About

Zygmunt S. Derewenda is a Professor of Molecular Physiology and Biological Physics at the University of Virginia School of Medicine. He holds an MSc in Biophysics and a PhD in Chemistry from the University of Lodz, Poland, and completed a postdoctoral fellowship in Protein Crystallography at the University of York, United Kingdom. His research disciplines include biochemistry, biophysics, structural biology, molecular biology, physiology, and biotechnology. His research interests focus on structure-function relationships in proteins. Currently, his laboratory is engaged in elucidating the molecular mechanisms controlling smooth muscle contractility, developing drugs against influenza and Ebola viruses, and studying the mechanisms of protein crystallization along with developing new methods in this field.

Research topics

• Crystallography
• Organic chemistry
• Chemistry
• Biochemistry
• Art history
• History
• Polymer science
• Stereochemistry

Selected publications

• How protein kinase inhibitors bind to the hinge region of the target protein.

  Structural Dynamics · 2025-09-01 · 1 citations

  articleOpen accessSenior author

  The human genome encodes 518 protein kinases which make up one of the most important families of regulatory proteins, catalyzing phosphorylation of hydroxyl-containing amino acids, i.e. tyrosine, serine and threonine [1]. These enzymes are involved, among others, in the regulation of cell cycle [2], cell growth and proliferation, as well as inflammation an immunological processes [3]. All of these phenomena are of significance in tumorigenesis and tumor growth and metastasis, and protein kinases are of-ten selectively overexpressed in a number of cancer types [4]. As a result, these regulatory enzymes now constitute one of the most important families of anti-cancer and immuno-logical disorder drug targets [3, 5-12]. As of January 2024, there were 80 protein kinase inhibitors approved by the FDA for clinical use, including 69 used to treat cancer an non-malignant neoplasm, and six for the treatment of inflammatory diseases [11]. More than 700 other protein kinase inhibitors are in clinical trials, promising to expand this class of drugs [13]. However, at this moment only ∼100 kinases are targeted, while the function of the remaining ones, and their potential as drug targets, are yet to be elucidated [14]. Protein kinases are typically large, multi-domain proteins, and the catalytic activity is harbored within a specific enzymatic module, which show significant amino acid and structural conservation, particularly within the ATP-binding pocket, which is targeted by most inhibitors [1, 15, 16]. The kinase catalytic domain is made up of two ‘lobes’: the N-terminal lobe (N-lobe), which is structurally highly malleable and subject to regulation, including phosphorylation, and the stable, α-helical C-lobe which is the substrate binding site [17]. The large solvent-accessible cavity between the two lobes is the binding site for ATP, organized in such a way that the adenine moiety penetrates most deeply, and is recognized via hydrogen bonds by a short stretch of the polypeptide chain linking the two lobes, i.e. the hinge motif [18]. Kinase inhibitors belong to six distinct types [11]. Types I and II, which include the vast majority clinical drugs, occupy the ATP-site and are therefore ATP competitors [19]. Type III and IV bind to other sites, one of which is a regulatory site adjacent to the ATP pocket and the other an allosteric site. Type V are bivalent inhibitors which may use the ATP site, while type VI are covalent inhibitors, that attach to a cysteine (or sometimes ly-sine) residues close to the ATP binding site, with an active module of the compound blocking irreversibly the latter. Thus, most clinically used inhibitors contain a module that acts as an anchor binding to the hinge motif. The ATP-pocket of kinases is selective for adenine, vs guanine, owing to a specific hydrogen bond (H-bond) network, involving the hinge. Three amino acids that make up the hinge are in extended, β-sheet conformation, denoted gk+1, gk+2 and gk+3 in accordance to their position downstream in the sequence from the ‘gatekeeper’ residue (or gk) [20]. It was first noted in 1997 [21] that a number of ATP-binding proteins share a main chain architecture found in kinases, that facilitates H-bonding of the gk+1 carbonyl to the 6-amino group, and the gk+3 peptide amide to N1 of adenine [22]. Subsequently, it was noted that many kinase inhibitors show a similar pattern of H-bonding, engaging the same main groups of the hinge, and thus mimicking ATP [18]. Recently, a structural database of 2,705 complexes of kinase domains with inhibitors was analyzed and helped to identify fifteen distinct binding modes [23]. Such studies are valuable for the design of alternative hinge binding motifs to assist in structure based drug design [24] (FIGURE 1). The above-cited studies focused on the canonical H-bonds between electronegative atoms in both inhibitors and enzymes. In reality, kinase- inhibitor interactions are mediated also by a plethora of C-H…O bonds, which—while weaker—have the potential to make significant contribution [25-28]. The existence of these bonds in proteins and protein-ligand interactions, as well as their energies, are well established [25, 29, 30]. In fact, adenine forms not two, but three H-bonds, with the third one mediated by the C2-H group of adenine which donates a proton to the main chain carbonyl of gk+3 (FIGURE 2). The presence of such bonds at kinase-inhibitor interfaces has been recognized for a long time [31, 32]. We have recently shown by mining the Protein Data Base, that the vast majority of kinase inhibitor scaffolds contain a specific core moiety—typically of an aromatic nature—which mimics the interaction of adenine with the hinge with three H-bonds, including canonical and C-H…O [33]. These core moieties contain H-bond donors (including methine groups, i.e. =CH-) and acceptors, which interact with the gk+1 and gk+3 carbonyl groups and the gk+3 amide [33]. Specifically, we identified three distinct templates where one or two of the potential H-bond sites involve a C-H donor from the inhibitor. However, our studies were limited to the analyses of the stereochemisty derived from crystal structures. Although as the presence of cohesive H-bonds can be inferred with reasonable accuracy from interatomic distances and angles, they fail to provide insight into energetic contributions, and therefore to the function and significance of specific interatomic interactions. It is also not clear, for example, if the core-hinge interaction is a major determinant of the binding affinity (or kD), or if such interactions are optimal in specific cases, or if are hindered by secondary enzyme-inhibitor contacts. Dissection of these questions is of paramount importance to open new approaches in structure-based drug design. To address these important questions, we resorted to quantum mechanics (QM) as implemented in Gaussian 16 [34] and Atoms-In- Molecules (AIM) [35], and the available X-ray crystallographic structures of a number of FDA-approved kinase-inhibitor complexes. We analyzed representative complexes and in each case dissected the energetic contribution of each of the H-bonds involved. We show that the energies of interactions of the cores of inhibitors and their target hinge motifs is mostly dependent on the three H-bonds, of which typically at least one is of the C-H…O type. Importantly, the outcome of the calculations is highly dependent on the quality of the crystallographic model, as well as the intrinsic limitations of each of the computational methods. Our results should are of importance to drug design/discovery projects.

  PublisherOA PDFDOI

• The occurrence of TrpCd1-H…O=Cbackbone hydrogen bonds in proteins

  Structural Dynamics · 2025-03-01

  articleOpen accessSenior author

  Tryptophan is the largest amino acid found in proteins, with multiple functional roles. Its side-chain is made up of the hydrophobic indole moiety, with two groups acting as donors in hydrogen bonds: the Nε-H group, which is a potent donor in canonical hydrogen bonds; and a polarized Cδ1-H group, capable of forming weaker, non-canonical hydrogen bonds. Due to adjacent electron withdrawing moieties, C-H…O hydrogen bonds are ubiquitous in macromolecules, albeit contingent on the polarization of the donor C-H group. Consequently, the Ca-H groups (adjacent to carbonyl and amino groups of flanking peptide bonds), as well as Ce1-H and Cd2-H groups of histidines (adjacent to imidazole nitrogen atoms) are known to serve as donors in hydrogen bonds, e.g. stabilizing parallel and antiparallel β-sheets. However, the nature and the functional role of interactions involving the Cd1-H group of the indole ring of tryptophan are not well characterized. We perform data mining of high- resolution (r £ 1.5 Å) crystal structures from the Protein Data Bank, and identified ubiquitous close contacts between the Cd1-H groups of tryptophan and a range of electronegative acceptors, specifically main-chain carbonyl oxygen atoms immediately up- and downstream in the polypeptide chain. Our stereochemical analysis shows that most interactions bear all the hallmarks of proper hydrogen bonds. At the same time, their cohesive nature is confirmed by quantum chemical calculations, which reveal energies of the interaction of 1.5–3.0 kcal/mol, depending on the specific stereochemistry. Our study demonstrates that the Cd1-H group of a tryptophan residue plays an important role in stabilizing unique structural motifs by engaging as an H-bond donor with main-chain carbonyl oxygen atoms nearby along the sequence. The most common such interactions involves residues one peptide unit downstream, i.e. i(+1) (Fig 1), or 1–4 peptide units upstream, i.e. i(-1) – i(-4) (Fig. 2). Interestingly, of the possible twelve conformers, Trp is found in these motifs in only six, with the i(+1) class containing only conformers p90, t0, and t-105, while the remaining four classes show Trp only in m105, m0 and p-90. The frequencies of the high-energy m0 and t0 conformers is increased significantly for those classes where the contacts are strongly restricted by short-range steric constraints, while m105, the most populous class found in proteins, is strongly enriched in the i(-3) and i(-4) classes. Our work helps explain the relatively common occurrence of the m0 and t0 classes. It is important to note that the function of Trp residues is intimately contingent on its conformation. For example, Trp in transmembrane helices occurs most often in m0, t0 and p-90 conformations, all of which have been characterized by our study. Of importance is our observation that in the -3 and -4 classes, Trp is often engaged in capping of the acceptor oxygen atom with H-bonds donated by both the amide and Cd1-H groups. We also present evidence based on quantum chemical calculations that the short Cd1-H….O=C contacts revealed by structural data mining are in fact invariably cohesive interactions of the order of approximately half a canonical H-bond, and less sensitive to specific stereochemistry, such as C-H…O and H…O=C angles, then previously thoughts. The critical factor is the position of the hydrogen atom close to the sp2 plane of the acceptor oxygen atom.

  PublisherDOI

• Differing Effects of Nonlinearity around the Proton Acceptor on CH··O and NH··O H-Bond Strength within Proteins

  The Journal of Physical Chemistry B · 2024-07-18 · 4 citations

  articleSenior author

  The effects of deviations from nonlinearity around the carbonyl proton acceptor of an amide group are assessed by DFT quantum chemical calculations for both CH··O and NH··O H-bonds. The proton donors are the imidazole functional group of His and the indole of Trp, which are paired respectively with N-methylacetamide and acetamide. The displacement of either CH or NH group toward the carbonyl O sp2 lone pairs stabilizes the system and strengthens the H-bond. But the two donor groups differ in their response to a shift out of the amide plane. While the NH··O H-bond is weakened by this displacement, a substantial strengthening is observed when the CH donor is moved out of this plane, in one direction versus the other. This pattern is explained on the basis of simple Coulombic considerations.

  PublisherDOI

• A structural role for tryptophan in proteins, and the ubiquitous Trp C ^δ1 —H...O=C (backbone) hydrogen bond

  Acta Crystallographica Section D Structural Biology · 2024-06-22 · 7 citations

  articleOpen accessSenior author

  Tryptophan is the most prominent amino acid found in proteins, with multiple functional roles. Its side chain is made up of the hydrophobic indole moiety, with two groups that act as donors in hydrogen bonds: the N ɛ —H group, which is a potent donor in canonical hydrogen bonds, and a polarized C δ1 —H group, which is capable of forming weaker, noncanonical hydrogen bonds. Due to adjacent electron-withdrawing moieties, C—H...O hydrogen bonds are ubiquitous in macromolecules, albeit contingent on the polarization of the donor C—H group. Consequently, C α —H groups (adjacent to the carbonyl and amino groups of flanking peptide bonds), as well as the C ɛ1 —H and C δ2 —H groups of histidines (adjacent to imidazole N atoms), are known to serve as donors in hydrogen bonds, for example stabilizing parallel and antiparallel β-sheets. However, the nature and the functional role of interactions involving the C δ1 —H group of the indole ring of tryptophan are not well characterized. Here, data mining of high-resolution ( r ≤ 1.5 Å) crystal structures from the Protein Data Bank was performed and ubiquitous close contacts between the C δ1 —H groups of tryptophan and a range of electronegative acceptors were identified, specifically main-chain carbonyl O atoms immediately upstream and downstream in the polypeptide chain. The stereochemical analysis shows that most of the interactions bear all of the hallmarks of proper hydrogen bonds. At the same time, their cohesive nature is confirmed by quantum-chemical calculations, which reveal interaction energies of 1.5–3.0 kcal mol −1 , depending on the specific stereochemistry.

  PublisherOA PDFDOI

• C—H...O bonds involving Trp sidechain in protein structures

  Acta Crystallographica Section A Foundations and Advances · 2023-07-07

  articleOpen accessSenior author

  C-H...O hydrogen bonds are increasingly recognized as significant interactions which contribute to protein folding, stability and catalytic function.They typically involve polarized C-H groups such as those within the imidazole ring of histidine.The tryptophan side chains are another candidate with the C(2)-H group the most polarized by the adjacent nitrogen.We surveyed the Protein Data Bank, using a subset of non-redundant structures at near-atomic and atomic resolution to identify potential C-H...O bonds based on the stereochemistry of the interactions.We will present data showing the C( 2)-H group acts as a donor of hydrogen bonds involving water and other molecules, occasional carbonyl groups from the main chain groups andimportantly -selected halide ions.Importantly, a number of such interactions occur within active sites of enzymes, such as haloalkane dehalogenases, suggesting an active role in the catalytic mechanisms.Our results reaffirm the importance of C-H...O interactions in proteins.

  PublisherOA PDFDOI

• p90RSK2, a new MLCK, rescues contractility in myosin light chain kinase null smooth muscle

  bioRxiv (Cold Spring Harbor Laboratory) · 2023-05-22

  preprintOpen access

  Abstract Background Phosphorylation of smooth muscle (SM) myosin regulatory light chain (RLC 20 ) is a critical switch leading to contraction or cell migration. The canonical view held that the only kinase catalyzing this reaction is the short isoform of myosin light chain kinase (MLCK1). Auxiliary kinases may be involved and play a vital role in blood pressure homeostasis. We have previously reported that p90 ribosomal S6 kinase (RSK2) functions as such a kinase, in parallel with the classical MLCK1, contributing ∼25% of the maximal myogenic force in resistance arteries and regulating blood pressure. Here, we take advantage of a MLCK1 null mouse to further test our hypothesis that RSK2 can function as an MLCK, playing a significant physiological role in SM contractility. Methods Fetal (E14.5-18.5) SM tissues were used as embryos die at birth. We investigated the necessity of MLCK for contractility, cell migration and fetal development and determined the ability of RSK2 kinase to compensate for the lack of MLCK and characterized it’s signaling pathway in SM. Results Agonists induced contraction and RLC 20 phosphorylation in mylk1 -/- SM, that was inhibited by RSK2 inhibitors. Embryos developed and cells migrated in the absence of MLCK. The pCa-tension relationships in WT vs mylk1 -/- muscles demonstrated a Ca 2+ -dependency due to the Ca 2+ -dependent tyrosine kinase Pyk2, known to activate PDK1 that phosphorylates and fully activates RSK2. The magnitude of contractile responses was similar upon addition of GTPγS to activate the RhoA/ROCK pathway. The Ca 2+ -independent component was through activation of Erk1/2/PDK1/RSK2 leading to direct phosphorylation of RLC 20 , to increase contraction. RSK2, PDK1, Erk1/2 and MLCK formed a signaling complex on the actin filament, optimally positioning them for interaction with adjacent myosin heads. Conclusions RSK2 signaling constitutes a new third signaling pathway, in addition to the established Ca 2+ /CAM/MLCK and RhoA/ROCK pathways to regulate SM contractility and cell migration.

  PublisherOA PDFDOI

• p90RSK2, a new MLCK mediates contractility in myosin light chain kinase null smooth muscle

  Frontiers in Physiology · 2023-09-13 · 2 citations

  articleOpen access

  Introduction: Phosphorylation of smooth muscle (SM) myosin regulatory light chain (RLC 20 ) is a critical switch leading to SM contraction. The canonical view held that only the short isoform of myosin light chain kinase (MLCK1) catalyzed this reaction. It is now accepted that auxiliary kinases may contribute to vascular SM tone and contractility. We have previously reported that p90 ribosomal S6 kinase (RSK2) functions as such a kinase, in parallel with MLCK1, contributing ∼25% of the maximal myogenic force in resistance arteries. Thus, RSK2 may be instrumental in the regulation of basal vascular tone and blood pressure. Here, we take advantage of a MLCK1 null mouse ( mylk1 −/− ) to further test our hypothesis that RSK2 can function as an MLCK, playing a significant physiological role in SM contractility. Methods: Using fetal (E14.5-18.5) SM tissues, as embryos die at birth, we investigated the necessity of MLCK for contractility and fetal development and determined the ability of RSK2 kinase to compensate for the lack of MLCK and characterized its signaling pathway in SM. Results and Discussion: Agonists induced contraction and RLC 20 phosphorylation in mylk1 −/− SM was attenuated by RSK2 inhibition. The pCa-tension relationships in permeabilized strips of bladder showed no difference in Ca 2+ sensitivity in WT vs mylk1 −/− muscles, although the magnitude of force responses was considerably smaller in the absence of MLCK. The magnitude of contractile responses was similar upon addition of GTPγS to activate the RhoA/ROCK pathway or calyculinA to inhibit the myosin phosphatase. The Ca 2+ -dependent tyrosine kinase, Pyk2, contributed to RSK2-mediated contractility and RLC 20 phosphorylation. Proximity-ligation and immunoprecipitation assays demonstrated an association of RSK2, PDK1 and ERK1/2 with MLCK and actin. RSK2, PDK1, ERK1/2 and MLCK formed a signaling complex on the actin filament, positioning them for interaction with adjacent myosin heads. The Ca 2+ -dependent component reflected the agonist mediated increases in Ca 2+ , which activated the Pyk2/PDK1/RSK2 signaling cascade. The Ca 2+ −independent component was through activation of Erk1/2/PDK1/RSK2 leading to direct phosphorylation of RLC 20 , to increase contraction. Overall, RSK2 signaling constitutes a new third signaling pathway, in addition to the established Ca 2+ /CaM/MLCK and RhoA/ROCK pathways to regulate SM contractility.

  PublisherOA PDFDOI

• C-H Groups as Donors in Hydrogen Bonds: A Historical Overview and Occurrence in Proteins and Nucleic Acids

  International Journal of Molecular Sciences · 2023 · 49 citations

  1st authorCorresponding
  • Chemistry
  • Crystallography
  • Biochemistry

  Hydrogen bonds constitute a unique type of non-covalent interaction, with a critical role in biology. Until fairly recently, the canonical view held that these bonds occur between electronegative atoms, typically O and N, and that they are mostly electrostatic in nature. However, it is now understood that polarized C-H groups may also act as hydrogen bond donors in many systems, including biological macromolecules. First recognized from physical chemistry studies, C-H…X bonds were visualized with X-ray crystallography sixty years ago, although their true significance has only been recognized in the last few decades. This review traces the origins of the field and describes the occurrence and significance of the most important C-H…O bonds in proteins and nucleic acids.

  PublisherOA PDFDOI

• <b>Secret of Life: Rosalind Franklin, James Watson, Francis Crick, and the discovery of DNA's Double Helix</b>. By Howard Markel. W. W. Norton &amp; Co., 2021. Hardback, pp. 608. ISBN 978-1324002239. Price USD 30.00.

  Acta Crystallographica Section D Structural Biology · 2022-11-29

  article1st authorCorresponding

  Secret of Life: Rosalind Franklin, James Watson, Francis Crick, and the discovery of DNA's Double Helix. By Howard Markel. W. W. Norton & Co., 2021. Hardback, pp. 608. ISBN 978-1324002239. Price USD 30.00.

  PublisherDOI

• Aurora A phosphorylates Ndel1 to reduce the levels of Mad1 and NuMA at spindle poles

  Molecular Biology of the Cell · 2022-11-09 · 3 citations

  articleOpen access

  extracts. Our data suggest that dynein/SAC complexes that are generated at kinetochores and then transported directionally toward poles on microtubules are inhibited by Aurora A before they reach spindle poles. These data suggest that Aurora A generates a spatial signal at spindle poles that controls dynein transport and spindle function.

  PublisherOA PDFDOI

Recent grants

• Cost Extension for Ramon Ayon Diversity Supplement

  NIH · $3.6M · 2020–2025

• NIH Grant U54GM074946

  NIH · $34.7M · 2011

• NIH Grant R01GM095847

  NIH · $1.6M · 2016

• Molecular Mechanism of RhoA-mediated Ca2+-sensitization in Vascular Smooth Muscle

  NIH · $5.4M · 2009–2019

• NIH Grant R01GM062615

  NIH · $1.0M · 2006

Frequent coauthors

• Urszula Derewenda

  University of Virginia

  114 shared

• Lora Swenson

  Vertex Pharmaceuticals (United States)

  43 shared

• Zbigniew Dauter

  National Cancer Institute

  23 shared

• Yau‐Huei Wei

  Lanzhou University

  23 shared

• Y. Devedjiev

  University of Virginia

  23 shared

• Rolf D. Joerger

  University of Delaware

  22 shared

• Guy Dodson

  The Francis Crick Institute

  19 shared

• P M Kobos

  University of Alberta

  18 shared