Can we engineer molecular computers? - Merkx, Brunsveld, de Greef
The Chemical Biology focus area of ICMS can count on three familiar faces of the Biomedical Engineering Department of TU/e: Luc Brunsveld is full professor of Chemical Biology, Maarten Merkx is full professor of Protein Engineering, and Tom de Greef is associate professor of Synthetic Biology.
That is the question for Tom de Greef. Living organisms such as the human body are based on living cells which are able to sense and adapt to their environment. In a very simplistic way, living cells are - in the words of De Greef – “nothing more than chemical computers with molecular software and predefined rules”.
De Greef : “For many decades, biologists have studied cells via a reverse engineering approach deconstructing complex living systems into their constitutive elements, in order to extract knowledge on the molecular software”. In the last 20 years, scientists started moving towards a ‘forward engineering approach’, building complex biological systems starting from lower-level details like molecular circuits and networks. “In our group, we use DNA as a programmable substrate to make those networks. We engineer biology from scratch”, explains De Greef.
DNA can spontaneously process molecular operations via hybridization, a phenomenon in which single stranded molecules pair together via hydrogen bonds. It’s no surprise then that researchers worldwide have started looking at DNA as an appealing material for computing. DNA molecules have been used to create simple logic gates and circuits, the basic building blocks of computing. And, as DNA computers can naturally interface with biological signals, DNA-based computers opened the way for biological applications that, to date, conventional silicon based computers haven’t been capable to address. In biological scenarios, DNA computers are “the most efficient machines to control molecular operations”, says Brunsveld, and the perfect reflection of our modus operandi. De Greef: “Human beings compute using molecular signals. Our immune system, for example, can be seen as a neural network that learns from chemical patterns and, based on those, takes decisions.”
Merkx: “In our group, we use protein engineering to make protein-based sensors. We would like those sensors to be intelligent or, in other words, capable of sensing and responding to different biological signals”. That is how Merkx got interested in DNA computing. “The incorporation of all those functions in a single protein was simply too difficult. We practically reached the limits of what you can do with a single protein”, explains Merkx. And that’s when Merkx and his colleagues started looking at DNA and its “inherent advantage of being programmable”. Combining protein engineering and DNA nanotechnology, Merkx and De Greef are currently busy with the development of intelligent biomolecular sensors for applications in intracellular imaging, diagnostics and, possibly, antibody-based therapies.
Merkx: “To date, most of the research on DNA computing is DNA-based only, with DNA both as an input and as an output. Instead, we are very much interested in connecting DNA computing to the real world, the world of proteins. To do so, we use proteins, small molecules or antibodies as inputs. These inputs are used to control the release of drugs for specific therapies”, says Merkx. His ultimate dream is an autonomous, closed-loop therapeutic system where “human intervention is no longer contemplated.”
Even though DNA-based computers are on the horizon, Merkx acknowledges their current limitations: “Current DNA computers are very slow. Also, they are usually tested in diluted solutions, in systems that are not confined. However, when it comes to complex diagnostic and therapeutic environments, you must confine these systems to some extent.” In a recent publication, Merkx and colleagues demonstrated how to do so using supramolecular polymers. Merkx: “We developed DNA-functionalized supramolecular polymers, which could be used as dynamic scaffolds for DNA-based molecular computing.” Organizing the DNA circuits within supramolecular architectures, Merkx and his colleagues were able to accelerate the kinetics of reactions a hundredfold, elevating supramolecular polymers to efficient autonomous systems for molecular sensing, computation, and actuation.
De Greef: “Typically, a disease cannot be classified based on a single biomarker. Patterns, more than single biomarkers, are the real signatures of a disease state.” In this respect, the potential of DNA computers is endless, especially when compared to current diagnostic methods. DNA-based computers could detect a disease state in its entireness and in one shot, by sensing patterns of biomarkers, and by processing and responding to this information in a fully autonomous way.
All the ingredients for the implementation of DNA-computers in clinical routine are undeniably there. But, the journey isn’t over yet. In biological environments DNA-computers still demonstrate their fragility. De Greef: “Take for example blood, which contains nucleases: enzymes that are capable of destroying DNA computers. This poses a major limitation which current research is still trying to overcome.”
The ‘fingerprint’ of the research of Brunsveld, De Greef and Merkx is their multidisciplinary approach which matches perfectly with the ICMS vision. Brunsveld: “ICMS brings together scientists with different backgrounds that interact in a very open way. We ask each other critical questions, which can making our projects stronger or result, sometimes, in completely new lines of thought.”