Parallel DNA triplexes, homologous recombination, and other homology-dependent DNA interactions

Abstract

R. Daniel Camerini-Otero and Peggy Hsieh Genetics and Biochemistry Branch National institute of Diabetes and Digestive and Kidney Diseases National institutes of Health Bethesda, Maryland 20892 Genetic recombination is the process, common to all forms of life, by which new combinations of genetic material or nucleic acid sequences are generated (for a historical per- spective see Whitehouse, 1982). Biologists have been fas- cinated with this phenomenon for almost a century. In 1905, while studying the inheritance of traits in the sweet pea (Lathyrus odoratus), Bateson and colleagues reported an exception to Mendel’s third law, that of independent segregation. Certain combinations of traits were observed more frequently and others less frequently than expected. The mechanism underlying this partial linkage (neither complete linkage nor independent segregation) is recom- bination. Morgan coined the term “crossing over” to ex- plain the exchange that gave rise to new combinations of linked traits. This recombination involving exchanges of genetic information at equivalent positions anywhere along the length of two chromosomes with substantial overall sequence identity is called general or homologous recombination. A remarkable feature of this process is that, in spite of its lack of specificity, it has exquisite fidelity. That is, in homologous recombination, there is neither the loss nor the gain of a single nucleotide at the joint. These joints are therefore quite unlike the imprecise joints observed, for example, between coding regions in V(D)J recombina- tion. In a second class of recombination, site-specific re- combination, the two chromosomes exchange information in a very precise manner at sites of which at least one is highly preferred, or specific (Craig, 1988). In this form of recombination, overall homology between the two chro- mosomes is not a factor. It is curious that mechanistically we know much more about site-specific recombination than about general re- combination. To a great extent, this disparity in our state of knowledge is a reflection of how much more amenable site-specific recombination is to biochemical dissection. For example, the reactions are well defined; that is, the biologically relevant substrates and products can be easily distinguished biochemically, and only a very few proteins are involved (one in some cases). Perhaps as important, most and usually all of the proteins involved are encoded by the smaller of the substrate DNAs, the largest of which is the size of a bacteriophage genome. As a consequence, the first of many complete site-specific recombination re- actions was achieved in a cell-free system over 15 years ago (Craig, 1988). These assays were then used to purify all the components involved in several of these reactions. By contrast, in meiotic general recombination, the de- tails are so elaborate that they border on the rococo. Even in bacteria (during conjugation or transduction, for exam- ple), the details of homologous recombination are quite daunting, and the reaction(s) is hard to define. First, by definition the substrates and products are virtually indistin- guishable. This rather featureless aspect of homologous recombination still remains a great stumbling block in es- tablishing the relevance of in vitro biochemical findings to events in cells. Second, in Escherichia coli at least 20 gene products are known that participate in three separa- ble pathways of general recombination (Smith, 1989). Almost as important, there is a fundamental mechanistic difference between these two forms of recombination that follows from the difference in specificity. In site-specific recombination, specific DNA sequences are brought to- gether by protein-protein bridges anchored on specific protein-DNA complexes. In essence, the paradigm used for molecular recognition is not unlike that used in all other specific DNA-protein interactions. That is, evolution has sculpted protein surfaces that recognize certain features usually present in the major groove of duplex DNA. In contrast, in general (homologous) recombination, a similar degree of precision, and in some cases an even greater degree of fidelity, has to be achieved in the absence sequence specificity. Thus, the biochemical machinery cannot draw on such carefully crafted specific protein- DNA interactions. This has been the alluring mechanistic challenge in understanding general recombination: how are any two homologous DNA sequences brought to- gether? An early idea for fidelity of homologous recombina- tion was the copy choice model, which postulated DNA template switching during replication. In this model, it was assumed that some unspecified previous pairing of the chromosomes brought them into sufficiently close proxim- ity that accurate switching between DNA duplexes would ensue. The break-and-join paradigm, based on the ob- served behavior of chromosomes at meiosis, was consid- ered to be too imprecise account for the fidelity of this kind of genetic exchange. Although in hindsight it appears obvious that the required fidelity in the face of nonspeci- ficity could take advantage of direct DNA-DNA interac-