Small chicken growth hormone (scGH) variant in the neural retina: Structure-function correlates

A novel variant of chicken growth hormone (cGH) that is severely truncated has recently been discovered in the neural retina. It is, however, unknown whether this protein binds to GH receptors (GHRs) and has biological activity. This possibility has therefore been addressed by homology modeling, using human (h)GH as a template because it is the only GH molecule with a crystal structure and because hGH binds to cGH receptors (cGHRs). Most of the residues of the small (s)cGH model fitted the hGH template, apart from two restricted regions from Ser 12 to Gln 20 and from Ser 55 to Val 58. The scGH model differs, however, from hGH in structure: hGH is composed of a four-helix bundle, whereas scGH has three main helices. Helices 2, 3, and 4 of hGH correspond to helices 1, 2, and 3 of scGH, but they are longer by one, four, and one residues, respectively. The secondary structure of the C-terminus of scGH is therefore similar to C-terminal hGH. The N-terminus of scGH is, however, severely truncated, lacking the residues of the full-length molecule derived from exons 1, 2, and 3. The N-terminus of scGH also includes 20 residues derived from intron C of full-length cGH. The predicted structure of its N-terminus has no classical secondary structure (α-helix or β-sheet), whereas the N-terminus of hGH is composed of helix 1 and two mini-helices located between helix 1 and 2. This difference in ribbon structure results in a difference in the overall shape of the scGH model and hGH. The possibility that scGH could bind to a GHR dimer was assessed by examining the primary and hypothetical tertiary structure of scGH. hGH binds the extracellular domain (ECD) of two GHRs sequentially at its binding site 1 (or high affinity site) then at its binding site 2 (or low affinity site). Sequence alignment of scGH with hGH demonstrates that scGH lacks three key residues (of 14) at site 1 and nine residues (of 15) at site 2. It is therefore unlikely that tight binding of ECD1 to the site 1 of scGH could occur. scGH also lacks most of the site 2 residues, suggesting that it is unlikely that ECD2 would bind to the scGH model. In summary, we have developed a novel, structural model of scGH, with implications for its putative actions through classical GHRs. Copyright © 2007 Humana Press Inc. All rights of any nature whatsoever are reserved.