Isochore families

The second discovery was that isochores belonged to a small number of families characterized by distinct average base compositions and by other sequence features (see Table 3.4). The major DNA components were responsible for the discontinuities revealed by the CsCl analysis of CS2SO4 ligand fractions (Fig. 3.4) and derived from isochore families that caused discontinuities in the CsCl profiles of very high molecular weight samples (Fig. 3.5). 

I- igurc 3.17. A. Isochore families from Xenopus. human, mouse and chicken, as deduced from density gradient centrifugation. B. Compositional patterns of coding sequences (represented by GCj values averaged per coding sequence) for Xenopus, mouse, chicken mid human. tModihed from Bernardi. 149.S). 

The similar compositional features exhibited by isochore families and by coding sequences of the genomes of Fig. 3.17 raised the question of the possible correlation between the composition of the isochores, which cover a 30 to 60% GC range in the case of the human genome, and the composition of the coding sequences embedded in them. This question could be answered by localizing coding sequences of known primary structure in compositional DNA fractions (or in isolated major DNA components). 

Figure 5-11. Density of A genes and B CpG islands in isochore families, Relative numbers of sequences over relative amounts of isochore families are presented in the histograms. (From Jabbari and Bernardi, 1998).

A more detailed analysis was possible on integrated HTLV-1 sequences, in which case 40 sequences from 12 cell lines and T-cell clones were investigated (Zoubak et ah, 1994). Indeed, the transcriptionally active HTLV-1 sequences were found to be localized in the GC-richest regions which corresponded to the H2 and H3 isochore families of the human genome, whereas the transcriptionally inactive HTLV-1 sequences were localized in GC-rich regions corresponding to the HI isochore family (Fig. 6.4). 

Figure 6.7. A. Numbers of HIV-1 sequences in each isochorc family, the GC intervals are LI (<37%), L2 (37 1%), HI (41 6 M ), H2 (46-52 /Ii) and H3 (>52%). The size of the genome belonging to individual isochore families in 100 kb windows is 571 Mb for LI, 977 Mb for L2, 706 Mb for HI, 321 Mb for H2and 85 Mb for H3. B. The vertical axis shows the probability of targeting the individual isochore family. This is calculated as the number of integrations per Gb whose 100 kb surrounding belongs to the corresponding isochore family. (Modified from Elleder et al., 2002).

Figure 6.13. Human genome distribution of Alu sequences in terms of numbers of sequenecs (A) and of density as calculated in 100 kb-long nonoverlapping segments (B). Isochore family intervals are as in Table 6.1. Distributions concern AluYa5 family (mean A = 2.26%) AluYbS family (mean A = 5.23%) AluY family (mean A = 7.15%) AluJo family (mean A = 16.8 /o). (Updated from Pavlicek et ah, 2001).

Compositional mapping can also be done at the chromosomal level by in situ hybridization of compositional DNA fractions corresponding to different isochore families. In the first investigation of this kind (Saccone et al., 1992), it was shown that the hybridization of a DNA fraction derived from H3 isochores produced the highest concentration of signals on two largely coincident subsets of R bands (Figs. 7.4, 7.5) (i) the T bands (Dutrillaux, 1973), which are the most heat-denaturation-resistant R bands and (ii) the chromomycin... 

Figure 7.7. Human chromosomes hybridized wilh the bioiin-iabeled DNA from the H3 isochore family, at dilferenl levels of resolution. The hybridized regions were visualized by tluorescein (yellow signals) and chromosomes were red-siained with propidium iodide. Each panel presents chromosomes with a band resolution ranging from about 300 to about 850. (From Saccone et al., 1999).

The analysis of the localization of GC-poor DNA on metaphase and prometaphase chromosomes will be illustrated in some detail in order to provide an example of the procedure followed. Fig. 7.10 displays the CsCl profiles, the relative DNA level, and the proportion of the isochore families in the DNA fractions (Saccone et al., 1996) used as probes for the in situ hybridizations of GC-poor DNA. LI isochores are only present in the pellet DNA L2 isochores are distributed in the pellet and in fractions 1-3 fractions 4 and 5 contain almost exclusively DNA from FIl isochores. In the case of fraction 5 (as well as in the following fractions shown in Fig. 1 of Saccone et al., 1996), a light satellite peak corresponding to DNA from centromeric heterochromatin is present. The pellet DNA, essentially formed by LI isochores, hybridized on a subset of G bands (Fig. 7.11), which were called LI bands, and corresponded to the G1 and G2 bands, as we will name the two most intensely staining subsets of Francke s R bands (1994). In contrast, the pellet DNA is almost absent in the large majority of the FI3 4oo and H3 4oo bands. This is especially evident in chromosomes 15, 17, 19, and 22 (see Fig. 7.11). 

Figure 7. ), CsCl profiles of the human DNA fractions toiuaining the LI, L2 and HI isochore families (modified from Sacconc et ai 1996) fraction numbers (P stands for pellet), rcUitivcDNA content, modal buojant density and relative amounts of isochorc famitics are indicated. (From Federico et a1 2000).

Chromosomes 21 and 22 arc very different from each other. Together, they exhibit the whole spectrum of chromosomal bands. While chromosome 21 is made of several compositional regions representing all isochore families, chromosome 22 is essentially formed by H2 and H3 isochores, comprises HI isochores, but contains very few L2 Lsochorcs and no... 

Jabbari K. and Bernard G. (1998). CpG doublets, CpG islands and Alu repeats in long human DNA sequences from different isochore families. Gene 224 123-128. 

---