Ingemar André

Amyloid-β (Aβ) fibrils primarily composed of Aβ_1-40 and Aβ_1-42 form the core of senile plaques in Alzheimer's disease. Aβ_1-40 fibrils may exhibit significant polymorphism influenced by sample preparation conditions, complicating atomic resolution structural characterization. To establish a reliable structural baseline, we developed a protocol for expressing and purifying recombinant Aβ_1-40 that forms monomorphic fibrils under physiological conditions (pH 7.4). We present a high-resolution structure of these unseeded, monomorphic Aβ_1-40 fibrils obtained using magic-angle spinning NMR spectroscopy (PDB ID 12GB). We obtained unambiguous chemical shift assignments for approximately 90% of the residues and measured over 500 distance and torsion angle restraints. The resolved structure, with a backbone RMSD of 0.63 ± 0.06 Å, shows two monomers per filament plane, with two distinct β-sheets (residues E11-E22 and K28-V39, respectively) running along the fibril axis with H-bonding between each plane, and the two strands linked by a flexible loop region. This structure reveals three continuous hydrophobic cores inside each filament which bury 24 hydrophobic side chains per filament plane: those of L17, F19, A21, V24, A30, I32, M35, V40 between the two β-strands within each monomer and I31, L34, V36, V39 between the two monomers. Small angle X-ray scattering reveals the size and geometry of the fibril cross-section, which is compatible with a two-filament arrangement with a total of 4 monomers per fibril plane.

The genetic code is highly redundant, with many synonymous codons encoding the same amino acid. Codon usage influences RNA structure, signaling, and translation rates. Differences in tRNA availability modulate elongation, with rare codons slowing translation and affecting co-translational folding and gene expression. Despite their functional importance and non-random distribution, rare codons are underrepresented in natural datasets, restricting the development of predictive models. We developed a transformer-based model that predicts codon sequences from amino acids, substantially improving rare codon prediction. The model learns codon signatures encoding species identity, RNA thermodynamic properties, and elongation constraints without explicit labels. Attention analysis shows that codon choice depends on both short and long-range sequence contexts, recovering dicodon effects and highlighting additional motifs. Finally, predictions correlate with experimental measurements of the impact of synonymous mutations on protein fitness, linking gene sequence to fitness and functional consequences, providing a framework to connect sequence variation, translation, and protein function.

Early proteins likely arose from a chemically limited set of amino acids available through prebiotic chemistry, raising a central question in molecular evolution: could such primitive compositions yield stable, functional folds? Using de novo design, we constructed three ancient protein architectures using a reduced, evolution-inspired alphabet of ten amino acids, e.g. lacking all basic and aromatic residues. The resulting structures adopted their intended topologies and showed exceptional resistance to thermal and chemical denaturation. Computational simulations further revealed that proteins built from this restricted alphabet were as mutation-resilient as those using all twenty canonical residues. Besides their evolutionary implications, our results provide a foundation for minimalist protein design and generation of simplified, robust systems in minimal cell engineering.

The pattern of substitutions at sites in proteins provides invaluable information about their biophysical and functional importance and what selection pressures are acting at individual sites. Amino acid site rates are typically estimated using phenomenological models where sequence variability is described by rate factors that scale the overall substitution rate in a protein to sites. In this study, we demonstrate that site rates can be calculated accurately from amino acid sequences from a multiple sequence alignment using a mutation-selection model in combination with a simple nucleotide substitution model. The method performs better than the standard phylogenetic approach on sequences generated by structure-based evolutionary dynamics simulations, robustly estimates rates for shallow multiple sequence alignments, and can be rapidly calculated also on larger sequence alignments. On natural sequences, site rates from the mutation-selection model are strongly correlated with rates calculated with the empirical Bayes methods. The model complements other work in providing a link between amino acid substitution rates and equilibrium frequency distributions at sites in proteins. We show how an ensemble of equilibrium frequency vectors can be used to represent the rate variation encoded in empirical amino acid substitution matrices. This study demonstrates that a rapid and simple method can be developed from the mutation-selection model to predict substitution rates from amino acid data, complementing the standard phylogenetic approach.

Chaperone proteins are essential for maintaining proteostasis. Their main role is to assist with the folding of other proteins and to prevent the aggregation of misfolded proteins. The molecular chaperone DNAJB6b efficiently suppresses amyloid formation of several peptides. This activity may rely on the same physicochemical properties as those driving chaperone self-assembly into large micellar-like oligomers. We have therefore undertaken a systematic study of DNAJB6b’s self-assembly under different solution conditions. Using complementary biophysical techniques, we probe variations in aggregation number distribution and hydrodynamic radius, upon variation of pH, temperature, ionic strength, or anions across the Hofmeister series. We find that DNAJB6b maintains its propensity to self-assemble under all solution conditions examined. The size and compactness of the micelles change upon unfolding of the C-terminal domain, although a folded C-terminal domain does not drive micelle formation, which can likely be ascribed to hydrophobic interactions in the linker region. Mass photometry reveals that monomers of DNAJB6b coexist at equilibrium with the micelles. Furthermore, the free energy barrier for micelle dissociation into monomers was estimated by measuring dissociation rate constants at different temperatures.

Proteins can misfold during cotranslational folding, but how codon sequences influence this process is not well understood. Here, we develop an in vivo assay to comprehensively study the impact of single synonymous substitutions on protein folding efficiency and apply it to the N-terminal domain of E. Coli protein ddlA. We map the influence of codons along the sequence and demonstrate that codons can substantially influence the folding efficiency and that the impact depends on the structure and topology of the protein. A cluster of codons associated with residues in the center of the domain fold strongly influences the folding efficiency. Moreover, substitutions to rarer codons generally lead to increased folding efficiency. A mRNA sequence exclusively made up of rare codons results in higher expression than one with only common codons. Our results highlight the importance of rare codons in cotranslational folding and the relationship between codon sequence and protein structure.

Virus capsid formation is one the most complex self-assembly processes in nature, involving the association of a large number of protein subunits into well-defined structural states. The ability of protein subunits to efficiently self-assembly is encoded in the protein sequence, and ultimately in the protein-protein interfaces within the growing capsid. The relative strengths of interfaces can be important in directing the self-assembly process, and determine which intermediates are formed early in the self-assembly process. In this study we characterize the interfaces in homomeric T=1 virus capsids to investigate to what extent interfaces within the assemblies are different from other protein-protein interfaces, and which interfaces are most critical for self-assembly. Interfaces were divided into dimers, trimers, and pentamers and compared to interfaces of non-viral homomeric dimers, trimers, and pentamers. The analysis suggests that viral interfaces are larger than non-viral counterparts, and differ in amino acid content, but are energetically similar in terms of the quality of intermolecular interactions. Trimers are predicted to be the most stable oligomers, which may imply that they form early in the self-assembly process. However, dimeric and pentameric interfaces are typically similar in terms of predicted stability suggesting that assembly formation in T=1 capsids may progress through many different routes, rather than progressing through a single dominant intermediate species. With symmetric docking calculations, the energy landscape of the assembled capsid was characterized, and the results highlight that the assemblies exhibit deeply funneled energy landscapes encoded by protein-protein interfaces that have a high degree of specificity.

Variation in mutation rates at sites in proteins can largely be understood by the constraint that proteins must fold into stable structures. Models that calculate site‐specific rates based on protein structure and a thermodynamic stability model have shown a significant but modest ability to predict empirical site‐specific rates calculated from sequence. Models that use detailed atomistic models of protein energetics do not outperform simpler approaches using packing density. We demonstrate that a fundamental reason for this is that empirical site‐specific rates are the result of the average effect of many different microenvironments in a phylogeny. By analyzing the results of evolutionary dynamics simulations, we show how averaging site‐specific rates across many extant protein structures can lead to correct recovery of site‐rate prediction. This result is also demonstrated in natural protein sequences and experimental structures. Using predicted structures, we demonstrate that atomistic models can improve upon contact density metrics in predicting site‐specific rates from a structure. The results give fundamental insights into the factors governing the distribution of site‐specific rates in protein families.

The pattern of substitutions at sites in proteins provides invaluable information about their biophysical and functional importance and what selection pressures are acting at individual sites. Amino acid site rates are typically estimated using phenomenological models in which the sequence variability is described by rate factors that scale the overall substitution rate in a protein to sites. In this study, we demonstrate that site rates can be calculated accurately from amino acid sequences using a mutation-selection model in combination with a simple nucleotide substitution model. The method performs better than the standard phylogenetic approach on sequences generated by structure-based evolutionary dynamics simulations, robustly estimates rates for shallow multiple sequence alignments, and can be rapidly calculated also on larger sequence alignments. On natural sequences, site rates from the mutation-selection model are strongly correlated to rates calculated with the empirical Bayes methods. The model provides a link between amino acid substitution rates and equilibrium frequency distributions at sites in proteins. We show how an ensemble of equilibrium frequency vectors can be used to represent the rate variation encoded in empirical amino acid substitution matrices. This study demonstrates that a rapid and simple method can be developed from the mutation-selection model to predict substitution rates from amino acid data, complementing the standard phylogenetic approach.

Variation in mutation rates at sites in proteins can largely be understood by the constraint that proteins must fold into stable structures. Models that calculate site-specific rates based on protein structure and a thermodynamic stability model have shown a significant but modest ability to predict empirical site-specific rates calculated from sequence. Models that use detailed atomistic models of protein energetics do not outperform simpler approaches using packing density. We demonstrate that a fundamental reason for this is that empirical site-specific rates are the result of the average effect of many different microenvironments in a phylogeny. By analyzing the results of evolutionary dynamics simulations, we show how averaging site-specific rates across many extant protein structures can lead to correct recovery of site-rate prediction. This result is also demonstrated in natural protein sequences and experimental structures. Using predicted structures, we demonstrate that atomistic models can improve upon contact density metrics in predicting site-specific rates from a structure. The results give fundamental insights into the factors governing the distribution of site-specific rates in protein families.