Research
My laboratory is interested in problems at the interface of chemistry and biology. I combine theoretical and computational approaches to explore mechanistic details of chemically or biologically relevant reactions. Such quantitative details are essential in developing novel catalysis and smart materials for modern technological applications in industry and academia. My research is conducted in close collaboration with well-established experimental research groups in Europe, Japan, and the USA.
Software
QM/MM: PyQMMM, SICTWO
Electronic structure: Gaussian16, ADF, ORCA
Molecular dynamics: DFTB, TINKER
Periodic DFT: ADF-BAND
Computing
In-house servers:
HPE ProLiant DL360 Gen10 | 40, 64, or 80 CPUs | 256 GB or 320 GB memory | 1.2 TB HDD.
Supercomputers:
C3SE (Gothenburg)
ICR (Kyoto)
Development of computational methods
The chemical reactions are very fast processes (i.e., fs scale), and the lifetime of the intermediates is very short. Therefore, experimental characterisation of the atomic-scale chemical processes is challenging. Modern quantum chemical methods, employing ab-initio computations, offer a way to overcome these limitations. In this direction, I develop an Unbiased Reaction Path Search (URPS) approach to study complex molecular systems. My current focus is to combine URPS approach with a neural network, the so-called Artificial Intelligence-Guided Reaction Path Refinement Algorithm (AI-RPRA) to determine complex molecular structures or complex reaction mechanisms.
Also, I develop quantum mechanics/molecular mechanics (QM/MM) methods. In the QM/MM approach, the electronically important part of the molecular system is calculated using a QM method, while the remaining part is calculated using a MM method. The polarizable force fields are an advanced type of computational model used in QM/MM calculations to accurately simulate molecular interactions by incorporating the ability of atoms to change their electronic distributions in response to their environment. The QM/MM implementation in the in-house PyQMMM and SICTWO programs supports various modern polarizable and non-polarizable force fields.
Design novel biomimetic catalysis for energy applications and pharmaceutical industry
The Haber-Bosch process, the most successful industrial method for NH3 production from N2 and H2 gases, requires high temperatures and pressures. Thus, developing novel catalysts for NH3 synthesis under ambient conditions has become a prominent research area. In this pursuit, biomimetic Fe-S clusters play a key role, as the Fe-S clusters in the active site of the nitrogenase enzyme catalyse the reduction of N2 to NH3. I investigate metal-sulphur clusters as biomimetic models of the nitrogenase enzyme. These Fe-S clusters function as catalysts for the silylation of N2 and the reduction of CO2 to CH3OH under mild conditions. My long-term goal is to employ the URPS approach to design novel biomimetic catalysts for energy applications.
Collaborations: Yasuhiro Ohki (Kyoto).
The cross-coupling reactions play a key role in the synthesis of the chemically or biologically relevant compounds. In recent years, I have contributed to the development of transition metal catalysts for C–X bond formation reactions (where X = C, B). I employ DFT to elucidate the mechanisms of C–X bond formation reactions and use the URPS approach to understand the origin of selectivity. The URPS approach is vital for optimising catalyst selectivity, particularly through ligand design. My long-term goal is to utilise the UPRS approach to design novel ligands for homogeneous catalysts to synthesize drug-like molecules in a selective fashion for the pharmaceutical industry.
Experimental collaborations: Youhei Takeda (Osaka University, Japan), Arundhati Nag (Clark University, USA). Pedro J. Perez (Huelva).
Origin of life in the Universe
Radical-driven reactivity: More than 300 molecules have been found in the interstellar medium (ISM). However, the origin of the molecules in the ISM and the chemical evolution in the Universe that leads to the origin of life are unknown. One of the plausible theories for the origin of molecules in the ISM is the radical-driven chemistry of ice dust particles. Further, the molecular building blocks of the molecules in the ISM, particularly small radicals may adsorb on the interstellar ice grains in the ISM, diffuse, and react at low temperatures (e.g., 10 K). These chemical processes are not fully rationalized. I use computational methods to determine reaction mechanisms for the synthesis of biologically relevant molecules (E.g., amino acids, sugars) in the ISM.
Anion-driven reactivity: My recent ab-initio computations showed the effectiveness of the OH- (anion) diffusion via the proton hole transfer in ice at low temperatures. Thus, driving the OH- (anion) in ice bulk increases the possibility of reacting with the molecules trapped in ice and opens the OH- (anion)-driven reactivity to synthesize molecules in the interstellar ice bulk at low temperatures. I aim to establish anion-driven reactivity using the URPS scheme to determine known, unknown, and unexpected anion-driven reaction mechanisms to construct chemical networks.
Collaborations: Francois Dulieu (CY Cergy Paris Université),Bethmini Senevirathne (Gothenburg), Stefan Andersson (SINTEF), Gunnar Nyman (Gothenburg)
Design luminescent compounds for probe and sensor technology and cellular imaging applications
The design of luminescent compounds for probe and sensor technology, particularly in cellular imaging applications, are essential in biological research, diagnosis, or treatment of diseases. This field explores how physical principles are applied to understand and enhance biological processes at the molecular level, enabling the development of advanced imaging and sensing tools. Recent advances in computational chemistry provide experimental researchers with quantitative insights into the ground and excited state potential energy surfaces and their contributions to the emission properties. These insights are crucial for optimising luminescent materials to achieve high emission quantum yield and for tuning their emission colour.
I study novel luminescent complexes that respond to macroscopic gentle stimuli (e.g. vapour exposure, rubbing, and rotation), and exhibit visually remarkable changes such as luminescence and optical properties. Also, through ligand design, I develop luminescent compounds for various cellular imaging applications. For example, NS1 protein is used as a diagnostic to detect Dengue virus infection in a patient at the early stages of an infection. Moreover, I develop small molecule fluorescent binding partners that can be developed as diagnostic tools using in silico methods.
Collaborations: Masako Kato (Kwansei), Chathura Abeywickrama (Connecticut)
Computational biology
Molecular dynamics (MD) simulations are a powerful computational tool used in biology to study the physical movements of atoms and molecules over time. By applying principles of physics and chemistry, MD simulations provide detailed insights into the dynamic behavior of biomolecules like proteins, nucleic acids, and membranes at an atomic level. These simulations help researchers understand complex biological processes, such as protein folding, enzyme-substrate interactions, and conformational changes, which are challenging to capture experimentally. With advances in computational power, MD simulations are increasingly used to predict molecular behavior, support drug discovery, and explore fundamental mechanisms in cellular function.
Antibiotic resistance and host toxicity have limited the clinical use of many antibiotics. I combine experimental and computational methods to design novel antibiotic combination regimens to facilitate treatment efficacy while lowering their effective dose to improve the safety margins.
Collaborations: Gayathri N. Silva (Colombo)
Dengue virus is a pathogen that affects a significant fraction of the global population. NS1 protein is used as a diagnostic to detect dengue infection in a patient at the early stages of an infection. I develop small molecule fluorescent binding partners that can be developed as diagnostic tools using in-silico methods.
Collaborations: Keveesha Wijesinghe (Colombo)