Reinvestigation of the proposed folding and self-association of the Neuropeptide Head Activator

Abstract

The Neuropeptide Head Activator (HA), pGlu-Pro-Pro-Gly-Gly-Ser-Lys-Val-Ile-Leu-Phe (pGlu is pyroglutamic acid), is involved in head-specific growth and differentiation processes in the freshwater coelenterate Hydra attenuata. Peptides of identical sequence have also been isolated from higher-organism tissues such as human and bovine hypothalamus. Early studies by molecular sieve chromatography suggested that HA dimerizes with high affinity (K(d) approximately 1 nM). This dimerization was proposed to occur via antiparallel beta-sheet formation between the Lys(7)-Phe(11) segments in each HA molecule. We conducted biophysical studies on synthetic HA in order to gain insight into its structure and aggregation tendencies. We found by analytical ultracentrifugation that HA is monomeric at low millimolar concentrations. Studies by (1)H-NMR revealed that HA did not adopt any significant secondary structure in solution. We found no NOEs that would support the proposed dimer structure. We probed the propensity of the Lys(7)-Phe(11) fragment to form antiparallel beta-sheet by designing peptides in which two such fragments are joined by a two-residue linker. These peptides were intended to form stable beta-hairpin structures with cross-strand interactions that mimic those of the proposed HA dimer interface. We found that the HA-derived fragments may be induced to form intramolecular beta-sheet, albeit only weakly, when linked by the highly beta-hairpin-promoting D-Pro-Gly turn, but not when linked by the more flexible Gly-Gly unit. These findings suggest that the postulated mode of HA dimerization and the proposed propensity of the molecule to form discrete aggregates with high affinity are incorrect.

Figure 1.

(A) Proposed HA homodimer interface (Bodenmuller et al. 1986), where Z is pyroglutamic acid. Postulated hydrogen bonds are indicated by dotted lines. The dimerization is proposed to be driven by specific contact between residues on opposing strands, and a symmetrical pair of salt-bridges between the ɛ-amino of Lys₇ on one HA molecule with the C-terminal carboxylate of the other HA molecule. (B) Hairpin conformations of peptides (HA-f)₂-GG and (HA-f)₂-pG that would mimic the interactions that are proposed to stabilize the dimer interface. Hydrogen bonding pattern between specific residues would remain in registry if residues 10 and 11 adopted a 2:2 reverse turn. In (HA-f)₂-GG, X₁₀ = Gly, and in (HA-f)₂-pG, X₁₀ = d-Pro.

Figure 2.

Representative equilibrium analytical ultracentrifugation for 1.0 mM HA in aqueous 50 mM acetic acid, pH 5.0 at 297 K. Absorbance data were acquired at 257 nm. Presence of a single species is indicated by linearity of the data. The slope is directly proportional to the molecular weight. Linear least-square regression analysis yielded molecular weight estimates consistent with HA monomer. The data shown were acquired at a rotor speed of 56 krpm, resulting in an experimental mass of ∼1024 g mole⁻¹ (theoretical mass of HA is 1126 g mole⁻¹). Discrepancies between observed and predicted masses may be due to nonideality effects. The fits were judged as adequate by randomness of the residuals about zero (top). Fit is shown as a solid line. Theoretical plots for monomer and dimer are also shown, as dashed lines.

Figure 3.

(A) Deviation of α-proton chemical shifts for HA from random coil values (Wuthrich 1986) plotted against residue. Literature random coil values for pGlu are not available, therefore this residue is not plotted. The datum for the Phe₁₁ is also not plotted due to effects from the C-terminal carboxylate. The lack of consistently large deviations from random coil (δδαH > 0.1 ppm) implies that the peptide does not adopt significant secondary structure. Analysis was performed at 277 K in 90% H₂O/10% D₂O containing 100 mM acetic acid-d₆, pH 5.0 (uncorrected). (B) Chemical shift analysis for regions of (HA-f)₂-GG (open bars) and (HA-f)₂-pG (solid bars) that contain the HA-related sequences. Data acquired under the conditions described in A. Turn positions [Xaa₁₀ and Gly_11, where Xaa = Gly in (HA-f)₂-GG and Xaa = d-Pro in (HA-f)₂-pG] are not plotted. Data for Phe₁₆ of both peptides are also not included due to C-terminal carboxylate effects. Significant positive α-proton chemical shift deviations from random coil for (HA-f)₂-pG suggest some β-sheet content.

Figure 3.

(A) Deviation of α-proton chemical shifts for HA from random coil values (Wuthrich 1986) plotted against residue. Literature random coil values for pGlu are not available, therefore this residue is not plotted. The datum for the Phe₁₁ is also not plotted due to effects from the C-terminal carboxylate. The lack of consistently large deviations from random coil (δδαH > 0.1 ppm) implies that the peptide does not adopt significant secondary structure. Analysis was performed at 277 K in 90% H₂O/10% D₂O containing 100 mM acetic acid-d₆, pH 5.0 (uncorrected). (B) Chemical shift analysis for regions of (HA-f)₂-GG (open bars) and (HA-f)₂-pG (solid bars) that contain the HA-related sequences. Data acquired under the conditions described in A. Turn positions [Xaa₁₀ and Gly_11, where Xaa = Gly in (HA-f)₂-GG and Xaa = d-Pro in (HA-f)₂-pG] are not plotted. Data for Phe₁₆ of both peptides are also not included due to C-terminal carboxylate effects. Significant positive α-proton chemical shift deviations from random coil for (HA-f)₂-pG suggest some β-sheet content.

Figure 4.

Long-range NOEs from 200 msec NOESY and ROESY spectra of (HA-f)₂-pG that are consistent with β-hairpin formation. One signal could not be unambiguously assigned due to spectral overlap (between the α-protons of Leu₁₅ and Val₆ or Ile₇) and is indicated by a double-headed arrow. Extremely weak NOEs are depicted by dotted lines.