Tyler O'Neal, Staff Editor ENGINEERING

Chemists from Russia use machine learning, molecular modeling to discover the next generation of anti-cancer drugs

Chemists from RUDN University, located in Moscow, along with their colleagues in China, have successfully developed several machine-learning models to identify potential drugs that can restrict the activity of an enzyme, Cyclin-dependent kinase 2 (CDK 2), which is responsible for uncontrolled cell division. Although CDK 2 is not necessary for healthy cells, it plays a crucial role in the uncontrolled growth of cancer cells. Inhibiting the activity of CDK 2 can restrain tumor growth, making it crucial to find effective CDK 2 inhibitors. The chemists from RUDN University and their colleagues in China used a combination of machine learning and molecular modeling techniques to identify several potential inhibitors.

"Cyclin-dependent kinase 2 is a promising target for cancer treatment. The development of its inhibitors is important in antitumor therapy. The participation of this enzyme in tumor formation remains incompletely studied, but it is already clear that its inhibition is useful in the treatment of cancer. Several inhibitors have already undergone clinical trials, but a selective inhibitor specifically for this enzyme has not yet been found," said Alexander Novikov, Ph.D. in Chemistry, senior researcher at the Joint Institute of Chemical Research of RUDN University.

Chemists utilized machine learning methods to identify a potential candidate drug. The authors of the study developed multiple models to find active inhibitors of CDK 2. Additionally, they built a molecular model using the molecular docking method, which can identify the most favorable molecular orientation for forming a stable complex.

Using machine learning models, the team identified 25 potential active CDK 2 inhibitors with an accuracy of 98%. Chemists then tested each of the identified inhibitors using molecular docking. Out of the 25, three substances proved to work better than the rest. For the top three, a computer simulation was created using the molecular dynamics method and compared with the reference compound, dalpiciclib. The results showed that all three of the chosen inhibitors were more stable and more compact than the reference compound.

"Compared to the control drug dalpiciclib, the three calculated compounds showed more stable behavior and compactness. Despite the promising results, our study has several limitations. We need in-depth clinical trials in vitro and in vivo to confirm inhibitory activity and potential therapeutic efficacy. In addition, when developing drugs, it will be necessary to study the effect of compounds on off-target interactions and their toxicity," Alexander Novikov, Ph.D. in Chemistry, senior researcher at the Joint Institute of Chemical Research of RUDN University.

The research concludes that Russian chemists have utilized machine learning and molecular modeling effectively to detect new potential anticancer drugs. This is a significant breakthrough in cancer research that could save countless lives by leading to the development of new treatments. The research also exemplifies the power of combining modern technology with traditional scientific methods, demonstrating the potential of machine learning and molecular modeling to revolutionize the field of medicine.

Illustration of the structure of the nanozymes obtained, with details on how the tyrosine amino acids (in red) coordinate the metal ions (in orange).
Illustration of the structure of the nanozymes obtained, with details on how the tyrosine amino acids (in red) coordinate the metal ions (in orange).

Unlock the full potential of CO2 capture with minimal molecules, the revolutionary solution to creating a greener future

Tyler O'Neal, Staff Editor ENGINEERING

Researchers at the Universitat Autonoma de Barcelona (UAB) have developed enzymes capable of capturing carbon dioxide (CO2) emitted in industrial processes and other environmental remediation processes. These enzymes are based on artificial molecular structures formed by peptides of only seven amino acids. The new molecules can also act as metalloenzymes, which opens up new possibilities in biotechnology research. Furthermore, the study provides a new contribution to the origin of catalytic activity at the beginning of life.

The study was coordinated by Salvador Ventura, and Susanna Navarro was the first author. Both are researchers at the Institute of Biotechnology and Biomedicine and the UAB Department of Biochemistry and Molecular Biology. They collaborated with researchers from the UAB Department of Chemistry and the Research Centre bioGUNE.

In the study, the researchers used a combination of experiments and simulations, including spectrophotometry, fluorescence, electron microscopy, electron diffraction, and supercomputational modeling.

In 2018, researchers at UAB successfully created short molecules that can self-assemble. These molecules were inspired by the natural ability of amyloid fibrils to self-assemble and were based on a specific sequencing of prion proteins. These artificial amyloids demonstrate catalytic activities and have several advantages over natural enzymes, including modularity, flexibility, stability, and ease of use. Recently, researchers discovered that these molecules can effectively bind to metal ions and act as storage elements for metal and metalloenzymes.

“These peptides were particular, since they did not contain the typical amino acids, such as histidine, which is often considered essential for the coordination of metal ions in enzymes, and which were thought to be essential for catalytic activity. In contrast, they were enriched with residues from tyrosine, an element which although less known in this context, can also have the unique capacity of binding to metal ions if it finds itself in the correct structural context. Tyrosine’s ability to do so is what we used to create our nanozymes,” Salvador Ventura points out.

The results of the study have wide-ranging applications. Firstly, nanozymes exhibit excellent stability and can be employed for environmental remediation purposes, including wastewater treatment and decontamination of soils, due to their remarkable ability to sequester metal ions. Secondly, they can act as metalloenzymes, catalyzing reactions in conditions where current enzymes would be incapable of functioning due to their instability. This creates exciting new possibilities for biotechnology research, such as catalyzing reactions in extreme temperatures and pH values.

Researchers have developed a minimalistic variant of a carbonic anhydrase enzyme that can efficiently store CO2 emitted by greenhouse gases. They are convinced that their enzyme is highly capable and can be produced at a much lower cost than natural enzymes.

Researchers have developed new nanozymes by exploring the catalytic activity of short, low-complexity peptides that self-assemble into structures similar to amyloids. This hypothesis suggests that such structures acted as the primal ancestral enzymes, playing a vital role in the origin of life.

“Showing that these molecules have catalytic action without the need for conventional histidine-based coordination represents a significant change in how we understand the origin of catalytic activity at the start of life. We now know that this activity can be achieved if the ancestral peptides contain tyrosine. Therefore, we suggest that it is highly probable that the ancestral enzymes based on amyloids also used this second amino acid in their chemical reactions,” Salvador Ventura concludes.

The potential of minimal molecules in capturing CO2 is a promising development in the battle against climate change. By leveraging the unique properties of these molecules, we can create cost-effective solutions to reduce CO2 emissions and safeguard our planet. With further research and development, minimal molecules can become a potent tool in combating global warming and the consequences of climate change. With the appropriate resources and commitment, we can build a greener, cleaner, and more sustainable future for generations to come.

Herbivorous parrotfish feeding in the shallows on Palmyra Atoll Credit: Brian Zgliczynski
Herbivorous parrotfish feeding in the shallows on Palmyra Atoll Credit: Brian Zgliczynski

Researchers in Bangor show how human activities are contributing to the destruction of communities of reef fish

Tyler O'Neal, Staff Editor ENGINEERING

A recent study has provided evidence that supports one of the ecological theories developed in the 1950s and 1960s. This theory has been used to predict how different species are distributed in various environments. The study raises questions about whether these models need to be updated to account for the impact of humans on natural systems. The coral reef zonation theory is one of the earliest examples of these models, and it explains how different types of fish and corals are found at different depths in coral reefs. The accuracy of these theories has been previously tested on a small scale, and they have proven to be reliable across a wide range of variables such as food supply and temperature.

Modern supercomputing capabilities now allow testing theories at larger scales.

Scientists from Bangor University and the US Government National Oceanic and Atmospheric Administration (NOAA), led by Dr. Laura Richardson, validated the depth zonation model on coral reefs. They collected data from 5525 surveys at 35 Pacific Ocean islands to determine the distribution of different fish species according to depth. The results showed that the model is accurate but only on uninhabited islands where there is no human interference.

However, the pattern was not as clear and predictable on islands and reefs with human habitation. The findings suggest that traditional models of the natural world may no longer be valid due to increasing human impact.

Dr. Laura Richardson of Bangor University’s School of Ocean Sciences, the lead author of the study, suggests that we need to revise our understanding of the natural world in light of these findings.

“Science is cumulative, building on past work. Now that we have greater computing capabilities, we should be testing these widely accepted but spatially under-validated theories at scale. Moreover, the intervening years have seen human impacts on the environment increase to such an extent that these models may no longer predict the ecological distribution patterns we see today.

“This leads to more questions, both about the usefulness of models which represented a world less impacted by human activity, and about how to quantify or model our impact on the natural environment.”

“The results show that now is the time to consider whether and how to include human impacts into our understanding of the natural world today,” said Dr. Richardson.

The study concludes that human activities at a local level are significantly and adversely affecting the depth-dependent zonation of tropical reef fish communities. This has resulted in a decline in the abundance and diversity of fish species, leading to an overall decline in the health of the reef ecosystem. This trend is worrying and emphasizes the need for increased conservation measures to safeguard these delicate ecosystems.