Appendix

Alignment of partial denaturation maps of circularly permuted DNA by computer☆

Hexanucleotides containing modified bases (5-methylcytosine and 2-aminoadenine instead of cytosine and adenine) with increased capacities to bind complementary DNA sequences were used to map the distribution of their complementary sequences in a DNA target using electron microscopy. The method used hexamers to initiate DNA polymerse directed DNA synthesis at complementary sequences along a template. DNA synthesis was limited to about 200 residues by using a low concentration of deoxynucleotide precursors. During DNA synthesis a biotin ligand was incorporated to facilitate the subsequent binding of an electron-dense label (streptavidin-labeled colloidal gold particles) into newly synthesized DNA chains. The method can be implemented with commercially available products. The results demonstrate that the approach can be used to compare primary structural features of DNA fragments. The principles of the method can be adapted to a variety of single molecule detection methods such as electron, scanning tunneling, or atomic force microscopies.

A procedure for the testing of Electron Microscope (EM) mapping data for DNA molecules with site-specific bound ligands is suggested. The difficulty of distinguishing DNA molecule ends on electron micrographs indicates that their true orientations are not known. This in turn presents problems in obtaining correct maps relating to their alignment, and complicates checking the maps' validity. For these reasons a computer simulation of the EM study of double-stranded DNA molecules with site-specific bound ligands was carried out. The knowledge of the true orientations of the simulated DNA molecules allowed us to examine their final orientations after alignment. We used the number of improper-oriented molecules as the quantitative measure of the map quality. Detailed investigation based on this parameter permitted us to invent the criterion for the map validity, and to suggest the procedure for the testing of alignment of real DNA molecules. This procedure implies multiple randomization of initial orientations of the DNA molecules and minute analysis of the final maps. Most of the molecular, statistical and experimental parameters inherent to EM investigation of site-specific binding, such as the number of specific binding sites (N), the mean number of bound ligands (A), the length of the DNA molecules (L), the specific/non-specific ratio of binding (K), together with the standard deviation of DNA molecule lengths (H_L were tested for their influence upon the quality of EM mapping data. An empirical equation for the ultimate values of these parameters has been found, allowing us to predict the success of EM mapping.

An algorithm for aligning partially denatured DNA molecules was applied to various collections of molecules, using a Hewlett Packard desk-top calculator. The algorithm consists of two stages. The first is analogous to existing methods in that molecules are incorporated in a predetermined order into a reference map. The second refines the map via iteration which terminates upon convergence, resulting in a final map. Molecule order-dependence and the effects of algorithmic parameters on computation time and reproducibility of denaturation maps were studied. Differences between reference and final maps varied, as did differences between final maps derived from various molecule orders. Iteration consistently led to increased area overlap between histograms, hence, order-independence. Histogram quality is expressed in terms of a parameter which measures the correlation of all molecules of a population with their partial denaturation map. A preferential molecule order is suggested. Application to a unique (non-permuted) population showed that molecular, statistical and experimental factors inherent to all populations are crucial for the outcome of alignments. The present algorithm negotiates these more successfully than a pairwise cross-correlation alignment procedure used before. However, if these factors are too large, resolution is impaired but not necessarily due to algorithm quality. It is concluded that the present algorithm is operationally simple, generates maps with a high degree of reproducibility, produces optimal correlation of all molecules with their partial denaturation map, is time-saving and yields improved resolution compared to other methods. It also produces reliable results as judged from marker mapping capability.

In an approach to the functional anatomy of the mitochondrial genome and its evolution, we have compared buoyant densities, contour lengths, and denaturation maps in circular mitochondrial DNAs of the genus Drosophila. Mitochondrial DNAs from three representatives of the subgenus Drosophila (D. virilis, D. hydei, D.funebris) are similar in size (approx. 5 μm or 1 · 10⁷ daltons) and buoyant density (approx. 1.685 g/ml), while in two members of the subgenus Sophophora (D. melanogaster, D. simulans), mitochondrial DNAs are longer (approx. 6 μm or 12.4 · 10⁶ daltons) and have a lower buoyant density (approx. 1.681 g/ml). The latter mitochondrial DNAs also share one distinctly large early melting region, which in D. melanogaster is equivalent to 1.54 μm of native DNA. The corresponding (A + T)-rich region in D. virilis or D. hydei mitochondrial DNA is 1 μm shorter. Except for this region, denaturation maps of D. melanogaster and D. virilis mitochondrial DNAs are indistinguishable. The addition or deletion of a single block of (A + T)-rich sequences can fully account for the differences in buoyant density and size between the mitochondrial DNAs we have examined.

In an appendix, we show that there is an equivalent discrepancy between the extent of strand separation determined by electron microscopy and the actual extent of DNA denaturation, whether this is determined from absorbance changes or inferred from the reduction in contour lengths of individual circular molecules. The reduction in contour length appears to result exclusively from the uniform foreshortening of single-stranded DNA, not only in regions of visible strand separation but also in denatured regions hidden within putatively native segments of molecules. For molecules showing 15–45% strand separation, we estimate that putatively native segments are approximately 50% denatured.

This report presents evidence for the existence of two types of structural intermediates when λ DNA replicates as a circle. The first seems to consist of a partially replicated, covalently closed circle. The second was identified from a variety of properties, including electron microscope analysis after partial denaturation. It consists of a molecule with two interlocked rings, of which one is covalently closed and the other contains an interruption in one strand. A model for circle replication explaining how these structures might be related is presented.

Mutants in phage P22, displaying an increased frequency in generalized transduction, map in the locus of maturation gene 3 on the phage P22 genetic map (Raj et al., 1974). These mutants are able to transduce all markers of the host chromosome with the same increased frequency. DNA extracted from such a high transduction mutant is randomly permuted, which is different from the restricted permutation of DNA found in wild-type phase P22. A modified DNA packaging mechanism probably explains the abnormally high transduction frequency and randomly permuted DNA found in these high transduction mutants.