Left-Handed DNA and RNA editing: Z-DNA is an alternative left-handed conformation of the DNA double helix. It was first visualized 20 years ago in this laboratory in an X-ray diffraction analysis. The immediate question asked at the time was: Does it have a biological function? An important advance was made by publication of the crystal structure of Z-DNA bound to a protein domain of an editing enzyme. The crystal structure reveals a detailed fitting of the protein to the DNA, held together by a system of 11 hydrogen bonds and 5 distinct van der Waals interactions. The protein binds to 5 successive phosphate residues on one strand of the Z-DNA double helix and also has a van der Waals contact which can only occur if the base is in the syn conformation which is a hallmark of Z-DNA. This protein domain is one of a family of proteins in which the important residues involved in Z-DNA interaction are strongly conserved.
Conversion of B-DNA to Z-DNA requires an input of energy. That input is provided by movement of RNA polymerase which generates underwound DNA and stabilizes the Z form. Z-DNA is largely found in transcribing genes, and it is there only transiently as the cessation of transcription results in rapid conversion of Z-DNA to right-handed B-DNA through the action of topoisomerases.
The editing enzyme is double-stranded RNA adenosine deaminase. It deaminates adenine to produce inosine which codes as if it were guanine. Thus, it changes amino acids in proteins. Many of these occur in the central nervous system, such as in the glutamate or serotonin receptors. The editing enzyme binds to a hairpin or foldback RNA structure, which forms a double-stranded segment. The editing enzyme consists of three functional domains, one which carries out deamination, one which binds double-stranded RNA and a third which binds left-handed Z-DNA. Double-stranded RNA is often formed by an intronic sequence pairing with an exonic sequence, and an adenine in the exon is then deaminated. This must happen before the removal of introns by the splicing apparatus. Since Z-DNA is found almost exclusively in transcribing genes, it is believed that the Z-DNA binding domain targets the enzyme to the transcribed gene so that it can carry out its editing activities before the introns are removed. The Z-DNA binding domain has a binding constant near 10 nM, and thus may function as a targeting domain.
A new development in the biological activities of the Za domain derived from an analysis of the effects of measles virus infection. Measles is a negative strand RNA virus which goes into the cytoplasm of cells and elicits the interferon response. A number of proteins are upregulated, including the full-length ADAR1 containing the Za domain. As a result, the ADAR1 enzyme begins to accumulate in the cell's cytoplasm. Sometimes measles results in encephalitis, and in these cases the virus has been sequenced. Sequencing revealed that the virus had been subjected to extensive hypermutation, in that many As were changed to Gs and many U's were changed to Cs. This clearly represents the activity of the editing enzyme in a replicating viral system.
The question we posed is: "What is the Za domain doing in the cytoplasm where there is no Z-DNA?" However, double-stranded RNA is there due to viral replication. This prompted us to examine in greater detail the possible interaction of Za with left-handed Z-RNA. The investigation proved quite fruitful. Although reaction of Za with Z-RNA is slower than its reaction with Z-DNA, it occurs at a significant speed at 37×C. The conversion was monitored both by changes in the Raman spectrum of RNA as well as its circular dichroism. Using small probes of r(CG)n, it was shown that, when Za is titered into solution, it gradually converts RNA oligomers entirely from the right-handed A form into the left-handed Z conformation. Study of the kinetics revealed additional activation energy for the conversion of r(CG)n to Z-RNA compared to Z-DNA. This involves an additional activation energy of approximately 1.2 kcal/bp for conversion from the A form to the left-handed Z form. This is in the anticipated range of energy required to change the pucker of the ribose ring from its normal C3' endo conformation to the C2' endo conformation. Both Z-RNA and Z-DNA have C2' endo and C3' endo pucker alternating along the sugar phosphate chain. This set the stage for co-crystallization efforts of Z-RNA and Za. It raises the possibility that Za on the editing enzyme may use Z-RNA as a target for binding to replicating double-stranded RNA that may be under negative torsional strain in analogy with that generated during transcription.
Za is the only protein known to bind tightly to both double-stranded RNA and DNA. In the right-handed form, these molecules look quite different and therefore, their binding proteins are different. However, in the left-handed Z form, both DNA and RNA duplexes are very similar and the same protein is used to recognize both of them.
Brown, B.A.,II, Lowenhaupt, K., Wilbert, C.M., Hanlon, E.B., and Rich, A. The Za domain of the editing enzyme dsRNA adenosine deaminase binds left-handed Z-RNA as well as Z-DNA. Proc. Nat'l. Acad. Sci., USA 97: 13531-13586 (2000)
Kim, Y.-G., Lowenhaupt, K., Maas, S., Herbert, A., Schwartz, T. and Rich, A. The Zab domain of the human RNA editing enzyme ADAR1 recognizes Z-DNA when surrounded by B-DNA. J. Biol. Chem. 275: 26828-26833 (2000)
Schwartz, T., Rould, M.A., Lowenhaupt, K., Herbert, A. and Rich, A. Crystal structure of the Za domain of the human editing enzyme ADAR1 bound to left-handed Z-DNA. Science 284: 1841-1845 (1999)