Recombinant Expression and Chemical Amidation of Isotopically Labeled Native Melittin

Abstract

Post-translational modifications are ubiquitous in the eukaryotic proteome. However, these modifications are rarely incorporated in NMR studies of eukaryotic proteins, which are typically produced through recombinant expression in E. coli. Melittin is the primary peptide in honey bee venom. Its native C-terminal amide significantly affects its equilibrium structure and dynamics in solution and is thus a prerequisite for studying its native structure and function. Here, we present a method for producing triply isotopically labeled ( 2 H, 13 C, and 15 N) native melittin through recombinant expression followed by chemical amidation. We then show that structural models produced with AlphaFold-Multimer are in even better agreement with experimental residual dipolar couplings than the 2.0 Å resolution X-ray crystal structure for residues G3−K23.

From Milliseconds to Minutes: Melittin Self-Assembly from Concerted Non-Equilibrium Pressure-Jump and Equilibrium Relaxation Nuclear Magnetic Resonance

“…Melittin, the active component of honeybee venom, is a 26-residue peptide containing a post-translational C-terminal amidation (section II of the Supporting Information) that undergoes exchange between an intrinsically disordered monomer and an α-helical tetramer at millimolar peptide concentrations. Residual dipolar couplings (RDCs) measured in solution showed the same helical structure for the tetramer as observed via X-ray crystallography and predicted by AlphaFold-Multimer . The melittin tetramer forms on the millisecond time scale from monomers with residue P14 in the trans conformation, while there is also a much slower off-pathway exchange between disordered monomers with P14 in the trans and cis conformations.…”

“…Molecular dynamics simulations have suggested that melittin forms partially disordered tetramers by sequential addition of monomers in <300 ns that later relax into a well-ordered crystal-like structure. , We see no evidence from chemical shifts, RDCs, or relaxation measurements for exchange between a rigid and partially disordered tetramer at early time points . Furthermore, dimer–dimer addition has been observed to be ubiquitous in nature for homotetramer formation, possibly due to selective evolutionary pressure to avoid unwanted buildup of trimer species .…”

“…Residual dipolar couplings (RDCs) measured in solution showed the same helical structure for the tetramer as observed via X-ray crystallography and predicted by AlphaFold-Multimer. 16 The melittin tetramer forms on the millisecond time scale from monomers with residue P14 in the trans conformation, while there is also a much slower offpathway exchange between disordered monomers with P14 in the trans and cis conformations. Bioinformatics and mutation studies showed that P14 is required for efficient membranelytic activity, 17,18 but its isomerization appears to be unproductive, even though important for understanding the full kinetic pathway of melittin tetramerization.…”

“…31,32 We see no evidence from chemical shifts, RDCs, or relaxation measurements for exchange between a rigid and partially disordered tetramer at early time points. 16 Furthermore, dimer−dimer addition has been observed to be ubiquitous in nature for homotetramer formation, possibly due to selective evolutionary pressure to avoid unwanted buildup of trimer species. 33 It is plausible that the increased temperature (310 K), a high peptide concentration (10 mM), and short time scales (<300 ns) studied in these simulations probed a qualitatively different tetramerization process to what we observed in solution at 288 K.…”

Experimental NOE, Chemical Shift, and Proline Isomerization Data Provide Detailed Insights into Amelotin Oligomerization

“…Novel approaches for predicting a protein’s structure from its amino acid sequence frequently can generate atomic resolution models that rival or exceed the quality of experimentally determined structures. − An extension of one of these computational methods, AlphaFold, yielded the powerful AlphaFold2 and AlphaFold-Multimer programs, , which have been shown successful at correctly predicting the atomic resolution structures of well-defined oligomeric complexes. − However, no unique predictions are generally available for proteins that do not adopt a distinct oligomeric state of a well-defined size, as is often seen during the initial aggregation process of proteins implicated in amyloid diseases, prior to the formation of highly ordered fibril species.…”