Reciprocal-space mapping of diffuse scattering by serial femtosecond crystallography reveals analog-specific disorder in insulin analogs

Abstract

Insulin detemir and insulin aspart are clinically complementary analogs engineered for distinct pharmacokinetic behavior, yet their comparative structural heterogeneity across temperature regimes remains insufficiently resolved. Here, we present a multi-scale crystallographic analysis integrating near-physiological serial femtosecond crystallography (SFX) with previously reported cryogenic and ambient multicrystal datasets for both analogs. Across conventional quality metrics, reciprocal-space intensity-field reconstructions, model-derived diffuse-scattering representations, Ramachandran stereochemical validation, solvent-accessibility coupling (SAArea-MSArea), and residue-level BDamage (a packing-normalized B-factor metric highlighting local mobility outliers) profiling, we identify a coherent ambient-versus-cryogenic contrast. Ambient datasets show broader reciprocal-space heterogeneity and more diffuse model-space distributions, consistent with increased conformational sampling outside cryogenic trapping. Despite this shared trend, disorder partitioning is analog-specific: detemir exhibits strong pseudo-translational signatures with moderate twinning, whereas aspart shows weak pseudo-translation but pronounced merohedral twinning approaching the theoretical twinned limit in ambient conditions. Importantly, backbone stereochemistry remains globally stable across all datasets, indicating that the observed differences reflect structured heterogeneity rather than model deterioration. Collectively, these findings support an ensemble-aware interpretation of insulin crystallography and provide transferable structural descriptors for analog comparison, stability assessment, and formulation-oriented design.