Did Demerec Discover Intragenic Recombination in 1928?

STURTEVANT'S multiple allelism hypothesis (Sturtevant 1913) led to the assumption by geneticists during much of the first half of the twentieth century that genes were units indivisible by recombination (reviewed in Lipshitz 2004, pp. 13–18). Disproof of this hypothesis required the discovery of the cis-trans position effect in Drosophila and the recovery of both wild-type and double-mutant recombinants from flies heterozygous for the Star and asteroid rough-eye mutations (Lewis 1942, 1945). A decade later, fine-structure mapping of the rII locus in phage provided definitive proof of intragenic recombination (Benzer 1955). In fact, Oliver (1940) had reported the occurrence of wild-type flies from females heterozygous for the mutations of the lozenge rough-eye series. He interpreted these as revertants, as the title of his article indicates, but he does discuss the possibility that they were the result of recombination. Green and Green (1949) later obtained double and even triple lozenge mutant combinations, thus establishing recombination as the source of the wild-type flies that Oliver had discovered.

There are now many examples in which a gene is made up of not only a protein-coding region but also enhancer regions that, remarkably, recombine as if they are separate genetic loci. Not only do such regions recombine to produce wild-type crossovers, but also there are examples in which double- and even triple-mutant enhancer regions were derived by crossing over, such as in the case of the bithorax gene of the bithorax gene complex (Lewis 1955, 1967; reviewed in Lipshitz 2004, pp. 23–29).

Demerec's name may be familiar to many modern Drosophilists because of the 1994 reprinting of the book, Biology of Drosophila, which Demerec edited (Demerec 1950). Others may know of him through his tenure as director of the Cold Spring Harbor Laboratories from 1941 to his forced retirement in 1960. However, many do not know that he was a highly accomplished Drosophila geneticist in his own right.

Commencing in 1926, Demerec reported the existence of three mutable genes in Drosophila virilis: reddish-α (Demerec 1926a), miniature-α (Demerec 1926b), and magenta-α (Demerec 1927). In 1928 he provided an extensive analysis of two strains carrying the reddish-α body-color mutation, abbreviated here as re. Demerec had introduced into both strains the closely linked flanking markers sepia eye color (se) located at 0.2 and scute (sc) at 3.0 on the X chromosome. In one strain re behaved as an allele of the yellow body-color mutation, y, at 2.4 on the X chromosome. Among 2252 male progeny of re/se y sc females there were no wild-type recombinants between re and y. In the second strain, however, Demerec observed a remarkably high frequency of reversions of re: namely 171 reversions (17.1%) among a total of 9988 male progeny of re/se y sc females. Among the 171 revertants of re, 148 were wild type with respect to se and sc and hence were true revertants of re, 41 of which Demerec confirmed by actual progeny testing. The remaining 23 revertants of the original 171 carried the flanking marker sc and constituted 0.23% of the total offspring. Revertants occurred during the meiotic divisions in the female and were reported not to occur in the male germline or in somatic cells of either sex.

Demerec found that, unless he selected for instability of re, the frequency of revertants fell to nearly zero after about seven generations. The results involving two such lines that had stabilized have direct relevance to this Perspective.

1. From one such stabilized line among 24,823 male progeny of re/se y sc females, Demerec (1928)(p. 376) states: “there were only three reversions. It might be questioned, however, whether these three reversions did not originate in a different way to the others, since all three were crossovers in the yellow-scute region.”

2. In a second stabilized line Demerec sought to determine if there were modifier genes that induced the high frequency of reversions. Among a total of 9101 male offspring of re/se y sc females, he observed two reversions.

Thus from the two stabilized lines Demerec actually had found a total of five revertants among 33,924 offspring, or 0.015% that were wild-type revertants of re. They must have carried the sc marker, since he states that all of them were associated with crossing over in the y-sc region. There is therefore a strong hint that “in a different way” was by crossing over between y and re.

Even though the salivary gland chromosomes are the same length as the corresponding arms of D. melanogaster, in D. virilis the genetic length is much greater. In her review of D. virilis genetics, Mary Alexander mentions that a factor of 2.8 more recombination occurs than in D. melanogaster (Alexander 1976). That factor can vary widely; for example, although there is 0.6 MU between sc and y in D. virilis, the map distance in D. melanogaster is effectively zero. For example, the double mutation y² sc was first obtained as a rare male crossover from large-scale mating of y² sc females in which all of the other major chromosomes were structurally heterozygous for chromosomal rearrangements (E. B. Lewis, unpublished results).

The high frequency with which re reverted may have been the result of the jumping of a transposable element near if not at the locus of re. Any explanation for the high reversion rate must also account for the higher frequency of the sc-bearing revertants in the unstable lines (156/14,599) on the basis of a summary of two experiments presented in Tables 2 and 3 of Demerec (1928), relative to the frequency observed in the stable line (7/2252). A possible explanation of the high reversion frequency is the presence of a mobile element near the unstable re. It is known that in D. melanogaster the P transposable element not only jumps at high frequency in dysgenic crosses but also stimulates recombination in the vicinity of its insertion site (Sved et al. 1991). An element in D. virilis that may possess similar properties to the P element is Penelope (Evgenev et al. 1997). Together with the higher overall recombination frequency in D. virilis, local stimulation of recombination by Penelope may underlie any explanation of Demerec's results.

Demerec did not pursue his findings on re, perhaps because he was preoccupied with determining the cause of the instability of reddish-α. He also had to maintain many lines of re to be sure that some were still unstable.

I knew about Demerec's work on mutable genes from Oliver, not having read any of Demerec's papers. I was fortunate to have spent the summer of 1939 at Cold Spring Harbor. At the time, I brashly thought that, if I had the raw data for the experiments that Demerec had carried out with re, I would be able to figure out what had caused the high reversion frequency. Demerec generously provided me with the data, which formed a considerable pile, as I remember. I spent many days poring over them but eventually gave up. I mention this episode as an example of Demerec's legendary generosity and strong interest and support of young scientists. He certainly helped one beginner on his way!

In conclusion, Demerec in 1928, a dozen years before Oliver, probably had the first documented case of intragenic recombination.