Identification Of Key Residues At The C-terminus Of Adrenomedullin That Are Responsible For Receptor Affinity And Selectivity Using Alanine Scanning Mutagenesis

Maggie Au, Geoff M Williams, Harriet A Watkins, Margaret A Brimble, Debbie L Hay. University of Auckland, Auckland, New Zealand.

Human adrenomedullin (AM) has gained much recent attention because of its angiogenic properties. Mice which lack the AM or receptor subunit genes died in utero with vascular abnormalities (Fritz-Six et al., 2008). Angiogenesis also occurs in tumours, promoting growth and metastasis. Tumours treated with an AM receptor antagonist, AM22-52, had reduced tumour vascularisation and tumour growth (Iimuro et al., 2004).

AM acts on two receptors; AM1 and AM2 receptors. The AM1 receptor, expressed in endothelial cells, appears to mediate the angiogenic effects of AM. Therefore, blocking the AM pathway may open up the development of new anticancer drugs. AM22-52, a truncated form of full length AM is a weak and non-selective AM receptor antagonist. Therefore, more effective AM1 receptor antagonists are needed to confirm the AM1 receptor as a cancer target.

Robinson et al., (2009) have generated high affinity but non-selective AM receptor antagonists through modifying AM22-52. They speculated that the extreme C-terminus of AM22-52 was a key AM receptor binding region. Therefore, we hypothesised that high affinity and selective AM1 receptor antagonists can be generated by strategic amino acid substitutions in this region of AM22-52.

We explored the role of the individual residues in the C-terminus of AM22-52 in mediating binding to AM receptors. An alanine scan was performed in this region where each of nine residues was individually replaced with an alanine while leaving the rest of the peptide unmodified. These novel antagonists, along with AM22-52 were synthesised using Fmoc solid phase peptide synthesis methodologies. Peptides were purified by RP-HPLC and characterised by LC-MS. Peptides were tested in Cos7 cells transiently transfected with AM1, AM2 or the related CGRP receptor, and cAMP assays were used to determine their affinity and selectivity. Data were analysed using global Schild analysis in Graphpad Prism 4. Schild slopes were constrained to 1 to calculate the pKB value for each antagonist. Statistical analysis was by one-way ANOVA followed by Dunnett’s test where the pKB for each novel antagonist was compared to AM22-52. Circular dichroism (CD) spectroscopy was used to investigate the secondary structure of the peptides in 50% trifluoroethanol. The secondary structure content were analysed using the K2D programme.

Alanine mutants, Y52A, G51A, and I47A were detrimental to binding at all three AM receptors with at least 10-fold reduction of affinity compared to AM22-52 (P<0.01 at the AM1 receptor and total loss of affinity at the AM2 and CGRP receptor). S48A showed a differential effect where it retained affinity at the AM1 receptor, had reduced affinity at the AM2 receptor (P<0.01) and loss total affinity at the CGRP receptor (Table 1). The CD data showed that all the peptides had similar secondary structure content as AM22-52. Therefore, any loss of antagonist activity was not due to misfolding of the peptide.

Table 1. pKB summary for AM22-52 and S48A at the AM1, AM2, and CGRP receptor

Data are shown as pKB ± s.e.m. Numbers in brackets are the number of independent experiments performed in triplicates. **, p < 0.01

Y52, G51 and I47 are essential for AM22-52 binding to all three receptors. S48 appears to contribute to the selectivity of AM22-52 between AM receptors. This study has identified key residues in the C-terminus of AM22-52 that mediate receptor affinity and selectivity.

Fritz-Six K.L. et al., (2008). J. Clin. Invest., 118, 40-50

Iimuro S. et al., (2004). Circ. Res., 95, 415-423

Robinson S.R. et al., (2009). J. Pharmacol Exp. Ther., 331, 513-521