The authors regret that there are errors in Figs. 10 and 111 of the article referenced above. In subsequent experimentation, the results of the original study generated by J. Amstrup (GenBank: AY02676) could not be reproduced. In the laboratory of E.H. Larsen, this was discovered by Svend Erik Westh Hansen, who subsequently cloned and verified a significant piece of the gene as listed in GenBank: AY02676[1.2] (submitted by Westh Hansen, S.E. and Hviid Larsen, E.). The wrong data has been rectified in this corrigendum by replacing the original Figs. 10 and 11 by the revised versions given below. To accommodate these changes in the text, revised versions of the abstract and Section 3.5 are also included. The corrections do not affect the conclusions of the article, which are still well supported and unchanged. The authors apologize for any inconvenience the errors in the original version of this article may have caused.

In the heterocellular toad skin epithelium the β-adrenergic receptor agonist isoproterenol activates cyclic AMP dependent Cl⁻ channels that are not located in the principal cells. With four experimental approaches, in the present study we tested the hypothesis that the signaling pathway targets apical CFTR-chloride channels of mitochondria-rich cells. (i) Serosal application of isoproterenol (log₁₀EC50 = − 7.1 ± 0.2; Hill coefficient = 1.1 ± 0.2), as well as nor-adrenaline, activated an anion pathway with an apical selectivity sequence, G_Cl > G_BR ≥ G_NO3 > G_I comparable to the published selectivity sequence of cloned human CFTR expressed in Xenopus oocytes. (ii) Known modulators of human CFTR, glibenclamide (200 μmol/l) and genistein (50 μmol/l), depressed and activated, respectively, the receptor stimulated G_Cl. Genistein did not modify the anion selectivity. (iii) Transcellular voltage clamp studies of single isolated mitochondria-rich cells revealed functional β-adrenergic receptors on the basolateral membrane. With ~ 60,000 mitochondria-rich cells per cm², the saturating activation of 11.9 ± 1.6 ns/cell accounted for the measured isoproterenol-activated transepithelial conductance of 600–900 μs/cm². In forskolin stimulated cells, glibenclamide (200 μmol/l) reversibly inhibited the transcellular conductance by 9.6 ± 1.6 ns/cell. (iv) A nucleotide sequence of one third of the Bufo bufo CFTR gene corresponding to the R-domain and part of the first nucleotide binding domain (NBD1) including its Walker motif was amplified from gallbladder epithelium. Somewhat smaller sequences of the BbCFTR were cloned from lung and isolated skin epithelium. The above new results taken together with our previously identified small-conductance CFTR-like Cl⁻ channel in the apical membrane of isolated mitochondria-rich cells provide compelling evidence that the toad's CFTR gene codes for a functional Cl⁻ channel in the apical plasma membrane of this minority cell type.

Epithelia contain few mRNA copies of CFTR, e.g., about 1 copy, or less, of mRNA per cell of human lung. Thus, a cloning strategy was developed by S. E. Westh Hansen (unpublished) for cloning BbCFTR from chromosomal DNA and subsequently to clone the gene from different tissues at the RNA level. The regulatory R-domain of CFTR, unique for ABC proteins, was chosen as target. The structural gene from Xenopus tropicalis was cloned ‘in silico’ from JGI X. tropicalis genome database by BLAST, the exon-intron boundaries, and alignment against CFTR cDNA from Xenopus laevis. The CFTR gene of Xenopus tropicalis spans about 80,000 bp and contains 28 exons. A gene map was constructed by assuming conserved exon/intron boundaries between the species. Multiple alignments indicated that the R-domain has an average length of 732 bp. For generating full-length R-domain PCR primers were designed from the gene map. With this method, also primers on each side of an intron could be constructed that helped to verifying that a cloned CFTR is from RNA. By Clustal W multiple alignments of CFTR cdDNA, including AAK07685, a number of degenerated PCR primers were constructed and several independent clones from chromosomal DNA as well as from mRNA were isolated from gallbladder and skin. All clones had identical DNA sequences. From gallbladder mRNA clones were isolated that span 1400 bp of CFTR, including the R-domain and the first nucleotide binding domain, NBD1. The cloned DNA was assembled into a consensus sequence of CFTR coding regions (3-1394) and listed in GenBank with accession number AY02676[1.2] (authors: Svend Erik Westh Hansen and Erik Hviid Larsen). New Fig. 10 shows alignment of the primary structure of the coned toad gene (BbCFTR) and human CFTR (hCFTR) indicating that the primary structure of the cloned gene codes for the toad's copy of CFTR. Similar cDNA sequences were obtained from skin epithelium and lung epithelium (not shown). New Fig. 11 shows consensus sequences of the cloned BbCFTR and CFTR of 3 other amphibians demonstrating that Bufo clone codes for an amphibian CFTR.