Given the role of intermediate filaments (IFs) in normal cell physiology and scores of IF-linked diseases, the importance of understanding their molecular structure is beyond doubt. Research into the IF structure was initiated more than 30 years ago, and some important advances have been made. Using crystallography and other methods, the central coiled-coil domain of the elementary dimer and also the structural basis of the soluble tetramer formation have been studied to atomic precision. However, the molecular interactions driving later stages of the filament assembly are still not fully understood. For cytoplasmic IFs, much of the currently available insight is due to chemical cross-linking experiments that date back to the 1990s. This technique has since been radically improved, and several groups have utilized it recently to obtain data on lamin filament assembly. Here, we will summarize these findings and reflect on the remaining open questions and challenges of IF structure. We argue that, in addition to X-ray crystallography, chemical cross-linking and cryoelectron microscopy are the techniques that should enable major new advances in the field in the near future.

The molecular architecture and assembly mechanism of intermediate filaments have been enigmatic for decades. Among those, lamin filaments are of particular interest due to their universal role in cell nucleus and numerous disease-related mutations. Filament assembly is driven by specific interactions of the elementary dimers, which consist of the central coiled-coil rod domain flanked by non-helical head and tail domains. We aimed to investigate the longitudinal ‘head-to-tail’ interaction of lamin dimers (the so-called A_CN interaction), which is crucial for filament assembly. To this end, we prepared a series of recombinant fragments of human lamin A centred around the N- and C-termini of the rod. The fragments were stabilized by fusions to heterologous capping motifs which provide for a correct formation of parallel, in-register coiled-coil dimers. As a result, we established crystal structures of two N-terminal fragments one of which highlights the propensity of the coiled-coil to open up, and one C-terminal rod fragment. Additional studies highlighted the capacity of such N- and C-terminal fragments to form specific complexes in solution, which were further characterized using chemical cross-linking. These data yielded a molecular model of the A_CN complex which features a 6.5 nm overlap of the rod ends.

The limited information available on the structure of complexes involving transcription factors and cognate DNA response elements represents a major obstacle in the quest to understand their mechanism of action at the molecular level. We implemented a concerted structural proteomics approach, which combined hydrogen-deuterium exchange (HDX), quantitative protein-protein and protein-nucleic acid cross-linking (XL), and homology analysis, to model the structure of the complex between the full-length DNA binding domain (DBD) of Forkhead box protein O4 (FOXO4) and its DNA binding element (DBE). The results confirmed that FOXO4-DBD assumes the characteristic forkhead topology shared by these types of transcription factors, but its binding mode differs significantly from those of other members of the family. The results showed that the binding interaction stabilized regions that were rather flexible and disordered in the unbound form. Surprisingly, the conformational effects were not limited only to the interface between bound components, but extended also to distal regions that may be essential to recruiting additional factors to the transcription machinery. In addition to providing valuable new insights into the binding mechanism, this project provided an excellent evaluation of the merits of structural proteomics approaches in the investigation of systems that are not directly amenable to traditional high-resolution techniques.

The configuration of the first prototype timsOmni instrument, which integrates an Omnitrap linear ion trap into a timsTOF platform, is presented and applied to antibody analysis. A modified electrode design for the electron-based fragmentation (ExD) section of the Omnitrap platform is introduced, enhancing both the robustness and performance of the method. Optimal characterization of antibodies requires characterizing light and heavy chains as pairs in addition to sequencing their variable domains and identifying any modifications. This is best addressed using protein centric proteomics as heterogeneity information such as the specific clonal origin of each identified fragment can be retained. Furthermore, by acting on intact proteins that retain part of their structure such as disulfide bonds, it is possible to target selected regions for fragmentation such as some of the hypervariable complementarity determining regions (CDRs) that are unique for each clone and necessary for target recognition. Using electron-based fragmentation methods, we employed the prototype timsOmni mass spectrometer to obtain full CDR3 sequences for paired heavy and light chains with high confidence. Optimal results were obtained by performing Electron Induced Dissociation (EID) at ~35 eV electron energy, on native-like fragment antigen-binding (Fab) precursor ions. This approach yields both (a, x) as well as (c, z) fragment ion pairs with the potential to enhance both sequence coverage and annotation confidence. Overall, the timsOmni mass spectrometer presented here serves as an advanced and versatile platform for protein-centric proteomics, demonstrating exceptional sequencing power for comprehensive protein characterization.