C OMMENT TIG FEBRUARY 1998 VOL. 14 NO. 2 43 Copyright © 1998 Elsevier Science Ltd. All rights reserved. 0168-9525/98/$19.00 PII: S0168-9525(97)01365-6 Gene conversion in mitotically dividing cells: a view from Drosophila GREGORY B. GLOOR* AND DIRK-HENNER LANKENAU ‡ ggloor@julian.uwo.ca • d.lankenau@dkfz-heidelberg.de *DEPARTMENT OF BIOCHEMISTRY, UNIVERSITY OF WESTERN ONTARIO, LONDON, ONTARIO, CANADA N6A 5C1. ‡ DEPARTMENT OF DEVELOPMENTAL GENETICS, GERMAN CANCER RESEARCH CENTER, IM NEUENHEIMER FELD 280, 69120 HEIDELBERG, GERMANY. It was about 25 years ago that Hiraizumi noticed that the males of some strains of flies transmitted re- combinant progeny 1 . This was in apparent contrast to Morgan’s 1914 observation that recombination was absent in the Drosophila melanogaster male 2 . Were the rules of recombi- nation changing in Drosophila, or was there an external agent at work? Kidwell et al. linked the male recombi- nation phenomenon to a syndrome of other traits including sterility, chromosome rearrangements and mu- tations 3 . Investigation of mutations caused by this syndrome led to the identification of the P family of trans- posable elements, but it was not clear how P elements related to male recombination 4 . Two groups led by Sved and Engels recently solved this enigma 5–7 and, as so often happens in science, the answer to one prob- lem gives insights into seemingly un- related areas. The work initiated by Hiraizumi is providing a new under- standing of how somatic genomes might maintain their integrity. Double-strand breaks and P-element transposition The first real clue to the cause of male recombination came when Engels et al. 8 showed that reversion of a P-element allele, caused by the precise excision of a P element from the white gene, increased 100-fold if the P element jumped in the pres- ence of a homologous wild-type white gene 8 . To explain this result, they hypothesized that P-element excision made a double-strand break (DSB) in the genome that was repaired in a gene conversion-like process. Their initial model for P excision is shown in Fig. 1. In this process, gene con- version without flanking exchange was the predominant outcome. Pre- vious work, done mainly in Sac- charomyces cerevisiae, showed that DSB repair could initiate gene con- version as well as reciprocal ex- change 9 , and so it was hypothesized that DSB repair was inducing male recombination 8 . The gene-conversion hypothesis was tested in an experiment in which a white gene with several sequence alterations was inserted at various places in the Drosophila genome and used as a template for the repair of DSBs made by P excision at the white locus 10 . Almost all the precise P- element excisions were gene conver- sions in which sequence was copied from the ectopic template into the break site, as shown in Fig. 1. The conversion tracts extended in both directions, and contained an average of 1.4 kb of template sequence. The template was not altered, nor were chromosome aberrations induced by the repair process, showing that most of the gene conversions occurred without reciprocal exchange. This confirmed the original hypothesis and suggested that P elements jump by a cut-and-paste transposition pathway, in which the transposon first excises from its insertion site and inserts else- where in the genome (Fig. 1). Shortly afterward, an in vitro transposition system described by Kaufman and Rio substantiated this transposition pathway for P-element mobility 11 . Even though significant levels of reciprocal exchange were not ob- served, the DSB-repair events found in the initial experiments were con- sistent with several models for gene conversion in which reciprocal ex- change was a possible outcome. Further observations that led to our current understanding of DSB repair following P excision were made by Nassif et al. 12 They used an ectopic template, in which partial P-element ends were flanked by white gene sequence, and found that almost 15% of the conversion events derived from this template were inconsistent with the models entertained previ- ously 8,10,13 . The synthesis-dependent strand annealing (SDSA) model shown in Fig. 2, in which there is base pair- ing between the newly synthesized DNA strands on both sides of the break, was proposed to explain these events. This was an attractive model because it could easily explain sev- eral features of DSB repair following P-element excision; it explained the preponderance of long, continuous, bi-directional conversion tracts and the relative lack of exchange. Most impor- tantly, it explained the production of duplicated sequences during repair and the changes in P-element end sequences observed following DSB repair in many experiments 6,8,10,12,14 . This model was derived mainly from work on DNA replication in bacterio- phage T4 (Ref. 15) and on single- strand annealing in yeast 16 . However, similar models had previously been proposed for mating-type switching in S. cerevisiae 17–20 and for mitotic P element inserts elsewhere P element P excision and gap enlargement Invasion Synthesis FIGURE 1. The initial model for double-strand break (DSB) repair following P-element excision that was proposed by Engels et al. 8 In this model the P element (red) excises and inserts elsewhere in the genome. The excision event leaves a DSB behind in the chromosome that can be enlarged. The broken ends are used for a homology search, invade a homologous sequence and are used as primers for DNA synthesis. Synthesis across the gap results in the repair of the break in which the sequence at the gap site is derived from the template. In the situation pictured above, sequence is copied from a template lacking a P element, resulting in precise loss of the P element. If, on the other hand, the ends invade the sister chromatid (which contains the P element) a copy of the P element will remain at the white locus 13 .