Three-dimensional electron diffraction (3D ED)/microcrystal electron diffraction (MicroED) techniques are gaining in popularity. However, the data processing often does not fit existing graphical user interface software, instead requiring the use of the terminal or scripting. Scipion-ED, described in this article, provides a graphical user interface and extendable framework for processing of 3D ED/MicroED data. An illustrative project is described, in which multiple 3D ED/MicroED data sets collected on tetragonal lysozyme were processed with DIALS through the Scipion-ED interface. The ability to resolve unmodelled features in the electrostatic potential map was compared between three strategies for merging data sets.

Eukaryotic TIR (Toll/interleukin-1 receptor protein) domains signal via TIR-TIR interactions, either by self-association or by interaction with other TIR domains. In mammals, TIR domains are found in Toll-like receptors (TLRs) and cytoplasmic adaptor proteins involved in pro-inflammatory signaling. Previous work revealed that the MAL TIR domain (MALTIR) nucleates the assembly of MyD88TIR into crystalline arrays in vitro. A microcrystal electron diffraction (MicroED) structure of the MyD88TIR assembly has previously been solved, revealing a two-stranded higher-order assembly of TIR domains. In this work, it is demonstrated that the TIR domain of TLR2, which is reported to signal as a heterodimer with either TLR1 or TLR6, induces the formation of crystalline higher-order assemblies of MyD88TIRin vitro, whereas TLR1TIR and TLR6TIR do not. Using an improved data-collection protocol, the MicroED structure of TLR2TIR-induced MyD88TIR microcrystals was determined at a higher resolution (2.85Å) and with higher completeness (89%) compared with the previous structure of the MALTIR-induced MyD88TIR assembly. Both assemblies exhibit conformational differences in several areas that are important for signaling (for example the BB loop and CD loop) compared with their monomeric structures. These data suggest that TLR2TIR and MALTIR interact with MyD88 in an analogous manner during signaling, nucleating MyD88TIR assemblies uni­directionally.

This study examines various methods for modelling the electron density and, thus, the electrostatic potential of an organometallic complex for use in crystal structure refinement against 3D electron diffraction (ED) data. It focuses on modelling the scattering factors of iron(III), considering the electron density distribution specific for coordination with organic linkers. We refined the structural model of the metal-organic complex, iron(III) acetylacetonate (FeAcAc), using both the independent atom model (IAM) and the transferable aspherical atom model (TAAM). TAAM refinement initially employed multipolar parameters from the MATTS databank for acetylacetonate, while iron was modelled with a spherical and neutral approach (TAAM ligand). Later, custommade TAAM scattering factors for Fe-O coordination were derived from DFT calculations [TAAM-ligand-Fe(III)]. Our findings show that, in this compound, the TAAM scattering factor corresponding to Fe3+has a lower scattering amplitude than the Fe3+charged scattering factor described by IAM. When using scattering factors corresponding to the oxidation state of iron, IAM inaccurately represents electrostatic potential maps and overestimates the scattering potential of the iron. In addition, TAAM significantly improved the fitting of the model to the data, shown by improved R1 values, goodness-of-fit (GooF) and reduced noise in the Fourier difference map (based on the residual distribution analysis). For 3D ED, R1 values improved from 19.36% (IAM) to 17.44% (TAAM-ligand) and 17.49% (TAAM-ligand-Fe3+), and for singlecrystal X-ray diffraction (SCXRD) from 3.82 to 2.03% and 1.98%, respectively. For 3D ED, the most significant R1 reductions occurred in the low-resolution region (8.65-2.00 A ), dropping from 20.19% (IAM) to 14.67% and 14.89% for TAAM-ligand and TAAM-ligand-Fe(III), respectively, with less improvement in high-resolution ranges (2.00-0.85 A ). This indicates that the major enhancements are due to better scattering modelling in low-resolution zones. Furthermore, when using TAAM instead of IAM, there was a noticeable improvement in the shape of the thermal ellipsoids, which more closely resembled those of an SCXRD-refined model. This study demonstrates the applicability of more sophisticated scattering factors to improve the refinement of metal-organic complexes against 3D ED data, suggesting the need for more accurate modelling methods and highlighting the potential of TAAM in examining the charge distribution of large molecular structures using 3D ED.

Determining macromolecular structures is crucial for understanding biological mechanisms and advancing drug discovery. Recent advancements have highlighted the potential of electron diffraction using continuous sample rotation (MicroED) for resolving these structures from sub-micrometre-sized crystals. However, the achievable data quality of MicroED on proteins is limited by radiation damage of the crystal during collection. Serial electron diffraction (SerialED) avoids this issue by merging single-shot diffraction patterns from thousands of crystals. Despite its potential, the widespread use of SerialED has been limited by the complexity and rarity of the required equipment to acquire these single shot data. Here, we show a new continuous SerialED protocol that is simple, robust and universally accessible. It quickly and efficiently collects all diffraction data from an area without the need for prior crystal identification and enables structure determination of proteins at atomic resolution (0.83 Å) on widely available hardware. Furthermore, we found this technique to produce virtually radiation damage free structures, allowing us to measure radiation sensitive local chemical details and investigate protein-ligand interactions to state-of-the-art accuracy. Continuous SerialED significantly enhances the applicability of ED in structural biology by providing convenient, fast, and high-resolution data collection with minimal radiation damage. These advancements position continuous SerialED as a transformative tool in structural biology by providing high-quality protein structures using tools already available to scientists. We anticipate our protocol to enable a wide range of studies that require high quality diffraction data of radiation sensitive materials.

Understanding oxidation state (OS) variations in enzyme metal-ion cofactors is essential for elucidating enzymatic mechanisms. While traditional spectroscopic techniques can be used to determine oxidation states in materials, they lack spatial resolution. Electron diffraction provides a promising alternative by probing the electrostatic potential (ESP), which is sensitive to valence changes. However, modelling charged species in ESP maps remains challenging, as changes in oxidation states influence the ESP map in a complex manner. In this study, we assessed the impact of OS variations on ESP maps from electron diffraction data. Using serial electron diffraction (SerialED), we determined the structures of two redox states of the iron-containing protein ribonucleotide reductase R2 subunit (R2a).

Isomorphous difference maps computed between the experimental data from the two redox states revealed a signal at the iron sites, which could be attributed to OS changes. Model-derived intensities supported this interpretation, indicating that OS differences contributed ~12-14% to isomorphous difference peaks, while the remainder resulted from atomic displacement between redox states. These findings suggest that differences in scattering amplitude due to oxidation state changes are already detectable within the current accuracy and precision of the data.

To compute structure factors using the transferable aspherical atom model (TAAM), we developed the Python-based wrapper pyDiSCaMB, enabling communication between the MATTS databank and the functionalities available in the cctbx framework. This integration is a crucial step toward implementing TAAM scattering factors in phenix.refine (Afonine et al., 2012), which should enhance phase accuracy and reduce map noise. All in all, this study lays the foundation for oxidation state determination of metal-ion co-factors in metalloenzymes from electron diffraction data.