Heredity 64 (1990) 255—262 The Genetical Society of Great Britain Received 6 September 1989 Developmental anomalies in Drosophila hybrids are apparently caused by loss of microchromosome H. Allen Orr Department of Ecology and Evolution, University of Chicago, 1103 E. 57th Street, Chicago, IL 60637, U.S.A. Hybrids produced by crossing Drosophila virilis females to D. lummei males suffer from many developmental anomalies; the reciprocal hybridization yields normal offspring. Genetic analysis reveals that these anomalies involve a maternal effect: whether or not an individual will show an anomaly depends upon his mother's nuclear genotype. Several lines of evidence suggest that the proximal cause of the anomalies is the elimination of the D. lummei microchromosome (chromosome 6) from hybrids. Loss of the D. lum,nei microchromosome in this hybridization is known to involve a maternal effect (Evgen'ev, 1973), as mitosis in early development is under the control of maternally-acting genes. INTRODUCTION architecture" of morphological anomalies will resemble the architecture of normal morphological Developmentaland morphological anomalies are differences between species. common among species hybrids. Drosophila During a study of the basis of postzygotic isola- melanogaster-D. simulans hybrids, for example, tion in the yin/is group of Drosophila (Orr and often suffer from missing bristles and incomplete Coyne, 1989), we noticed that hybrids between D. scierotization of the abdomen (Sturtevant, 1920, virilis and D. lummei frequently showed several 1929). Unfortunately, almost nothing is known morphological anomalies. The following about the genetic causes of such hybrid anomalies. anomalies were observed: greatly reduced eye Elucidation of the basis of these anomalies may resembling the "eyeless" mutation of D. throw some light on speciation. In particular, it tnelanogaster (usually only one eye affected), will be useful to determine whether the genetic unequal wing lengths, broken or incomplete wing basis of developmental anomalies more closely veins, twisted abdomen, and missing or reduced resembles the basis of hybrid inviability/sterility thoracic bristles. The first three anomalies were or that of normal morphological differences particularly common. Anomalies only appeared between species. We know that the former when D. virilis females were crossed to D. lummei phenomena affect hybrid males more than females males; the reciprocal cross yielded normal pro- (Haldane's rule), involve a large effect of the X geny. This result was previously reported by chromosome (Coyne and Orr, 1989), and often Sokolov (1948, 1959) (although he misidentified show maternal effects (Orr, 1989). The genes D. lummei as "D. littoralis" (Throckmorton, underlying morphological differences between 1982)). As the D. lummei strains that Sokolov species, however, do not map disproportionately found to produce the most frequent anomalies to the X (see Coyne and Orr, 1989), and do not have been lost, little additional work on these seem to involve frequent maternal effects. As anomalies has been undertaken (however, as this Coyne and Orr (1989) note, certain explanations paper goes to press, an abstract by Heikkinen and of these different patterns predict that morpho- Lumme (1989) confirms the non-reciprocal nature logical/developmental anomalies in hybrids will of these anomalies; see below). Fortunately, the behave like postzygotic isolation (e.g., obey D. lummei strain we possessed produced frequent Haldane's rule and show a large effect of the X), anomalies when hybridized with D. yin/is. Also, while other explanations predict that the "genetic fortunately, male and female hybrids between D. 256 H. ALLEN ORR virilis and D. lummei are viable and fertile, allow- autosomes. The following anomalies were scored: ing genetic analysis of the basis of these anomalies. (1) reduced eye ("eyeless"); (2) obvious wing I report the results of this analysis here. length asymmetries; (3) wing vein anomalies (usually broken or missing veins; more rarely, small extra vein-segments appeared and were MATERIALSAND METHODS included in this category). As preliminary results demonstrated that the Thefollowing D. virilis and D. lummei stocks were D. virilis cytoplasm plays a large role in the produc- used: D. yin/is Pasadena, D. yin/is white (w: 1—105 tion of anomalies, most crosses were designed to [all map positions from Alexander (1976)]), D. determine whether this effect results from a mater- virilis peach; glossy (pe: 5—203; gl: 6—1.0), D. virilis nal effect (i.e., a mother's nuclear genotype deter- varnished (va 2—231.5), D. vinilis broken; tiny mining her progeny's phenotype) or from a true bristles, gap; cardinal; peach (b 2—1880; tb 3— cytoplasmic effect (e.g., an endosymbiont or 1040, gp 3-1180; cd 4-322; pe) and D. lummei mitochondrial gene). The role of an endosymbiont Finland. A few crosses involving D. americana Red was tested by rearing flies on tetracycline, follow- Cloud, D. novamexicana San Antonio, and D. ing the protocol of Hoffman and Turelli (1988). texana Morrilton were also made. Unless otherwise indicated, all crosses were The method used to determine the genetic basis made for one week at 18°C and then transferred of the hybrid anomalies is similar to that used by to fresh vials at 22°C. Orr and Coyne (1989) to analyse the basis of hybrid sterility in the D. virilis group: D. virilis flies carry- ing morphological markers are hybridized with RESULTS wild-type D. lummei flies, and the fertile F1 hybrids are backcrossed to either parental species. By cor- Individualsfrom the pure species stocks of D. relating the presence of hybrid anomalies with virilis and D. lummei show no developmental particular markers, one can determine which anomalies (table 1). However, the D. yin/is w chromosomes play a large role in the production female x D. lummei male hybridization produces of the anomalies. Inmost cases only the X chromo- males and females who frequently suffer develop- some carried a marker, allowing one to separate mental anomalies (table 1). Hybrid females and the effects of the X from that of the unmarked males are affected at the same rate (27.8 vs. 261 Table 1Number of individuals showing various developmental anomalies in pure species, F1 hybrids, and hybrid backerosses. In this and all subsequent tables, V =D.yin/is and L =D.lummei. In a cross, the species origin of the maternal species is given first, followed by the paternal species, e.g., VL D. virilis 9 x D. lurnmei 5. (VL)L= F19 (D. virilis9 x D. Iummeid)x D. lummeid Winglength Wing vein Genotype Eyeless asymmetry anomaly Normal Total D. virilis w 9 0 0 0 200 200 5 0 0 0 200 200 D. lummej Finland 9 0 0 0 195 195 5 0 0 0 211 211 VL 9 17 38 24 205 284 5 16 23 18 161 218 LV 9 0 1 0 327 328 5 0 0 0 325 325 (VL)L 9 0 0 13 358 371 wS 0 2 1 102 105 +5 0 1 2 152 155 HYBRID ANOMALIES 257 per cent anomalous, respectively; x2= 121,df= 3, involved: although these backcross hybrids carry P>O•75). Therefore, the hybrid anomalies do not mitochondria and any endosymbionts from D. obey Haldane's rule. yin/is on a largely D. lummei genetic background, The reciprocal cross does not produce they show a greatly reduced frequency of anomalous progeny (table 1). The difference anomalies. This result instead suggests that the between the frequency of anomalies in the two anomalies involve a maternal effect: these back- reciprocal hybridizations is highly significant (for cross hybrids show few anomalies because their females, x2= 10144,df= 3, P <O0OOl; for males, mothers had an F1 hybrid, not a pure species, df=3,P<O.000l). Because the F1 genotype. If a maternal effect is involved, the females from the two reciprocal hybridizations maternally-acting genes from D. virilis are nearly possess the same nuclear genotype, production of recessive: D. yin/is mrjjs/ mrrjj,, females produce the developmental anomalies must involve the many anomalous offspring, while M,ummei/mvjrj,i, cytoplasm. There are two ways this can occur: F1 hybrid females produce very few. production of anomalies may depend upon the Proof that maternal genotype—not the species nuclear genotype of a hybrid's mother (a "maternal source of cytoplasm—causes the anomalies effect"), or may involve some autonomous cyto- requires showing that hybrids carrying cytoplasm plasmic agent, e.g.,amitochondrial gene or an from the "wrong" species, i.e., D. lummei, can also endosymbiont to which D. yin/is is resistant (a suffer anomalies. As the above results suggest that "cytoplasmic effect"). hybrid anomalies are most common when mothers To test for the presence of an endosymbiont, are homozygous for some putative maternally- the D. virilis w female x D. lummei male hybridiz- acting allele from D. virilis, I backcrossed F1 ation was performed, and progeny reared, on females from a D. lurnmei female x D. virilis w medium containing tetracycline. Control crosses male cross to D. yin/is w males. The resulting were run on medium lacking tetracycline. Tetra- backcross females carry cytoplasm from D. lum- cycline treatment did not cure the hybrid anomalies mei, but, on average, are homozygous at 50 per (table 2). Indeed, more anomalies appeared among cent of their loci for alleles from D. virilis (they hybrids receiving the tetracycline treatment (x2= are D. virilis/D. lummei heterozygotes at their 5473,df=3, P<00001). While this result does remaining loci). not rule out the endosymbiont-hypothesis (a The results are unambiguous: these hybrid tetracycline-insensitive microorganism could be females do produce anomalous offspring when involved), it suggests that it is worthwhile to test crossed to D. himmei males, despite the fact that for maternal effects. their cytoplasm ultimately derives from D.