Visit from Vilnius University

Prof. Daumantas Matulis and Dr. Vytautas Petrauskas as a part of ongoing collaborative discussions that began with the visit of the Lithuanian delegation to York.

Prof. Daumantas Matulis and Dr. Vytautas Petrauskas are from the Department of Biothermodynamics and Drug Design, Life Sciences Center, Vilnius University. Their research interests are closely related to biomolecular interactions, biophysical analysis, and drug discovery, so I expect the presentation and discussion to be scientifically engaging.

—

Drug Design and Protein–Ligand Binding Database for AI-ML
Prof. Daumantas Matulis
Head of Department of Biothermodynamics and Drug Design, Institute of Biotechnology, Vilnius
University, Saulėtekio 7, Vilnius LT-10257, Lithuania Email: [email protected],
[email protected]

Abstract
Drug molecules exert their effects on disease-associated proteins by binding to them with high
affinity and selectivity. However, most small molecules interact only weakly with the majority of
sites on protein surfaces, and only a limited number of chemical compounds can selectively and
strongly bind to desired sites. In addition, even minor modifications to a compound’s chemical
structure may lead to substantial changes in binding affinity. We designed and investigated the
binding of over 1,000 small-molecule compounds to Carbonic Anhydrase IX, a transmembrane
protein overexpressed in hypoxic solid tumors that promotes metastasis and invasiveness, and
supports cancer cell survival under hypoxic conditions by acidifying the tumor microenvironment.
We characterized the thermodynamics of binding of these compounds across the carbonic
anhydrase enzyme family and identified the determinants of selectivity for CA IX inhibition (1) .
Fluorescent CA IX-selective compounds enabled visualization of CA IX expression in cancer cells
and tumors and demonstrated efficacy against Neuroblastoma in mouse xenograft models.
The data were assembled into the Protein-Ligand Binding Database (PLBD) to apply artificial
intelligence – machine learning approaches to drug design and are available at https://plbd.org
(2) . The database contains 8233 binding datasets of 792 compounds interacting with human CA
isozymes, Hsp90, and COVID proteases, as determined by inhibition of enzymatic activity, TSA, ITC,
and SPR techniques. The PLBD emphasizes intrinsic parameters and enthalpy, enhancing the
understanding of reaction mechanisms. The binding data are linked to 138 crystal structures of CA
complexes with ligands. The database has been built using the FAIR data principles to contain the
raw data with revision and versioning systems. The PLBD led to a deeper understanding of the
protein-ligand recognition principles and enabled the achievement of femtomolar affinities in
small-molecule drug design (3) .

References
1. Matulis, D. 2019. Carbonic anhydrase as drug target: thermodynamics and structure of inhibitor
binding. SPRINGER NATURE.
2. Lingė, D., M. Gedgaudas, A. Merkys, V. Petrauskas, A. Vaitkus, A. Grybauskas, V. Paketurytė, A.
Zubrienė, A. Zakšauskas, A. Mickevičiūtė, J. Smirnovienė, L. Baranauskienė, E. Čapkauskaitė, V.
Dudutienė, E. Urniežius, A. Konovalovas, E. Kazlauskas, K. Shubin, H.B. Schiöth, W.-Y. Chen, J.E.
Ladbury, S. Gražulis, and D. Matulis. 2023. PLBD: protein–ligand binding database of thermodynamic
and kinetic intrinsic parameters. Database 2023:baad040, doi: 10.1093/database/baad040.
3. Zubrienė, A., M. Kurtenoka, V. Paketurytė-Latvė, J. Leitans, E. Manakova, M. Žvirblis, A. Kazaks, V.
Eimonta, K. Tars, S. Gražulis, V. Petrauskas, J. Matulienė, V. Dudutienė, K. Shubin, and D. Matulis. 2026.
Achieving femtomolar affinities in structure-based drug design. Eur. Biophys. J., doi: 10.1007/s00249-
025-01812-5.