Scientists tweaked the first few building blocks of the growth‑hormone‑releasing peptide (GRF‑1‑29) and found that swapping them for mirror‑image (D‑) amino acids can make the peptide 12‑ to 38‑fold more powerful at triggering growth hormone release in rats. Some specific changes (like D‑Ala at position 2) were especially effective, while others killed the activity.

In previous reports illustrating the effects of conformational restriction of the N-terminal region of human pancreatic growth hormone releasing factor, we demonstrated that D-amino acid substitutions in either of positions 1, 2, or 3 resulted in greatly increased growth hormone releasing activity both in vivo and in vitro. The most active compound, [D-Ala-2]GRF(1-29)NH2, was 51 times more active than the parent 29 amino acid peptide in the sodium pentobarbital anesthetized rat. These observations have now been extended to analogues containing multiple D-amino acid replacements in these three positions. Once again, peptides with superagonist potencies ranging from 1200% to 3800% were obtained after solid-phase synthesis and purification by medium-pressure reverse-phase liquid chromatography. In addition, it was found that [D-Asn-8]- and [D-Ala-4]GRF(1-29)NH2 were, respectively, 2.43 and 1.1 times more active than GRF(1-29)NH2 itself. In contrast, [D-Phe-6] and [D-Thr-7] analogues were virtually inactive. Chou-Fasman structural predictions suggest that the first three residues of the peptide assume no fixed type of conformation but that a reverse turn could be present between residues 6 and 10. Attempts are made to rationalize the biological results with these calculations. The effects of other side chains on the D-amino acid in position 2 were also investigated. Both the Ac-[D-Phe-2]- and Ac-[D-Arg-2]peptides had very low activity. Several of the inactive peptides were tested as possible antagonists of GRF; however, none was able to block the stimulatory effects of GRF(1-29)NH2 after combined administration.