Biology subjects

Schmid-Burgk, J. L.

Publications and source records attributed to Schmid-Burgk, J. L..

9 recordsLinked to original sources

bioRxiv2026

Evolutionary algorithms accelerate de novo design of potent Nectin-4-specific cancer biologics

Recent advances in AI-based structural biology have made de novo protein binder design increasingly effective, yet performance remains highly target dependent. For instance, the cancer surface antigen Nectin-4, an immunoglobulin-like cell adhesion protein, proved particularly challenging for RFdiffusion-based minibinder generation, yielding substantially fewer high-quality candidates than related targets. To address this bottleneck, we integrated an evolutionary genetic algorithm (GA) with AI-driven design. GA selection with tunable stringency was coupled with diversification via partial diffusion or direct sequence editing, enabling efficient exploration of sequence-structure space and rapid enrichment of promising candidates. This AI-GA pipeline quickly produced large and diverse minibinder panels with very good in silico quality metrics and is compatible with inputs from multiple design algorithms. Pooled, large-scale experimental screening identified highly stable Nectin-4 minibinders with single-digit nanomolar down to subnanomolar affinities. Lead binders were further engineered into Nectin-4-specific flow cytometry detection reagents and potent bispecific T cell engagers, demonstrating functional activity beyond binding. Together, these results show that evolutionary refinement can unlock challenging targets and accelerate de novo protein design for next-generation cancer biologics.

bioengineering↗

bioRxiv2025

Development of AI-designed protein binders for detection and targeting of cancer cell surface proteins

Artificial intelligence (AI)-based protein design opens new avenues for the rapid generation of new research tools and therapeutics, but experimental validation lags behind the computational design throughput. Here, we present a scalable workflow for the discovery and validation of AI-designed protein binders. Leveraging the RFdiffusion protein design pipeline with a custom filter for stable alpha-helical bundle folds, we construct libraries of thousands of AI-binders against cancer-associated surface proteins. Mammalian cell-surface and phage display screening yield multiple high-affinity PD-L1 binders but fewer hits for CD276 (B7-H3) and VTCN1 (B7-H4), reflecting the target-dependent efficiency of RFdiffusion in generating high-quality designs. Using our experimentally validated AI-designed binder libraries, we benchmark freely available structure prediction models. We find that interface predicted template modelling (ipTM) scores by Chai-1 with ESM embedding correlate well with experimental success and even predict deleterious effects of binding interface mutations. To demonstrate the versatility of AI-binders as research tools, we deploy them in CAR-T cells and also assemble them with fluorophore-labeled streptavidin into tetravalent quattrobinders, which achieve antibody-comparable staining of endogenous PD-L1 by flow cytometry. With high production yields and accessible structural models, AI-designed quattrobinders are versatile and cost-effective research tools amenable to community-driven validation and optimization.

bioengineering↗

bioRxiv2025

Transposon-Display of AI-designed binders enables manipulation of the proteome in human cells

Transposon-Display is a highly scalable screening method that links proteins to their encoding DNA during expression in E. coli via a mutant transposase. Leveraging this system, we identified AI-designed binders targeting four intracellular human proteins, saturation-mutagenized top candidates, and charge-balanced variants to improve compatibility with the intracellular environment while preserving target binding. The resulting neutral binders can be fused to EGFP allowing antibody-free intracellular staining. Expressing neutral binders in human cells as fusions with functional domains can drive small molecule-controlled protein aggregation and trigger proteasomal degradation of endogenous target proteins in living cells. Manipulating the proteome of living cells using libraries of AI-designed binders may provide a new avenue to screen for disease-relevant protein functions and large-scale functional data may help to refine protein design algorithms.

bioengineering↗

bioRxiv2024

Sensing endogenous RNA in living human cells using a CRISPR-activated protease

Most techniques used to detect specific mRNAs in eukaryotic cells require to extract nucleic acids and thereby kill the cells. A programmable sensor for monitoring endogenous transcripts in living cells, in contrast, would enable to enrich living cells based on a specific transcription or splicing event, and studying these cells by live microscopy or sequencing methods requiring intact cells. We have engineered CRISPR-READ, a live cell RNA detector based on the CRISPR-associated Lon protease CalpL and a cA4-producing Type III CRISPR system. Upon RNA-programmable RNA sensing, CRISPR-READ produces an orthogonal second messenger, which leads to the cleavage of a dual FRET / localization reporter compatible with FACS sorting and live microscopy. Using this genetically encoded sensing circuit as a readout for a genome-wide CRISPR perturbation screen, we identified an extended Type-I interferon signaling cascade; RNA-Seq on sensor-sorted cells enabled unbiased identification of correlated stochasticity in gene expression across single cells.

bioengineering↗

bioRxiv2024

Cell Type-Agnostic Optical Perturbation Screening Using Nuclear In-Situ Sequencing (NIS-Seq)

Genome-scale perturbation screening is widely used to identify disease-relevant cellular proteins serving as potential drug targets. However, most biological processes are not compatible with commonly employed perturbation screening methods, which rely on FACS- or growth-based enrichment of cells. Optical pooled screening instead uses fluorescence microscopy to determine the phenotype in single cells, and subsequently to identify individual perturbagens in the same cells. Published methods rely on cytosolic detection of endogenously expressed barcoded transcripts, which limits application to large, transcriptionally active cell types, and often relies on local clusters of clonal cells for unequivocal barcode assignment, thus precluding genome-scale screening for many biological processes. Nuclear In-Situ Sequencing (NIS-Seq) solves these shortcomings by creating bright sequencing signals directly from nuclear genomic DNA, enabling screening any nucleus-containing cell type at high density and high library complexity. We benchmark NIS-Seq by performing three genome-scale optical screens in live cells, identifying key players of inflammation-related cellular pathways.

genomics↗

bioRxiv2020

LAMP-Seq: Population-Scale COVID-19 Diagnostics Using a Compressed Barcode Space

The ongoing SARS-CoV-2 pandemic has already caused devastating losses. Exponential spread can be slowed by social distancing and population-wide isolation measures, but those place a tremendous burden on society, and, once lifted, exponential spread can re-emerge. Regular population-scale testing, combined with contact tracing and case isolation, should help break the cycle of transmission, but current detection strategies are not capable of such large-scale processing. Here we present a protocol for LAMP-Seq, a barcoded Reverse-Transcription Loop-mediated Isothermal Amplification (RT-LAMP) method that is highly scalable. Individual samples are stabilized, inactivated, and amplified in three isothermal heat steps, generating barcoded amplicons that can be pooled and analyzed en masse by sequencing. Using unique barcode combinations per sample from a compressed barcode space enables extensive pooling, potentially further reducing cost and simplifying logistics. We validated LAMP-Seq on 28 clinical samples, empirically optimized the protocol and barcode design, and performed initial safety evaluation. Relying on world-wide infrastructure for next-generation sequencing, and in the context of population-wide sample collection, LAMP-Seq could be scaled to analyze millions of samples per day.

molecular biology↗

bioRxiv2019

Experimental and stochastic models of melanoma T-cell therapy define impact of subclone fitness on selection of antigen loss variants

Antigen loss is a key mechanism how tumor cells escape from T-cell immunotherapy. Using a mouse model of melanoma we directly compared antigen downregulation by phenotypic adaptation with genetically hardwired antigen loss. Unexpectedly, genetic ablation of Pmel, the melanocyte differentiation antigen targeted by adoptively transferred CD8+ T-cells, impaired melanoma cell growth in untreated tumors due to competitive pressure exerted by the bulk wild-type population. This established an evolutionary scenario, where T-cell immunotherapy imposed a dynamic fitness switch on wild-type melanoma cells and antigen loss variants, which resulted in highly variable enrichment of the latter in recurrent tumors. Stochastic simulations by an individual-based continuous-time Markov process suggested variable fitness of subclones within the antigen loss variant population as the most likely cause, which was validated experimentally. In summary, we provide a framework to better understand how subclone heterogeneity in tumors influences immune selection of genetic antigen loss variants through stochastic events.

systems biology↗

bioRxiv2019

Priming enables a NEK7-independent route of NLRP3 activation

The NLRP3 inflammasome plays a central role in antimicrobial defense, as well as in sterile inflammatory conditions. NLRP3 activity is governed by two independent signals. The first signal primes NLRP3, allowing it to respond to its activation signal. In the murine system, the mitotic spindle kinase NEK7 has been identified as a crucial factor in relaying the activation signal to NLRP3. Here we show that the requirement for NEK7 can be bypassed by TAK1-dependent post-translational priming. Under pro-inflammatory conditions that activate TAK1, NEK7 was dispensable for NLRP3 inflammasome formation in human and murine cells. Intriguingly, dissecting the NEK7 requirement in iPSC-derived primary human macrophages revealed that this NEK7-independent mechanism constitutes the predominant NLRP3 priming pathway in these cells. In summary, our results suggest that NEK7 functions as an NLRP3 priming - rather than activation - factor that can work in synergy or redundancy with other priming pathways to accelerate inflammasome activation.

immunology↗

bioRxiv2018

Pooled optical screens in human cells

Large-scale genetic screens play a key role in the systematic discovery of genes underlying cellular phenotypes. Pooling of genetic perturbations greatly increases screening throughput, but has so far been limited to screens of enrichments defined by cell fitness and flow cytometry, or to comparatively low-throughput single cell gene expression profiles. Although microscopy is a rich source of spatial and temporal information about mammalian cells, high-content imaging screens have been restricted to much less efficient arrayed formats. Here, we introduce an optical method to link perturbations and their phenotypic outcomes at the singlecell level in a pooled setting. Barcoded perturbations are read out by targeted in situ sequencing following image-based phenotyping. We apply this technology to screen a focused set of 952 genes across >3 million cells for involvement in NF-{kappa}B activation by imaging the translocation of RelA (p65) to the nucleus, recovering 20 known pathway components and 3 novel candidate positive regulators of IL-1{beta} and TNF-stimulated immune responses.

genomics↗