Wing shape in the Drosophila simulans species complex. 

Macdonald, Stuart J.  Department of Ecology & Evolutionary Biology, University of California, Irvine, CA 92697-2525. Tel: 949-824-5994. Fax: 949-824-2181. Email: sjm@uci.edu

Introduction

      The Drosophila wing is seen as an ideal organ with which to examine developmental processes, and increasingly as an important quantitative genetic model system. Weber (1990) has demonstrated that substantial selectable genetic variation exists for wing shape in D. melanogaster, and QTL (Quantitative Trait Locus) mapping studies have shown that the trait is under the control of many small-effect genes (Weber et al., 1999, 2001), which can have different effects on separable components of wing shape (Zimmerman et al., 2000).

      Two recent studies have measured geographic variation in wing shape in natural populations: Gilchrist et al. (2000) measured shape changes along three body size clines on different continents in D. melanogaster, and Hoffmann and Shirriffs (2002) examined a single Australian cline in D. serrata. Both studies found that wing size and shape vary independently, and that shape changed within each cline, but that the change was different across clines. However, while shape variation was significant, the range of values was very small, possibly indicating stabilizing selection (Gilchrist et al., 2000). It is unclear what relevance these subtle differences in shape have in natural populations, as very little is known about the functional implications of variation in wing morphology (Grodnitsky, 1999).

      Thus far wing shape has only been investigated within Drosophila species, and the degree to which different Drosophilid species differ is unknown. Here I assess interspecies variation in wing shape using the three species of the D. simulans complex – D. simulans, D. mauritiana, and D. sechellia.

Materials and Methods

      Isofemale lines:  D. simulans – simiso1, simiso24 (obtained from A.G. Clark), S130, S131, S132, S134, S145 (European Drosophila Stock Center, Umea); D. mauritiana – 14021-0241.1, 14021-0241.2, 14021-0241.4, 14021-0241.5 (Tucson Stock Center), maur535 (A.G. Clark); D. sechellia – JDsec (J. David), 14021-0248.8, 14021-0248.15, 14021-0248.19 (Tucson Stock Center), S-9, S-32 (J. Roote).

      Fly culture:  Each of the 18 laboratory-adapted lines was maintained in vial culture on standard cornmeal/yeast/sugar medium, at 25°C on a 12:12 h light:dark cycle. To generate test flies a large population was allowed to lay, and the eggs collected into vials at a density of 30 eggs per vial. Parents of the test flies were cultured in the same way, such that the density was controlled for two generations.

      Wing size/shape data:  Ten males and 10 females were collected from each line (aside from the D. sechellia line S-32, for which only 5 males and 5 males could be collected due to low egg hatching success). Single wings were removed from each fly and mounted on microscope slides in DPX (Merck, Ltd).

      Images of wings were taken with a digital camera connected to a light microscope, and the PC software SigmaScan Pro v.4 (Jandel Corporation) was used to map six wing vein landmarks about the perimeter of the wing. Wing size was defined as the area of the polygon formed by joining the landmark coordinates with line segments.

      To investigate shape variation, landmarks from all test wings were subjected to a full Procrustes superimposition, implemented using the public domain software tpsSuper (version 1.07, F.J. Rohlf, http://life.bio.sunysb.edu/morph). This process uses iterative reflection, rotation and scaling of the original shape data to produce the best least square fit of all the landmarks of all the specimens, to minimize deviation from a consensus shape (described fully in Rohlf and Slice, 1990). In order to reduce the dimensionality of the wing shape data, the resulting size-independent Procrustes coordinates were rotated to their principal components, computed using the covariance matrix, using the free statistical programming language R (http://www.r-project.org).

Results and Discussion

      PCA (Principal Components Analysis) yields a number of independent PC (Principal Components) that account for different proportions of the total variation in the data.  Figure 1 plots PC1 against PC2, which account for 64.2% and 11.8%, respectively, of the total shape variation. The remaining components each account for less than 9% of the variation, and as such are less likely to be biologically meaningful.

      Figure 1 clearly shows that within each species the sexes are separable along the PC1 axis, with females consistently having higher values of PC1 than males. Plotting the mean of each of the six Procrustes xy-coordinates for each species and sex shows that males have very slightly broader wings than females (not shown), confirming previous results in D. melanogaster (Gilchrist et al., 2000).

      The PC1 axis (Figure 1) also shows that the D. sechellia lines cluster separately from all other lines (aside from the D. mauritiana line maur535 at the top of the figure). The shape difference between D. simulans and D. sechellia male wings is shown in Figure 2, highlighting the broader wings of D. sechellia (the same is also true for female wing shape). D. sechellia is endemic to the Seychelles archipelago, and has reduced molecular variation compared to its sibling species (Cariou et al., 1990;  Hey and Kliman, 1993). Therefore, it is possible the species originated by a founder event, with genetic drift largely responsible for the wing shape difference. However, without knowing any of the ecological factors that might influence Drosophila wing shape in nature, one cannot rule out the action of some form of selection.

      The PC2 axis (Figure 1) seems to represent within-species variation. How-ever, insufficient vari-ation exists along this axis to assess what aspect of shape this component may repre-sent.

      It is unfortunate that the ecological consequences, if any, of variation in wing shape are unknown. However, any trait that varies both between populations and species is clearly of interest from an evolutionary perspective. One of the attractive features of morphological variation in the D. simulans complex is its amenability to genetic analysis. The species can be intercrossed to yield fertile hybrids, enabling standard QTL mapping techniques to be employed, for instance backcross designs (Macdonald and Goldstein, 1999;  Zeng, 2000), and RI (Recombinant Inbred) line designs (Civetta et al., 2002). Studies of this nature would allow one to test whether morphological variation between species shows similar genetic architecture to variation within species.

      References:  Cariou, M.L., M. Solignac, M. Monnerot, and J.R. David 1990, Experientia 46: 101-104;  Civetta, A., H.M. Waldrip-Dail, and A.G. Clark 2002, Genet. Res. 79: 65-74;  Gilchrist, A.S., R.B.R. Azevedo, L. Partridge, and P. O’Higgins 2000, Evol. Dev. 2: 114-124;  Grodnitsky, D.L., 1999 Form and Function of Insect Wings, Johns Hopkins University Press;  Hey, J., and R.M. Kliman 1993, Mol. Biol. Evol. 10: 804-822;  Hoffmann, A.A., and J. Shirriffs 2002, Evolution 56: 1068-1073;  Macdonald, S.J., and D.B. Goldstein 1999, Genetics 153: 1683-1699;  Rohlf, F.J., and D. Slice 1990, Syst. Zool. 39: 40-59;  Weber, K.E., 1990, Genetics 126: 975-989;  Weber, K.E., R. Eisman, L. Morey, A. Patty, J. Sparks, M. Tausek, and Z.B. Zeng 1999, Genetics 153: 773-786;  Weber, K.E., R. Eisman, S. Higgins, L. Morey, A. Patty, M. Tausek, and Z.B. Zeng 2001, Genetics 159: 1045-1057;  Zeng, Z.B., J. Liu, L.F. Stam, C.H. Kao, J.M. Mercer, and C.C. Laurie 2000, Genetics 154: 299-310;  Zimmerman, E., A. Palsson, and G. Gibson 2000, Genetics 155: 671-683.