Protein structure elucidation from NMR proton densities

The NMR-generated foc proton density affords a template to which the molecule has to be fitted to derive the structure. Here we present a computational protocol that achieves this goal. H(N) atoms are readily recognizable from (1)H/(2)H exchange or (1)H/(15)N heteronuclear single quantum correlation (HSQC) experiments. The primary structure is threaded through the unassigned foc by leapfrogging along peptidyl amide H(N)s and the connected H(α)s. Via a Bayesian approach, the probabilities of the sequential connectivity hypotheses are inferred from likelihoods of H(N)/H(N), H(N)/H(α), and H(α)/H(α) interatomic distances as well as (1)H NMR chemical shifts, both derived from public databases. Once the polypeptide sequence is identified, directionality becomes established, and the foc N and C termini are recognized. After a similar procedure, side chain H atoms are found, including discriminated cis/trans proline loci. The folded structure then is derived via a direct molecular dynamics embedding into mirror image-related representations of the foc and selected according to a lowest energy criterion. The method was applied to foc densities calculated for two protein domains, col 2 and kringle 2. The obtained structures are within 1.0–1.5 Å (backbone heavy atoms) and 1.5–2.0 Å (all heavy atoms) rms deviations from reported x-ray and/or NMR structures.