozymes, rather than as random samples of the populations. Therefore, the selective significance of Amy gene regions in D. melanogaster remained untested.

In this study, we randomly sampled D. melanogaster individuals from natural populations from Japan and Kenya and examined the selective neutrality of DNA sequence variation in two Amy coding regions and 5' flanking regions, which include cis-regulatory elements. To evaluate the possible geographical differentiation between the two populations, we also examined the allelic variation of DROYANETSB (LAI and RUBIN 1992 ), which is a microsatellite locus known to be highly variable with respect to the repeat number (GOLDSTEIN and CLARK 1995 ; SCHLOTTERER et al. 1998 ). This locus is located in an intron of the yan gene (genetic map position 2-12) and is supposed to be randomly associated with Amy loci (genetic map positions 2-77.3 and 2-77.9).

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Population samples:
The tested flies were collected in Akayu, Yamagata, Japan in 1996 by T. Yamazaki, and in Nairobi, Kenya by R. Woodruff in August 1988 (WOODRUFF et al. 1990 ). The Kenyan samples were supplied to us by S. D. Irvin (Cornell University). The amylase isozymes were analyzed by acrylamide gel electrophoresis, using 185 and 24 isofemale lines from the Japanese and Kenyan populations, respectively. The CyO/Pm balancer stock was used to obtain isogenic lines. Ten Japanese and 12 Kenyan lines, all of which were homozygous for the second chromosome and free of both lethal and sterile genes, were selected at random for sequencing.

Polyacrylamide gel electrophoresis:
One adult fly was homogenized by sonication in 0.1 M Tris-borate/5 mM MgCl₂/10% sucrose buffer (pH 8.9). Samples of the homogenate were immediately applied to polyacrylamide gels (5% acrylamide/0.2% Bis-acrylamide/20 mM CaCl₂/0.1 M Tris-borate) and were subjected to electrophoresis in 0.1 M Tris-borate (pH 8.9) for 3 hr at 4°, 300 V. After electrophoresis, the gels were incubated in 2% starch solution for 1 hr at 37°, washed with water, and stained in an I₂-KI solution (1:2:10 solution of I₂:KI:H₂O, diluted 30 times). The relative mobilities of six amylase isozymes found in natural populations of D. melanogaster, AMY¹ to AMY⁶, were designated 1.00, 0.92, 0.87, 0.80, 0.75, and 0.68, respectively. The band mobilities of the sample were compared with those of marker strains KO123 (AMY^1,2), AO168 (AMY^1,3), 1420#1 (AMY^4,6), and L16 (AMY^5,6).

DNA preparation and sequencing:
Genomic DNA from all lines was extracted according to protocol 48 in ASHBURNER 1989 , with minor modifications. A fragment of 2.2 kb, including the Amy coding and the 5' flanking region, was amplified by the PCR method of SAIKI et al. 1985 for each of the proximal and distal Amy gene regions. The primers employed for PCR amplification were MPRO600 5'-GCACTGCGACAGGAAGG-3' and –7P 5'-GGACTTTAGACCTGGAC-3' for the proximal gene region, and MDIS600 5'-CTGTAGCGTGAGATTCC-3' and –7D 5'-GCCATATTATGCGTAAC-3' for the distal gene region. For sequencing, several primers, amplifying regions of both strands of the sequences investigated and comprising 250 bp, were selected on the basis of the conserved regions in the coding and 5' flanking sequences. The amplified fragments were sequenced and separated using a Perkin-Elmer (Norwalk, CT) ABI PRISM 377 automated DNA sequencer. The sequence data presented in this article have been submitted to the DDBJ database under the accession nos. ABO42862–42883 and ABO43027–43048. Microsatellite sequences of DROYANETSB alleles were also determined using the same method, except that the forward and reverse primers used were YANF 5'-GCAAAAGCAGCGTAAAT-3' and YANR 5'-CTGCTCCTCCATCTTCC-3', respectively, and the PCR products were 212 bp long on average.

Southern blot analysis:
Southern blot analysis was performed according to SAMBROOK et al. 1989 . Genomic DNA was digested with SalI, and a 1-kb fragment of the Amy 5' flanking region derived from D. melanogaster (described in OKUYAMA et al. 1997 ) was labeled using a MEGAPRIME DNA labeling kit (Amersham, Tokyo), and then used as a probe.

Nucleotide data analysis:
We used SeqPup program version 0.6f (GILBERT 1996 ) and ProSeq program pre-beta version 2.6 (D. FILATOV, personal communication; ProSeq v 2.56 is available at http://helios.bto.ed.ac.uk/evolgen/filatov/proseq.html) for editing the sequences, and CLUSTAL X (THOMPSON et al. 1997 ) for multiple alignment. The TREECON program (VAN DE PEER and DE WACHTER 1994 ) was used to construct neighbor-joining (NJ) trees involving 1000 bootstrap replications, under the assumption that substitutions followed KIMURA's (1980) two-parameter model. Most of the statistical analyses were performed using the DnaSP version 3.14 program (ROZAS and ROZAS 1999 ). Programs for performing WALL's (1999) test and FAY and WU's (2000) test were kindly provided by J. D. Wall and J. C. Fay. For tests that require an outgroup, Amy sequences of D. simulans (SHIBATA and YAMAZAKI 1995 ; accession nos. D17733–D17734) were used. In several neutrality tests where multiple comparisons were involved, we employed the Bonferroni correction method.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Amylase variants at the isozyme level:
Initially we examined variation among isozymes according to differences in mobility in acrylamide gel electrophoresis. In 185 and 24 isofemale lines from Japanese and Kenyan populations, the frequencies of the most common haplotype, AMY¹ (AMY^1,1 or AMY^1,null), as estimated by the maximum-likelihood method (MATSUO and YAMAZAKI 1984 ), were 0.90 and 0.63, respectively. In isogenic lines, 9 out of 10 Japanese lines (90%) and 9 out of 12 Kenyan lines (75%) displayed an AMY¹ haplotype, 1 Japanese line displayed AMY^2,3, and 3 Kenyan lines displayed AMY^5,6 haplotypes (Table 1). These values agree well with the frequencies of AMY¹ from isofemale lines; thus the isogenic lines are typical of their populations.

Nucleotide sequences of 5' flanking regions:
We examined the DNA sequence variation in a total of 4241 bp and found that 22 sequences contained 229 segregating sites and 10 insertion/deletion (indel) polymorphisms. There were 40 segregating sites and 2 indels in the proximal 5' flanking region, and 47 segregating sites and 5 indels were located in the distal 5' flanking region.

One of the insertion/deletion polymorphisms in the 5' flanking region was of particular interest. In two Kenyan lines, KN-3 and KN-10, we found a significant rearrangement of the 5' flanking region of the distal gene (Fig 1). These two lines are identical with respect to all sequences in Amy gene regions, and we refer to them as "rearranged" in contrast to the other "standard" sequences. In the rearranged sequence, the 59-bp region immediately upstream of the translation initiation site of the distal gene is replaced with an inserted sequence. The length of the inserted sequence was shown to be 2 kb by genomic Southern blot analysis (data not shown). We then sequenced a 550-bp region upstream of the inserted sequence (Seq-A in Fig 1) and a 660-bp inserted region upstream of the distal coding region (Seq-B in Fig 1) to analyze the 5' flanking region of the rearranged distal gene. Interestingly, Seq-B showed a high degree of homology (on average 98%) to the 5' flanking region of the Amy "proximal" gene region. Furthermore, the sequence downstream of Seq-A (the left end of the inserted sequence in Fig 1) showed a high level of homology to the sequence 2 kb upstream of the translation initiation site of the proximal gene. This indicates that the whole inserted sequence was transferred from the proximal gene region. Fig 2 shows the NJ tree of the 5' flanking regions, Seq-A and -B, with sequences of D. simulans as an outgroup. Fig 2 clearly shows the differentiation between the proximal and distal 5' flanking regions, except for Seq-B in KN-3 and KN-10. Thus, in their distal gene regions, KN-3 and KN-10 have sequences that are homologous to the distal 5' flanking region (Seq-A) and to the proximal 5' flanking region (Seq-B). In spite of the significant rearrangement, the product of the distal gene (AMY⁶) in KN-3 and KN-10 was enzymatically active in polyacrylamide gels (data not shown; see Table 1).

View larger version (8K): In this window In a new window Download PPT slide   Figure 1. The pattern of Amy gene regions and the difference between two kinds of arrangements, designated "standard" and "rearranged" sequences. The boxes indicate Amy coding regions and the arrows indicate the sequenced region. In the rearranged sequence, the 59-bp region immediately upstream of the distal coding sequence was replaced by an 2-kb-long inserted sequence. We defined Seq-A as a region upstream of the inserted sequence and Seq-B as a distal coding region in the rearranged sequence.

View larger version (19K): In this window In a new window Download PPT slide   Figure 2. The phylogenetic relationships produced by the neighbor-joining method from the nucleotide sequence data of the 5' flanking regions in D. melanogaster. The regions in which proximal and distal sequences can align (positions -1 to -488) were used. -p and -d represent proximal and distal gene regions, respectively. Seq-A and -B, defined in Fig 1, are also included in the relationships. The numbers above the main branches indicate bootstrap values (1000 replicates).

Nucleotide sequences of coding regions:
There were 66 segregating sites and no indels in the proximal coding regions. In the distal coding regions there were 76 segregating sites and three indels. Fig 3 shows the NJ tree derived from the coding sequences. In the coding regions, proximal and distal gene sequences within species showed strong similarities. Most of the topology was unchanged, even when the NJ tree was drawn using only sequences at synonymous sites (results not shown). This observation is consistent with previous studies, indicating that the Amy coding regions evolved in a concerted fashion, which contrasts with the process in 5' flanking regions (HICKEY et al. 1991 ; SHIBATA and YAMAZAKI 1995 ). Indeed, three different conversion tracts, measured as the distance between the outermost informative site within the tract (see BETRAN et al. 1997 ), were detected in coding sequences in the two populations, while no conversion tract was detected in 5' flanking sequences. The average lengths of the observed tracts were 600 bp and 375 bp in the Japanese and Kenyan populations, respectively.

View larger version (24K): In this window In a new window Download PPT slide   Figure 3. The phylogenetic tree constructed using the neighbor-joining method from the nucleotide sequence data of the coding regions in D. melanogaster. Out of the entire 1490 bp, only comparable sequences without gaps (797 bp) were used. Proximal and distal coding regions of Amy in D. simulans were used as an outgroup. The numbers associated with branches are bootstrap values of >500 out of 1000 replicates.

Three Kenyan lines, KN-7, KN-22, and KN-28, had independent insertions or deletions in the coding region of the distal gene. KN-22 had a deletion of 3 bp at position +16, which was considered to be in the coding sequence of a signal peptide (BOER and HICKEY 1986 ), and the reading frame was not shifted. In contrast, KN-7 had an insertion of 8 bp at position +217 and KN-28 had a deletion of 689 bp at +201, both of which resulted in frameshifts and generated stop codons at +586 in KN-7 and at +274 in KN-28. Thus, the function of their distal genes would have been lost; therefore, we presumed that the distal genes of KN-7 and KN-28 were null alleles. For these reasons, the distal 5' flanking region of KN-3 and KN-10 and the distal coding regions of KN-7 and KN-28 were excluded from the following analysis. However, we found that the exclusion of these lines had little effect on the results.

Levels of polymorphism in Amy gene regions:
Table 2 gives a summary of polymorphism in the Amy regions. The level of polymorphism was estimated as , the average number of pairwise nucleotide differences (NEI 1987 ). To investigate the pattern of substitutions in 5' flanking regions in detail, we divided them into three subregions on the basis of the work of OKUYAMA et al. 1996 . Region 1 was defined as the sequences from positions -1 to -333 for the proximal gene region and from positions -1 to -296 for the distal gene region. In region 1, the proximal and distal regions are relatively similar and include several putative cis-regulatory elements. Region 2 was defined as the sequences from positions -334 to -447 for the proximal and from positions -297 to -432 for the distal genes. In region 2, less similarity was found between the proximal and distal genes, and there was much more interspecific divergence than in region 1 (see OKUYAMA et al. 1996 for details). The cited authors also found that there was little or no paralogy between the proximal and distal 5' flanking regions upstream of region 2; therefore, we named this region 3.

The levels of polymorphism shown in Table 2 were qualitatively consistent with the values estimated for the Amy gene from nonrandom samples by INOMATA et al. 1995B and MORIYAMA and POWELL 1996 . The synonymous sites had the highest nucleotide diversity, followed by the whole 5' flanking region and the total coding region. Considering the data in Table 2 in detail, however, it can be seen that three 5' flanking regions exhibited large differences in the amount of nucleotide diversity. Regions 2 and 3 exhibited unexpectedly low levels of polymorphism, except for the proximal region 2 in the Kenyan population. Fig 4 shows sliding-window plots of the estimates of polymorphism and divergence across the regions studied. Because we have no sequences comparable to region 3 in D. simulans, we excluded this region from analyses in which sequence data from other species were required. Fig 4 shows the notably high level of divergence in contrast with the low level of polymorphism in both the proximal and distal region 2. In total in the Amy gene regions, the nucleotide diversity in the distal region was 1.52 and 1.12 times as large as that in the proximal gene regions for the Japanese and Kenyan populations, respectively. Comparing the two populations, the nucleotide diversity of the Kenyan lines is 1.36 times greater than that of the Japanese lines.

View larger version (26K): In this window In a new window Download PPT slide   Figure 4. Sliding window plot of the nucleotide diversity, , of the two populations (JP and KN) and the nucleotide divergence between D. melanogaster and D. simulans, (interspecific), across the Amy gene regions. (A) Window plot for the proximal gene region; (B) window plot for the distal gene region. R1 and R2 represent region 1 and region 2, respectively; Amy-p and Amy-d represent the coding regions of proximal and distal genes, respectively. Arrows and open arrowheads indicate the positions of fixed replacement and fixed synonymous substitutions in the D. melanogaster branch, respectively (see text). The window size was 50 bp and the step size 5 bp.

Gene flow between the two populations:
Fig 2 and Fig 3 indicate the incomplete genetic differentiation of the lines from the two geographically distant populations. Because we used random samples from the two populations, we were able to estimate the level of gene flow by calculating the fixation index, F_st, and Nm, where N is the effective population size and m is the fraction of migrants in each subpopulation in each generation (HUDSON et al. 1992 ). The values of F_st for the Amy gene region ranged between 0.026 and 0.090, 0.084 in total. These values were relatively low in contrast to corresponding values comparing African and Asian populations at other loci. For example, BEGUN and AQUADRO 1995 estimated that F_st values between Kenyan and Beijing populations and between Kenyan and Taiwanese populations at the vermilion locus in D. melanogaster were 0.176 and 0.158, respectively. To examine whether the low value is specific to Amy gene regions, we also determined the allelic variation of the number of TG repeats at DROYANETSB. The results are shown in Table 1. G^'_st values for the populations, the estimator equivalent to F_st for haplotype frequencies (NEI 1987 ), were notably low: (-0.018), smaller than the value derived for the Amy gene regions (G^'_st = 0.019), while average heterozygosities of DROYANETSB were high (0.80 and 0.79 for the Japanese and Kenyan populations, respectively). Although many more loci need to be examined to determine the population structure of these two geographically distant populations, these data suggest that the two populations are genetically more similar than previously supposed.

Neutrality tests:
In neutrality analyses, we first examined the selective neutrality using TAJIMA's (1989), and FU and LI's (1993) tests using an outgroup. Both of these tests compare measures of = 4N_eu, where N_e is the effective population size and u is the mutation rate. Tajima's D (D_Tajima) and Fu and Li's D (D_{Fu and Li}) values are shown in the first and second columns, respectively, for each population in Table 3. Another statistic used to test neutrality, Wall's Q parameter (WALL 1999 ), is based on a measure of linkage disequilibrium between adjacent pairs of segregating sites. The rejection probabilities for observed Q values, P(Q), are listed for each population in the third column of Table 3. To test these statistics, 10⁵ coalescent simulations were run using observed values for each region, i.e., the sample size, the length of sequence, and the number of segregating sites. WALL 1999 found that Q does not decrease monotonically with increasing recombination rate, and that no-recombination simulations would not be conservative to obtain critical values for Q. We then applied the estimated value of the recombination rate from the pairwise comparisons (HUDSON 1987 ) for each region to the coalescent simulations for Q. For each population, the fourth column in Table 3 represents P(H), i.e., the P value of Fay and Wu's H statistics (FAY and WU 2000 ). This compares an estimator of weighted by the homozygosity of the new variants to , and thus indicates variations linked to the selective sites exposed to hitchhiking events. We also examined FU and LI's (1993) F and WALL's (1999) B statistics (results not shown). These statistics produced no significant evidence for the nonneutral distribution of the observed nucleotide frequencies. This was the case even before Bonferroni corrections for multiple comparisons were applied, except for Fay and Wu's H values for the distal region 2 in the Japanese population (see Table 3).

We also examined the populations by means of the MCDONALD and KREITMAN 1991 test. Application of an exact McDonald and Kreitman test to each of the Amy genes would be undesirable because there are no polymorphism data for this gene in sibling species, which decreases the power of the test. Indeed, no results of tests, in which only one sequence was used as an outgroup, were statistically significant (data not shown). Instead, we employed a modified McDonald and Kreitman test for duplicated genes with gene conversion. This is conceptually identical to the method of SHIBATA and YAMAZAKI 1995 , although they applied this test to the comparison of interspecific variation in D. melanogaster species subgroups. In this test, we regarded all proximal and distal coding regions as alleles. This modification resulted in another classification of "polymorphic" and "fixed" substitutions. In this case, fixed substitution was defined as a substitution between species at sites where all alleles in the duplicated loci in each species shared an identical nucleotide. The rest of the substitutions were classified as polymorphic. Even with this modification, the ratio of replacement to synonymous fixed substitutions should be the same as the ratio of replacement to synonymous polymorphic substitutions where selective neutrality is assumed (see MCDONALD and KREITMAN 1991 ; SHIBATA and YAMAZAKI 1995 ). The results are shown in Table 4. The ratio of replacement to synonymous fixed substitutions (0.875 for both populations) was significantly greater than that of replacement to synonymous polymorphic substitutions (0.174 and 0.250 for the Japanese and Kenyan populations, respectively).

Furthermore, we separated fixed substitutions on the D. melanogaster and D. simulans branches to identify the branches in which nonneutral substitutions occurred. This was achieved using an estimate of ancestral state derived from an Amy coding sequence in D. teissieri and D. yakuba. Interestingly, all seven fixed replacement substitutions and seven out of eight fixed synonymous substitutions appeared on the D. melanogaster branch. Significant heterogeneity was found between the ratio of replacement to synonymous fixed substitutions and between replacement to synonymous polymorphic substitutions in the D. melanogaster branch (1.00 for fixed substitutions, and 0.250 and 0.220 for polymorphic substitutions in the Japanese and Kenyan populations, P = 0.034 and 0.015, respectively). Thus, a significantly higher number of fixed replacement substitutions was found in the D. melanogaster branch. This could have been caused by the acceleration of replacement substitutions during or immediately after the speciation of D. melanogaster, if we assume that synonymous substitution occurs in a selectively neutral manner.

The seven fixed replacement substitutions in the D. melanogaster branch were located at the following sites: +164 (Tyr Phe), +277 (Gln Glu), +467 (Asn Ser), +487 (Gln Glu), +1051 (Ser Thr), +1105 (Val Ile), and +1397 (Ser Thr). The seven fixed synonymous substitutions in the D. melanogaster branch were located at +174, +819, +900, +933, +1059, +1062, and +1464. These changes are plotted in Fig 4. The functional significance of these changes is not clear, but the observation that the -amylase activity of D. melanogaster is much higher than that of other sibling species (SHIBATA and YAMAZAKI 1994 ) may be due to one or more of these substitutions.

Another neutrality test, the Hudson-Kreitman-Aguadé (HKA) test (HUDSON et al. 1987 ), which compares two distinct regions, was performed to determine whether the amount of nucleotide variation within and between species was compatible with a strictly neutral mutation model. We compared six Amy regions (regions 1, 2, and the coding region for each of the proximal and distal regions) to the 5' flanking region of Adh (KREITMAN and AGUADE 1986 ; KREITMAN and HUDSON 1991 ), which was assumed to be undergoing selectively neutral evolution (Table 5). As a result, only the proximal region 1 for the Kenyan population showed significant departure from neutrality after the application of Bonferroni correction at the 5% probability level. The value for the corresponding region for the Japanese population was nearly significant (P = 0.011 before the correction). Because the interspecific divergence in proximal region 1 is severely restricted (Fig 4), these results could be interpreted as the result of a severe selective constraint on this region in the between-species branch.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic differentiation between Japanese and Kenyan populations:
In this study, we found low levels of genetic differentiation in the Amy gene regions and DROYANETSB between Japanese and Kenyan populations. Although this observation cannot be extrapolated to the genome-wide perspective, it would be interesting if these two populations are genetically similar, because it would contradict the orthodox view. Although an ancient introduction from Africa to Asia and a long separation of the species in the two regions has been generally accepted (see DAVID and CAPY 1988 ), HALE and SINGH 1991 have suggested the possibility that Japan was recently colonized or recolonized by this species. They found that the haplotype variation of mitochondrial DNA in a Japanese population was very different from other Asian populations. The similarity between Japanese and Kenyan populations is consistent with this hypothesis, but the high levels of polymorphism observed in Amy and DROYANETSB of Japanese samples do not support the hypothesis. As described above, studies of larger numbers of loci and populations are needed to clarify the history of the two populations.

Level and effect of genetic exchange in Amy coding regions:
We have confirmed the concerted evolution of the Amy coding sequences (Fig 3). The presence of the rearranged sequence also indicates that genetic exchange has recently occurred between duplicated gene regions (see Fig 1 and Table 1). OHTA 1982 defined three identity coefficients at the equilibrium state: f, the average identity probability of alleles; c₁, the average identity of genes at different loci of the duplicated gene on the same chromosome; and c₂, the average identity of genes taken from different loci of two homologous chromosomes within the population. From our data, values of f, c₁, and c₂ in two Amy coding regions were estimated as the average homozygosity rather than allelic identity (Table 6). All of these values were very close to one. We tried to estimate the rate of genetic exchange between two Amy genes using the equation of OHTA 1982 . However, this was not possible because in both populations c₁ c₂ (probably due to sampling error), which prevented us from solving the equation directly. Accordingly, we simplified the equation by approximation to fit the case of Amy genes. When we assume c₁ = c₂ = and f 1, where represents the weighted mean of c₁ and c₂, OHTA's (1982) equations can be simplified; thus

where and represent estimates of the genetic exchange rate and the mutation rate, respectively. Because could be estimated from the sequence data, we estimated the rate of gene exchange between two Amy coding regions as 1.22 x 10^-4 and 9.73 x 10^-5 per gene per generation for the Japanese and Kenyan populations, respectively. This was based on the assumption that the mutation rate per nucleotide site is in the order of 10^-9 per site per generation, consistent with recent understanding of molecular evolution. Such a high level of genetic exchange may seriously bias the result of neutrality tests based on genealogy, and the degree of bias will be relevant for future studies. However, the majority of our findings from the neutrality tests were unaffected by gene exchange, because we used a modified version of the McDonald and Kreitman test for duplicated genes with gene conversion, and the significant results from the HKA test were based on a comparison of flanking regions, where the proximal and distal sequences were clearly divergent.

Molecular evolution of Amy 5' flanking regions in D. melanogaster:
Our study demonstrated that the level of interspecific divergence in proximal region 1 was significantly lower than it was in other Amy regions (Fig 4). Four putative cis-acting regulatory elements were found in this region (OKUYAMA et al. 1996 ). MAGOULAS et al. 1993 showed that a 109-bp sequence in this region was required for overall activity and for glucose repression of the Amy gene. The severe selective constraint on these elements would presumably have affected the level of divergence of the whole of proximal region 1 throughout evolutionary history. The distal region 1, conversely, showed a relatively high level of intra- and interspecific divergence, although this region also includes several cis-acting regulatory elements identical to those of the proximal gene region. One possible explanation is that the selective constraint is relaxed in the distal gene region due to functional redundancy, as has been demonstrated for several other duplicated genes (see, for example, CADIGAN et al. 1994 ). The observations in this study that all rearrangement and null alleles were located in the distal gene region, and that the level of polymorphism in the distal coding region was higher than in the proximal coding region, also support the relaxed selection hypothesis.

The low level of polymorphism in region 2 of the Amy flanking region in the Japanese population (Table 2 and Fig 4) contrasts with the much higher level of interspecific divergence. There are several possible explanations for this observation. For distal region 2 in the Japanese population, the low value of Tajima's D (-1.56) parameter compared with values for Amy gene regions and the low probability detected by Fay and Wu's test (P[H] = 0.021) support the suggestion that there has been recent hitchhiking by the selective site linked to this region (see Table 3).

Molecular evolution of Amy coding regions in D. melanogaster:
The levels of polymorphism in the Amy coding regions (Table 2) are generally consistent with previous suggestions (INOMATA et al. 1995B ; MORIYAMA and POWELL 1996 ). The polymorphisms in the Amy coding regions are two to four times greater than average values for 24 genes estimated by MORIYAMA and POWELL 1996 . The levels of polymorphism in these regions could be affected by the mutation rate, a relaxed selective constraint and balancing selection, and frequent genetic exchange between duplicated genes (see Table 6), resulting in a significant effect on the levels of polymorphism in the Amy coding regions.

Classification of the nucleotide substitutions as "synonymous" or "replacement" demonstrated significant heterogeneity (Table 4), and seven out of eight fixed replacement substitutions were estimated to occur in the D. melanogaster branch. These observations clearly indicate the occurrence of adaptive evolution during or immediately after the speciation of D. melanogaster.

	ACKNOWLEDGMENTS

We thank R. R. Hudson, D. Filatov, D. Wall, and J. C. Fay for supplying computer programs; S. D. Irvin for supplying the Kenyan samples of Drosophila melanogaster; and S. Ikeda for technical support. We are also grateful to H. Tachida and A. E. Szmidt for useful discussions, and we thank N. Takahata and three anonymous referees for comments on an earlier version of this manuscript. This work was supported, in part, by research grants to T.Y. and N.I. from the Ministry of Education and Science and Culture of Japan and by Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists to H.A.

Manuscript received July 24, 2000; Accepted for publication October 23, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ABRAHAM, I. and W. W. DOANE, 1978  Genetic regulation of tissue-specific expression of amylase structural genes in Drosophila melanogaster.. Proc. Natl. Acad. Sci. USA 75:4446-4450[Abstract/Free Full Text].

ASHBURNER, M., 1989 Drosophila, A Laboratory Handbook. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

BAHN, E., 1967  Crossing over in the chromosomal region determining amylase isozymes in Drosophila melanogaster.. Hereditas 58:1-12[Medline].

BEGUN, D. J. and C. F. AQUADRO, 1995  Molecular variation at the vermilion locus in geographically diverse populations of Drosophila melanogaster and D. simulans.. Genetics 140:1019-1032[Abstract].

BENKEL, B. F. and D. A. HICKEY, 1986  Glucose repression of amylase gene expression in Drosophila melanogaster.. Genetics 114:137-144[Abstract/Free Full Text].

BETRÁN, E., J. ROZES, A. NAVARRO, and A. BARBADILLA, 1997  The estimation of the number and the length distribution of gene conversion tracts from population DNA sequence data. Genetics 146:89-99[Abstract].

BOER, P. M. and D. A. HICKEY, 1986  The alpha-amylase gene in Drosophila melanogaster: nucleotide sequence, gene structure and expression motifs. Nucleic Acids Res. 14:8399-8411[Abstract/Free Full Text].

CADIGAN, K. M., U. GROSSNIKLAUS, and W. J. GEHRING, 1994  Functional redundancy: the respective roles of the two sloppy paired genes in Drosophila segmentation. Proc. Natl. Acad. Sci. USA 91:6324-6328[Abstract/Free Full Text].

DA LAGE, J. L., and M. L. CARIOU, 1993 Organization and structure of the amylase gene family, pp. 171–181 in Drosophila ananassae, Genetical and Biological Aspects, edited by Y. N. TOBARI. Japan Societies Press, Karger, Tokyo.

DAÏNOU, O., M. L. CARIOU, J. M. GOUX, and J. R. DAVID, 1993  Amylase polymorphism in Drosophila melanogaster: haplotype frequencies in tropical African and American populations. Genet. Sel. Evol. 25:133-151.

DAVID, J. R. and P. CAPY, 1988  Genetic variation of Drosophila melanogaster natural populations. Trends Genet. 4:106-111[Medline].

DE JONG, G. A. and W. SCHARLOO, 1976  Environmental determination of selective significance or neutrality of amylase variants in Drosophila melanogaster.. Genetics 84:77-94[Abstract/Free Full Text].

DOANE, W. W., 1969  Amylase variants in Drosophila melanogaster: linkage studies and characterization of enzyme extracts. J. Exp. Zool. 171:321-342.

DOANE, W. W., 1980  Selection for amylase allozymes in Drosophila melanogaster: some questions. Evolution 34:868-874.

FAY, J. C. and C. I. WU, 2000  Hitchhiking under positive Darwinian selection. Genetics 155:1405-1413[Abstract/Free Full Text].

FU, Y. X. and W. H. LI, 1993  Statistical tests of neutrality of mutations. Genetics 133:693-709[Abstract].

GILBERT, D. G., 1996 SeqPup, biosequence editor and analysis platform. Bionet. Software.

GOLDSTEIN, D. B. and A. G. CLARK, 1995  Microsatellite variation in North American populations of Drosophila melanogaster.. Nucleic Acids Res. 23:3882-3886[Abstract/Free Full Text].

HALE, L. R. and R. S. SINGH, 1991  A comprehensive study of genetic variation in natural populations of Drosophila melanogaster. IV. Mitochondrial DNA variation and the role of history vs. selection in the genetic structure of geographic populations. Genetics 129:103-117[Abstract].

HICKEY, D. A., 1977  Selection for amylase allozymes in Drosophila melanogaster.. Evolution 31:800-804.

HICKEY, D. A., 1979  The geographical pattern of an enzyme polymorphism in D. melanogaster.. Genetica 51:1-4.

HICKEY, D. A., L. BALLY-CUIF, S. ABUKASHAWA, V. PAYANT, and B. F. BENKEL, 1991  Concerted evolution of duplicated protein-coding genes in Drosophila.. Proc. Natl. Acad. Sci. USA 88:1611-1615[Abstract/Free Full Text].

HUDSON, R. R., 1987  Estimating the recombination parameter of a finite population model without selection. Genet. Res. 50:245-250[Medline].

HUDSON, R. R., M. KREITMAN, and M. AGUADÉ, 1987  A test of neutral molecular evolution based on nucleotide data. Genetics 116:153-159[Abstract/Free Full Text].

HUDSON, R. R., M. SLATKIN, and W. P. MADDISON, 1992  Estimation of levels of gene flow from DNA sequence data. Genetics 132:583-589[Abstract].

INOMATA, N. and T. YAMAZAKI, 2000  Evolution of nucleotide substitutions and gene regulation in the amylase multigenes in Drosophila kikkawai and its sibling species. Mol. Biol. Evol. 17:601-615[Abstract/Free Full Text].

INOMATA, N., K. KANDA, M. L. CARIOU, H. TACHIDA, and T. YAMAZAKI, 1995a  Evolution of the response pattern to dietary carbohydrates and the developmental differentiation of gene expression of -amylase in Drosophila.. J. Mol. Evol. 41:1076-1084[Medline].

INOMATA, N., H. SHIBATA, E. OKUYAMA, and T. YAMAZAKI, 1995b  Evolutionary relationships and sequence variation of -amylase variants encoded by duplicated genes in the Amy locus of Drosophila melanogaster.. Genetics 141:237-244[Abstract].

INOMATA, N., H. TACHIDA, and T. YAMAZAKI, 1997  Molecular evolution of the Amy multigenes in the subgenus Sophophora of Drosophila.. Mol. Biol. Evol. 14:942-950[Abstract].

KIKKAWA, H., 1964  An electrophoretic study on amylase in Drosophila melanogaster.. Jpn. J. Genet. 39:401-411.

KIMURA, M., 1980  A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16:111-120[Medline].

KLARENBERG, A. J., A. J. S. VISSER, M. F. M. WILLEMSE, and W. SCHARLOO, 1986  Genetic localization and action of regulatory genes and elements for tissue-specific expression of -amylase in Drosophila melanogaster.. Genetics 114:1131-1145[Abstract/Free Full Text].

KREITMAN, M. and M. AGUADÉ, 1986  Excess polymorphism at the Adh locus in D. melanogaster.. Genetics 114:93-110[Abstract/Free Full Text].

KREITMAN, M. and R. R. HUDSON, 1991  Inferring the histories of the Adh and Adh-dup loci in Drosophila melanogaster from patterns of polymorphism and divergence. Genetics 127:565-582[Abstract].

LAI, Z. C. and G. M. RUBIN, 1992  Negative control of photoreceptor development in Drosophila by the product of the YAN gene, an ETS domain protein. Cell 70:609-620[Medline].

LANGLEY, C. H., A. E. SHRIMPTON, T. YAMAZAKI, N. MIYASHITA, and Y. MATSUO et al., 1988  Naturally occurring variation in the restriction map of the Amy region of Drosophila melanogaster.. Genetics 119:619-629[Abstract/Free Full Text].

LEVY, J. N., R. M. GEMMILL, and W. W. DOANE, 1985  Molecular cloning of -amylase genes from Drosophila melanogaster. II. Clone organization and verification. Genetics 110:313-324[Abstract/Free Full Text].

MAGOULAS, C., L. BALLY-CUIF, A. LOVERRE-CHYURLIA, B. BENKEL, and D. HICKEY, 1993  A short 5'-flanking region mediates glucose repression of Amylase gene expression in Drosophila melanogaster.. Genetics 134:507-515[Abstract].

MATSUO, Y. and T. YAMAZAKI, 1984  Genetic analysis of natural populations of Drosophila melanogaster in Japan IV. Natural selection on the inducibility, but not on the structural genes, of Amylase loci. Genetics 108:879-896[Abstract/Free Full Text].

MCDONALD, J. H. and M. KREITMAN, 1991  Adaptive protein evolution at the Adh locus in Drosophila.. Nature 351:652-654[Medline].

MORIYAMA, E. N. and J. R. POWELL, 1996  Intraspecific nuclear DNA variation in Drosophila.. Mol. Biol. Evol. 12:261-277.

NEI, M., 1987 Molecular Population Genetics. Columbia University Press, New York.

OHTA, T., 1982  Allelic and non-allelic homology of a supergene family. Proc. Natl. Acad. Sci. USA 79:3251-3254[Abstract/Free Full Text].

OKUYAMA, E., H. SHIBATA, H. TACHIDA, and T. YAMAZAKI, 1996  Molecular evolution of the 5'-flanking regions of the duplicated Amy genes in Drosophila melanogaster species subgroup. Mol. Biol. Evol. 13:574-583[Abstract].

OKUYAMA, E., H. TACHIDA, and T. YAMAZAKI, 1997  Molecular analysis of the intergenic region of the duplicated Amy genes of Drosophila melanogaster and Drosophila teissieri.. J. Mol. Evol. 45:32-42[Medline].

POWELL, J. R. and M. ANDJELKOVIC, 1983  Population genetics of Drosophila amylase. IV. Selection in laboratory populations maintained on different carbohydrates. Genetics 103:675-689[Abstract/Free Full Text].

POWELL, J. R., M. RICO, and M. ANDJELKOVIC, 1980  Population genetics of Drosophila amylase III. Interspecific variation. Evolution 34:209-213.

ROZAS, J. and R. ROZAS, 1999  DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15:174-175[Abstract/Free Full Text].

SAIKI, R. K., S. SCHARF, F. FALOONA, K. B. MULLIS, and G. T. HORN et al., 1985  Enzymatic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemia. Science 230:1350-1354[Abstract/Free Full Text].

SAMBROOK, J., E. F. FRITSCH and R. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SCHLÖTTERER, C., R. RITTER, and G. BRAM, 1998  High mutation rate of a long microsatellite allele in Drosophila melanogaster provides evidence for allele-specific mutation rates. Mol. Biol. Evol. 15:1269-1274[Abstract].

SHIBATA, H. and T. YAMAZAKI, 1994  A comparative study of the enzymological features of -amylase in the Drosophila melanogaster species subgroup. Jpn. J. Genet. 69:251-258[Medline].

SHIBATA, H. and T. YAMAZAKI, 1995  Molecular evolution of the duplicated Amy locus in the Drosophila melanogaster species subgroup: concerted evolution only in the coding region and an excess of nonsynonymous substitutions in speciation. Genetics 141:223-236[Abstract].

SINGH, R. S., D. A. HICKEY, and J. DAVID, 1982  Genetic differentiation between geographically distant populations of Drosophila melanogaster.. Genetics 101:235-256[Abstract/Free Full Text].

TAJIMA, F., 1989  Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123:585-595[Abstract/Free Full Text].

THOMPSON, J. D., T. J. GIBSON, F. PLEWNIAK, F. JEANMOUGIN, and D. G. HIGGINS, 1997  The CLUSTAL X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25(24):4876-4882[Abstract/Free Full Text].

VAN DE PEER, Y. and R. DE WACHTER, 1994  TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Comput. Applic. Biosci. 10:569-570[Free Full Text].

WALL, J. D., 1999  Recombination and the power of statistical tests of neutrality. Genet. Res. 74:64-79.

WOODRUFF, R. C., J. N. THOMPSON, JR., H. N. B. GOPALAN, W. A. NGURE, E. S. NORRIS et al., 1990 Genetic and molecular analyses of P DNA elements sampled from natural populations of Drosophila melanogaster in Kenya, pp. 117–126 in Molecular Evolution: Proceedings of a UCLA Colloquium, edited by M. T. CLEGG and S. J. O'BRIEN. Wiley-Liss, New York.

YAMAZAKI, T., 1986  Genetic analysis of natural populations of Drosophila melanogaster in Japan. V. Genetic variabilities of amylase activities in different developmental stages and their relation to fitness. Jpn. J. Genet. 61:329-336.

YAMAZAKI, T. and Y. MATSUO, 1984  Genetic analysis of natural populations of Drosophila melanogaster in Japan. III. Genetic viability of amylase inducibility and fitness. Genetics 108:223-235[Abstract/Free Full Text].

YAMAZAKI, T., and Y. MATSUO, Y. INOUE, Y. MATSUO, 1984  Genetic analysis of natural populations of Drosophila melanogaster in Japan. I. Protein polymorphism, lethal gene, sterility gene, inversion polymorphism, and linkage disequilibrium. Jpn. J. Genet. 59:33-49.

YARDLEY, D. G., W. W. ANDERSON, and H. E. SCHAFFER, 1977  Gene frequency changes at the -amylase locus in experimental populations of Drosophila pseudoobscura.. Genetics 87:357-369[Abstract/Free Full Text].

This article has been cited by other articles:

				H. Araki, S. Yoshizumi, N. Inomata, and T. Yamazaki Genetic Coadaptation of the Amylase Gene System in Drosophila melanogaster: Evidence for the Selective Advantage of the Lowest AMY Activity and of Its Epistatic Genetic Background J. Hered., July 1, 2005; 96(4): 388 - 395. [Abstract] [Full Text] [PDF]

				K. Thornton and M. Long Excess of Amino Acid Substitutions Relative to Polymorphism Between X-Linked Duplications in Drosophila melanogaster Mol. Biol. Evol., February 1, 2005; 22(2): 273 - 284. [Abstract] [Full Text] [PDF]

				D. Charlesworth, C. Bartolome, M. H. Schierup, and B. K. Mable Haplotype Structure of the Stigmatic Self-Incompatibility Gene in Natural Populations of Arabidopsis lyrata Mol. Biol. Evol., November 1, 2003; 20(11): 1741 - 1753. [Abstract] [Full Text] [PDF]

				H. Innan The Coalescent and Infinite-Site Model of a Small Multigene Family Genetics, February 1, 2003; 163(2): 803 - 810. [Abstract] [Full Text] [PDF]

				H. Innan A Method for Estimating the Mutation, Gene Conversion and Recombination Parameters in Small Multigene Families Genetics, June 1, 2002; 161(2): 865 - 872. [Abstract] [Full Text] [PDF]

				N. Inomata and T. Yamazaki Nucleotide Variation of the Duplicated Amylase Genes in Drosophila kikkawai Mol. Biol. Evol., May 1, 2002; 19(5): 678 - 688. [Abstract] [Full Text] [PDF]