Enzyme research is evolving with the integration of computational tools, and at the core of this progress is the OXIPRO project. A key part of this EU-funded initiative is the development of computational platforms designed to support enzyme engineering. These platforms are being developed to make enzyme discovery more accessible and efficient for various industries.
Enzymes play a fundamental role in industrial biotechnology, catalysing reactions that are essential for producing sustainable products, from bio-based detergents to pharmaceutical compounds. However, finding the right enzymes and optimizing them for specific industrial applications remains challenge due to the complexity of biological systems. OXIPRO’s computational platforms aim to streamline this process, making enzyme discovery and engineering faster and more cost-effective.
Figure 1: Phosphotriesterase (PTE) enzyme from B diminuta (1HZY). Enzyme used to validate enzyme engineering tools such as PROSS1 and FuncLib2. Image made with Horus.
Addressing the Challenges in Enzyme Discovery
Enzymes act as natural catalysts, enhancing biological reactions and making them useful in industries such as medicine, food production, textiles, and household products. However, finding and refining new enzymes for industrial applications can be a complex and time-consuming process.
Traditional enzyme discovery relies on extensive laboratory testing, requiring significant resources. Scientists must screen large enzyme libraries, test their activity, and modify their structures to enhance performance. This requires high levels of expertise, expensive reagents, and sophisticated laboratory equipment. Scaling up enzyme production while maintaining stability and efficiency also poses challenges.
Our brand-new graphical user interface aims to address these issues by integrating biomodelling tools and high-performance computing (HPC) to improve enzyme selection and optimization. By simulating enzyme behaviour in a virtual environment, researchers can reduce the need for physical testing and prioritize the most promising enzyme candidates for further development.
The Team Behind the Platforms
The Barcelona Supercomputing Center (BSC) is leading the development of these computational tools. Under the guidance of Victor Guallar and his team, BSC has created a workflow management system called Horus. This platform is designed to provide researchers with a structured approach to enzyme bioprospecting and engineering among other computational tools.
“Our goal with Horus is to provide a functional and user-friendly platform that allows researchers to model, test, and refine enzymes in a centralized environment before moving to the lab.” explains Guallar. “This can help streamline the enzyme discovery process and reduce experimental costs.” Horus is not just a tool for computational biologists; it is designed to be accessible to a wider scientific community, including biochemists, molecular biologists, and industrial researchers who may not have extensive computational expertise. By providing an intuitive interface and user-friendly workflows, Horus lowers the barriers to adopting computational methods in enzyme research.
How Horus Works
Horus is a modular application that can integrate as many applications as the scientist needs. This modular approach allows for the use of plug-ins, which brings more functionality to the application based on the needs of the researcher.
- Customizable Workflows: Provides a flexible environment for designing and executing computational workflows. Each block is independent, which allows for the ultimate customizable pipeline.
- High-Performance Computing (HPC) Integration: Allows to stablish remote connections between different HPC clusters in order to send simulations simultaneously.
- Collaboration-Friendly Design: Can function as both a standalone application and a web-based tool, supporting team-based research.
Some plugin-driven applications used in the OXIPRO’s workshop are:
- AI-Driven Enzyme Design: Structure prediction is tightly integrated into Horus, making it easy for non-experienced users to generate de novo 3D data of their engineered sequences. In this workshop, we will be using ProteinMPNN3 to generate new sequences that will then be folded using ESMFold4. A demo workflow can be seen on figure 2 and 3.
- Docking, Molecular Modelling & Simulations: Predicts enzyme behaviour under different conditions, allowing for preliminary testing before laboratory validation using computational tools like docking, molecular dynamics or Montecarlo algorithms. The rDock5 plugin for Horus is one of the many approaches that we can use to validate the newly generated structure, as seen on figure 4.
One of the standout features of Horus is its ability to visualize enzyme structures and interactions dynamically using its embedded 3D molecular visualizer. This enables researchers to explore how enzymes interact with their substrates, identify potential modifications to improve performance, and predict how changes in temperature, pH, and other conditions affect enzyme activity.
Figure 2: ProteinMPNN3 workflow in Horus with 6MRR and 5L33 sequences.
Figure 3: ESMFold3 and PDB alignment workflow in Horus.
Figure 4: Example workflow for a docking simulation with rDock5. On the left top, the Horus flow. On the left bottom, the 3D molecular visualizer. On the right, the table of results.
Why Horus? The Name’s Significance
The name Horus reflects the platform’s role in providing researchers with better insight and clarity in modelling studies. Just as the Egyptian god Horus was associated with vision, this system helps researchers navigate complex molecular processes visually with greater efficiency.
A Step Forward in Enzyme Research
The development of Horus and other computational tools within OXIPRO represents progress in enzyme research. While computational tools will not replace laboratory research, they are becoming an essential part of the scientific toolkit, helping researchers work more effectively and achieve meaningful results. The OXIPRO project is contributing to this shift, ensuring that enzyme research keeps pace with the growing demand for sustainable and innovative biotechnological solutions.
Figure 5: Horus can be downloaded from the official webpage at https://horus.bsc.es
References
Automated Structure- and Sequence-Based Design of Proteins for High Bacterial Expression and Stability. Goldenzweig, Adi et al. Molecular Cell, 63, 337 – 346
DOI: doi.org/10.1016/j.molcel.2016.06.012
Automated Design of Efficient and Functionally Diverse Enzyme Repertoires. Khersonsky, Olga et al. Molecular Cell, 72, 178 – 186.e5
DOI: 10.1016/j.molcel.2018.08.033
Robust deep learning–based protein sequence design using ProteinMPNN. J. Dauparas et al. Science 378, 49-56.
DOI: 10.1126/science.add2187
Zeming Lin et al. ,Evolutionary-scale prediction of atomic-level protein structure with a language model. Science 379, 1123-1130.
DOI:10.1126/science.ade2574
Ruiz-Carmona S, Alvarez-Garcia D, Foloppe N, Garmendia-Doval AB, Juhos S, et al. (2014) rDock: A Fast, Versatile and Open Source Program for Docking Ligands to Proteins and
Nucleic Acids. PLoS Comput Biol 10.
DOI: 10.1371/journal.pcbi.1003571