Publications by Sarah Wehrle | EPFL Graph Search

De novo design of protein interactions with learned surface fingerprints

Michael Bronstein, Stephen M Buckley, Bruno Emanuel Ferreira De Sousa Correia, Beat Fierz, Pablo Gainza Cirauqui, Sandrine Madeleine Suzanne Georgeon, Casper Alexander Goverde, Zander Harteveld, Anthony Marchand, Jane Marsden, Dongchun Ni, Elisa Oricchio, Aaron Simone Petruzzella, Charlène Mireille Raymonde Raclot, Stéphane Rosset, Andreas Scheck, Henning Paul-Julius Stahlberg, Freyr Sverrisson, Alexandra Teslenko, Didier Trono, Priscilla Turelli, Alexandra Krina Van Hall-Beauvais, Sarah Wehrle

Physical interactions between proteins are essential for most biological processes governing life. However, the molecular determinants of such interactions have been challenging to understand, even as genomic, proteomic and structural data increase. This knowledge gap has been a major obstacle for the comprehensive understanding of cellular protein–protein interaction networks and for the de novo design of protein binders that are crucial for synthetic biology and translational applications. Here we use a geometric deep-learning framework operating on protein surfaces that generates fingerprints to describe geometric and chemical features that are critical to drive protein–protein interactions10. We hypothesized that these fingerprints capture the key aspects of molecular recognition that represent a new paradigm in the computational design of novel protein interactions. As a proof of principle, we computationally designed several de novo protein binders to engage four protein targets: SARS-CoV-2 spike, PD-1, PD-L1 and CTLA-4. Several designs were experimentally optimized, whereas others were generated purely in silico, reaching nanomolar affinity with structural and mutational characterization showing highly accurate predictions. Overall, our surface-centric approach captures the physical and chemical determinants of molecular recognition, enabling an approach for the de novo design of protein interactions and, more broadly, of artificial proteins with function.

2023

Targeting molecular surfaces to engineer novel protein-based immunogens and inhibitors

Sarah Wehrle

Proteins are the basic building blocks necessary for the operation and regulation of virtually all functions in living organisms. Over millions of years, evolution has created a vast repertoire of proteins finely tuned to execute their biological functions. However, biomedical and biotechnological applications often require protein properties and functions beyond the natural scope. To overcome this limitation, protein engineering methods have emerged which has enabled the targeted alteration and optimization of proteins. Driven by advances in structural biology and growing understanding of the structure-function relationship, the development of computational design methods has opened new possibilities for the design of functional proteins. Since many proteins carry out their functions through physical interactions with other proteins, the alteration or creation of protein-protein interactions (PPIs) represents a crucial aspect of functional protein design. The design of PPIs from scratch remains a substantial challenge, particularly when no structural data is available. To tackle this challenge, we showcase a novel strategy for the design of de novo PPIs by identifying complementary buried surfaces to a target interface using geometric deep learning, leveraging our knowledge from known protein complexes. Using this approach, we were able to show the design of de novo protein binders to two highly relevant targets. Computation-based functional protein design also holds great potential for the development of novel and more efficient vaccines to combat pathogens, like the influenza virus, that have yet been resisting traditional vaccine development efforts. The isolation of broadly neutralizing antibodies targeting the influenza hemagglutinin (HA) have guided vaccine development towards a conserved epitope on the stem domain. However, the low immunogenicity and complexity of this epitope hinders the robust and focused induction of these antibodies. Computational design of immunogens allows us to spotlight this epitope to elicit a targeted antibody response. By using a design approach that focuses on the mimicry of the molecular surface of the epitope, we were able to design an immunogen that elicits focused immune response cross-reactive to HA and able to activate cellular cytotoxicity in vitro. While this immunogen design focused on the induction of antibodies against the conserved stem domain of influenza HA, we also explored design routes to engineer antigens based on the variable head domain, known to induce very potent neutralizing antibodies. By exploiting the high structural similarity between head domains variable in sequence, we aim to construct novel chimeric head domains combining semi-conserved sites from two or more relevant HA subtypes. These immunogens bear great potential to induce heterosubtypic antibody responses by a single molecule, but also to further elucidate the immunodominance hierarchies on the HA head domain. Altogether, this work demonstrates the utilization of different computation-guided design strategies in combination with experimental fine-tuning and validation to develop new or re-engineered protein binders. The implemented strategies represent general frameworks that can be applied to the development of immunogens and other protein-based therapeutics.

EPFL2022

Cathepsin S Regulates Antigen Processing and T Cell Activity in Non-Hodgkin Lymphoma

Elena Battistello, Giovanni Ciriello, George Coukos, Elie Charles Eugène Dheilly, Bruno Emanuel Ferreira De Sousa Correia, Natalya Katanayeva, Marco Mina, Elisa Oricchio, Julien Racle, Jessica Sordet-Dessimoz, Stephanie Jocelyne Sungalee, Sarah Wehrle

Genomic alterations in cancer cells can influence the immune system to favor tumor growth. In non-Hodgkin lymphoma, physiological interactions between B cells and the germinal center microenvironment are coopted to sustain cancer cell proliferation. We found that follicular lymphoma patients harbor a recurrent hotspot mutation targeting tyrosine 132 (Y132D) in cathepsin S (CTSS) that enhances protein activity. CTSS regulates antigen processing and CD4+ and CD8+ T cell-mediated immune responses. Loss of CTSS activity reduces lymphoma growth by limiting communication with CD4+ T follicular helper cells while inducing antigen diversification and activation of CD8+ T cells. Overall, our results suggest that CTSS inhibition has non-redundant therapeutic potential to enhance anti-tumor immune responses in indolent and aggressive lymphomas.

2020

, , ,

The humoral responses of Ebola virus (EBOV) survivors mainly target the surface glycoprotein GP, and anti-GP neutralizing antibodies have been associated with protection against EBOV infection. In order to elicit protective neutralizing antibodies through vaccination a native-like conformation of the antigen is required. We therefore engineered and expressed in CHO cells several GP variants from EBOV (species Zaire ebolavirus, Mayinga variant), including a soluble GP Delta TM, a mucin-like domain-deleted GP Delta TM-Delta MUC, as well as two GP Delta TM-Delta MUC variants with C-terminal trimerization motifs in order to favor their native trimeric conformation. Inclusion of the trimerization motifs resulted in proteins mimicking GP metastable trimer and showing increased stability. The mucin-like domain appeared not to be critical for the retention of the native conformation of the GP protein, and its removal unmasked several neutralizing epitopes, especially in the trimers. The soluble GP variants inhibited mAbs neutralizing activity in a pseudotype transduction assay, further confirming the proteins' structural integrity. Interestingly, the trimeric GPs, a native-like GP complex, showed stronger affinity for antibodies raised by natural infection in EBOV disease survivors rather than for antibodies raised in volunteers that received the ChAd3-EBOZ vaccine. These results support our hypothesis that neutralizing antibodies are preferentially induced when using a native-like conformation of the GP antigen. The soluble trimeric recombinant GP proteins we developed represent a novel and promising strategy to develop prophylactic vaccines against EBOV and other filoviruses.

FRONTIERS MEDIA SA2020

Structural basis for recognition of RALF peptides by LRX proteins during pollen tube growth

Sarah Wehrle

Plant reproduction relies on the highly regulated growth of the pollen tube for sperm delivery. This process is controlled by secreted RALF signaling peptides, which have previously been shown to be perceived by Catharanthus roseus RLK1-like (CrRLK1Ls) membrane receptor-kinases/LORELEI-like GLYCOLPHOSPHATIDYLINOSITOL (GPI)-ANCHORED PROTEINS (LLG) complexes, or by leucine-rich repeat (LRR) extensin proteins (LRXs). Here, we demonstrate that RALF peptides fold into bioactive, disulfide bond-stabilized proteins that bind the LRR domain of LRX proteins with low nanomolar affinity. Crystal structures of LRX2-RALF4 and LRX8-RALF4 complexes at 3.2- and 3.9-A resolution, respectively, reveal a dimeric arrangement of LRX proteins, with each monomer binding one folded RALF peptide. Structure-based mutations targeting the LRX-RALF4 complex interface, or the RALF4 fold, reduce RALF4 binding to LRX8 in vitro and RALF4 function in growing pollen tubes. Mutants targeting the disulfide-bond stabilized LRX dimer interface fail to rescue Inc infertility phenotypes. Quantitative biochemical assays reveal that RALF4 binds LLG5 and LRX cell-wall modules with drastically different binding affinities, and with distinct and mutually exclusive binding modes. Our biochemical, structural, and genetic analyses reveal a complex signaling network by which RALF ligands instruct different signaling proteins using distinct targeting mechanisms.

NATL ACAD SCIENCES2020

Design and characterization of near-native Ebola GP vaccine candidates: implications for immunogenicity

Bruno Emanuel Ferreira De Sousa Correia, Sarah Wehrle, Florian Maria Wurm

E M H SWISS MEDICAL PUBLISHERS LTD2018

Rosetta FunFolDes - A general framework for the computational design of functional proteins

Anne Marie-Thérèse Suzanne Billet, Jaume Bonet Martinez, Bruno Emanuel Ferreira De Sousa Correia, Roxanne Marine Lourman, Stéphane Rosset, Andreas Scheck, Karen Schriever, Fabian Sesterhenn, Freyr Sverrisson, Mélanie Evelyne Villard, Sabrina Susan Vollers, Sarah Wehrle, Che Yang

The robust computational design of functional proteins has the potential to deeply impact translational research and broaden our understanding of the determinants of protein function and stability. The low success rates of computational design protocols and the extensive in vitro optimization often required, highlight the challenge of designing proteins that perform essential biochemical functions, such as binding or catalysis. One of the most simplistic approaches for the design of function is to adopt functional motifs in naturally occurring proteins and transplant them to computationally designed proteins. The structural complexity of the functional motif largely determines how readily one can find host protein structures that are "designable", meaning that are likely to present the functional motif in the desired conformation. One promising route to enhance the "designability" of protein structures is to allow backbone flexibility. Here, we present a computational approach that couples conformational folding with sequence design to embed functional motifs into heterologous proteins-Rosetta Functional Folding and Design (FunFolDes). We performed extensive computational benchmarks, where we observed that the enforcement of functional requirements resulted in designs distant from the global energetic minimum of the protein. An observation consistent with several experimental studies that have revealed function-stability tradeoffs. To test the design capabilities of FunFolDes we transplanted two viral epitopes into distant structural templates including one de novo "functionless" fold, which represent two typical challenges where the designability problem arises. The designed proteins were experimentally characterized showing high binding affinities to monoclonal antibodies, making them valuable candidates for vaccine design endeavors. Overall, we present an accessible strategy to repurpose old protein folds for new functions. This may lead to important improvements on the computational design of proteins, with structurally complex functional sites, that can perform elaborate biochemical functions related to binding and catalysis.

2018