Two unrooted trees for the four Hox gene clusters. Topology no. 1 is proposed by Amores et al. (46) and is supported by our analysis of Na⁺ channel genes, topology no. 2 is proposed by Bailey et al. (45) and not supported.

Results of separate genomic DNA surveys across both introns X and Y (X and Y), and intron X alone (X). Each band corresponds to a distinct Na⁺ channel gene. Negative controls without template (Neg.) were included in the assay (the clean negative control lane is not shown in Left).

Schematic representation of primer positions for RT-PCR of RNA and PCR of genomic DNA. (A) Schematic Na⁺ channel in the membrane with four repeating domains (I–IV). Primers for RT-PCR were directed to the highly conserved regions of domains III and IV spanning the inactivation loop (loop III-IV) that links them (shown in red). The region of the gene used in phylogenetic analysis is shown in purple. The positions of the highly conserved intron/exon borders are denoted as X and Y. (B) Depiction of primer sites and intron locations for Na⁺ channels in the conserved pore region of domain III. Three semi-independent primers sets were used to amplify ≈380 bp of coding sequence and introns X and Y (primers 1 and 4); 255 bp of coding sequence and intron X (primers 1 and 2); and 125 bp of coding sequence and intron Y (primers 3 and 4).

Phylogenetic analysis of vertebrate Na⁺ channel nucleotide sequences based on an unrooted maximum likelihood tree. The analysis was done without rooting because rooting with sequences of distantly related animals, as invertebrates in this case, are not strongly supported and potentially misleading due to long-branch attraction (25). The longest available nucleotide sequences common to the pufferfishes, Fugu pardalis (Fugu p.) (AB030482) and Fugu rubripes (Fugu r.) (D37977), the electric eel, Electrophorus electricus (Eel) (M22252), the knife fish, S. macrurus, the Japanese newt, Cynops pyrrhogaster (Newt) (AF123593), and rat were aligned by using megalign software DNAstar, Madison, WI. The mammalian sequences are the same as used by Plummer and Meisler (11), except for hNa_v1.9 (AF109737) (53), which was thought to be a novel channel and which is now believed to be an ortholog of rNa_v1.9 (53). The common segment corresponds to the available SterNa5 sequence, which runs from just past IISI to IVS1. This region contains 2,190 nucleotides or ≈40% of the nucleotides in the coding sequence of the E. electricus channel. From the preliminary alignment, nine blocks, totaling 1,545 nucleotides, were unambiguously aligned across the taxa being considered. The final alignment represents an estimated 55% of what might have been included, had complete cDNAs been available for all 22 channels. The aligned blocks correspond to the nucleotide positions from the E. electricus channel 2146–2319, 2350–2529, 2572–2661, 2680–2775, 2788–2832, 3166–3192, 3259–3618, 3673–3858, and 3949–4335. Mammalian Na⁺ channel genes are in colored blocks; genes on the same chromosome are the same color. Sternopygus Na⁺ channel genes (GenBank accession nos. AF378139–AF378144) are underlined with the color used for their mammalian orthologs. To the right of the tree are boxes indicating the Hox gene cluster with which each mammalian Na⁺ channel gene is linked as well as the main tissue in which these genes are expressed.