An ultracentrifugation analysis of two hundred ﬁsh genomes Giuseppe Bucciarelli a , Giacomo Bernardi b , Giorgio Bernardi a, * a Laboratorio di Evoluzione Molecolare, Stazione Zoologica Anton Dohrn, Villa Comunale, Naples 80121, Italy b Department of Ecology and Evolutionary Biology, University of California, Santa Cruz, CA 95064, USA Received 31 March 2002; accepted 18 May 2002 Received by E. Olmo Abstract The goal of this study was to provide a comprehensive view of the compositional characteristics of ﬁsh genomes. We therefore expanded the number of ﬁsh species that we had explored so far in their DNAs by analytical ultracentrifugation in CsCl density gradient from 122 to 201. This study included representatives from three out of nine orders of Elasmobranchs (sharks and rays), both orders of dipnoan lungﬁshes, and both orders of chondrosteans (sturgeons and bichirs). We also studied 19 out of 38 teleostean orders, which represent all but four (minor) superorders of the subdivision Teleostei, a group comprising about 23,600 species (96% of all extant ﬁshes). This leaves for further studies two subclasses, Holocephali (chimaeras), and Coelacanthimorpha (gombessas). In spite of this substantial increase in the number of species and orders analysed, all average properties (the modal buoyant density, r 0 , the average buoyant density, krl, the CsCl proﬁle asymmetry, A, and the compositional heterogeneity, H), and all their ranges were unchanged compared to a previous study [J. Mol. Evol. 31 (1990) 265]. This suggests that, in all likelihood, the properties reported in the present paper can be considered as generally valid for all ﬁsh genomes. q 2002 Elsevier Science B.V. All rights reserved. Keywords: Analytical ultracentrifugation; Phylogeny; Genome; Base composition; Fish 1. Introduction Previous investigations from our laboratory analysed the compositional patterns of ﬁsh genomes by studying the proﬁles obtained in analytical CsCl ultracentrifugation experiments. This was done in three steps involving a total of two (Thiery et al., 1976), 34 (Hudson et al., 1980), 39 (Pizon et al., 1983) and 122 (Bernardi and Bernardi, 1990a) species, respectively. Since the early studies (Thiery et al., 1976) the compositional properties of ﬁsh genomes were shown to be strikingly different from those of birds and mammals. While birds and mammals display genomes with vast heterogeneities, 10–15% of their genomes containing GC-rich sequences, ﬁsh genomes, in general, display DNAs characterized by a lower heteroge- neity, GC-rich sequences being less GC-rich and less abun- dant that un warm-blooded vertebrates. In contrast, the distribution of genes is similar in all vertebrates, in that about half of the genes are located in the GC-rich sequences of the genomes, even if theses sequences are less GC-rich in ﬁshes compared to mammals and birds (Bernardi and Bernardi, 1990b). The transition in genome organization between cold- and warm-blooded vertebrates is therefore a crucial step in vertebrate DNA evolution. In order to fully understand this transition, it is essential to have a clear picture of the compositional properties of ﬁsh genomes. Since ﬁsh repre- sent a huge variety of vertebrate forms (half of all vertebrate species), it is essential to study a phylogenetically represen- tative sample of ﬁsh species. Thus we decided to expand the genomes explored so far to a total of 201 species. The reasons for exploring such a large number of ﬁsh genomes was (i) as mentioned above, to obtain a phylogen- etically representative sample of different orders as ﬁshes include a vast array of distantly related vertebrates with about 25,000 species, corresponding to about half of the extant vertebrate species (Nelson, 1994); and (ii) to extend as much as possible comparisons of ﬁsh orders, families, and genera, in order to investigate the compositional differences that arose over evolutionary time (see Bernardi and Bernardi, 1990b). Gene 295 (2002) 153–162 0378-1119/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved. PII: S0378-1119(02)00733-3 www.elsevier.com/locate/gene Abbreviations: A, asymmetry; GC, molar fraction of guanine and cyto- sine in DNA; H, heterogeneity; ,  H, mean heterogeneity (interspeciﬁc); kb, kilobase pair(s); MW, molecular weight; r, buoyant density; r 0 , modal buoyant density; , krl, mean buoyant density; s , standard deviation (inter- speciﬁc) * Corresponding author. Tel.: 1 39-081-583-3215; fax: 1 39-081-245- 5807. E-mail address: bernardi@alpha.szn.it (Giorgio Bernardi).