Posters

Download list of posters here and read the abstracts below. Abstracts of semi-confidential posters not shown.

The posters will be located in the corridors of the conference venue. The poster owners are encouraged to put up their posters when they arrive.

POSTER NO 1 - Alain André:

Biosynthesis within bio-inspired engineered artificial condensates

Alain André, Ankush Garg, Nikita Rehnberg, Magnus Kjærgaard

Aarhus University

Contact: alain.andre@mbg.au.dk

Biomolecular condensates offer a versatile platform for spatially organizing biochemical reactions. Yet their potential for creating an optimized micro-environment for multi-enzymatic reactions is largely unexplored in synthetic biology.  Here, we utilize resilin-like proteins, a bio-inspired repeat protein, to generate tuneable condensates as reaction crucibles for enzymatic cascades. Using fluorescence microscopy, we characterize the physical properties of these condensates and their capacity to recruit enzymes.

As a case study, we test a part of the anthocyanin pathway to investigate the conversion of para-coumarate to naringenin within the condensate environment, chosen for its relevance to natural product biosynthesis. Preliminary results confirm the formation of stable condensates with controlled recruitment of enzymes. Ongoing experiments seek to elucidate how these compartments influence enzymatic reactions, such as the potential metabolic channelling effects. Our findings highlight the innovative potential of engineered condensates for biocatalysis.

POSTER NO 2 - Anna Toft Sakamoto:

Investigating the structure and function of cable bacteria truncated hemoglobins

Anna Toft Sakamoto(1,2), Benjamin Smed Korsgaard(1,2), Max Theo Ben Clabbers(2,3), Thomas Boesen(1,2,3)

(1) Center for Electromicrobiology, Aarhus University, Ny Munkegade 116, 8000 Aarhus C, Denmark. (2) Department of Molecular Biology and Genetics, Universitetsbyen 81, 8000 Aarhus C, Denmark. (3) Interdiciplinary Nanoscience Center, Aarhus University, Gustav Wieds Vej 14, 8000 Aarhus C, Denmark

Contact: ats@mbg.au.dk

Cable bacteria are filamentous bacteria, with a unique ability to perform long-distance (centimeter) transport of electrons between spatially separated electron donors and acceptors. This is believed to occur through a conduction machinery consisting of periplasmic conductive fibers with putative associated factors involved in delivering to and extracting electrons from the fibers. An interesting protein family suggested to be involved in electron transport in cable bacteria is the pentaheme cytochrome (PHC) family. Certain members of this family have an N-terminal or/and C-terminal truncated hemoglobin (trHb) domain and were suggested to act as an alternative terminal oxidase.

To understand the role of trHbs in terminal oxidase activity of cable bacteria the truncated hemoglobin domains (trHbs) were expressed and purified in high quantities and purity and used for structural and functional studies. The experimental atomic structure of the N-terminal trHb (tN) was previously determined by X-ray crystallography. In contrast, it has so far not been possible to obtain the structure of the C-terminal trHbs (tC). This reason for this was that it was difficult to get crystals large enough for X-ray crystallography and the proteins are too small for single particle analysis by cryo-EM. We were able to produce tiny crystals suitable for microcrystal electron diffraction and are now pursuing the tC structure with this method The results obtained will be presented.

POSTER NO 4 - Bram Mylemans:

De novo designed bifunctional proteins for targeted protein degradation

Bram Mylemans1,2, Amanda Acevedo-Jake3, Boguslawa Korona4, Andrew J. Wilson3, Laura S. Itzhaki4, Derek N. Woolfson1,2,5,6

1. Novo Nordisk Foundation Center for Protein Design, Department of Biology/Department of Drug Design and Pharmacology, University of Copenhagen, Denmark
2. School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, UK
3. School of Chemistry, University of Birmingham, Edgbaston, Birmingham B15 2TT
4. Department of Pharmacology, University of Cambridge, Cambridge, CB2 1PD, UK
5. Max Planck-Bristol Centre for Minimal Biology, University of Bristol, Cantock’s Close, Bristol BS8 1TS
6. School of Biochemistry, University of Bristol, Medical Sciences Building, University Walk, Bristol BS8 1TD, UK

Contact: bram.mylemans@bio.ku.dk

Rapid advances in AI-based design and modelling methods have dramatically increased the success rates of protein design. The creation of novel binders to any protein target has become feasible. However, for many practical applications binding alone is not sufficient. Here, we describe the design of a bifunctional de novo miniprotein that binds to a specified protein of interest and directs its targeted degradation by the proteasome. First, we rationally designed a water-soluble helical hairpin scaffold. This scaffold is highly mutable and allows for binding sites to be added either on the outer faces of the helices or within a constrained loop. Motif grafting augmented with AI tools—ProteinMPNN1 and AlphaFold22—was used to design specific binding sites for the anti-apoptotic cancer target BCL-xL3. ITC measurements confirmed nanomolar binding and a crystal structure reveals that side chains in the binding site align with the design model. A similar approach was used to introduce a loop capable of recruiting the E3 ligase adaptor KLHL205.  When introduced into cells with elevated levels of BCL-xL, the resulting bifunctional protein caused significant degradation of this target. Moreover, the designed miniproteins induced apoptosis on par with a known PROTAC drug5 and at higher levels then the monofunctional binders.

POSTER NO 5 - Celia Fricke:

Thermodynamic Stability Modulates Chaperone-Mediated Disaggregation of α-Synuclein Fibrils

Celia Fricke[a], Antonin Kunka [a], Rasmus K. Norrild[a], Shuangyan Wang[a], Thi Lieu Dang[b], Jonas Folke[c], Mohammad Shahnawaz[d], Claudio Soto[d], Susana Aznar[c], Anne S. Wentink[e], Bernd Bukau[b] and Alexander K. Buell[a]

[a] Department of Biotechnology and Biomedicine,

Technical University of Denmark,

Søltofts Plads, Building 227, 2800 Kgs. Lyngby, Denmark

[b] Center for Molecular Biology of Heidelberg University (ZMBH)

DKFZ-ZMBH Alliance

Heidelberg, Germany

[c] Centre for Neuroscience and Stereology,

Department of Neurology, Copenhagen University Hospital, Bispebjerg and Frederiksberg Hospital,

Nielsine Nielsens Vej 6B, Entrance 11B, Copenhagen, DK-2400, Denmark.

[d] Mitchell Center for Alzheimer’s Disease and Related Brain Disorders, Department of Neurology,

University of Texas McGovern Medical School at Houston, Houston, TX, USA.

[e] Leiden Institute of Chemistry

Leiden University Einsteinweg 55, 2333 CC Leiden, Netherlands

Contact: celfri@dtu.dk

Aggregation of the intrinsically disordered protein alpha-synuclein into amyloid fibrils and their subsequent intracellular accumulation are hallmark features of several neurodegenerative disorders, including Parkinson’s disease, for which no curative treatments currently exist. In this study, we investigate the relationship between fibril morphology, thermodynamic stability, and susceptibility to disaggregation by the human chaperone system comprising HSP70, DNAJB1, and Apg2. By varying assembly conditions and incubation times, we generated alpha-synuclein fibrils with diverse morphological and biochemical properties, including a broad range of thermodynamic stabilities, which we quantified using a chemical depolymerization assay. The chaperone system effectively disaggregated three of the four fibril types, with efficiencies that correlated with their thermodynamic stabilities. One fibril type resisted disaggregation despite exhibiting a comparable stability to those that were disaggregated, suggesting that additional structural features influence chaperone susceptibility. Our findings establish a quantitative link between fibril stability and chaperone-mediated disaggregation for three in vitro αSyn fibril types as well as fibrils amplified from brain extracts of MSA and PD patients and underscore the importance of fibril thermodynamics in biologically relevant disaggregation processes and disease pathology.

POSTER NO 6 - Christa Kanstrup:

Cytoplasmic IDRs as Regulators of Plant POT Activity

Christa Kanstrup, Adam T. Henriksen, Meike Burow

Department of Plant and Environmental Sciences (PLEN), Faculty of Science, University of Copenhagen

Contact: cka@plen.ku.dk

The Major Facilitator Superfamily (MFS) represents one of the largest and most functionally diverse groups of transporters, facilitating the movement of a wide range of substrates across cellular membranes. Within this superfamily, the Proton-dependent Oligopeptide Transporter (POT) family is known for transporting di- and tri-peptides via proton symport. In plants, this family is referred to as the Nitrate and Peptide transporter Family (NPF) and has undergone a notable expansion in gene numbers. While most NPF members mediate the transport of primary metabolites such as nitrate and peptides, some have evolved to transport specialized compounds. The first NPF members identified as transporters of specialized metabolites were the glucosinolate transporters GTR1, GTR2, and GTR3. Their expression patterns, substrate specificities, and physiological roles in root–shoot translocation and seed loading are well understood. Extensive mutational studies have pinpointed key residues in the transmembrane cavity that determines specificity. Despite this functional insight, little is known about how the GTRs are regulated at the protein level. Structurally, NPF members - including the GTRs - consists of 12 transmembrane helices arranged in two bundles, separated by a long cytoplasmic loop between helices 6 and 7, and with both N- and C-terminal tails also facing the cytoplasm. These cytoplasmic regions are generally predicted to be intrinsically disordered, suggesting they may play a role in regulating transporter function. Specifically, in GTR2, we have in vitro evidence from Xenopus laevis oocyte uptake assays, that the C-terminal tail is important for transporter functionality. Truncation of 23 amino acids, from the 29 amino acid tail, almost completely abolishes transporter function, and specific substitution mutations in the tail render the transporter partially or even fully inactive. Building on these findings, we are currently investigating the function of the C-terminal tail in GTR2 in planta by complementing gtr2 knockout plants with GTR2 variants with truncated and mutated tails. In parallel, we are examining the subcellular localization of these variants to assess potential effects on membrane integration, and we are performing protein–protein interaction studies focused on the C-terminal tail. These efforts aim to deepen our understanding of how structural features of the C-terminal tail regulate transporter activity.

POSTER NO 7 - Christian Kofoed:

Programming Protein Function via Cell Surface Induced Proximity

Christian Kofoed, Girum Erkalo, Nicholas E. S. Tay, Xuanjia Yi, Yutong Lin, Tom W. Muir

UCPH, Department of Drug Design and Pharmacology

Contact: christian.kofoed@sund.ku.dk

ell surface landscapes vary with cell type and are frequently reshaped in disease. Harnessing these differences underpins many therapeutic strategies and the development of advanced diagnostic and research tools. Yet most approaches rely on single-antigen recognition, which rarely defines cell identity with high precision. To address this, we developed SMART (Splicing-Modulated Actuation upon Recognition of Targets), a programmable platform that uses proximity-gated protein trans-splicing to convert combinatorial surface proteins into user-defined outputs. By coupling caged split inteins to antibody fragments or mimetics (such as DARPins), peptides, or small molecules, SMART performs Boolean logic (AND, OR, NOT) on live-cell surfaces to achieve ligation of functional proteins. We demonstrate the cell-selective, in situ generation of the protein “glue” SpyCatcher for precision delivery, as well as antigen-gated activation and release of IL-1β, shaping the surrounding microenvironment. The platform’s modularity, tunable thresholds, and retain-versus-release control position SMART as a versatile induced-proximity modality for proteomics, diagnostics, and precision therapeutics.

POSTER NO 8 - Daniel Saar:

The Subtle Art of Nuclear Receptor Modulation Captured by NMR:  MTMR7 as a Non-Canonical PPARγ Modulator with Potentially Low Affinity

Daniel Saar1,2,3, Elisabeth G. K. Thomsen1,2,3, Philip Weidner4, Elke Burgermeister4, Birthe B. Kragelund1,2,3

1 The REPIN and 2 The Linderstrøm-Lang Centre for Protein Science, Department of Biology, University of Copenhagen, Copenhagen, Denmark. 3 Structural Biology and NMR Laboratory, Department of Biology, University of Copenhagen, Copenhagen, Denmark. 4 Department of Medicine II, Medical Faculty Mannheim, Heidelberg University, Mannheim, Germany.

Contact: daniel.saar@bio.ku.dk

Activation of human transcription factor PPARγ typically involves agonist binding in the ligand binding pocket and co-activator binding via an LXXLL motif after a ligand-induced conformational change [1, 2]. The phosphatase MTMR7 was predicted as a PPARγ co-activator through the action of its coiled-coil domain, impacting metabolic disorders, inflammation, and cancer, but the mechanism was unclear [3]. Our NMR data shows that MTMR7 binds PPARγ unconventionally, using an alternative surface without involving the co-activator binding site or ligand binding pocket and in the absence of a canonical binding motif. We identify two PPARγ binding sites for MTMR7 on the PPARγ ligand binding domain and show that the coiled-coil domain of MTMR7 stabilizes the ligand binding domain in its active conformation independently of ligand binding, potentially competing with the PPARγ AB domain. Rosiglitazone, a synthetic agonist for PPARγ enhances MTMR7's effect but is not required for MTMR7 binding.

POSTER NO 10 - Darian Yang:

Driving MD Simulations with Dynamic Experimental Data: Adaptive Refinement of Conformational Ensembles

Darian T. Yang & Kresten Lindorff-Larsen

Structural Biology and NMR Laboratory & the Linderstrøm-Lang Centre for Protein Science, Department of Biology, University of Copenhagen, Copenhagen, Denmark

Contact: darian.yang@bio.ku.dk

Understanding protein conformational flexibility is essential for structure prediction and design beyond static snapshots. While molecular dynamics (MD) simulations provide dynamic, atomistic insights, they often fail to fully capture experimental observables due to force field inaccuracies and insufficient sampling. Reweighting techniques can improve agreement with experiments but cannot recover unsampled states. To address this, we introduce a weighted ensemble (WE) simulation method that adaptively enhances sampling using dynamic experimental observables.

 

WE simulations improve sampling efficiency by running multiple unbiased trajectories in parallel, selectively replicating pathways-of-interest while merging redundant ones. Unlike conventional biasing approaches for accelerating MD simulations, our method directly incorporates time-dependent experimental observables, such as NMR relaxation rates, into the resampling step, balancing both trajectory diversity and experimental agreement. This strategy enables guided exploration of dynamically relevant conformations and is broadly applicable to any observable that depends on time correlation functions or multi-frame MD information.

 

We demonstrate our approach on T4 lysozyme simulations using NMR relaxation data as an example system. Our method samples a more diverse set of protein conformations and achieves faster convergence with experimental data compared to standard MD, capturing a more complete and valid conformational ensemble.

 

Our dynamic observable-guided WE approach uniquely enables the integration of experimental data involving time-dependent molecular motion, overcoming limitations of conventional MD and biasing methods. This framework is broadly applicable to refining protein structural ensembles and improving or validating the accuracy of structure prediction and design models.

POSTER NO 11 - Emil G P V Stender:

RAPID KD SCREENING WITH A SINGLE MEASUREMENTUSING FLOW INDUCED DISPERSION ANALYSIS

Philipp Willmer1,2, Adam C. Hundahl1, Rodolphe Marie2, Henrik Jensen1

1FIDA Biosystems ApS, 2860 Søborg, Denmark www.fidabio.com

2Technical University of Denmark, Department of Health Technology, 2800 Kongens Lyngby, Denmark

Contact: egps@fidabio.com

We present Continuous Titration Based Spectral Related Intensity Change (cSPRING), using FIDA-LD to measure a dissociation constant (KD) in seconds to minutes while reducing the sample preparation time 8-fold.

When a fluorophore is conjugated to a target protein, conformational changes or binding events can alter its emission spectrum. Detecting these spectrum-related intensity changes for different ligand concentrations can be used to determine a KD by performing a titration using different ligand concentrations.

To overcome labor-intensive sample preparation and accelerate time to result, we introduce a continuous titration method, where the target protein is exposed to a ligand concentration gradient generated by Taylor dispersion. This allows obtaining the KD and the hydrodynamic radius (Rh) of the ligand from a single experiment, making it ideal for quantitative screening of large libraries of small molecules.

POSTER NO 13 - Eric Beyerle:

Machine Learning Ligand Binding Pathways

Eric R. Beyerle and Kresten Lindorff-Larsen

University of Copenhagen

Contact: eric.beyerle@bio.ku.dk

There is no reason to expect ligands bind to proteins via a single pathway, in general. And, since the work done (or released) and transit time to bind are path-dependent, it is useful to study what pathways a ligand can take to bind to protein to understand why a pharmaceutical is effect or methods to develop new and effective drugs. Here, we propose a method to cluster ligand binding pathways using a combination of the dynamical time warping algorithm and machine learning. We use this method to characterize binding pathways of benzene to a T4 lysozyme mutant and rank the quality of coarse-grained force fields used to model this system. This technique helps us understand the relevant driving forces between ligand-protein binding and how changes in the model Hamiltonian affect pathway preference and population.

POSTER NO 14 - Fan Cao:

A coarse-grained model for disordered and multi-domain proteins

Fan Cao, Sören von Bülow, Giulio Tesei, Kresten Lindorff-Larsen

Structural Biology and NMR Laboratory & the Linderstrøm-Lang Centre for Protein Science, Department of Biology, University of Copenhagen, Copenhagen, Denmark

Contact: fan.cao@bio.ku.dk

Many proteins contain more than one folded domain, and such modular multi-domain proteins help expand the functional repertoire of proteins. Because of their larger size and often substantial dynamics, it may be difficult to characterize the conformational ensembles of multi-domain proteins by simulations. Here, we present a coarse-grained model for multi-domain proteins that is both fast and provides an accurate description of the global conformational properties in solution. We show that the accuracy of a one-bead-per-residue coarse-grained model depends on how the interaction sites in the folded domains are represented. Specifically, we find excessive domain–domain interactions if the interaction sites are located at the position of the Cα atoms. We also show that if the interaction sites are located at the center of mass of the residue, we obtain good agreement between simulations and experiments across a wide range of proteins. We then optimize our previously described CALVADOS model using this center-of-mass representation, and validate the resulting model using independent data. Finally, we use our revised model to simulate phase separation of both disordered and multi-domain proteins, and to examine how the stability of folded domains may differ between the dilute and dense phases. Our results provide a starting point for understanding interactions between folded and disordered regions in proteins, and how these regions affect the propensity of proteins to self-associate and undergo phase separation.

POSTER NO 15 - Federica Saraceno:

Characterising the phase separation and aggregation behaviour of  the C-terminal LCD of TDP-43

Federica Saraceno, Soumik Ray, Rasmus K. Norrild, Sophie Hertel, Baptiste A. Zanchet, Alexander K. Buell

Department of Biotechnology and Biomedicine, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark

Contact: fedsa@dtu.dk

TDP-43, a nucleic acid-binding protein central to RNA metabolism in our cells, exhibits a remarkable ability to undergo liquid-liquid phase separation (LLPS) through its intrinsically disordered C-terminal low complexity domain (LCD). This region harbours the majority of mutations and post-translational modifications sites associated with neurodegenerative disorders, such as amyotrophic lateral sclerosis and frontotemporal dementia.

Despite substantial progress in research, the molecular mechanisms underlying the transition of functional TDP-43 within condensates into pathogenic amyloid fibrils are not yet fully understood. Although numerous disease-associated mutations have been identified, the phase separation dynamics and aggregation behaviour of the wild-type TDP-43 LCD are yet to be fully elucidated. Thus, a thorough biophysical characterization of the wild type would represent the key to establish a foundational scaffold, paving the way for comparative analyses with pathological mutants.

In this study, we aim to systematically study the LLPS and aggregation behavior of the wild-type TDP-43 LCD, both in its native form and with engineered solubility tags. By employing a comprehensive approach that integrates highly quantitative in-house microfluidics, state-of-the-art biophysical and microscopy techniques, and interferometric light scattering microscopy, we aim to thoroughly characterise the thermodynamic and kinetic determinants of TDP-43 LCD self-assembly.

Combined, these approaches will enable an in-depth investigation of the complex aggregation mechanisms of the protein, ultimately providing critical insights into the molecular determinants of TDP-43 aggregation and its role in disease.

POSTER NO 16 - Frederik Oskar Graversgaard Henriksen:

Antitoxin CrlA of the CrlTA toxin-antitoxin system mediates defence against T5-like phages

Frederik O. G. Henriksen, Ragnhild B. Skjerning, Ditlev E. Brodersen

Department of Molecular Biology and Genetics, Aarhus University, Universitetsbyen 81, 8000 Aarhus C

Contact: fogh@mbg.au.dk

Bacteriophages are a large threat to bacteria in the wild, eradicating roughly 40% of all bacterial life every day. Bacteria, therefore, evolved many ways to defend themselves from these predators, and the discovery of highly specialised defence systems has skyrocketed in recent times. Many of these include so-called toxin-antitoxin systems, where the antitoxin keeps the toxin at bay and functions as a sensor for phage infection, releasing the toxin if a phage is detected.

However, we find that for the type II TA system CrlTA in Yersinia pestis, the antitoxin alone is sufficient to grant complete protection against T5-like phages of the BASEL collection. CrlA is a small 11 kDa protein expected to contain a HTH-type DNA-binding domain. Comparing homologous proteins shows a conservation of positively charged amino acids located around the HTH region, suggesting that DNA binding is a conserved trait in CrlA proteins.

Interestingly, T5 phages are generally known to be resistant to restriction systems, while maintaining their DNA unmodified and not encoding DNA-like proteins, which are common ways of overcoming restriction systems. 

In silico prediction of CrlA DNA binding using its own promoter as a bait predicts interaction with one specific sequence. This sequence is also enriched in the T5 genome ‘injection’ stop region. An essential component for T5 cell takeover.

This suggests that CrlA might work as a novel defence initiator against T5 phages. That could work through the inhibition of T5’s two-step genome injection process.

POSTER NO 17 - Georgia Nasi:

Amyloid fibril formation from rapeseed proteins through mild processing

Georgia Nasi1, Azad Farzadfard1, Thomas Oliver Mason1, Svitlana Mykolenko2, Raffaele Mezzenga2, Alexander Kai Buell1

1Department of Biotechnology and Biomedicine, Technical University of Denmark (DTU), 2800 Kongens Lyngby, Denmark

2Department of Health Sciences and Technology, ETH Zurich, 8092 Zurich, Switzerland

Contact: geona@dtu.dk

The growing demand for sustainable protein sources has also led to increased interest in plant-derived proteins, which are abundant in agricultural by-products and waste streams. In recent years, these proteins have gained attention not only as nutritional components but also as promising building blocks for protein-based biomaterials (PBMs). Among their structural states, the amyloid form of plant proteins has attracted particular interest due to its exceptional mechanical and functional properties. However, most existing methods to obtain functional PBMs are energy-consuming and rely on harsh thermochemical conditions to convert globular proteins into fibrillar structures. In this study, we investigated the main storage proteins of rapeseed cake, napin and cruciferin, as underutilized protein sources for the formation of amyloid-based biomaterials. We explored whether these proteins can assemble into amyloid-like fibrils under gentle, low-energy conditions, avoiding the conventional thermochemical routes typically used for amyloid conversion. Using biophysical characterization techniques, such as Thioflavin T fluorescence kinetic assays, transmission electron microscopy and circular dichroism, we demonstrated that napin can form amyloid-like fibrils, providing the first experimental evidence of its ability to fibrillate. In contrast to traditional treatments that induce random hydrolysis and structural degradation, our approach employs milder heating and controlled chemical environments that preserve the integrity of the protein chains. This allows a larger fraction of the protein mass to participate in the formation of ordered fibrillar structures, improving both energy efficiency and molecular-level utilization in the process. Overall, this work contributes to a broader vision of circular bioeconomy, transforming agricultural by-products into functional materials while promoting environmentally responsible protein valorization.

POSTER NO 19 - Hossein Mohammad-Beigi:

Taylor Dispersion Analysis of Micellization (TDAM) Reveals Distinct Assembly and Dissociation Pathways of α-, β-, and κ-Casein Micelles

Hossein Mohammad-Beigi, Thomas O. Mason, Tijs Albert Maria Rovers, Tanja Christine Jæger, Marie Sofie Møller, Richard Ipsen, Anni Bygvrå Hougaard, Birte Svensson, Alexander K. Buell

Department of Biotechnology and Biomedicine, Technical University of Denmark, Søltofts Plads, Building 227, DK-2800 Kgs. Lyngby, Denmark

Contact: hosmbe@dtu.dk

Casein micelles are essential protein structures in milk and are crucial for its stability and nutritional properties. Understanding their self-assembly and dissociation dynamics is vital for applications in food science, biotechnology, and pharmaceuticals. In this study, we introduce Taylor Dispersion Analysis of Micellization (TDAM), a microcapillary-based technique combining intrinsic and extrinsic fluorescence detection to investigate the Ca²⁺-dependent colloidal stability, viscosity, and association/dissociation kinetics of α-, β-, and κ-casein (αCN, βCN, κCN) micelles. Our results reveal distinctly different behaviors between the types of caseins: αCN micelles exhibit the lowest stability, rapidly dissociating in the absence of Ca²⁺ and pronounced sensitivity to Ca²⁺-induced size changes. βCN forms concentration-dependent micellar species, with complex dissociation patterns, while κCN micelles remain highly stable under all investigated conditions. TDAM, complemented by dynamic light scattering (DLS) and finite element modeling, uncovers the critical influence of Ca²⁺ on micelle association, viscosity, and dissociation kinetics. Although excessive Ca²⁺ can induce aggregation and thereby reduce the colloidal stability of casein micelles, Ca²⁺ also promotes the formation of micellar aggregates that are more resistant to dissociation upon dilution under flow conditions and exhibit reduced monomer exchange rates, particularly for κCN. This study provides fundamental insights into casein micellization mechanisms, advancing the understanding of their structural and functional properties. The TDAM methodology offers a powerful tool for characterizing protein self-assembly in complex matrices, with broad implications for dairy science, biomaterials, and protein engineering.

POSTER NO 21 - Ida Sjøgaard:

Regulation of the stress responsive transcription factor DREB2A through formation of a ternary complex between the repressor RCD1 and the MED25 subunit

Ida M Z Sjøgaard, Andreas Prestel, Steffie Elkjær, Frederik F Theisen, Birthe B Kragelund & Karen Skriver

University of Copenhagen

Contact: ida.zobbe@bio.ku.dk

The Arabidopsis thaliana transcription factor DREB2A interacts with subunit 25 of the transcriptional regulator complex, Mediator, and positive and negative co-regulators to guide transcriptional activity of target genes in response to abiotic stress signals. DREB2A binds to the ACID-domain of MED25 through two separate short linear motifs (SLiMs) engaging a novel ACID-binding motif (ABS) and the known RCD1-binding motif (RIM).

Here, we examine the bivalent ABS-RIM binding region as a molecular switch to integrate signals from co-regulators through formation of ternary complexes with the MED25-ACID domain.

POSTER NO 22 - Isabell Lindahl:

Who induces what? Exploring Membrane Curvature Generation and   Sensing of Mitochondrial Membrane Proteins using Coarse-Grained MD

Isabell Lindahl, Paulo Telles de Souza, Weria Pezeshkian

University of Copenhagen, Niels Bohr Institute

ENS de Lyon

Contact: isabell.lindahl@nbi.ku.dk

Often referred to as the 'powerhouse of the cell', mitochondria generate the majority of cellular ATP through respiration. A process facilitated by the respiratory chain (RC) consisting of the electron transport chain (ETC) and ATP-synthase, located in the cristae of the inner mitochondrial membrane (IMM). The cristae are the folded regions of the IMM, hosting most of the energy conversion machinery, the RC complexes. The shape of the cristae is dynamically organized and has important implications for mitochondrial function and energy production. Structural studies suggest that the ETC complexes and the formation of ETC super complexes contribute to the generation of membrane curvature. However, this has not been fully quantified, and the underlying mechanisms remain elusive. In this work, we employ coarse-grained molecular dynamics simulations using the Martini3 force field, TS2CG and GROMACS to investigate the curvature- inducing of individual ETC complexes and their super complex assembly within a lipid bilayer resembling the IMM. For the analysis we have developed a computational tool, CALM (Calibrate and Analyze Lipid Membranes), which enables efficient quantification of membrane curvature, membrane thickness, bending rigidity and membrane rigidifications induced by the protein. These parameters are often important for mesoscale computer simulations of biomembranes and allow for membrane protein interaction classification through only a few parameters.

In this poster, I will be presenting aspects of the tool CALM and our results on complex IV of the ETC-complexes from porcine hearts. Preliminary results indicate that Complex IV alone generates curvature in a membrane, suggesting that super complex formation isn ́t necessary for membrane shape deformation.

POSTER NO 23 - Jacob Aunstrup Larsen:

Solubility profiling of Peptides with minimal sample consumption

Jacob Aunstrup Larsen (1), Marie Østergaard Pedersen (2) & Alexander Kai Buell (1)

(1): Technical University of Denmark, Department of Bioengineering

(2): Global Research Technology, Novo Nordisk A/S, Måløv, Denmark

Contact: Jaula@dtu.dk

Peptide-based biologics are increasingly attractive pharmaceutical molecules, with GLP-1 agonists making a particular splash recently. Advances in peptide drug design and array-based synthesis now enable early-stage high-throughput screening. However, developability assays of critical parameters, such as solubility, often require up to a milligram of material, making them incompatible with small-scale synthesis and creates critical bottlenecks in drug discovery pipelines.

 

We developed a turnkey solution for high-throughput solubility profiling that requires less than 10 µg purified material. Our method is built on an automated microfluidic platform (Fida 1) that injects a small plug of molecules through a laminar flow channel. Here, diffusive species sample many flow-lines and form a Gaussian distribution – a Taylor Dispersion, whereas molecules with low diffusivity will distribute according to the laminar flow profile. This distinction enables quantification of diffusive vs. non-diffusive species or soluble vs. precipitated proteins.

We perform classical Ammonium Sulphate (AMS) precipitation assay in this framework by loading peptide samples into increasing AMS concentrations. Thereby, we can precipitate, separate and quantify the soluble (diffusive) peptide in a single 4 min experiment, while consuming only tens of nL sample or a few µg material per assay.

We can adapt this approach by varying buffer pH instead of AMS concentration, to rapidly evaluate pH-solubility profiles. Additionally, by modulating the flow-rate during the experiments, we can even study kinetically limited precipitation and probe the kinetic barriers of amorphous aggregation. Finally, we demonstrate that this same approach can characterize peptide solubility directly from 80% DMSO stock solutions, and thereby avoid laborious peptide solubilisation or buffer exchange.

 

This unified approach to study protein solubility profiles is directly compatible with high-throughput peptide synthesis technologies. Together, this establishes a platform to integrate solubility as a routine developability parameter in peptide discovery. We are now applying it to a set of 300 peptides, curated for diverse amino acid compositions, to generate a large, coherent training set for machine learning-based solubility prediction tools.

POSTER NO 24 - Jeppe Kari:

A Michaelis-Menten model for heterogeneous substrates: An Energy Distribution Approach

Poul Thrane^1,  Johanne Gudmand-Høyer^2, Morten Andersen^2, Ulf Rørbæk Pedersen^3, William O. Hancock^4 and Jeppe Kari^1

1)  Dept. of Science and Environment, RUC

2) Centre for Mathematical Modeling, RUC

3) Glass and Time, IMFUFA, RUC \samelineand

4) Dept. of Biomedical Engineering and Chemistry, Penn. State University

Contact: jkari@ruc.dk

Enzymes are typically analyzed under the assumption of homogeneous substrates, yet many biological, biotechnological, and industrial reactions involve chemically or physically heterogeneous substrates. Here we present a theoretical framework showing that mixtures of non-identical substrates at steady state still follow the Michaelis–Menten (MM) rate law, yielding apparent parameters  (K_M_app and k_cat_app) that depend on the mean and variance of the underlying energy distributions. This mathematical identity with the classical MM form can bias interpretation of fitted parameters and conceal mechanistic diversity.

We show that variance in substrate energetics can shift K_M_app and k_cat_app as strongly as changes in mean binding or activation energies, making substrate heterogeneity an independent but hidden design axis. Numerical simulations of 500 coupled Michaelis–Menten reactions - each representing a distinct substrate - validated the closed-form expressions for the apparent parameters. Seperate analysis of 43220 curated BRENDA entries quantified the kinetic spread in enzymology, and our framework shows that realistic substrate heterogeneity gives comparable variability. Our framework extends the MM theory to heterogeneous substrates by adding a single, physically interpretable variance term to the classical equations. This enables inference of hidden heterogeneity from bulk fits and positions substrate pretreatment as a complementary optimization axis alongside conventional enzyme engineering for enzymes acting on heterogeneous substrates.

POSTER NO 27 - Ketty  Tamburrini:

Deciphering the Disordered Interaction of Prothymosin α with S100A13

Catarina B. Fernandes1,2, Ketty C. Tamburrini1,2, Louise Pinet3, Freia Buus1,2, Johan G. Olsen1,2, Andreas Prestel1, Benjamin Schuler3, Cy M. Jeffries4, Birthe B. Kragelund1,2

1Structural Biology and NMR Laboratory, 2REPIN, Department of Biology, University of Copenhagen, Ole Maaløes Vej 5, 2200 Copenhagen, Denmark,3Department of Biochemistry, University of Zurich, Zurich, Switzerland and 4European Molecular Biology Laboratory, Hamburg Site, Notkestrasse 85, Building 25 A, Hamburg, 22607, Germany

Contact: ketty.tamburrini@bio.ku.dk

Prothymosin α (ProTα) is a small, highly negatively charged, intrinsically disordered protein. ProTα displays a dual role: intracellularly, it is involved in chromatin remodeling, transcription, cellular proliferation, and apoptosis; extracellularly, it can recruit and activate innate immune cells, promoting inflammation and cytokine secretion. Recently, S100A13 has been shown to mediate the extracellular release of ProTα. S100A13 is a calcium-binding homodimer from the EF-hand family, known for its role as a molecular hub with various binding partners. As an alarmin, S100A13 is crucial in signaling cellular damage and triggering inflammatory responses when released into the extracellular space. Despite the pivotal roles of ProTα and S100A13, the molecular details of their interaction remain unexplored. In this study, we characterized the S100A13-ProTα interaction using nuclear magnetic resonance (NMR) spectroscopy, single molecule Förster resonance energy transfer (smFRET), chemical cross-linking, and small-angle X-ray scattering (SAXS). Our findings reveal that ProTα dynamically binds to the positively charged surface of the S100A13 dimer without forming persistent residue-specific contacts. The binding occurs in a region on S100A13, previously unknown for ligand binding, suggesting that similar binding mechanisms may apply to other disordered, negatively charged proteins. We propose that S100A13 may act as a transport chaperone for ProTα under certain conditions, and we suggest potential fates for the complex in the extracellular space.

POSTER NO 28 - Kristoffer Johansson:

Global Analysis of Multi-Mutants to Improve Protein Function and Stability

Kristoffer E. Johansson, Kresten Lindorff-Larsen and Jakob R. Winther

The Linderstrøm-Lang Centre for Protein Science, Department of Biology, University of Copenhagen.

Contact: kristoffer.johansson@bio.ku.dk

Stabilizing proteins without otherwise hampering their function is a central task in protein engineering and design. Multiplexed assays can screen thousands of protein variants in a single high-throughput experiment, for properties and at conditions that are relevant for the design task. Using such assays, we have demonstrated that large libraries of protein variants with several amino acid substitutions each are highly informative for design purposes. To realize this, we have developed a global multi-mutant analysis (GMMA) that can disentangle the effects of individual amino acid substitutions. This approach has proven robust and effective in three different studies: The stabilization of a designed protein to >150 °C, enhancement of green fluorescent protein, and stabilization of a molecular sensor across more

functional constraints. In all cases, the optimization is achieved in a single round of experiments by introducing 5-9 mutations.

POSTER NO 29 - kyriakos tsatsis:

Exploring domain nesting in carbohydrate-active enzymes

Kyriakos Tsatsis, Marie Sofie Møller

DTU Bioengineering

Contact: kyrts@dtu.dk

Carbohydrate-active enzymes (CAZymes), such as glycoside hydrolases (GHs), often include carbohydrate-binding modules (CBMs) that enhance substrate targeting. While CBM fusion has been explored to create multimodular CAZymes , success has been limited due to poor understanding of their spatial architecture. The distance and spatial orientation between CBMs and catalytic domain (CD) active sites critically influence activity and stability but remain under investigated.

This project investigates domain nesting, a rare configuration in CAZymes where one domain is inserted within the CD, and its impact on structural and functional coupling. Using a combined experimental and bioinformatics approach, we aim to understand the functional advantages and disadvantages  of nested domains  and enhance the toolbox of domain fusion for designing enzymes with improved thermal stability, resistance to degradation, and broadened substrate specificity.

POSTER NO 31 - Louise Laursen:

A PROSS-designed lipoprotein lipase displays enhanced thermal stability and native regulation mechanism

Louise Laursen and Michael Ploug

Finsen Laboratory, Rigshospitalet, Copenhagen, Denmark

Biotech Research and Innovation Centre, University of Copenhagen, Copenhagen, Denmark

Contact: louise.laursen@finsenlab.dk

Familial Chylomicronaemia Syndrome (FCS) is a rare genetic disorder characterized by elevated plasma triglycerides (TG) levels leading to increased risk of acute pancreatis.  TG accumulates in the blood due to impaired or absent lipoprotein lipase (LPL) activity, that are caused by dysfunctional or non-functional LPL gene variants or genes that affect the folding and transport of LPL. The standard treatment of FCS patients is a strict diet of low-fat intake as no effective pharmacological treatment is available. Different strategies have been tested; inhibitor targeting and enzyme replacement by gene therapy. Inhibitor targeting is only available for FCS patients with impaired LPL activity, whereas enzyme replacement is in theory an efficacious treatment for all FCS patients. The first generation of enzyme replacement was based on the sequence of WT LPL, that are an inherent unstable protein, thus the treatment was later discontinued due to low outcome. Therefore, we designed, purified and tested several PROSS generated LPL variants to increase the thermostability of LPL. The increased thermostability of LPL reduced the enzymatic activity significant. We applied the FuncLib software on the thermostable LPL PROSS variant to gain flexibility in the active site and hopefully increase activity. The LPL variant developed by combining the PROSS and FuncLib software shows enhanced thermostability and decent enzymatic activity. The LPL variant is a promising candidate to develop an enzyme replacement treatment for FCS patients.

POSTER NO 32 - Luis I Gutierrez Rus:

Expanding all-alpha-helical protein space through rational computational design

Luis I. Gutierrez-Rus, Katherine I. Albanese, Joel J. Chubb, Xiyue Leng, Kathleen W. Kurgan, Bram Mylemans, Katarzyna Ożga, Rokas Petrenas, Andrey V. Romanyuk, Ross Anderson, Jonathan Clayden, Graham J. Leggett, Jennifer J. McManus, Thomas A. A. Oliver, N

NNF Center for Protein Design

Department of Drug Design and Pharmacology, University of Copenhagen

School of Chemistry, University of Bristol

Max Planck-Bristol Centre for Minimal Biology, University of Bristol

Contact: luis.rus@bio.ku.dk

Natural proteins achieve remarkable structural and functional complexity, often through the integration of multiple domains. Current AI-based de novo protein design methods generate new scaffolds but still fall short of reproducing the complexity of natural multidomain proteins. Here, we present an alternative strategy that leverages rational and computational design to overcome these limitations.

Building on sequence-to-structure rules for coiled-coil assemblies, we previously designed a range of single-chain coiled-coil scaffolds with controlled oligomeric states and topologies. In this work, we extend this framework to multidomain proteins. By combining rational design rules with computational modelling, we developed a pipeline for constructing higher-order architectures from modular α-helical domains.

The resulting multi modular coiled-coil scaffolds (mCCs) explore regions of the all-α structural space largely unpopulated by natural proteins. Structural similarity analysis shows that most designs lack detectable counterparts in experimental or predicted databases, highlighting their ability to expand the all α-helical structural space. Like natural multidomain proteins, mCCs achieve complexity through hierarchical assembly and controlled inter-domain connectivity, but under explicit sequence-to-structure control.

Importantly, these scaffolds are not only structurally complex but also modular, predictable, and amenable to functionalisation, including small-molecule and protein binding as well as catalytic activity. This enables predictable variation and orthogonal functional combinations within a single polypeptide chain.

Together, these results establish a generalisable framework for building novel classes of α-helical multidomain proteins, bridging the gap between evolutionary modularity and rational protein design.

POSTER NO 33 - Maria Haugaard Bohl Andersen:

Two mutations convert the first T-antigen specific sialidase into a trans-sialidase

Michael Jakob Pichler, Marina Corbella, Sebastian Meier, Maria Haugaard Bohl Andersen, Tine Sofie Nielsen, Sanchari Banerjee, Jens Preben Morth, Carme Rovira, Maher Abou Hachem

Section for Protein Chemistry & Enzyme Technology, DTU Bioengineering, Technical University of Denmark

Contact: mhban@dtu.dk

AmGH181 from Akkermansia muciniphila is the first known inverting sialidase with remarkable selectivity for the sialyl T-antigen, a cancer-associated glycan prevalent on tumor cell surfaces. We show that this specificity is governed by a flexible tryptophan–histidine pair that forms a “sugar tang”, precisely positioning the sialyl T-antigen for catalysis.

Remarkably, by introducing only two mutations — one of which installed a tyrosine as the catalytic nucleophile — we altered the reaction stereochemistry. The resulting retaining mutant acquired trans-sialidase activity and synthesized 3′-sialyllactose in exceptionally high yields.

This work provides mechanistic insight into the evolutionary divergence of inverting and retaining sialidases and sheds light on a novel innovative strategy to repurpose inverting glycosidases for transglycosylation, oligosaccharide synthesis, and glycoengineering applications.

POSTER NO 34 - Markéta Linhartová:

CooJ -The first nickel-binding protein purified from cable bacteria

Markéta Linhartová(1), Viktoria Svane(2), Nykola Jones(3), Cecilie Sølund Bolø (1), and Thomas Boesen (1,2,4)

(1) Center for Electromicrobiology, Institute for Biology, Aarhus University, Denmark

(2) Department of Molecular Biology & Genetics, Aarhus University, Denmark

(3) ISA, Department of Physics and Astronomy, Aarhus University, Denmark

(4) Interdisciplinary Nanoscience Center, Aarhus University, Denmark

Contact: linhart.marketa@bio.au.dk

Cable bacteria are filamentous Gram-negative bacteria that are intensively studied for their highly conductive periplasmic polymers, likely based on a protein backbone loaded with an unprecedented nickel-ligating cofactor. Due to a lack of experimental atomic structure, the available data indicate that the cofactor resembles synthetic nickel(II) bis-dithiolene polymers. Importantly, there is no evidence of a biosynthetic pathway for such polymers in living organisms; therefore, it is tempting to search for protein molecules that are: i) able to polymerise and ii) coordinate nickel in a tetra-sulfur structure mimicking the nickel(II) bis-dithiolene character. The current knowledge of nickel homeostasis in cable bacteria is largely based on genomic data analyses because proteomic studies are hampered by difficulties in cell collection and protein extraction. We aimed to identify the Ni-binding proteins from lysed cable bacteria using purification with Ni-affinity chromatography and mass spectrometry identification. CooJ from cable bacteria, a nickel chaperone supporting the activity of [NiFe] carbon monoxide dehydrogenase (CODH), was repeatedly identified as selectively interacting with Ni-NTA resin. Overexpression of this CooJ variant in E. coli cells led to production of a soluble protein that can be purified using a protocol for purification of His-tagged proteins based on the natural polyhistidine motif conserved in the cable bacteria CooJ sequence. An interesting observation was that purified CooJ formed filamentous homooligomers, which we documented by negative stain single particle analysis using transmission electron microscopy. Further data on CooJ stability in buffers and interactions with ions and metals obtained from thermal unfolding experiments and using synchrotron radiation circular dichroism spectroscopy are in accordance with the published description of other CooJ homologs (Darrouzet et al., 2021, Journal of Inorganic Biochemistry).

POSTER NO 37 - Mette Errebo Rønne:

Carbohydrate-binding module regulation of enzymatic transglycosylation

Mette Errebo Rønne and Marie Sofie Møller

Applied Molecular Enzyme Chemistry, Department of Biotechnology and Biomedicine, Technical University of Denmark, Denmark

Contact: meero@dtu.dk

Oligosaccharides and their conjugates are widely utilized in industry, particularly as health-promoting prebiotics and in the formulation of therapeutics and cosmetics. Industrial production primarily relies on two approaches: chemical synthesis from smaller precursors, which enables the generation of well-defined products and glycoconjugates, and the degradation of polysaccharides via chemical or enzymatic methods, typically yielding complex mixtures. However, the development of cost-effective and sustainable synthesis strategies remains a major challenge. Biotechnological methods, especially enzymatic synthesis, offer environmentally friendly alternatives to chemical processes and hold considerable promises for the production of structurally diverse and novel compounds.

Carbohydrate-active enzymes (CAZymes), especially glycoside hydrolases (GHs), can catalyze glycosidic bond formation through transglycosylation (TG). Optimizing the transglycosylation-to-hydrolysis (T/H) ratio requires careful engineering. While studies typically focus on the catalytic domain alone, emerging research suggests that. This study aims to clarify the role of carbohydrate binding domains (CBMs) in TG reactions and to offer strategies and method development for enhancing TG yields in the GH5 family with confirmed TG activity on mannan. We hypothesize that CBMs promote TG by increasing local substrate concentration, product elongation, or active site conformation.

We compare ten selected enzymes from the GH5_8 family, each containing carbohydrate-binding modules (CBMs) from six different CBM families (CBM2, CBM3, CBM10, CBM13, CBM35, and CBM59). We investigate the effects of CBMs on transglycosylation (TG), and in selected cases, examine the influence of CBM positioning relative to the catalytic domain. The insights gained from this study will be applied to the design of GH fusion constructs to enhance TG efficiency and substrate specificity through CBM engineering, with the aim of advancing enzyme engineering for the industrial synthesis of oligosaccharides.

POSTER NO 38 - Nadja Joachim:

Phosphorous recovery using optimised phosphate binding proteins

Nadja Joachim, Kaare Teilum

University of Copenhagen, Department of Biology, SBiNlab

Contact: nadja.joachim@bio.ku.dk

The overall aim of this project is to construct a phosphate-specific chromatographic column that can recover inorganic phosphate (Pi) from wastewater, using the bacterial phosphate binding protein PstS. Phosphorous is essential for plant growth and modern agriculture relies on its use in fertilizers. However, phosphorous reserves are limited, and by the use in farming, it ends up in sewage or washed out of the soil and into our waterways, lakes and oceans mainly as inorganic phosphate, Pi.1  PstS is a highly selective and high affinity binder of Pi, and could therefore serve as a tool to recover phosphate.2 By screening a library of 96 diverse PstS sequences, highly stable PstS variants were identified. To engineer a pH sensitive switch to release bound phosphate, we implemented a point mutation in the binding site, which leads to preferred binding of the monobasic compared to dibasic phosphate, and therefore disfavours binding above pH 8.3 As the binding site is highly conserved, this mutation can also be introduced in the stable homologues identified in the screen, which resulted in a more stable PstS variant with pH sensitive affinity for Pi, demonstrated by ITC. In order to engineer a more sensitive switch and even more stable variants, we study the dynamics of the protein, especially the hinge movement between Pi bound and Pi free state, using NMR methods. We aim to further optimise phosphate release properties as well as protein stability to produce a durable PstS column.

 

1 Cordell, D. and White, S. (2014) Life’s Bottleneck: Sustaining the World’s Phosphorus for a Food Secure Future. Annu. Rev. Env. Resour. 39, 161–188.

2 Luecke, H. and Quiocho, F. A. (1990) High specificity of a phosphate transport protein determined by hydrogen bonds. Nature 347, 402–406.

3 Wang, Z., Luecke, H., Yao, N. and Quiocho, F. A. (1997) A low energy short hydrogen bond in very high resolution structures of protein receptor-phosphate complexes. Nat. Struct. Biol.

POSTER NO 39 - Nicole Galenkamp:

Cooperative Protein Binding on simple and complex membranes

Nicole Stéphanie Galenkamp, Peter Jönsson, Sara Linse, Emma Sparr

Division of Physical Chemistry, Department of Chemistry, Lund University, 221 00 Lund, Sweden

Contact: nicole.galenkamp@fkem1.lu.se

Cellular membranes play a fundamental role in protein binding, membrane remodeling, and cellular signaling. Amphipathic proteins, such as α-synuclein (aSyn), have been shown to bind membranes cooperatively, forming dense patches rather than distributing uniformly. While this clustering is essential for physiological membrane remodeling, it may also promote protein aggregation, contributing to neurodegenerative diseases such as Alzheimer’s. However, the molecular factors governing cooperative binding and its consequences on membrane dynamics remain poorly understood.

This study aims to investigate the real-time dynamics of cooperative protein binding to extracellular vesicles (EVs) and model membranes to identify key factors regulating this process. We hypothesize that cooperative binding is modulated by lipid composition, protein conformation, and membrane curvature. Using a combined top-down and bottom-up approach, we will assess how membrane interfacial crowding, asymmetry, and lipid heterogeneity influence cooperative binding. By progressively simplifying EV membranes and reconstructing vesicles with extracted lipids, we will dissect the specific lipid-protein interactions driving cooperative behavior.

POSTER NO 43 - Peter Røgen:

Sequence-Similar Protein Domain Pairs With Structural or Topological Dissimilarity

Peter Røgen

DTU Compute

Contact: prog@dtu.dk

For a variety of applications protein structures are clustered by sequence similarity and sequence-redundant structures are disregarded. Sequence-similar chains are likely to have similar structures but significant structural variation, as measured with RMSD, has been documented for sequence-similar chains and found usually to have a functional explanation. Moving two neighboring stretches of backbone through each other may change the chain topology and alter possible folding paths. The size of this motion is compatible to a variation in a flexible loop. We search and find domains with alternate chain topology in CATH4.2 sequence families relatively independent of sequence identity and of structural similarity as measured by RMSD. Structural, topological, and functional representative sets should therefore keep sequence-similar domains not just with structural variation but also with topological variation.

 

We present BCAlign that finds Alignment and superposition of protein Backbone Curves by optimizing a user chosen convex combination of structural derivation and derivation between the structure-based sequence alignment and an input sequence alignment. Steric and topological obstructions from deforming a curve into an aligned curve are then found by a previously developed algorithm. For highly sequence-similar domains sequence-based structural alignment better represents the chains motion and generally reveals larger structural and topological variation than structure-based does. Fold-switching protein pairs have been reported to be most frequent between X-ray and NMR structures and estimated to be underrepresented in the PDB as the alternate configuration is harder to resolve. Here we similarly find chain topology most frequently altered between X-ray and NMR structures.

POSTER NO 44 - Ryan Cantwell Chater:

Leveraging cryo-EM to design allosteric inhibitors of GlyT2 that alleviate neuropathic pain without on-target side effects

Ryan P. Cantwell Chater1,2, Julian Peiser-Oliver2, Tanmay K. Pati3, Ada S. Quinn4,5, Irina Lotsaris2, Zachary J. Frangos6,7, Anna E. Tischer6, Billy J Williams-Noonan4,5, Karin R. Aubrey8, 9, Megan L. O’Mara4,5, Michael Michaelides6, Sarasa A. Mohammadi9,

1Department of Drug Design and Pharmacology, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark.

2School of Medical Sciences, University of Sydney, Sydney, NSW, Australia.

3Department of Chemistry and Chemical Biology, Rensselaer Polytechnic Institute, Troy, New York, USA.

4Australian Institute of Bioengineering and Nanotechnology, The University of Queensland, St Lucia, Queensland, Australia.

5ARC Industry Transformation Training Centre for Cryo-electron Microscopy of Membrane Proteins (CCeMMP).

6Biobehavioral Imaging & Molecular Neuropsychopharmacology Unit, Neuroimaging Research Branch, National Institute on Drug Abuse, Intramural Research Program, Baltimore, MD, USA.

7Medication Development Program, National Institute on Drug Abuse, Intramural Research Program.

8Pain Management Research Institute, Kolling Institute, Royal North Shore Hospital, St Leonards, New South Wales, Australia.

9Faculty of Medicine and Health, University of Sydney, Sydney, New South Wales, Australia.

10School of Pharmacy, University of Sydney, Sydney, NSW, Australia.

Contact: ryan.chater@sund.ku.dk

Chronic neuropathic pain, caused by nerve damage or disease, is increasing in prevalence, but current treatments are ineffective and over-reliant on opioids. The neuronal glycine transporter, GlyT2 (SLC6A5), regulates inhibitory glycinergic neurotransmission and represents a promising target for new analgesics. However, most GlyT2 inhibitors cause significant side effects, in part due to irreversible inhibition at analgesic doses. Here we develop a reversible inhibitor of GlyT2, RPI-GLYT2-82, and identify its binding site by determining the first cryo-EM structures of human GlyT2 (70kDa membrane protein). We capture three fundamental conformational states of GlyT2 in the substrate-free state (inward-open conformation) (2.97 Å), and bound to either glycine (inward-occluded conformation) (3.02 Å), RPI-GLYT2-82 (2.79 Å) or the pseudo-irreversible inhibitor ORG25543 (2.49 Å) (outward-open conformation). We have achieved determined the cryo-EM structures without the need of fiducial markers. We demonstrate that RPI-GLYT2-82 dissociates from GlyT2 faster than ORG25543, providing analgesia in mouse neuropathic pain models without on-target side-effects or addiction liability. Our data provide a mechanistic understanding of allosteric inhibition of glycine transport, enabling structure-based design of non-opioid analgesics.

POSTER NO 45 - Selene Forget:

FreeQuantum: Bridging Quantum Computing and Machine Learning for Accurate Biomolecular Binding Free Energy Predictions

Selene Forget, Beatriz Piniello, Gemm Solomon, Kresten Lindorff-Larsen

NNF Quantum Computing Programme, Niels Bohr Institute, University of Copenhagen, Denmark

Structural Biology and NMR Laboratory & the Linderstrøm-Lang Centre for Protein Science, Department of Biology, University of Copenhagen,

Contact: selene.forget@nbi.ku.dk

Interactions of biomolecules and metallic elements are central to many biological functions and underpin the design of metallodrugs. So far, the characterization of these interactions, and notably the quantification of the corresponding binding free energies, remains highly challenging. The existing in silico prediction tools, typically based on classical molecular dynamics (MD) and alchemical free energy (AFE) simulations, fail to describe the complex electronic structure of metal-containing systems. Only a higher level of resolution, i.e. the use of quantum mechanical (QM) simulations, can describe these metalloorganic interactions accurately. However, these methodologies remain far too computationally expensive, given the extensive sampling required to simulate biomolecular systems.

The rising development of quantum computing (QC) offers a promising route to overcome this accuracy-versus-computation-cost dilemma, with the possibility of evaluating the electronic ground exactly thanks to the inherent properties of quantum bits.

 

Here, we present FreeQuantum [1,2,3], an automated hierarchical framework that integrates quantum computing of ground-state energies into AFE calculations via a three-level embedding quantum computing/quantum mechanics/molecular mechanics (QM/QM/MM) scheme. The workflow consists of an alchemical free energy perturbation (FEP) corrected via non-equilibrium (NEQ) switching simulations with system-specific machine learning potentials (MLPs). These MLPs are first trained on QM/MM data, then refined via transfer learning using embedded QM/QM/MM energies, within an active learning procedure.

This multilevel pipeline has been applied to two representative systems: the binding of the ruthenium-based anticancer drug NKP1339 to the chaperone GRP78 [1,3], and the small-molecule inhibitor 19G bound to the apoptosis regulator MCL1 [1,2]. In both cases, the ML-refined QM/QM/MM models yield binding free energies consistent with experimental affinities while drastically reducing computational cost. Looking ahead, FreeQuantum is being extended to enable the evaluation of covalent binding, the treatment of larger QM regions, and ultimately, the exploration of reactive biochemical processes. Beyond transition-metal complexes, we believe this framework defines a practical route toward quantum-enhanced, machine-learning-driven simulations applicable across the full spectrum of electronic and entropic complexity.

 

[1] Bensberg, Moritz, Marco Eckhoff, F. Emil Thomasen, et al. « Machine Learning-Enhanced Calculation of Quantum-Classical Binding Free Energies ». Journal of Chemical Theory and Computation 21, no 16 (2025): 8182 98.

[2] Bensberg, Moritz, Marco Eckhoff, Raphael T. Husistein, et al. « Hierarchical Quantum Embedding by Machine Learning for Large Molecular Assemblies ». Journal of Chemical Theory and Computation 21, no 15 (2025): 7662 74.

[3] Günther, Jakob, Thomas Weymuth, Moritz Bensberg, et al. « How to use quantum computers for biomolecular free energies ». Version 1. Prépublication, arXiv, 2025.

POSTER NO 46 - Vili Lampinen:

Design of protein binders to Probe Neurotrophin Signaling in Neuronal Connectivity and Memory

Vili Lampinen, Magnus Kjærgaard

Aarhus University

Contact: vili.lampinen@mbg.au.dk

The formation of memories and learning is based on accurate regulation of the connection between neurons. This is mainly orchestrated by neurotrophin receptors that sit atop neurons and transmit neurotrophin signaling. We aim to develop genetically encodable protein binders to study and manipulate the most elusive of these neuroreceptors, p75NTR. We use the machine learning tools RFDiffusion, ProteinMPNN, and BindCraft to design binders targeting both extra- and intracellular domains of p75NTR. We screen for successful binders and measure their affinities towards their target with biolayer interferometry and flow induced dispersion analysis. Finally, the binders are used in cell culture assays to activate or block p75NTR. This can lead to activation of apoptosis, differentiation, or survival signals, which we can measure in vitro. Protein binders are genetically encodable, so they provide a whole new kind of tool set for studying intracellular neuron signaling.

POSTER NO 47 - Westley Pawloski:

Evaluating the use of lanthanide containing dendrimers for solvent paramagnetic relaxation enhancement NMR spectroscopy

Westley Pawloski *, James Grushchus *, Ana Opina #, Olga Vasalattiy #, Nico Tjandra *

* Biochemistry and Biophysics Center, National Heart Lung and Blood Institute, US National Institutes of Health

 

# Chemistry and Synthesis Center, National Heart, Lung, and Blood Institute, US National Institutes of Health

Contact: wespawloski@gmail.com

Paramagnetic relaxation enhancement (PRE) is utilized in biomolecular NMR spectroscopy to obtain long-range distance and orientational information for intra- or intermolecular interactions of biomolecules. In contrast to conventional PRE measurements, which require tethering small molecules containing either a radical or paramagnetic ion to specific sites on the target protein, solvent PRE (sPRE) experiments utilize paramagnetic cosolutes to induce a delocalized PRE effect. This PRE effect is similarly utilized in magnetic resonance imaging (MRI) to generate high contrast images between water in the vasculature and those that interact with tissues and fats. Conventional MRI contrast agents typically consist of Gd3+ chelated by a small molecule. Tethering these Gd-containing small molecules to larger complexes has been shown to increase the PRE-effect and produce more effective contrast agents in MRI. Inspired by their use as MRI contrast agent, in this work we evaluate the effectiveness of using a functionalized polyamidoamine (PAMAM) dendrimer for sPRE measurements in NMR spectroscopy measurements. Using ubiquitin as a model system, we measured the sPRE effect from a generation 5 PAMAM dendrimer (G5-Gd) as a function of temperature and pH and this was contrasted with conventional relaxation agents. We also demonstrate the utility of G5-Gd in sPRE studies to monitor changes in the structures of two proteins as they bind their ligands. These studies highlight the attractive properties of these macromolecular relaxation agents in biomolecular sPRE.

POSTER NO 48 - William Pallisgaard Olsen:

Enzymatic Detoxification of Deoxynivalenol (DON) for Low-Impact Agriculture

William Pallisgaard Olsen, 1

Jacob Gjøderum Koch, 1

Valeria Della Gala, 1

Marko Vasiljević, 2

Jog Raj, 2

Hunor Farkaš, 2

Onur Kırtel, 1

Ditte Welner, 1

1: Technical University of Denmark, Lyngby, Denmark

2: Patent-Co, Mišićevo, Serbia

Contact: wilols@biosustain.dtu.dk

Deoxynivalenol (DON) is a mycotoxin produced by Fusarium spp. fungi on crops used for both human and animal consumption. The current global annual crop production of the major crops affected by mycotoxins is approximately 2819M tons, which at current European crop prices equates to $750B [1]. On a global scale, it is estimated that over 60% of all food crops are affected by mycotoxin contamination [2]. In addition to the economic losses, the environmental impact includes increased land use, loss of biodiversity, and increased CO2-equivalent emissions. These losses translate directly into losses in livestock agriculture through increased crop prices and a reduction in livestock production.

 

In these crop infections, DON is the most frequently detected toxin. The symptoms of DON exposure in animals include reduced feed intake and reduced gain of mass, vomiting, and immune system suppression, which further increases the toxin-related losses. Currently there are only limited ways to minimize the toxin, either through physical separation of affected crops or through strictly regulated mixing with non-affected crops. A sustainable and easy-to-implement method for removing or detoxifying DON and other mycotoxins is sorely missed.

 

We propose that the detoxification of DON in crop products can be addressed through the discovery, engineering, and development of enzymes that are specifically optimized to target and degrade toxins through a 2-step epimerization reaction. Such enzymes have already been discovered but require substantial engineering to function at an industrial scale. Our vision of a feed detoxification technology requires the enzymes to remain stable and functional under three critical conditions: (1) high activity at pH 7 and 35C, (2) a flash heat treatment of >80C for 10 seconds, (3) retaining activity after incubation in low pH of the stomach. While we have reached a relatively fast detoxification in simple in vitro systems, we are currently focused on engineering the enzymatic system to meet the harsh pH and temperature conditions through an effective engineering-and-testing pipeline. We have likewise shown that freeze-dried lysate containing enzyme can be added and mixed directly with DON extracted directly from contaminated corn, where it detoxifies DON upon addition of water. This is an ongoing project, and the current status will be presented.

 

[1] European Commision - Directorate-General for Agriculture and Roral Development - Cereal Prices

[2] Mari Eskola, Gregor Kos, Christopher T. Elliott, Jana Hajšlová, Sultan Mayar & Rudolf Krska (2020) Worldwide contamination of food-crops with mycotoxins: Validity of the widely cited ‘FAO estimate’ of 25%, Critical Reviews in Food Science and Nutrition, 60:16, 2773-2789

POSTER NO 49 - Ying Xia:

Heterogeneous sampled subgraph neural networks with knowledge distillation to enhance double-blind compound-protein interaction prediction

Ying Xia, Xiaoyong Pan, Hong-Bin Shen

University of Copenhagen, Shanghai Jiao Tong University

Contact: ying.xia@bio.ku.dk

Identifying binding compounds against a target protein is crucial for large-scale virtual screening in drug development. Recently, network-based methods have been developed for compound-protein interaction (CPI) prediction. However, they are difficult to be applied to unseen (i.e., never-seen-before) proteins and compounds. In this study, we propose SgCPI to incorporate local known interacting networks to predict CPI interactions. SgCPI randomly samples the local CPI network of the query compound-protein pair as a subgraph and applies a heterogeneous graph neural network (HGNN) to embed the active/inactive message of the subgraph. For unseen compounds and proteins, SgCPI-KD takes SgCPI as the teacher model to distillate its knowledge by estimating the potential neighbors. Experimental results indicate: (1) the sampled subgraphs of the CPI network introduce efficient knowledge for unseen molecular prediction with the HGNNs, and (2) the knowledge distillation strategy is beneficial to the double-blind interaction prediction by estimating molecular neighbors and distilling knowledge.