Conformational origin of a difficult coupling in a human growth hormone releasing factor analog.

During the solid-phase synthesis of the human growth hormone releasing factor (GRF) analog [Ala15, Leu27, Asn28] -GRF(1-32)-OH, incorporation of Boc-Gln16 was determined to be incomplete. While aggregation of growing resin-bound peptide chains with concomitant beta-sheet formation and "precipitation" has been proposed to account in general for such "difficult coupling," no feature of sequence in the Gln16 region of this GRF analog provided an immediate rationale for this result. We now report 500 MHz 1H NMR spectra of a series of resin-bound GRF segments surrounding the Gln16 position (19-32 through 14-32), swelled in dimethylsulfoxide-d6 solutions [GRF(14-32) = Leu14-Ala-Gln-Leu-Ser(Bzl)-Ala-Arg(Tos)-Lys(CIZ)-Leu- Leu-Gln-Asp(OcHex)-Ile-Leu-Asn-Arg(Tos)-Gln-Gln-Gly32-PAM resin]. While relatively sharp spectra are observed for GRF(19-32), components with resonances broadened by an order-of-magnitude appear in spectra of the 18-32 and 17-32 peptide-resin, and the entire spectrum of 16-32 is ill-resolved and highly broadened. Subsequent spectra sharpen again (15-32, 14-32). These combined synthesis/spectroscopic experimental results, in conjunction with predictive analyses using standard Chou-Fasman 2 degrees structure parameters, suggest that the completeness of the Gln16 coupling is hindered by formation of a specific, folded beta-sheet/beta-turn structure in GRF(16-32) (with the turn located at 18-21, "upstream" of the difficult coupling site), and accompanying aggregation of peptide chains. This analysis suggests that awareness of such potential beta-sheet/beta-turn sequences can guide analog choices, and/or facilitate pre-programming of synthesis steps in anticipation of problem couplings.