Manipulation of gap junction proteins: effect on cellular coupling and impulse propagation

www.elsevier.com/locate/jelectrocard Manipulation of gap junction proteins: effect on cellular coupling and impulse propagation S.M. Taffet, S. Shah, R.L. Lewandowski, W. Coombs J. Shibayama, P. Sorgen, M. Delmar (Upstate Medical University, Syracuse, NY, USA; University of Nebraska, Omaha, NE, USA) The cardiac gap junction protein, connexin43 (Cx43), is not a passive pore but an actively regulated channel. Under conditions of cellular stress, such as acidification, the pore closes resulting in alterations in impulse propagation. We have previously demonstrated that the carboxyl terminal (CT) domain of Cx43 (Cx43CT) is involved in various intraand intermolecular interactions that regulate gap junctions. Here, we used a nonbiased phage display library to identify novel peptidic sequences that bind Cx43CT and hopefully will alter the properties of the channel. A library of phage containing a randomized series of 12–amino acid peptides (approximately 10 different sequences) was presented to immobilized Cx43CT (pH, 6.5). Analysis of the selected phage revealed a total of 48 Cx43CTinteracting peptides. One third of the peptides had a preserved arginine-X-proline (RXP) sequence. Surface plasmon resonance confirmed binding of 6 of these peptides to Cx43CT. Two of these peptides (RXP-1 and RXP-4) were further analyzed. The interaction of RXP-4 with the CX43-CT was further analyzed by NMR revealing 0022-0736/$ – see front matter D 2005 Published by Elsevier Inc. doi:10.1016/j.jelectrocard.2005.06.091 that this peptide modifies the structure of region 375-379 of Cx43CT. Dual-patch clamp analysis revealed that RXP-4 (but not RXP-1) reduced the frequency with which Cx43 channels transit into the residual state. Unitary conductances were not modified by this peptide. To further analyze the optimal RXP sequences for binding and modifying the function of Cx43, we generated a biased phage display library where the RXP sequence was flanked on each side by 4 random amino acids. Approximately 40000 distinct sequences were presented to Cx43CT. A minor fraction of the inserts resulted from concatenation of 2 peptides in tandem (a bdoublet Q). No doublets were found in a random sampling of 63 inserts obtained before binding. Yet, of the 485 clones sequenced after selection by binding, 85% were doublets, suggesting a high selectivity of Cx43CT for doublets during the binding step. This increased affinity of doublets for CT interaction was confirmed by surface plasmon resonance. Further analysis of the sequences obtained by phage display indicated that there was an increased occurrence of acidic amino acids at position 3 (relative to RXP) but an absence of acidic residues at positions 2 and +1. Basic amino acids were absent from position 1. Overall, these data show that RXP peptides can lead to structural modifications in Cx43CT and changes in the function of Cx43 channels. These results open the door for further development of peptide-based strategies to modify the regulation of Cx43-containing gap junctions in native tissues such as heart. ardiology 38 (2005) 51