Geographic Selection in the Small Heat Shock Gene Complex ...

Journal of Heredity 2012:103(3):400–407 doi:10.1093/jhered/esr150 Advance Access publication February 16, 2012

Ó The American Genetic Association. 2012. All rights reserved. For permissions, please email: [email protected].

Geographic Selection in the Small Heat Shock Gene Complex Differentiating Populations of Drosophila pseudoobscura ALLIE M. GRAHAM, JENNIFER D. MERRILL, SUZANNE E. MCGAUGH,

AND

MOHAMED A. F. NOOR

From the Biology Department, Duke University, Durham, NC 27708. Address correspondence to Allie M. Graham at the address above, or e-mail: [email protected].

Abstract Environmental temperature plays a crucial role in determining a species distribution and abundance by affecting individual physiological processes, metabolic activities, and developmental rates. Many studies have identified clinal variation in phenotypes associated with response to environmental stresses, but variation in traits associated with climatic adaptation directly attributed to sequence variation within candidate gene regions has been difficult to identify. Insect heat shock genes are possible agents of thermal tolerance because of their involvement in protein folding, traffic, protection, and renaturation at the cellular level in response to temperature stress. Previously, members of the Drosophila small heat shock protein (sHSP) complex (Hsp23, Hsp26, Hsp27, Hsp67Ba) have been implicated as candidate climatic adaptation genes; therefore, this research examines sequence variation at these genes in 2 distant populations of Drosophila pseudoobscura. Flies from Tempe, AZ (n 5 30) and Cheney, WA (n 5 17) were used in the study. We identify high differentiation in the heat-shock complex (FST : 0.219**, 0.262*, 0.279***, 0.166 not significant) as compared with neighboring genes and Tajima’s D values indicative of balancing selection (Mann–Whitney U 5 38, n1 5 10 n2 5 4, P , 0.05 two-tailed), both of which are suggestive of such climatic adaptation. Key words: balancing selection, environmental adaptability

Environmental temperature impacts the distribution and abundance of organisms by simultaneously affecting physiological processes, biophysical structures, and metabolic activities, as well as developmental rates and growth (see review by Sinclair et al. 2003). Many species possess genetic variation in response to environmental stress, and heritable differences among populations may reflect local adaptation (Hoffmann, Scott, et al. 2003; Hoffmann, Sørenson, et al. 2003). Adaptation to higher or lower temperatures can allow organisms to expand their range or to acclimate to a changing environment. The ability of a species to adapt to new environments depends on the nature and amount of variation in genes affecting environmental sensitivity. Drosophila species have served as a model system for identifying candidate genes with allele frequencies correlated to specific changes in the environment (e.g., Van Delden and Kamping 1989; McKechnie et al. 1998; Bettencourt et al. 2002). One way to find such genes is to examine field populations for an association of trait variation with environmental variables. Many studies have investigated, and found, variation in response to environmental stresses (cold, heat, starvation, desiccation) in Drosophila (James et al. 1997; McColl and

400

McKechnie 1999; Van’t Land et al. 2000; Verrelli and Eanes 2001; Hoffmann et al. 2002; Frydenberg et al. 2003; Schmidt and Paaby 2008). However, few studies have identified variation at the sequence level for known environmental stress resistance traits in Drosophila species (e.g., Durvernell et al. 2003; Frydenberg et al. 2003, 2010; Fry et al. 2007). Heat shock genes may affect clinal thermal tolerance because they alter protein folding, traffic, protection, and renaturation at the cellular level in response to temperature stress (Neupert 1997; Feder and Hofmann 1999; Hartl and Hayer-Hartl 2002; Morrow et al. 2006). This class of genes is comprised of 5 distinct conserved groups and is classified according to size: small heat shock proteins (sHSPs), HSP60, HSP70, HSP90, and HSP100. In Drosophila melanogaster, the sHSP family consists of 7 genes at Locus 67B: Hsp22, Hsp23, Hsp26, Hsp27, Hsp67Ba, Hsp67Bb, and Hsp67Bc (Corces et al. 1980; Ayme and Tissieres 1985). This group is the least conserved of the Hsp families and range in molecular weight from 10 to 40 kDa. This group is also characterized by the presence of a C-terminal alphacrystallin domain and a native oligomeric structure (de Jong et al. 1998). Previous research in other organisms suggests that sHSPs act like molecular chaperones (Ehrnsperger

Graham et al. Geographic Selection in the Small Heat Shock Gene Complex

et al. 1997; Haslbeck et al. 1999; Fernando and Heikkila 2000) and can confer thermotolerence (Landry et al. 1989; Rollet et al. 1992; Kitagawa et al. 2002). Three members of the sHSP complex (Hsp23, Hsp26, Hsp27) have been implicated as candidate climatic adaptation genes (Frydenberg et al. 1999, 2003). Because of their extensive variation and important functions, Drosophila sHSP genes are excellent candidates for affecting thermal tolerance and local adaptation. Here, we test for differentiation among 2 D. pseudoobscura populations (Tempe, AZ and Cheney, WA) at 4 genes within this complex. The 2 collection sites exhibit widely different environmental conditions (Supplementary Table 2): December average daily low/high temperatures of 38–68 °F and July average 75–105 °F in Tempe, AZ in contrast with December averages of 25–34 °F and July average of 56–83 °F in Cheney, WA.

Materials and Methods Study Population The study population consisted of 30 wild-caught flies in 1996 from Tempe, AZ and 17 flies caught from Washington (WA). There were also 3 additional flies from Goldendale, WA collected from 1996, which were included in some aspects of data analyses ( Tajima’s D). Eight of the Washington flies were lab reared from isofemale lines established in 1996 from Turnbull Wildlife Refuge in Cheney, WA (Old-WA); the remaining 9 were wild-caught from Cheney, WA in September 2010 (New-WA). For the New-WA sample, a total of 46 wild-caught flies were sent from Cheney, WA, and their species identified based on sex comb morphology (Beckenbach and Prevosti 1986). Based on physical identification as belonging to the D. pseudoobscura subgroup, the flies were then genotyped using speciesspecific primers for D. pseudoobscura and D. persimilis. Of 13 identified, 12 were confirmed as D. pseudoobscura (New-WA). All flies used in this study were males. Study Genes Based on Muller’s element homologies and sequence similarity, the 4 focal sHSP genes (Hsp23, Hsp26, Hsp27, and Hsp67Ba) were inferred to be on the XR-chromosome arm in D. pseudoobscura (Flybase; Richards et al. 2005). For comparison with the sHSP gene complex, we sequenced and analyzed variation at a total of 5 control genes, 2 of which were flanking the sHSP region. The first control gene, GA11728, is in XR_group3a (1356325. . .1355496) and is 88 225 bp away from Hsp27, which is at the beginning of the sHSP complex (1443718. . .1451417). There are 12 genes in between Hsp27 and GA11728. The second control gene, GA10408, is in XR_group8 (3852168. . .3853226), however, the exact distance to the sHSP complex is unknown due to a break in the assembly. These 2 control genes are uncharacterized and free of any transposon repeats. All genes were free of introns except for Hsp67Ba

and GA11728. The 3 remaining control genes all had known gene function and were on the XR-chromosome arm: GA23426 (XR_group6: 10,150,108..10,150,867), GA23862 (XR_group6: 9,865,266..9,865,719), and GA20470 (XR_group6: 118,281..118,637). These genes were selected for 1) their locations in regions of moderately low recombination (,4 cM/Mb) for this species’ genome, yet 2) moderately high variation (p . 0.01) in a sample of 10 individuals from diverse populations (McGaugh SM, personal communication). As such, these additional control genes have general characteristics similar to what we observed in the sHsp genes studied (see below). Sequencing and Alignment The 4 sHSP gene regions and 5 control genes were sequenced through the Genome Sequencing and Analysis Core Facility at Duke University on the Applied Biosystems 3730 XL. The output was transferred into DNASTAR Lasergene v6-SeqMan (Swindell and Plasterer 1996) to assemble the contigs and was checked for discrepancies in sequence between forward and reverse reactions. The discrepancies were then labeled as ‘‘N’’ if the sequences for a particular strain were unclear. Assembled contigs were imported to BioEdit (Hall 1999), where a ClustalW alignment was performed, and the finished alignment was then imported to DnaSP v 5.10.01 (Librado and Rozas 2009). Any individuals with large sequence gaps (N or -) or single nucleotide polymorphism (SNP) regions after alignment were resequenced, and the new sequences combined with previous sequencing data for that individual to ensure maximum coverage. The average sequence lengths for the study genes are 555 bp for Hsp23, 615 bp for Hsp27, 1479 bp for Hsp67Ba, and 495 bp for Hsp26. The average sequence lengths for the control genes are 1024 bp for GA10408, 700 bp for GA11728, 317 bp for GA20470, 643 bp for GA23426, and 316 bp for GA23862. Sequences were deposited in the EMBL/GenBank under accession numbers JQ176688–JQ177059. Test for Concerted Evolution/Nonhomologous Gene Conversion The sequences for the sHSP genes in D. melanogaster and D. pseudoobscura were all compared using NCBI—Blast2Seq (blastn). If the D. pseudoobscura sequence was more similar to the D. melanogaster ortholog than to the D. pseudoobscura paralogs, then we inferred little concerted evolution had occurred. In contrast, the D. pseudoobscura sequence being more similar to a D. pseudoobscura paralogs would constitute evidence for concerted evolution. Levels of nonallelic gene conversion in the sample sets were also inferred using GENECONV v.1.81 (http:// www.math.wustl.edu/~sawyer/geneconv; Sawyer 1989). This program establishes significance of highly similar tracts, which represent conversion events, using permutation. The program was run using default settings except for the options required to display pairwise P values (-ListPair) and to include

401

Journal of Heredity 2012:103(3) Table 1

Polymorphic statistics for comparisons of populations of Drosophila pseudoobscura from Arizona and Washington Hsp23

All-WA New-WA Old-WA AZ AZ/All-WA Old-WA/New-WA

Hsp26

Seq. #

p

Seg. sites

FST

Seq. #

p

Seg. sites

FST

16 11 5 25 — —

0.0034 0.00315 0.00406 0.00738 — —

5 4 6 7 — —

— — — — 0.219** 0.057

17 12 5 25 — —

0.00552 0.00389 0.00715 0.00509 — —

13 10 9 10 — —

— — — — 0.262* 0.266***

Hsp27

All-WA New-WA Old-WA AZ AZ/All-WA Old-WA/New-WA

Hsp67Ba

Seq. #

p

Seg. sites

FST

Seq. #

p

Seg. sites

FST

17 12 5 23 — —

0.00077 0.00067 0.00283 0.00321 — —

2 2 3 5 — —

— — — — 0.279*** 0.177 ns

9 4 5 22 — —

0.00372 0.00318 0.00436 0.00805 — —

15 7 14 36 — —

— — — — 0.166 ns 0.104 ns

All-WA 5 all individuals from Washington, Old-WA 5 all individuals collected prior to 2011, New-WA 5 all individuals collected from 2011, ns 5 nonsignificant, Seq. # 5 sequence number, Seg. sites 5 segregating sites. *P , 0.05, **P , 0.01, and ***P , 0.001.

monomorphic sites (-Include-monosites). All fragments identified by pairwise comparisons with P , 0.05 were regarded as gene conversion tracts.

‘‘rapid permutations’’ was selected; therefore, a fixed number of permutations were calculated, or permutations were performed until 10 significant permutations were found that were higher than the observed P value (Bradbury 2007).

Data Analysis Pairwise measures of FST were calculated in Arlequin (Excoffier et al. 2005), and significance was assessed by comparing the results with 5000 permutations of haplotypes among populations calculate the P values associated with the FST values. DnaSP was also used to perform a Tajima’s D test. A negative D statistic indicates an excess of low frequency polymorphism consistent with purifying selection or population expansion, whereas a positive statistic could suggest balancing selection (Tajima 1989). The data set included 3 individuals from Goldendale, WA in conjunction with the individuals from Cheney, WA. The 4 coding regions were artificially concatenated in the order in which they are found in the genome. DnaSP was then used to identify linkage disequilibrium by calculating D, D#, R, and r2. To compute significance for associations between polymorphic sites, two-tailed Fisher’s Exact test and chi-squared test were calculated. For this study, instances of linkage disequilibrium were characterized by chi-square values that were significant by the Bonferroni procedure for multiple comparisons (P , 0.001). TASSEL v3 (Trait Analysis by aSSociation, Evolution and Linkage) was used to identify linkage disequilibrium (D#, r2, and P value) though pairwise combinations of SNPs. The sites in the data set were filtered, as required by the program. The P values for this data set were determined by a two-sided Fisher’s Exact test. Both r2 and D# values range from 0 to 1, where 0 represents linkage equilibrium and 1 indicates complete linkage disequilibrium. The option for

402

Results Test for Concerted Evolution/Nonhomologous Gene Conversion Two of the genes in D. pseudoobscura, Hsp23 and Hsp26, exhibited some significant sequence similarity (4E-61, 79%). However, overall the D. pseudoobscura sequences were more similar to the D. melanogaster ortholog than to the D. pseudoobscura paralogs (Supplementary Table 2). Therefore, we inferred little concerted evolution occurred in the recent history of these genes. Similarly, GENECONV found no globally significant similarity (inner or outer) among any of the genes. These data suggest there is no evidence of any recent conversion events among the genes. Gene Flow and Genetic Differentiation Since the D. pseudoobscura from Washington were collected at 2 different time periods, the groups were tested against each other to check for genetic differentiation. A large FST value between them would suggest that they fall into 2 separate groups; however, the values were not indicative of genetic differences (Table 1). As such, we pooled the old and new samples from Washington for further analysis. The FST measurements for the Washington samples versus the Arizona samples were 0.279 for Hsp27, 0.220 for Hsp23, 0.166 for Hsp67Ba, and 0.262 for Hsp26 (Table 1). The FST values for the control genes were well within


inferred FST values on the X chromosome, 0.01–0.09 (Noor et al. 2000): GA10408 (0.086, P , 0.05), GA11728 (0.098, P , 0.05), GA20470 (0.105, P , 0.05), GA23862 (0.051, P , 0.05), and GA23426 (0.105, P , 0.05). Thus, we observed evidence of strong differentiation between the Washington and Arizona samples at the experimental loci but weaker differentiation at the control loci. This differentiation is unusually high compared with that observed in other population genetic studies of D. pseudoobscura (e.g., Prakash et al. 1969; Schaeffer and Miller 1992). These findings were not a result of differentiation between the old and new Washington samples, as we saw no significant genetic differences between these samples FST values, except for Hsp26. Linkage Disequilibrium DnaSP identified a total of 130 nt positions exhibiting significant LD (P , 0.05). However, there were 11 instances of linkage disequilibrium between Hsp27/23, 4 instances between Hsp27/26, 9 instances between Hsp26/23, 11 instances between Hsp23/26, 1 instance within Hsp27, 7 instances within Hsp26, and 11 instances within Hsp23 (P , 0.001). Curiously, DnaSP failed to detect LD associated with Hsp67Ba and Hsp23, Hsp26, and Hsp27, despite it being in the middle of this complex. Nonetheless, there is extensive LD between and within Hsp23, Hsp26, and Hsp27 (Table 2). TASSEL also indicated a high level of LD within and between all of the Hsp genes (Figure 1). The discrepancy could be because of the use of multiple permutations used within TASSEL as compared with DnaSP. Tajima’s D Test Estimates of Tajima’s D from nucleotide sequences are influenced by both natural selection (with positive values indicating balancing selection) and demographic effects (with negative values resulting from population expansion). Previous studies of nucleotide sequence data in D. pseudoobscura have consistently observed negative Tajima’s D values (ranging from 0.6977 to 1.9945: see Machado et al. 2002), suggesting that the population has expanded in size. In testing for balancing selection on sHsps, evaluating statistical significance of a ‘‘positive’’ Tajima’s D value would be overly conservative given the demographic bias toward negative values. Hence, instead, we compare the Tajima’s D values for the sHsps with those for various control genes for statistical significance. Tajima’s D statistics showed a general positive trend in all 4 sHSPs compared with the control genes within D. pseudoobscura. The sHSPs had D values of 0.86505 for Hsp23, 0.00094 for Hsp26, 0.05499 for Hsp27, and 0.10752 for Hsp67Ba. The control genes had D values of 0.73928 for GA10408, 1.50523 for GA11728, 0.67291 for GA20470, 0.27871 for GA23862, and 0.04886 for GA23426. The moderately negative D values for the control genes are within range of typical values for D. pseudoobscura on the X chromosome (Machado et al. 2002). Hence, although not significantly greater than zero, the sHSP genes have

significantly higher Tajima’s D values than average compared with published gene sequences and with the control genes on either side of the complex (Mann–Whitney U 5 38, n1 5 10 n2 5 4, P , 0.05 two-tailed).

Discussion The results of the analysis suggest that Hsp23, Hsp26, Hsp27, and Hsp67Ba are genetically differentiated between D. pseudoobscura populations as a result of local natural selection. Supporting this conclusion, we observed higher FST at these sHSP loci than are typical for D. pseudoobscura populations (Machado et al. 2002), higher FST than neighboring genes on both sides, and higher FST than additional control genes with higher than expected within species diversity. We also observed extensive linkage disequilibrium within and between the sHSP complex genes. Linkage disequillibrium in D. pseudoobscura decays rapidly over a short distance (Schaeffer and Miller 1992); therefore, the persistence of high levels of LD over this 7.7 kB region seems indicative of selection. Furthermore, the contrast between the positive Tajima’s D estimate in this complex and the negative Tajima’s D values in most D. pseudoobscura genes implies that these genes are experiencing balancing selection. Balancing selection is known to preserve variation across heterogeneous landscapes, though only if fitness differences are large enough to overcome gene flow (Felsenstein 1976; Endler 1997). Such local selection may be expected given the sHSPs’ role in cellular response to temperature (Bettencourt et al. 2008; Colinet et al. 2010) and connection to certain developmental pathways (Berger 1984; Amin et al. 1991; Dubrovsky et al. 1994, 1996). Unlike other major HSP genes, members of the sHSP complex are not constitutively expressed but are inducible through temperature stress. The promoter regions of the sHSP genes are also similar in sequence, suggesting a similarity in function (Sirotkin and Davidson 1982; Mestril et al. 1984; Ayme and Tissieres 1985). In D. melanogaster, Hsp23, 26, 27, 67Ba were strongly induced via heat shock (Bettencourt et al. 2008; Gonsalves et al. 2011) but are also inducible by cold (Colinet et al. 2010), suggesting a general role in temperature dependant tolerance. The different sHSPs also show a cell-and stage-specific pattern of expression at developmental periods that do not necessarily overlap with stress events (Michaud et al. 1997, 2002). During puparium formation, genes in the sHSP complex are transcribed in elevated quantities at a time when the concentration of insect molting hormone ecdysterone is high (Ireland et al. 1982; Sirotkin and Davidson 1982; Zimmerman et al. 1983; Mason et al. 1984). Specifically, some of the genes in the sHSP complex have a role in primary (Hsp27) or secondary (Hsp23) response to ecdysone, a major prohormone of ecdysterone (Amin et al. 1991; Dubrovsky et al. 1996). This may explain the high linkage disequilibrium between Hsp27 and Hsp23 (Figure 1). The link to ecdysone could also be a link to

403

Journal of Heredity 2012:103(3) Table 2 DnaSP linkage disequilibrium with linked sites, genes associated with linked sites, distance between linked sites, D#, Fisher and Chi-Squared values Site 1

Gene (s1)

Site 2

Gene (s2)

Distance

D#

Fisher

Chi-square

16 16 16 16 16 16 16 16 99 237 237 237 237 237 237 237 598 667 868 868 868 868 868 868 882 882 882 882 882 952 952 952 952 970 970 970 1018 1018 2770 2847 2940 2940 2946 3018 3060

Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp27 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26

237 868 882 952 970 1018 3060 3066 904 868 882 952 970 1018 3060 3066 3111 700 882 952 970 1018 3060 3066 952 970 1018 3060 3066 970 1018 3060 3066 1018 3060 3066 3060 3066 2847 2946 3018 3167 3099 3167 3066

Hsp27 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp26 Hsp26 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp26 Hsp26 Hsp26 Hsp23 Hsp23 Hsp23 Hsp23 Hsp23 Hsp26 Hsp26 Hsp23 Hsp23 Hsp23 Hsp26 Hsp26 Hsp23 Hsp23 Hsp26 Hsp26 Hsp23 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26 Hsp26

221 852 866 936 954 1002 2938 2944 805 631 645 715 733 781 2717 2723 2407 33 14 84 102 150 2086 2092 70 88 136 2072 2078 18 66 2002 2008 48 1984 1990 1936 1942 75 99 78 227 153 149 6

1 1 1 1 1 1 0.725 1.725 1 1 1 1 1 1 1 1 1 0.876 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0.725 1 0.843 1

0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000*** 0.000*** 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.030* 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.000***B 0.006** 0.002** 0.000***B 0.000*** 0.001*** 0.000***B 0.000***B

21.656***B 25.109***B 25.109***B 25.109***B 25.109***B 25.109***B 15.249***B 15.249***B 26.518***B 28.463***B 28.463***B 28.463***B 28.463***B 28.463***B 19.038***B 19.038***B 33.000***B 25.320***B 33.000***B 33.000***B 33.000***B 33.000***B 22.074***B 22.074***B 33.000***B 33.000***B 33.000***B 22.074***B 22.074***B 33.000***B 33.000***B 22.074***B 22.074***B 33.000***B 22.074***B 22.074***B 23.074***B 22.074***B 21.290***B 18.480***B 22.074***B 15.249**B 15.714***B 17.837***B 33.000***B

B, significant by the Bonferroni procedure. *P , 0.05, **P , 0.01, and ***P , 0.001.

juvenile hormone, which regulates metamorphosis and diapause (Flatt et al. 2005). Entering into diapause would delay development, provide a means of evading stressful conditions, and allow the fly to survive unfavorable environmental extremes, such as temperature. The nature and occurrence of diapause induction has the potential for quick evolutionary shifts that might help to expand physiological limits of the species (Hoffmann 2010). According to Schmidt et al. (2005), diapause in populations of D. melanogaster is often polymorphic; though the ability of

404

D. pseudoobscura to undergo diapause is currently unknown. The sHSPs’ distinctive potential interactions with the major insect developmental hormones would make them more likely targets of selection for diapause induction (Schiesari et al. 2011) compared with major HSP genes, which are solely involved in combating temperature stress. The sHSP complex is still largely uncharacterized compared with other major heat shock genes in Drosophila. This is the first published research on the D. pseudoobscura sHSP complex, as well as the first characterization of


Figure 1. Linkage disequilibrium map for the sHsp (Hsp27, Hsp23, Hsp67Ba, Hsp26) complex created by TASSEL v3. SNPs are represented as vertical gray lines within and between genes. Sites 1–495 are Hsp27, 496–1057 are Hsp23, 1058–2557 are Hsp67Ba, and 2558–3282 are Hsp26. The matrix plot indicates significance by Fisher’s Exact test and includes correlation coefficients (r2) above the diagonal and significance values below the diagonal.

geographic selection within any of the sHSP genes for D. pseudoobscura. To elucidate thermal adaptability and heat protection, a combination of genetic, ecological, and physiological studies are needed. Additionally, the reason why Drosophila needs 4 structurally and functionally similar sHSPs is currently unclear (Morrow et al. 2006). The evidence presented in this article, in conjunction with what is known about the sHSPs’ cellular and developmental interactions, adds another interesting facet to the complexity of the sHSPs. A larger sample size and more populations between Arizona and Washington could be incorporated into future studies to best deduce the nature and extent of geographic variation. Including additional members of the sHSP complex (Hsp22, Hsp67Bb, and Hsp67Bc) in analyses would allow a better depiction of the selection on the complex as a whole. Also, functional tests of heat tolerance in these 2 populations could be performed in the future.

Supplementary Material Supplementary material can be found at http://www. jhered.oxfordjournals.org/.

Funding National Institutes of Health (GM076051 and GM086445). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Acknowledgments We would like to thank Dr L. Matos for providing some of the wild-caught flies from Washington and Dr C. Casola for his help on using GENECONV. Three anonymous reviewers and Dr A. Caccone provided helpful feedback on the manuscript. Additional thanks go out to all members of the Noor lab for their camaraderie and general advice through all stages of this research.

References Amin J, Mestril R, Voellmy R. 1991. Genes for Drosophila small heat shock proteins are regulated by ecdysterone. Mol Cell Biol. 11:5937–5944. Ayme A, Tissie`res A. 1985. Locus 67Ba of Drosophila melanogaster contains seven, not four, closely related heat shock genes. EMBO J. 4(11): 2949–2954.

405

Journal of Heredity 2012:103(3) Beckenbach AT, Prevosti A. 1986. Colonization of North America by the European species, Drosophila subobscura and D. ambigua. Am Midl Nat. 115:10–18.

Frydenberg J, Pierpaoli M, Loeschke V. 1999. Drosophila melanogaster is polymorphic for a specific repeated (CATA) sequence in the regulatory region of hsp23. Gene. 236(2):243–250.

Berger EM. 1984. The regulation and function of small heat-shock protein synthesis. Dev Genet. 4:255–265.

Gonsalves SE, Moses AM, Razak Z, Robert F, Westwood JT. 2011. Wholegenome analysis reveals that active heat factor binding sites are mostly associated with heat shock genes in Drosophila melanogaster. PLoS One. 6(1):1–21.

Bettencourt BR, Hogan CC, Nimali M, Drohan BW. 2008. Inducible and constitutive heat shock gene expression modification of Hsp70 copy number in Drosophila melanogaster but does not compensate for loss of thermotolerance in Hsp70 null flies. BMC Biol. 6:5. Bettencourt BR, Kim IY, Hoffmann AA, Feder ME. 2002. Response to natural and laboratory selection at the Drosophila Hsp70 genes. Evolution. 56:1796–1801. Bradbury PJ. 2007. TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics. 23:2633–2635. Colinet H, Lee SF, Hoffmann A. 2010. Knocking down expression of Hsp22 and Hsp23 by RNA interference affects recovery from chill coma in Drosophila melanogaster. J Exp Biol. 213:4148–4150. Corces V, Holmgren R, Freund R, Morimoto R, Meselson M. 1980. Four heat shock proteins of Drosophila melanogaster coded within a 12-kilobase region in chromosome subdivision 67B. Proc Natl Acad Sci U S A. 77(9):5390–5393. de Jong WW, Caspers GJ, Leunissen JA. 1998. Genealogy of the alpha-crystallinsmall heat-shock protein superfamily. Int J Biol Macromol. 22:151–162. Dubrovsky EB, Dretzen G, Bellard M. 1994. The Drosophila Broad-complex regulates developmental changes in transcription and chromatin structure of the 87B heat-shock gene cluster. J Mol Biol. 241(3):353–562. Dubrovsky EB, Dretzen G, Berger EM. 1996. The broad-complex gene is tissue specific modulator of the ecdysone response of the Drosophila hsp23 gene. Mol Cell Biol. 16:6542–6552. Durvernell DD, Schmidt PS, Eanes WF. 2003. Clines and adaptive evolution in the methuselah gene region in Drosophila melanogaster. Mol Ecol. 12:1277–1285. Ehrnsperger M, Graber A, Gaestel M, Buchner J. 1997. Binding of nonnative protein to Hsp25 during heat shock creates a reservoir of folding intermediates for reactivation. EMBO J. 16:221–229. Endler JA. 1997. Geographic variation, speciation and clines. Princeton (NJ): Princeton University Press. Excoffier L, Laval G, Schneider S. 2005. Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evol Bioinform Online. 1:47–50. Feder ME, Hofmann AA. 1999. Heat-shock proteins, molecular chaperones and the stress response: evolutionary and ecological physiology. Annu Rev Physiol. 61:243–282. Felsenstein J. 1976. The theoretical population genetics of variable selection and migration. Annu Rev Genet. 10:253–280. Fernando P, Heikkila JJ. 2000. Functional characterization of Xenopus small heat shock protein, Hsp30C: the carboxyl end is required for stability and chaperone activity. Cell Stress Chaperones. 5:48–59. Flatt T, Tu M-P, Tatar M. 2005. Hormonal pleiotropy and juvenile hormone regulation of Drosophila development and life history. BioEssays. 27:999–1010.

Hall T. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser. 41:95–98. Hartl FU, Hayer-Hartl M. 2002. Molecular chaperones in the cytosol: from nascent chain to folded protein. Science. 295:1852–1858. Haslbeck M, Walke S, Stromer T, Ehrsperger M, White HE, Chen S, Saibil HR, Buchner J. 1999. Hsp 26: a temperature-regulated chaperone. EMBO J. 18:6744–6751. Hoffmann AA. 2010. Physiological climatic limits in Drosophila: patterns and implications. J Exp Biol. 213:870–880. Hoffmann AA, Anderson A, Hallas R. 2002. Opposing clines for high and low temperature resistance in Drosophila melanogaster. Ecol Lett. 5:614–618. Hoffmann AA, Scott M, Patridge L, Hallas R. 2003. Overwintering in Drosophila melanogaster: outdoor field cage experiments on clinal and laboratory selected populations help to elucidate traits under selection. J Evol Biol. 16:614–623. Hoffmann AA, Sørenson JG, Loeschcke V. 2003. Adaptation of Drosophila to temperature extremes: bringing together quantitative and molecular approaches. J Therm Biol. 28:175–216. Ireland R, Berger E, Sirotkin K, Yund M, Osterbur D, Fristrom J. 1982. Ecdysterone induces the transcription of four heat-shock genes in Drosophila S3 cells and imaginal discs. Dev Biol. 93:498–507. James AC, Azevedo RBE, Patridge L. 1997. Genetic and environmental reponses to temperature of Drosophila melanogaster from a latitudinal cline. Genetics. 146:881–890. Kitagawa M, Miyakawa M, Matsumura Y, Tsuchido T. 2002. Escherchia coli small heat shock proteins, IbpA and IbpB, protect enzymes from inactivation by heat and oxidants. Eur J Biochem. 269:2907–2917. Landry J, Chretien P, Lambert H, Hickey E, Weber LA. 1989. Heat shock resistance conferred by expression of the human HSP27 gene in rodent cells. J Cell Biol. 109:7–15. Librado P, Rozas J. 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics. 25:1451–1452. Machado CA, Kliman RA, Makert JA, Hey J. 2002. Inferring the history of speciation from multilocus DNA sequence data: the case of Drosophila pseudoobscura and close relatives. Mol Biol Evol. 19(4):472–488. Mason P, Hall L, Gauz J. 1984. The expression of heat shock genes during normal development in Drosophila melanogaster (heat shock/abundant transcripts/developmental regulation). Mol Gen Genet. 194:73–78. McColl G, McKechnie SW. 1999. The Drosophila heat shock hsr-omega gene: an allele frequency cline detected by quantitative PCR. Mol Biol Evol. 16:1568–1574.

Fry JD, Donion K, Saweikis M. 2007. A worldwide polymorphism in aldehydedehydrogenase in Drosophila melanogaster: evidence for selection mediated by dietary ethanol. Evolution. 62(1):66–75.

McKechnie SW, Halford MM, McColl G, Hoffmann AA. 1998. Both allelic variation and expression of nuclear and cytoplasmic transcripts of Hsr-omega are closely associated with thermal phenotype in Drosophila. Proc Natl Acad Sci U S A. 95:2423–2428.

Frydenberg J, Barker SF, Loeschcke V. 2010. Characterization of the shsp genes in Drosophila buzzatii and association between the frequency of Valine mutations in hsp23 and climatic variables along the longitudinal gradient in Australia. Cell Stress Chaperones. 15:271–280.

Mestril R, Schiller P, Amin J, Klapper H, Ananthan J, Voellmy R. 1984. Heat shock and ecdysone activation of the Drosophila melanogaster hsp23 gene; a sequence element implied in developmental regulation. EMBO J. 5(7):1667–1673.

Frydenberg J, Hoffmann A, Loeschke V. 2003. DNA sequence variation and latitudinal associations in hsp23, hsp26, and hsp27 from natural populations of Drosophila melanogaster. Mol Ecol. 12:2025–2032.

Michaud S, Marin R, Tanguay RM. 1997. Regulation of heat shock gene induction and expression during Drosophila development. Cell Mol Life Sci. 53:104–113.

406

Graham et al. Geographic Selection in the Small Heat Shock Gene Complex Michaud S, Morrow G, Marchand J, Tanguay RM. 2002. Drosophila small heat shock proteins: cell and organelle specific chaperones? Prog Mol Subcell Biol. 28:79–101.

Schmidt PS, Paaby AB. 2008. Reproductive diapause and life history clines in North American populations of Drosophila melanogaster. Evolution. 62:1204–1215.

Morrow G, Heikkila JJ, Tanquay RM. 2006. Differences in the chaperonelike activities of the four main small heat shock proteins of Drosophila melanogaster. Cell Stress Chaperones. 11(1):51–60.

Sinclair BJ, Vernon P, Klok CJ, Chown SL. 2003. Insects at low temperatures: an ecological perspective. Trends Ecol Evol. 18(5): 257–262.

Neupert W. 1997. Protein import into mitochondria. Annu Rev Biochem. 66:863–917.

Sirotkin K, Davidson N. 1982. Developmentally regulated transcription from Drosophila melanogaster chromosomal site 67B. Dev Biol. 89:196–210.

Noor MAF, Schug MD, Aquadro CF. 2000. Microsatellite variation in populations of Drosophila pseudoobscura and Drosophila persimlis. Genet Res. 75:25–35. Prakash S, Lewontin RC, Hubby JL. 1969. A molecular approach to the study of genic heterozygosity in natural populations. IV. Patterns of genic variation in central, marginal and isolated populations of Drosophila pseudoobscura. Genetics. 145:311–323. Richards S, Yue L, Bettencourt BR. 2005. Comparative genome sequencing of Drosophila pseudoobscura: chromosomal, gene, and cis-element evolution. Genome Res. 15:1–18. Rollet E, Lavoie JN, Landry J, Tanguary RM. 1992. Expression of Drosophila’s 27 kDa heat shock protein into rodent cells confers thermal resistance. Biochem Biophys Res Commun. 185:116–120.

Swindell SR, Plasterer TN. 1996. SEQMAN: contig assembly. Methods Mol Biol. 70:75–89. Tajima F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics. 123:585–595. Van Delden W, Kamping A. 1989. The association between polymorphisms at the Adh and the aGpdh loci and the In(2 L)t. Drosophila melanogaster in relation to temperature. Evolution. 43:775–793. Van’t Land J, Van Putten WF, Villarroel H, Kamping A, Van Delden W. 2000. Latitudinal variation for two enzyme loci and an inversion polymorphism in Drosophila melanogaster from central and south America. Evolution. 54:201–209.

Sawyer SA. 1989. Statistical tests for detecting gene conversion. Mol Biol Evol. 6:526–538.

Verrelli BC, Eanes WF. 2001. Clinal variation for amino acid polymorphism at the Pgm locus is consistent with adaptive protein evolution in Drosophila melanogaster. Genetics. 156:1649–1663.

Schaeffer SW, Miller EL. 1992. Estimates of gene flow in Drosophila pseudoobscura determined from nucleotide sequence analysis of the alcohol dehydrogenase region. Genetics. 132:471–480.

Zimmerman J, Petri W, Meselson M. 1983. Accumulation of a specific subset of D. melanogaster heat shock mRNA’s in normal development without heat shock. Cell. 32:1161–1170.

Schiesari L, Kyriacou CP, Costa R. 2011. The hormonal and circadian basis for insect photoperiodic timing. FEBS Lett. 585(10):1450–1460. Schmidt PS, Matzkin L, Ippolito M, Eanes WF. 2005. Geographic variation in diapause incidence, life-history traits, and climatic adaptation in Drosophila melanogaster. Evolution. 59:1721–1732.

Received June 28, 2011; Revised December 1, 2011; Accepted December 1, 2011 Corresponding Editor: Adalgisa Caccone

407