Triplet repeats, over-expanded in neuromuscular diseases, are under-represented in mammalian DNA: a survey of models.

Simple tandem repeats represent more than 1% of the human genome: occasionally they exhibit intergenerational expansibility and are associated with neuromuscular diseases. In transgenic mice the same sequences elicit similar symptoms, but do not expand. We have searched for di-, tri-, and tetra-repeats in the published DNA sequences of chromosomes 21 and 22 of Homo sapiens, as well as in more than five megabases of Mus musculus DNA. Human and murine DNA sequences show a shortage in frequency and base coverage of tri-repeats as compared to di- and tetra-repeats. In murine sequences the cumulative frequency of di-, tri-, and tetra-repeats and their overall base coverage are about threefold higher than in human. Models for both the shortage of tri-repeats found in man and mouse and for their dynamic expansions are discussed. We propose that some of the 10 possible tri-repeats may be more prone than others to assume unusual structures capable of interfering with DNA synthesis: hence the shortage of tri-repeats. If such repeats are located at the 3'end of a chain growing and thus approaching another chain annealed to the same template, as Okazaki fragments do during discontinuous and encumbered replication, a combination of strand displacement, template switch, and branch migration may produce structures resistant to removal, hence the expansion of tri-repeats.

A model for the involvement of Okazaki fragments maturation in the expansion of short tandem repeats.

We propose a model for the expansion of short tandem repeats (ESTR), a phenomenon which has been found to occur in human DNA and is associated with a dozen of neuromuscular diseases. The model is based mainly on theoretical considerations and recovers experimental data from the literature; it also finds support in preliminary results obtained by us in multiprimed polymerase chain reactions designed to assess the effects of a downstream primer on the fidelity of the elongation of an upstream one. The model links the occurrence of the ESTR to a defective maturation of the Okazaki fragments (OF), and in particular to an improper processing of their 3' termini. This may occur when the last OF approaches the 5' terminus of the previous one in a susceptible region of the template. It is postulated here that when a growing OF has progressed past the priming region and its main portion has been synthesized, upon approaching its conclusion, the final elongation may take place in a region of the template where certain triplets are repeated: in that case a series of aberrations on the elongation mechanism may occur. These aberrations could involve (a) the displacement of the 5' terminus of the penultimate, properly matured OF, enacted by the incoming 3' terminus of the last OF, (b) the switch of the latter to the displaced strand of the former as template, (c) the fold-back on itself of the growing 3' terminus of the last OF, (d) its assumption of an unusual structure because of the repetition, and (e) some impairment of its removal by structure-specific exo-endonuclease(s). Derangements of this last part of the process may trigger the ESTR.

A procedure for cloning genomic DNA fragments with increasing thermoresistance.

Genomic DNAs have been cleaved by restriction or sonication, and the resulting double-stranded fragments have been exposed to increasing temperatures. This treatment may induce the helix-coil transition either in a single or in several steps, depending on the size and composition of the duplexes. Eventually, a critical temperature is reached at which each duplex melts completely and the two constitutive single strands separate. A transition interval can thus be defined for each duplex by the temperature at which the earliest strand separation takes place and that at which the most resistant double-stranded core collapses. If solutions containing a mixture of DNA duplexes are exposed to temperatures within their transition intervals, three kinds of molecules should originate: (1) duplexes that have not yet initiated the melting phase; (2) duplexes that have undergone only partial melting; and (3) single strands that derive from fully melted duplexes. If the heated solutions are quickly cooled to 0 degreesC, only the molecules from the first two classes can be ligated to a compatibly ended vector and cloned: class (1) are intact duplexes, and class (2) are molecules that snap immediately back to fully duplex structures: both are double-stranded. Conversely, the single strands of class (3) may not reanneal and thus be neither ligated nor cloned. We have tested the procedure on restricted coliphage lambda DNA, in view of its compartmentalized organization and known sequence. Then, we have applied it to human genomic DNA fragmented by sonication. After cloning of the available duplexes in a bacterial plasmid, libraries of molecules endowed with a progressively higher thermoresistance can be prepared for thermodynamic and genomic studies.

Proteins from the prokaryotic nucleoid: primary and quaternary structure of the 15-kD Escherichia coli DNA binding protein H-NS.

The primary sequence of H-NS (136 amino acid residues, Mr = 15,402), an abundant Escherichia coli DNA-binding protein, has been elucidated and its quaternary structure has been investigated by protein-protein cross-linking reactions. It was found that H-NS exists predominantly as a dimer, even at very low concentrations, but may form tetramers at higher concentrations and that the protein-protein interaction responsible for the dimerization is chiefly hydrophobic.

Proteins from the prokaryotic nucleoid. Interaction of nucleic acids with the 15 kDa Escherichia coli histone-like protein H-NS.

The interaction between nucleic acids and Escherichia coli H-NS, an abundant 15 kDa histone-like protein, has been studied by affinity chromatography, nitrocellulose filtration and fluorescence spectroscopy. Intrinsic fluorescence studies showed that the single Trp residue of H-NS (position 108) has a restricted mobility and is located within an hydrophobic region inaccessible to both anionic and cationic quenchers. Binding of H-NS to nucleic acids, however, results in a change of the microenvironment of the Trp residue and fluorescence quenching; from the titration curves obtained with addition of increasing amounts of poly(dA)-poly(dT) and poly(dC)-poly(dG) it can be estimated that an H-NS dimer in 1.5 x SSC binds DNA with an apparent Ka approximately equal to 1.1 x 10(4) M-1.bp-1. H-NS binds to double-stranded DNA with a higher affinity than the more abundant histone-like protein NS(HU) and, unlike NS, prefers double-stranded to single-stranded DNA and DNA to RNA; both monovalent and divalent cations are required for optimal binding.

Proteins from the prokaryotic nucleoid. A protein-protein cross-linking study on the quaternary structure of Escherichia coli DNA-binding protein NS (HU).

Escherichia coli DNA-binding proteins NS1, NS2 and NS (NS1 + NS2) react with the protein-protein bifunctional cross-linking reagents dimethylsuberimidate and dimethyladipimidate to yield oligomers up to hexamers. The former reagent, with the longer arm, is more efficient than the other shorter one. Both one- and two-dimensional gel electrophoreses show that the cross-linked trimers are homogeneous, while the dimers appear heterogeneous, suggesting that at least two types of dimers but geometrically equivalent trimers are formed. In the presence of DNA, the cross-linking reaction with either reagent yields fewer dimers and more of the larger products. The yield of cross-linked products of various sizes was determined for NS1, NS2 and NS as a function of the protein concentration (0.03-3000 microM). From the results obtained in these experiments, we derived a model of quaternary structure in which dimers and tetramers are predominant in very solutions of the proteins. Above a critical concentration (10-50 microM), interactions among tetramers become increasingly important, yielding octamers and perhaps larger products. Our data do not support a recently proposed model in which the DNA is packaged around a protein disc consisting of 8-10 NS dimers.

Proteins from the prokaryotic nucleoid. High-resolution 1H NMR spectroscopic study of Escherichia coli DNA-binding proteins NS1 and NS2.

The 1H-NMR spectra of the two Escherichia coli basic, low-Mr (approximately equal to 9000) DNA-binding proteins NS1 and NS2 and of their native complex NS were studied at 400 MHz and a number of resonances and resonance peaks were assigned. As in the case of some eukaryotic histones, the presence of a large number of high-field perturbed Phe resonances, several shielded and deshielded methyl resonances and backbone NH protons quite inaccessible to the solvent clearly indicate the existence of extensive tertiary and, even more so, quaternary structures involving hydrophobic interactions. These structures are lost upon heating, but readily reform upon cooling. Spectral differences between NS1, NS2 and NS and the greater thermal stability of NS indicate that molecules of the heterologous subunits (NS1 and NS2) aggregate (dimerize) preferentially in comparison to the self-aggregation of the homologous subunits. Unlike those of the eukaryotic histones, the tertiary and quaternary structures of NS are insensitive to extensive variations of the ionic strength.