Biology subjects

Villiger, L.

Publications and source records attributed to Villiger, L..

7 recordsLinked to original sources

bioRxiv2024

Rapid protein evolution by few-shot learning with a protein language model

Directed evolution of proteins is critical for applications in basic biological research, therapeutics, diagnostics, and sustainability. However, directed evolution methods are labor intensive, cannot efficiently optimize over multiple protein properties, and are often trapped by local maxima. In silico-directed evolution methods incorporating protein language models (PLMs) have the potential to accelerate this engineering process, but current approaches fail to generalize across diverse protein families. We introduce EVOLVEpro, a few-shot active learning framework to rapidly improve protein activity using a combination of PLMs and protein activity predictors, achieving improved activity with as few as four rounds of evolution. EVOLVEpro substantially enhances the efficiency and effectiveness of in silico protein evolution, surpassing current state-of-the-art methods and yielding proteins with up to 100-fold improvement of desired properties. We showcase EVOLVEpro for five proteins across three applications: T7 RNA polymerase for RNA production, a miniature CRISPR nuclease, a prime editor, and an integrase for genome editing, and a monoclonal antibody for epitope binding. These results demonstrate the advantages of few-shot active learning with small amounts of experimental data over zero-shot predictions. EVOLVEpro paves the way for broader applications of AI-guided protein engineering in biology and medicine.

bioengineering↗

bioRxiv2024

Treatment of a genetic liver disease in mice through transient prime editor expression

Prime editing is a versatile genome editing technology that does not rely on DNA double-strand break formation and homology-directed repair (HDR). This makes it a promising tool for correcting pathogenic mutations in tissues consisting predominantly of postmitotic cells, such as the liver. While recent studies have already demonstrated proof-of-concept for in vivo prime editing, the use of viral delivery vectors resulted in prolonged prime editor (PE) expression, posing challenges for clinical application. Here, we developed an in vivo prime editing approach where we delivered the pegRNA using self-complementary adeno-associated viral (scAAV) vectors and the prime editor using nucleoside-modified mRNA encapsulated in lipid nanoparticles (LNPs). This methodology led to transient expression of the PE for 48h and 26% editing at the Dnmt1 locus using AAV doses of 2.5x1013 vector genomes (vg)/kg and a single dose of 3mg/kg mRNA-LNP. When targeting the pathogenic mutation in the Pahenu2 mouse model of phenylketonuria (PKU), we achieved 4.3% gene correction using an AAV dose of 2.5x1013 vg/kg and three doses of 2 mg/kg mRNA-LNP. Editing was specific to the liver and the intended locus, and was sufficient to reduce blood L-phenylalanine (Phe) levels from over 1500 {micro}mol/l to below the therapeutic threshold of 600 {micro}mol/l. Our study demonstrates the feasibility of in vivo gene correction in the liver with transient PE expression, bringing prime editing closer to clinical application.

bioengineering↗

bioRxiv2022

Programmable eukaryotic protein expression with RNA sensors

The diversity of cell types and states can be scalably measured and defined by expressed RNA transcripts. However, approaches to programmably sense and respond to the presence of specific RNAs within living biological systems with high sensitivity are lacking. RNA sensors that gate expression of reporter or cargo genes would have diverse applications for basic biology, diagnostics and therapeutics by enabling cell-state specific control of transgene expression. Here, we engineer a novel programmable RNA-sensing technology, Reprogrammable ADAR Sensors (RADARS), which leverages RNA editing by adenosine deaminases acting on RNA (ADAR) to gate translation of a protein payload on the presence of endogenous RNA transcripts. In mammalian cells, we engineer RADARS with diverse payloads, including luciferase and fluorescent proteins, with up to 164-fold activation and quantitative detection in the presence of target RNAs. We show RADARS are functional either expressed from DNA or as synthetic mRNA. Importantly, RADARS can function with endogenous cellular ADAR. We apply RADARS to multiple contexts, including RNA-sensing induced cell death via caspases, cell type identification, and in vivo control of synthetic mRNA translation, demonstrating RADARS as a tool with significant potential for gene and cell therapy, synthetic biology, and biomedical research. One Sentence SummaryA new technology utilizing ADAR mediated RNA-editing enables robust reprogrammable protein expression based on target RNA transcripts in mammalian cells, leading to broad applications in basic science research, cell engineering, and gene therapy.

genomics↗

bioRxiv2021

Drag-and-drop genome insertion without DNA cleavage with CRISPR-directed integrases

Programmable and multiplexed genome integration of large, diverse DNA cargo independent of DNA repair remains an unsolved challenge of genome editing. Current gene integration approaches require double-strand breaks that evoke DNA damage responses and rely on repair pathways that are inactive in terminally differentiated cells. Furthermore, CRISPR-based approaches that bypass double stranded breaks, such as Prime editing, are limited to modification or insertion of short sequences. We present Programmable Addition via Site-specific Targeting Elements, or PASTE, which achieves efficient and versatile gene integration at diverse loci by directing insertion with a CRISPR-Cas9 nickase fused to both a reverse transcriptase and serine integrase. Without generating double stranded breaks, we demonstrate integration of sequences as large as [~]36 kb with rates between 10-50% at multiple genomic loci across three human cell lines, primary T cells, and quiescent non-dividing primary human hepatocytes. To further improve PASTE, we discover thousands of novel serine integrases and cognate attachment sites from metagenomes and engineer active orthologs for high-efficiency integration using PASTE. We apply PASTE to fluorescent tagging of proteins, integration of therapeutically relevant genes, and production and secretion of transgenes. Leveraging the orthogonality of serine integrases, we engineer PASTE for multiplexed gene integration, simultaneously integrating three different genes at three genomic loci. PASTE has editing efficiencies comparable to or better than those of homology directed repair or non-homologous end joining based integration, with activity in non-dividing cells and fewer detectable off-target events. For therapeutic applications, PASTE can be delivered as mRNA with synthetically modified guides to programmably direct insertion of DNA templates carried by AAV or adenoviral vectors. PASTE expands the capabilities of genome editing via drag-and-drop gene integration, offering a platform with wide applicability for research, cell engineering, and gene therapy. One Sentence SummaryA new technology combining CRISPR-mediated genome editing and site-specific integrases enables efficient programmable gene integration at any targeted genomic locus without double-strand DNA breaks, leading to broad applications in basic science research, cell engineering, and gene therapy.

bioengineering↗

bioRxiv2021

Treatment of a metabolic liver disease by in vivo prime editing in mice

Prime editing is a highly versatile CRISPR-based genome editing technology with the potential to correct the vast majority of pathogenic mutations (1). However, correction of a disease phenotype in vivo in somatic tissues has not been demonstrated thus far. Here, we establish proof-of-concept for in vivo prime editing and repair the metabolic liver disease phenylketonuria (PKU) in mice. We first developed a size-reduced SpCas9 prime editor (PE) lacking the RNaseH domain of the reverse transcriptase (PE2{Delta}RnH), and a linker- and NLS-optimized intein-split PE construct (PE2 p.1153) for delivery by adeno-associated virus (AAV) vectors. Systemic dual AAV-mediated delivery of this variant into the liver of neonatal mice enabled installation of a transversion mutation at the Dnmt1 locus with an average efficiency of 15%, and delivery of unsplit PE2{Delta}RnH using human adenoviral vector 5 (AdV5) further increased editing rates to 58%. PE2{Delta}RnH-encoding AdV5 was also used to correct the disease-causing mutation of the phenylalanine hydroxylase (Pah)enu2 allele in phenylketonuria (PKU) mice with an average efficiency of 8% (up to 17.3%), leading to therapeutic reduction of blood phenylalanine (L-Phe) levels. Our study demonstrates in vivo prime editing in the liver with high precision and editing rates sufficient to treat a number of metabolic liver diseases, emphasizing the potential of prime editing for future therapeutic applications. One Sentence SummaryIn vivo prime editing corrects phenylketonuria in mice.

genetics↗

bioRxiv2020

Replacing the SpCas9 HNH domain by deaminases generates compact base editors with an alternative targeting scope

Base editors are RNA-guided deaminases that enable site-specific nucleotide transitions. The targeting scope of these Cas-deaminase fusion proteins critically depends on the availability of a protospacer adjacent motif (PAM) at the target locus and is limited to a window within the CRISPR-Cas R-loop, where single stranded (ss)DNA is accessible to the deaminase. Here, we reason that the Cas9-HNH nuclease domain sterically constrains ssDNA accessibility, and demonstrate that omission of this domain expands the editing window. By exchanging the HNH nuclease domain with an adenosine deaminase we furthermore engineer adenine base editor variants (HNHx-ABE) with PAM-proximally shifted editing windows. This work expands the targeting scope of base editors, and provides base editor variants that are substantially smaller. It moreover informs of potential future directions in Cas9 protein engineering, where the HNH domain could be replaced by other enzymes that act on ssDNA.

bioengineering↗

bioRxiv2020

Predicting base editing outcomes with an attention-based deep learning algorithm trained on high-throughput target library screens

Base editors are chimeric ribonucleoprotein complexes consisting of a DNA-targeting CRISPR-Cas module and a single-stranded DNA deaminase. They enable conversion of C•G into T•A base pairs and vice versa on genomic DNA. While base editors have vast potential as genome editing tools for basic research and gene therapy, their application has been hampered by a broad variation in editing efficiencies on different genomic loci. Here we perform an extensive analysis of adenine- and cytosine base editors on thousands of lentivirally integrated genetic sequences and establish BE-DICT, an attention-based deep learning algorithm capable of predicting base editing outcomes with high accuracy. BE-DICT is a versatile tool that in principle can be trained on any novel base editor variant, facilitating the application of base editing for research and therapy.Competing Interest StatementThe authors have declared no competing interest.View Full Text

genomics↗