A collaboration between researchers in the United States, the United Kingdom and Hungary has mutated the amino acid sequence of the AppA BLUF domain, a blue-light photoreceptor, and succeeded in capturing the proton-coupled electron transfer (PCET) radical intermediates that had never been observed in this protein. By introducing a double mutation (H44N, H78R) that adds a second hydrogen bond at the C2=O position of the flavin (FAD), the team steered AppA's photoactivation pathway toward the PixD model, in which radicals are observed. The work shows that a photoreceptor's "on" pathway (photoactivation) and its "off" pathway (dark state recovery) can be manipulated separately, while the authors also raise the possibility that these radicals were not newly created but were already present below an observability threshold and have simply been pushed above it. [Quantum Biology Society] Proton-coupled electron transfer (PCET), in which an electron and a proton move together and generate radicals, is a core mechanism of biological energy conversion and redox regulation. The BLUF (Blue Light Using FAD) domain, a photoreceptor that senses blue light, is understood to make use of it as well. Within the BLUF family, however, different proteins have given conflicting results, and the argument has persisted. In the cyanobacterial protein PixD, flavin radicals (FAD•⁻ and FADH•) are clearly observed during the photocycle, which established the PCET model. In the closely related protein AppA, from the purple bacterium Rhodobacter sphaeroides, no radical intermediates appeared in the ultrafast infrared spectra at all, and the same was true of BlsA. A study by a US, UK and Hungarian collaboration (corresponding authors Peter Tonge, Stephen Meech and Andras Lukacs), published in the Journal of the American Chemical Society in issue 1 of 2025, tackles that long-standing dilemma head-on. Instead of continuing the argument, the team chose experimental demonstration: deliberately reshaping the protein structure so that AppA switches over to the PixD mechanism. ■ Adding a hydrogen bond: in search of the missing intermediate The decisive difference the team focused on was the number of hydrogen bonds interacting with the C2=O position of the FAD molecule. In PixD two residues form hydrogen bonds there, while AppA has only one. The limits of a single mutation: when H44 in AppA was first replaced with other amino acids, the C2=O vibrational band shifted down from 1650 cm⁻¹ (to 1642 cm⁻¹ in H44N), but the radical marker near 1520 cm⁻¹ still did not appear. Capturing the radical with a double mutation: to mimic the role of R65 in PixD, the team replaced H78 in AppA with arginine, producing the double variant H44N,H78R. Time-resolved infrared (TRIR) spectra showed the C2=O band dropping to 1634 cm⁻¹, confirming that a second hydrogen bond had formed. Notably, the single H78R variant on its own showed no radical intermediate, so both substitutions were needed. Evidence for stepwise PCET: within 1 ps of illumination, new transient absorption bands appeared at 1507 cm⁻¹ and 1521 cm⁻¹, corresponding to the tyrosine cation radical (Tyr•⁺) and the flavin anion radical (FAD•⁻). At 8 ps, FAD* and Tyr•⁺ decayed as a band at 1532 cm⁻¹, assigned to the neutral semiquinone radical FADH•, appeared. Because decay of the 1521 cm⁻¹ band precedes that of the 1532 cm⁻¹ band, the authors take the electron and proton transfer to be stepwise rather than simultaneous. ■ Rewiring the activation pathway while the reverse reaction stays put To establish that the radicals observed were not incidental by-products but lie on the obligatory path to photoactivation, the team replaced the conserved tyrosine Y21 in the double variant with fluoro-tyrosines of higher acidity, which lower the phenol pKa from 9.9 in tyrosine to 8.4 for the monofluoro analogue and 6.4 for the trifluoro one. In the most acidic variant, 2,3,5F₃Y21, the transient at 1532 cm⁻¹ was absent, which the authors read as the increased acidity of Y21 preventing formation of the FADH• state and halting the photocycle at FAD•⁻. The protein vibrational mode at 1618 cm⁻¹, which appears when the final signalling state forms, also disappeared. Together this indicates that the radicals found in the double variant are key intermediates on the way to the light state. More striking still is the rate of recovery to the dark state, the reverse reaction. Despite the forward photoactivation pathway having been thoroughly rebuilt, the recovery rate of the double variant rose only about twofold relative to wild-type AppA. Its sensitivity to fluoro-tyrosine substitution also kept the wild-type AppA pattern rather than the PixD one: a 34-fold acceleration in 3FY21 H44N,H78R, close to the 51-fold seen in 3FY21 wild-type AppA, against only 5.8-fold in the corresponding PixD variant. This suggests that a photoreceptor's path into the light state and its path back out are chemically separable and can be manipulated independently. ■ Significance and open questions The study is a notable achievement in showing that manipulating specific amino acid residues alone can dramatically change how a PCET pathway inside a protein is activated. The authors do not gloss over how complicated the interpretation is. Recent quantum mechanics/molecular mechanics (QM/MM) calculations suggest that the PCET process is not a single clean step but a set of multiple relaxation pathways deeply entangled with the dynamics of the surrounding protein. Accordingly, the authors write that this work is probably not a matter of "creating PCET that wild-type AppA did not have", but rather of the C2=O interactions perturbing protein dynamics and charge-transfer state stabilisation so as to push the population and decay rate of metastable radical intermediates, previously hidden below an observability threshold, above that threshold. Work on another BLUF protein, OaPAC, has meanwhile reported that the forward PCET there proceeds by a concerted mechanism, so the detailed mechanism of the BLUF photocycle remains an open question requiring further study. #QuantumBiology #ProtonCoupledElectronTransfer #PCET #BLUF #Photoreceptor #Flavin #FAD #AppA #PixD #RadicalIntermediates #TimeResolvedInfrared #Fluorotyrosine #HydrogenBondNetwork #Optogenetics #Rhodobacter https://pmc.ncbi.nlm.nih.gov/articles/PMC11726546/
inquantio· Zenodo (CERN European Organi...· 0 citations
Cold-start drug–target affinity (DTA) prediction requires evaluation on genuinely unseen chemical and protein entities. We considered QVGAT-DPI (Quantum Variational Graph Attention Network for Drug–Protein Interaction) as a residual hybrid model combining full molecular graphs, permutation-invariant Breaking of Retro-synthetically Interesting Chemical Substructures (BRICS) fragment queries, ESM-2 and ChemBERTa representations, fingerprints, cross-attention, and two simulated six-qubit residual modules. QVGAT-DPI was compared with specification-based implementation of Q-BAFNet, the closest structured hybrid architecture, under identical local fold manifests. Because of the limitation of the computational resource, the planned ten paired folds were completed only for DAVIS S2 (unseen drugs) and S4 (unseen drugs and targets). For S2, paired differences were + 0.060 CI (bootstrap 95% interval − 0.000 to 0.123; exact p = 0.0996) and − 0.094 RMSE (− 0.145 to − 0.046; p = 0.0098). For S4, differences were + 0.087 CI (0.032–0.143; p = 0.0195) and − 0.135 RMSE (− 0.294 to − 0.005; p = 0.1191). After Holm adjustment, only S2 RMSE remained below 0.05, and no protocol supported both endpoints. The prespecified aggregate S2–S4 conclusion was unevaluable because S3 was incomplete. These results support further controlled evaluation, not system-wide superiority or quantum advantage.
TNG961 induces potent, selective CRBN-dependent degradation of HBS1L. A, Endogenous HBS1L degradation and PELO destabilization measured by Western blot after 24 hours of treatment with TNG961 in the indicated cell lines. The top dose was 10 µmol/L, tested as a 9-point, 3-fold dilution series. Representative images are from N = 2 biological replicates. B, CRBN dependence of HBS1L degradation and PELO destabilization assessed by Western blot after 24 hours of treatment with the indicated concentrations of TNG961 in parental and CRBN knockout (sgCRBN) MIAPACA2 cells. Representative images are from N = 2 biological replicates. C, Neddylation dependence assessed by Western blot after 24 hours of treatment with TNG961 alone or in combination with 100 nmol/L MLN4924 in MIAPACA2 cells. Representative images are from N = 2 biological replicates. D, Degron-loop dependence assessed by Western blot after 24 hours of treatment with the indicated concentrations of TNG961 in parental MIAPACA2 cells and in engineered lines expressing either HBS1L G625N (covering endogenous HBS1L knockout) or GSPT1 G575N (covering endogenous GSPT1 knockout). Representative images are from N = 3 biological replicates. E, Global quantitative proteomics in MM.1s cells after 6 hours of treatment with 500 nmol/L TNG961 or DMSO. Data are shown as relative protein abundance (TNG961/DMSO); HBS1L, PELO, and common IMiD neosubstrates are annotated. F, Western blot analysis of common CRBN neosubstrates after 24 hours of treatment with TNG961 at the indicated concentrations. Representative images are from N = 2 biological replicates.
Hilary E. Nicholson, Douglas A. Whittington, Frank J. Bruzzese et al.· 0 citations
Abstract This sub-paper, "Fractal Compact Manifold Theory Core Mechanics" serves as a concise, standalone distillation of FCMT's foundational elements, emphasizing the atemporal symmetry breaking, field emergence, Lagrangian formulation, projection operator, and foliation map. By focusing solely on these core mechanics—without delving into the full theory's extensions like general relativity recovery, timeless quantum applications, gauge structures, or empirical predictions—it aims to provide a more accessible entry point, reducing the reading commitment from the comprehensive 200-page main document to a targeted exploration of the theory's essence. Fractal Compact Manifold Theory (FCMT) is a unified framework for quantum mechanics, gravity, electromagnetism, and consciousness, built from atemporal symmetry breaking in a multifractal configuration space. The theory begins in a pre-perturbation state: a stable unified field on a compact manifold whose maximum multifractal dimension is *d*_max ≈ 4.01–4.02. That slight excess over four is required both to recover ordinary four-dimensional physics in the infrared and to supply the scale-dependent measure that keeps loop integrals finite. The field is governed by the quadratic potential V(φ_unified) = ½ m² φ_unified² with m² > 0. An atemporal quantum fluctuation then triggers symmetry breaking and differentiates the unified field into four fields: the informational field φ_info, which encodes self-similar structure; the consciousness field ψ_c, which carries the primordial awareness substrate; the entanglement field φ_e, which supplies non-local correlations; and the Higgs field φ_H, which generates mass. There is one variational structure. It is the action. Inside that action already sits the Shannon term *S*_info = −∑ *p*_ij log *p*_ij. Consciousness-Modulated Informational Entropy Minimization is that term, not a second principle standing beside the action. Isolating ψ_c as the field that weights the Shannon term produces the control Hamiltonian H = λ · f(p, ψ_c) − *S*_info[p]. Pontryagin’s necessary conditions are the Euler–Lagrange equation of the same action written so that the role of the control is explicit. When multifractal spectra rather than a single Shannon functional are required, the same construction applies to the Rényi family *H*_q. The principle does not change. A renormalized projection operator Proj_d^R then foliates the atemporal configuration space into ordinary four-dimensional Lorentzian hypersurfaces. Relational time emerges from the renormalization-group flow that accompanies the projection. Gravity arises as the geometric response to ψ_c-orchestrated clustering of the stress-energy; in the low-consciousness limit the multifractal corrections vanish and the classical Einstein equations are recovered exactly. The same residual symmetry of the entanglement field that produces its transverse-traceless two-point function also yields a massless vector mode whose projected dynamics reproduce Maxwell’s equations, so electromagnetism appears as a controlled consequence of the identical breaking that generates φ_e. In the high-energy sector the multifractal measure and the running intermittency γ(k) generate a Gaussian hard form factor that renders the spectrum finite. There is no infinite linear Regge trajectory. The effective cutoff Λ_R is restricted by three matching conditions, all built from functions already present in the architecture, to a window of roughly 3–30 TeV: the form factor is anchored to the same intermediate dimensionality already used for the consciousness-field length; suppression is required once γ(k) falls below 10⁻³; and suppression is required once the running projection kernel has narrowed enough that non-local comparison ceases to be effective. Those three conditions share the architecture. They are not three independent theories of the cutoff. The projection framework also generates the principal structural features of the Standard Model — three fermion generations from discrete scale bands, hierarchical Yukawa couplings from the running kernel, and the gauge group from residual transformations of the entanglement field — rather than inserting them by hand. Three faces of the same Lagrangian, plus one empirical lock, return one characteristic length for the consciousness field. The effective mass of ψ_c on the infrared slice *d*_i ≈ 3.3, the running width of the projection kernel on that slice, and the feedback coupling κ_c / *v*_IR² ≈ 0.06 share those two anchors and give the Compton length λ_ψc ≈ 1.5 μm (window 1.1–1.9 μm). A fourth contact is empirical rather than calculational: the optical and near-infrared member of the microtubule resonance hierarchy, and the Fröhlich condensate it supports, already sit at that length. That is a lock, not a fourth independent derivation. The length coincides with the characteristic size of large protein complexes, cytoskeletal bundles, and dendritic spines — the regime in which living systems must maintain order against thermal noise. FCMT therefore treats consciousness as a fundamental field that participates in the generation of spacetime, the emergence of electromagnetism, the finiteness of the high-energy spectrum, and the selection of low-informational-entropy configurations. Ordinary quantum phenomena and classical gravity appear as controlled projections of one atemporal stationarity condition. The framework yields sharp, testable signatures: a Higgs-consciousness Yukawa coupling |*y*_h| = 0.0153 ± 0.0022, consistent with public LHC limits as of November 2025; essentially null running of the CMB spectral index α_s ≈ 0; neural coherence times of order τ_d = 10⁻⁴ s; and a size-scanned search for enhanced order-maintenance or coherence in the window 0.5–3 μm, with the predicted peak at 1.5 μm. Consciousness is not emergent. It is the field that folds the universe.
Zachary Hosack· Zenodo (CERN European Organi...· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Many biological processes rely on mechanical forces, with protein molecules acting as key mediators. Understanding how proteins respond to mechanical stress is essential for conditions including cardiomyopathy and muscular dystrophy. Natural proteins such as dystrophin and utrophin are composed of heterogeneous folding domains with distinct mechanical properties; deciphering domain-level behavior provides insights into disease mechanisms and informs therapeutic strategies. Single-molecule force spectroscopy (SMFS) enables probing the mechanical properties of entire proteins, yet current approaches struggle to identify heterogeneous folding domains, particularly without prior knowledge. Here, we present the first automated framework to identify heterogeneous folding domains in SMFS data, applying both existing clustering methods and a novel physics-aware deep clustering architecture, LatentUnfold. LatentUnfold learns complementary latent representations from force magnitude and the force-extension physical relationship through dual autoencoders, jointly optimized for clustering assignments. We apply our framework to experimental SMFS data collected from a synthetic two-domain protein (ddFLN4-Titin I27) as well as natural protein constructs of dystrophin and utrophin, with Monte Carlo simulated datasets serving as controlled validation. For the synthetic protein, we recover mechanical properties consistent with previously reported values for each domain. For the natural proteins, we uncover two mechanically distinct domain populations - corresponding to the N-terminal domain and spectrin-like repeats - with differences in both unfolding force and contour length increase, and reveal different unfolding order between them for the first time. This work enables domain-level biological inference, overcoming prior limitations that relied on averaging and overlooked heterogeneity, thus advancing the understanding of mechanical behavior in protein unfolding.
Cailong Hua, Yiyuan Zhang, Vinitendra Singh et al.· bioRxiv (Cold Spring Harbor...· 0 citations
Abstract Saccharomyces cerevisiae yeast is a widely used recombinant protein production host. Recombinant protein expression requires adaptation of the host cell proteome to accommodate the increased biosynthetic and folding demands. However, this underlying proteomic changes remain poorly understood. In this study, we quantified the proteome of a laboratory S. cerevisiae strain over four days during batch cultivation for recombinant laccase production to characterize the resulting proteomics remodeling. Whereas a substantial portion of the proteome changed in response to nutrient depletion during batch growth, only a smaller subset of proteins was affected by laccase expression. By comparing yeast strains of different origins and laccase production capacities, we found that each strain displayed a distinct response to heterologous expression, regardless of the origin of the laccase. For example, the chaperones Hsp26 and Kar2 were specifically elevated in a whey-derived strain upon laccase expression. Nonetheless, the higher capacity to produce active recombinant laccase in some strains appears to be associated primarily with small groups of proteins that are constitutively expressed at different levels. These results indicate that strains of different origins each provide a unique cellular milieu that, in some cases, is more favorable for the expression of a given recombinant protein. This study provides the first insights into the dynamic proteome remodeling that occurs during recombinant laccase expression and highlights the potential of exploiting naturally occurring yeast diversity, rather than relying solely on strain engineering, to improve recombinant protein yields. Key points • Proteomes of S. cerevisiae strains during recombinant laccase expression determined • Ribosomal and metabolic protein levels change during recombinant expression • Unique cellular milieu, rather than proteome shifts, is linked to higher yields
Ryan Wong, Sahil Chandhok, Elizabeth Hui et al.· Applied Microbiology and Bio...· 0 citations
AIM
To establish a multicellular three-dimensional (3D) pancreatic ductal adenocarcinoma (PDAC) spheroid model in a pancreas-derived extracellular matrix environment and determine how cancer-associated fibroblasts (CAFs) and endothelial CD93 affect spheroid architecture and endothelial organization.
METHODS
Porcine pancreatic decellularized extracellular matrix (dECM) was processed into a hydrogel and characterized. PANC-1 cells and mCherry-labelled human umbilical vein endothelial cells (HUVECs) were embedded with or without patient-derived CAFs. Microscopy, immunofluorescence staining, and image analysis were used to evaluate spheroid formation, cellular composition, HUVEC distribution, CD31/CD44 double-positive staining, and matrix-remodeling-associated proteins. CD93-overexpressing HUVECs were used to examine the effects of endothelial CD93.
RESULTS
The dECM hydrogel retained major matrix components and supported PDAC spheroid formation. HUVECs exhibited elongated and interconnected network-like distributions around and within PANC-1 aggregates. Compared with PANC-1/HUVEC co-cultures, CAF-containing tri-cultures formed more spheroids and showed greater mean and total spheroid volumes, higher cell density, and increased estimated PANC-1 and HUVEC abundance. The CD31/CD44 double-positive signal was 1.46-fold higher in tri-cultures. Endothelial CD93 overexpression increased spheroid number and cell density and produced a more intermixed cellular distribution, but reduced the number of large spheroids, mean spheroid volume, total spheroid burden, estimated PANC-1 abundance, and MMP2 and MMP9 immunofluorescence signals. The CD31/CD44 double-positive signal increased by 1.3-fold without a significant change in estimated HUVEC abundance.
CONCLUSION
This model enables analysis of tumor-endothelial-stromal interactions in a pancreas-derived dECM environment. CAFs and endothelial CD93 differentially altered spheroid architecture and the CD44-associated endothelial phenotype. The formation of continuous and functional microvascular structures remains to be established.
Han-Rong Li, Yi-Zhou Zhang, Chuhan Ma et al.· Translational Research: The...· 0 citations
Chronic Myeloid Leukemia (CML) is a well-defined hematological malignancy driven principally by the BCR::ABL1 fusion oncogene and aberrant tyrosine kinase signaling. Computational methods to predict leukemia suffer from the limitation of using single-modality data or black-box models, which cannot adequately incorporate complementary molecular evidence and deliver interpretable biomarker support. In this study, we introduce an explainable multi-omics fusion approach that combines protein language model embeddings, gene-expression data, mutation-level data, and pathway-informed representations to classify CML-positive and control samples and identify candidate biomarkers. The pre-trained protein language models encode context from the sequence, and dense encoders represent modalities based on transcriptomic, mutational, and pathway information. They are fused adaptively for classification using an attention module, and biomarkers are ranked by SHAP, Integrated Gradients, and attention attribution to support biological interpretation. The proposed framework achieved an accuracy of (98.14%), precision (98.0%), recall (98.4%), F1-score (98.14%), ROC–AUC (0.991), and PR–AUC (0.987), outperforming classical machine learning, deep learning, single-modality, and conventional fusion baselines. Reliability analysis yielded a Brier score of 0.031, and an Expected Calibration Error of 0.024, and stable performance on five-fold cross-validation (97.9 ± 0.4% accuracy and 0.988 ± 0.003 ROC–AUC). Biologically relevant biomarkers identified by explainability analysis included the BCR::ABL1 fusion gene, ABL1 kinase-domain mutations, BCL2, HSP90, RUNX1, ASXL1, PARP1 and RB1, and key pathways were identified: JAK–STAT, PI3K–AKT, RAS–MAPK, apoptosis and DNA repair. The findings indicate that the proposed model achieved improved predictive performance under the evaluated experimental conditions and provided interpretable identification of candidate biomarkers in CML and control samples.
Atiq Ur Rehman, Ali Sayyed, Muhammad Ismail Mohmand et al.· Scientific Reports· 0 citations
This repository accompanies a computational analysis decomposing protein variant fitness into folding-stability and function components across 151 deep mutational scanning (DMS) assays from ProteinGym. For each assay, the agreement of every predictor with measured fitness was quantified, structure- and sequence-based stability predictors (RaSP and DDGun) were benchmarked against a graded ladder of trivial baselines (BLOSUM62, hydrophobicity change, side-chain volume change, and relative solvent accessibility) and against five function-aware models (ESM2, ESM-1v, ESM-C, ESM3, and EVE), and partial-correlation and cross-validated variance-partition analyses were used to separate signal shared with, versus independent of, folding stability. A clinical arm evaluated pathogenic-versus-benign separation on expert-annotated ClinVar variants in disease genes, using predictors that had not been trained on clinical labels. Contents: all analysis scripts (numbered in execution order), derived per-assay result tables from which every figure and reported value is computed, the six manuscript figures, and the software-environment specifications for both the data-preparation and analysis pipelines. Raw input data (ProteinGym assays and zero-shot scores, the Tsuboyama et al. 2023 mega-scale stability dataset, and the ClinVar variant summary) are not included; they are available from their original sources, with MD5 checksums listed in the manuscript Methods, and are reproduced by the deposited scripts. Computation was performed on the PARAM Shakti supercomputing facility at the Indian Institute of Technology Madras.
Prion diseases are transmissible, neurodegenerative diseases caused by misfolded, protease-resistant, and infectious aggregates of the mammalian prion protein (PrPSc) that replicate by converting properly folded prion protein (PrPC) into PrPSc. Spongiform change and cellular loss in the brain are hallmarks of prion disease, but our understanding of how prions alter cellular fitness remains incomplete. Here we characterized changes in mitochondrial redox state and respiration in neural cells following uptake of two different PrPSc strains, 22L and 87V. Only 22L PrPSc induced changes in cellular respiration and mitochondrial redox state, even in cells that did not produce PrPC. These effects were disrupted by detergent and dependent upon endo-lysosomal acidification, suggesting that both PrPSc membrane association and lysosomal degradation are involved. Interestingly, cells chronically infected with 22L appeared to adapt to infection, showing no signs of mitochondrial dysfunction, but were more susceptible to oxidative stress even though mitochondrial respiration was normal. Thus, during initial prion infection, PrPSc drives mitochondrial dysfunction in a manner that is both strain dependent and independent of PrPC expression, while persistent prion infection increases mitochondrial sensitivity to cellular stress.
Daniel Shoup, L. Varner, B. Race et al.· Journal of Cell Science· 0 citations
White and brown adipocytes show therapeutic potential, yet their exosomes mediated cell-specific effects remain poorly understood. This study aimed to assess the impact of conditioned media (CM) or exosomes derived from brown and white adipocyte-like cells (BALCs and WALCs, respectively) on four human cell types, including ovarian granulosa cells (OGCs), MCF-7 breast cancer cells, human adipose-derived stem cells (hADSCs), and human umbilical cord mesenchymal stem cells (hUCMSCs). Isolated hADSCs were differentiated into BALCs and WALCs using specific induction media. Corresponding exosomes were isolated from BALCs or WALCs derived CM and characterized. Human cell types—representing stem, somatic, and cancer cells—were treated with CM or exosomes. Cellular responses and lipidomic profiling were evaluated using molecular biology techniques and mass spectrometry, respectively. BALCs-derived exosomes significantly increased estradiol secretion from ovarian granulosa cells (p < 0.05). In MCF-7 cells, CM had a stronger pro-apoptotic effect than exosomes (>2-fold). In hADSCs, both CM and exosomes promoted osteogenic differentiation (2.3-fold) and reduced stemness markers proteins Oct4 and Sox2 (p < 0.05). When treated with BALCs-derived exosomes, hADSCs displayed the highest number of differentially expressed lipids, primarily upregulated hits. Pathway analysis indicated that BALCs-derived exosomes promoted lipid storage and membrane remodeling in hADSCs, enhanced energy metabolism in granulosa cells, reduced fatty acid oxidation in MCF-7 cells, and decreased lipid storage in hUCMSCs. Our results highlight a potential modulatory role of adipocyte-derived exosomes in intercellular communication. The findings of this study may be applicable in regenerative medicine, metabolic regulation and cancer therapeutic approaches.
Zeinab Ghesmati, Mohsen Rashid, S. Fayezi et al.· PLoS ONE· 0 citations
The widespread emergence of antibiotic resistance necessitates the development of novel agents with unique mechanisms of action. Obafluorin (OB), a natural β-lactone antibiotic, is a covalent inhibitor of threonyl-tRNA synthetase (ThrRS), but the high conservation of the active site between prokaryote and eukaryote ThrRSs results in minimal selectivity, hindering the therapeutic potential of OB. Here, we report a structure dynamics-based design strategy that transforms OB into a selective antibacterial agent. OB inhibits human and bacterial ThrRSs with nearly equal potency due to identical binding modes. The nitrophenyl moiety of OB is proposed as a 'kinetic sensor' that discriminates between sensitive and resistant ThrRS paralogs. Guided by this insight, we designed a series of OB analogs through rational modification of this moiety. Among them, OB-D4 bearing a para-methoxyphenyl group in place of the nitrophenyl group, exhibited a 241-fold selectivity for bacterial over human ThrRS, along with a markedly improved safety profile with minimal cytotoxicity. In a murine skin infection model, OB-D4 effectively eradicated pathogens, resolving inflammation, and promoting wound healing. Together, this work establishes a 'kinetic sensor' strategy for achieving species selectivity, turning a fundamental challenge in drug discovery-high active-site conservation-into an exploitable opportunity based on dynamic differences.
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.
