Abstract of some important

During a research career spanning more than 50 years, Dr. Bernardi’s work has been centered since 1959 on two major areas, molecular genetics and molecular evolution, in which he has played a major role. His detailed analysis of the organization of the eukaryotic genome and, in particular, of the vertebrate genome has led him to investigate genome evolution from an original point of view.

While the basic results have been outlined above, it is relevant to stress the general interest of these discoveries. Indeed, the results obtained on the organization of the vertebrate genome could not be accounted by any mechanism essentially based on stochastic changes, like the fixation of mutations by random drift proposed by the neutral theory. Yet, the results are compatible with the neutral theory, provided that natural selection operates not only on single-nucleotide changes (as it is the case in coding sequences) but also at a regional level in the non-coding sequences that form 97-98% of most vertebrate genomes.

In other words natural selection controls not only the “classical phenotype” of form and function (or, in molecular terms, of the proteins and of their expression), as generally accepted, but also the “genome phenotype”, the compositional properties of the genome and all their functional implications. This occurs in such a way (by regional selection) that most of the changes are in fact neutral as proposed by Kimura. This leads to a picture in which the neutral theory and the regional selection are complementary facets of molecular evolution; this neo-selectionist theory of evolution was recently presented in a book (G. Bernardi, 2004; reprinted in 2005. Structural and Evolutionary Genomics: Natural Selection in Genome Evolution. Elsevier, Amsterdam) and in an article (Bernardi G., 2007. The neo-selectionist theory of genome evolution. Proc. Natl. Acad. Sci. USA 104: 8385-8390).

Abstract: Chromatography of nucleic acids on hydroxyapatite.

The development of this methodology (Bernardi, 1965) led to the separation of single and double stranded DNA, so opening the door to the study of the reassociation kinetics of DNA and the discovery of repeated sequences in eukaryotic genomes. It could be shown that in the case of both nucleic acids and proteins, separations were based on the fact that native structures had more binding groups on their surfaces compared to denatured structures (see Bernardi, 1971, for a review article).

Abstract: Acid deoxyribonuclease

First purified in Bernardi’s Laboratory, acid DNase was the first enzyme shown to be endowed with a specificity towards DNA sequences. A dimeric, allosteric enzyme, cutting both DNA strands at the same time, acid DNase was a prefiguration of restriction enzymes (purified ten years later) and was used to demonstrate sequence differences among different DNAs, such as yeast mitochondrial DNA, satellite DNAs, and compositional DNA fractions from mammalian genomes. This work was summed up in three review articles (Bernardi, 1968, 1971; Bernardi et al., 1973).

Abstract: The nuclear genome of warm-blooded vertebrates.

The major discoveries in this field were (i) the demonstration of a striking compositional heterogeneity in the genomes of warm-blooded vertebrates, which could be described as mosaics of isochores (long, relatively homogeneous regions belonging to a small number of families); (ii) the demonstration that, in contrast, cold-blooded vertebrates were endowed with genomes characterized by a much less striking heterogeneity, so raising the problem of a compositional transition between the genomes of cold-and warm-blooded vertebrates; (iii) the existence of strong compositional correlations between coding and flanking non-coding sequences as well as between different codon positions, the latter being a universal correlation valid from prokaryotes to human; (iv) the demonstration that gene distribution was strikingly non-uniform in the vertebrate genome, about 65% of the genes being located in the 15% of the genome characterized by an open chromatin structures, and a weakly or strongly higher GC level (in cold- and warm-blooded vertebrates, respectively); (v) the distribution of isochores and genes in human and mouse chromosomes and in interphase nuclei; the results obtained in the nuclei showed an extremely decondensed chromatin structure for the GC-rich, gene-rich regions of the genome, and a very compact chromatin structure for the GC-poor, gene-poor regions

Abstract: The mitochondrial genome of yeast.

Bernardi and coworkers provided the first demonstration that the cytoplasmic “petite colonie” mutation of yeast (Ephrussi, 1949) was due to large deletions in the mitochondrial genome (see Bernardi, 1979, for a review). The major discoveries on this model genome were the demonstrations of (i) abundant intergenic sequences (made up of long AT spacers and short GC clusters), which were responsible for the excisions leading to the defective mitochondrial genomes of “petite” mutants; (ii) physical recombination in the mitochondrial genome; (iii) the effect of flanking non-coding sequences on the replicative efficiency of ori sequences; (iv) the effect of temperature on the replicative efficiency of ori sequence in petite mutants, in which a crucial stem-and-loop structure was only made of AT base pairs; this provided the first evidence for reversible transconformations of a genome by an environmental factor, temperature (Goursot et al., 1988); (v) the fact that the replication efficiency of petite genomes explains the outcome of crosses with wild-type cells (the phenomenon of suppressivity): the more efficient the replication of the petite mutant entering the cross, the higher the number of petite mutants in the progeny.

Abstract: The nuclear genome of warm-blooded vertebrates.

The major discoveries in this field were (i) the demonstration of a striking compositional heterogeneity in the genomes of warm-blooded vertebrates, which could be described as mosaics of isochores (long, relatively homogeneous regions belonging to a small number of families); (ii) the demonstration that, in contrast, cold-blooded vertebrates were endowed with genomes characterized by a much less striking heterogeneity, so raising the problem of a compositional transition between the genomes of cold-and warm-blooded vertebrates; (iii) the existence of strong compositional correlations between coding and flanking non-coding sequences as well as between different codon positions, the latter being a universal correlation valid from prokaryotes to human; (iv) the demonstration that gene distribution was strikingly non-uniform in the vertebrate genome, about 65% of the genes being located in the 15% of the genome characterized by an open chromatin structures, and a weakly or strongly higher GC level (in cold- and warm-blooded vertebrates, respectively); (v) the distribution of isochores and genes in human and mouse chromosomes and in interphase nuclei; the results obtained in the nuclei showed an extremely decondensed chromatin structure for the GC-rich, gene-rich regions of the genome, and a very compact chromatin structure for the GC-poor, gene-poor regions

Abstract: The integration of retroviral sequences in the mammalian genomes.

 Starting in 1979, a series of investigations led to the demonstration that stably integrated, transcribed proviral sequences were localized in isochores matching their base composition. In other words, stability of integration and transcription was associated with a localization in the mammalian genome that mimicked that of host genes (a review article is Rynditch et al., 1998). Viral sequences integrated in compositionally non-matching regions were rare and not transcribed. These findings showed the effect of the composition of flanking non-coding sequences on the expression of integrated viral sequences. As mentioned above, similar effects on transcription and replication were found in the mitochondrial genome of yeast. These results obviously go against the idea of non-coding sequences being junk DNA.

Abstract: The genomes of other eukaryotes and prokaryotes. 

The key conclusions of these investigations (covering the genomes of plants, insects, unicellular eukaryotes and prokaryotes) were (i) that a genome compartmentalization is not at all restricted to vertebrates, being quite widespread among eukaryotes; (ii) that, at least in a number of cases, the compositional compartmentalization was the result of regional GC increases associated with temperature (see the following section); and (iii) that, in the case of Gramineae, genes were concentrated in regions (collectively called the gene space) encompassing a very narrow GC range; this is due to an invasion of transposons in the gene-rich regions of the genome; the narrow GC range is determined by the abundant transposons flanking the genes.

Abstract: Molecular evolution.

The major discoveries made on the vertebrate genome, the compositional compartmentalization into a mosaic of isochores, the genome phenotypes (the different compositional patterns e.g. of cold- and warm-blooded vertebrates), the genomic code (the compositional correlations between coding and non-coding sequences, as well as between the codon positions), the bimodal gene distribution and its correlation with functional properties, could not be accounted for by the neutral theory of Kimura. This raised two problems. The first one was how to explain the formation and maintenance of the mosaic isochore organization of warm-blooded vertebrates. The explanation provided was that natural selection (more precisely, negative or purifying selection) was responsible for both phenomena. The advantage of increasing the stability of the gene-dense regions of the genome by increasing their GC level could be due to the increased body temperature following the appearance of homeothermy. Several lines of evidence on both vertebrate and prokaryotic genomes have very recently confirmed this explanation.

The second problem was to reconcile natural selection with the overwhelming numbers of neutral or nearly neutral mutations. This problem was solved by proposing a regional selection mechanism in which the average composition of an isochore must be kept within certain thresholds in order to avoid a deleterious effect on the expression of genes embedded in that isochore. This model of genome evolution, the neo-selectionist model, inspired by the results obtained on the expression of integrated viral sequences (see above), only requires negative selection on the average composition of a region.