Predicting Serotonin Detection with DNA-Carbon Nanotube Sensors across Multiple Spectral Wavelengths.

Owing to the value of DNA-wrapped single-walled carbon nanotube (SWNT)-based sensors for chemically specific imaging in biology, we explore machine learning (ML) predictions DNA-SWNT serotonin sensor responsivity as a function of DNA sequence based on the whole SWNT fluorescence spectra. Our analysis reveals the crucial role of DNA sequence in the binding modes of DNA-SWNTs to serotonin, with a smaller influence of SWNT chirality. Regression ML models trained on existing data sets predict the change in the fluorescence emission in response to serotonin, ΔF/F, at over a hundred wavelengths for new DNA-SWNT conjugates, successfully identifying some high- and low-response DNA sequences. Despite successful predictions, we also show that the finite size of the training data set leads to limitations on prediction accuracy. Nevertheless, incorporating entire spectra into ML models enhances prediction robustness and facilitates the discovery of novel DNA-SWNT sensors. Our approaches show promise for identifying new chemical systems with specific sensing response characteristics, marking a valuable advancement in DNA-based system discovery.

Directed Evolution of Near-Infrared Serotonin Nanosensors with Machine Learning-Based Screening.

In this study, we employed a novel approach to improve the serotonin-responsive ssDNA-wrapped single-walled carbon nanotube (ssDNA-SWCNT) nanosensors, combining directed evolution and machine learning-based prediction. Our iterative optimization process is aimed at the sensitivity and selectivity of ssDNA-SWCNT nanosensors. In the three rounds for higher serotonin sensitivity, we substantially improved sensitivity, achieving a remarkable 2.5-fold enhancement in fluorescence response compared to the original sequence. Following this, we directed our efforts towards selectivity for serotonin over dopamine in the two rounds. Despite the structural similarity between these neurotransmitters, we achieved a 1.6-fold increase in selectivity. This innovative methodology, offering high-throughput screening of mutated sequences, marks a significant advancement in biosensor development. The top-performing nanosensors, N2-1 (sensitivity) and L1-14 (selectivity) present promising reference sequences for future studies involving serotonin detection.

Genetically encoded discovery of perfluoroaryl macrocycles that bind to albumin and exhibit extended circulation in vivo.

Peptide-based therapeutics have gained attention as promising therapeutic modalities, however, their prevalent drawback is poor circulation half-life in vivo. In this paper, we report the selection of albumin-binding macrocyclic peptides from genetically encoded libraries of peptides modified by perfluoroaryl-cysteine SNAr chemistry, with decafluoro-diphenylsulfone (DFS). Testing of the binding of the selected peptides to albumin identified SICRFFC as the lead sequence. We replaced DFS with isosteric pentafluorophenyl sulfide (PFS) and the PFS-SICRFFCGG exhibited KD = 4-6 µM towards human serum albumin. When injected in mice, the concentration of the PFS-SICRFFCGG in plasma was indistinguishable from the reference peptide, SA-21. More importantly, a conjugate of PFS-SICRFFCGG and peptide apelin-17 analogue (N3-PEG6-NMe17A2) showed retention in circulation similar to SA-21; in contrast, apelin-17 analogue was cleared from the circulation after 2 min. The PFS-SICRFFC is the smallest known peptide macrocycle with a significant affinity for human albumin and substantial in vivo circulation half-life. It is a productive starting point for future development of compact macrocycles with extended half-life in vivo.

Characterizing the Interactions of Cell-Membrane-Disrupting Peptides with Lipid-Functionalized Single-Walled Carbon Nanotubes.

Lipid-functionalized single-walled carbon nanotubes (SWNTs) have garnered significant interest for their potential use in a wide range of biomedical applications. In this work, we used molecular dynamics simulations to study the equilibrium properties of SWNTs surrounded by the phosphatidylcholine (POPC) corona phase and their interactions with three cell membrane disruptor peptides: colistin, TAT peptide, and crotamine-derived peptide. Our results show that SWNTs favor asymmetrical positioning within the POPC corona, so that one side of the SWNT, covered by the thinnest part of the corona, comes in contact with charged and polar functional groups of POPC and water. We also observed that colistin and TAT insert deeply into the POPC corona, while crotamine-derived peptide only adsorbs to the corona surface. In separate simulations, we show that three examined peptides exhibit similar insertion and adsorption behaviors when interacting with POPC bilayers, confirming that peptide-induced perturbations to POPC in conjugates and bilayers are similar in nature and magnitude. Furthermore, we observed correlations between the peptide-induced structural perturbations and the near-infrared emission of the lipid-functionalized SWNTs, which suggest that the optical signal of the conjugates transduces the morphological changes in the lipid corona. Overall, our findings indicate that lipid-functionalized SWNTs could serve as simplified cell membrane model systems for prescreening of new antimicrobial compounds that disrupt cell membranes.

BinderSpace: A package for sequence space analyses for datasets of affinity-selected oligonucleotides and peptide-based molecules.

Discovery of target-binding molecules, such as aptamers and peptides, is usually performed with the use of high-throughput experimental screening methods. These methods typically generate large datasets of sequences of target-binding molecules, which can be enriched with high affinity binders. However, the identification of the highest affinity binders from these large datasets often requires additional low-throughput experiments or other approaches. Bioinformatics-based analyses could be helpful to better understand these large datasets and identify the parts of the sequence space enriched with high affinity binders. BinderSpace is an open-source Python package that performs motif analysis, sequence space visualization, clustering analyses, and sequence extraction from clusters of interest. The motif analysis, resulting in text-based and visual output of motifs, can also provide heat maps of previously measured user-defined functional properties for all the motif-containing molecules. Users can also run principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE) analyses on whole datasets and on motif-related subsets of the data. Functionally important sequences can also be highlighted in the resulting PCA and t-SNE maps. If points (sequences) in two-dimensional maps in PCA or t-SNE space form clusters, users can perform clustering analyses on their data, and extract sequences from clusters of interest. We demonstrate the use of BinderSpace on a dataset of oligonucleotides binding to single-wall carbon nanotubes in the presence and absence of a bioanalyte, and on a dataset of cyclic peptidomimetics binding to bovine carbonic anhydrase protein. BinderSpace is openly accessible to the public via the GitHub website: https://github.com/vukoviclab/BinderSpace.

Characterizing the Interactions of Cell Membrane-Disrupting Peptides with Lipid-Functionalized Single-Walled Carbon Nanotube Systems for Antimicrobial Screening.

Lipid-functionalized single-walled carbon nanotubes (SWNTs) have garnered significant interest for their potential use in a wide range of biomedical applications. In this work, we used molecular dynamics simulations to study the equilibrium properties of SWNTs surrounded by the phosphatidylcholine (POPC) corona phase, and their interactions with three cell membrane disruptor peptides: colistin, TAT peptide, and crotamine-derived peptide. Our results show that SWNTs favor asymmetrical positioning within the POPC corona, so that one side of the SWNT, covered by the thinnest part of the corona, comes in contact with charged and polar functional groups of POPC and water. We also observed that colistin and TAT insert deeply into POPC corona, while crotamine-derived peptide only adsorbs to the corona surface. Compared to crotamine-derived peptide, colistin and TAT also induce larger perturbations in the thinnest region of the corona, by allowing more water molecules to directly contact the SWNT surface. In separate simulations, we show that three examined peptides exhibit similar insertion and adsorption behaviors when interacting with POPC bilayers, confirming that peptide-induced perturbations to POPC in conjugates and bilayers are similar in nature and magnitude. Furthermore, we observed correlations between the peptide-induced structural perturbations and the near-infrared emission of the lipid-functionalized SWNTs, which suggest that the optical signal of the conjugates transduces the morphological changes in the lipid corona. Overall, our findings indicate that lipid-functionalized SWNTs could serve as simplified cell membrane model systems for pre-screening of new antimicrobial compounds that disrupt cell membranes.

Discovery of DNA-Carbon Nanotube Sensors for Serotonin with Machine Learning and Near-infrared Fluorescence Spectroscopy.

DNA-wrapped single walled carbon nanotube (SWNT) conjugates have distinct optical properties leading to their use in biosensing and imaging applications. A critical limitation in the development of DNA-SWNT sensors is the current inability to predict unique DNA sequences that confer a strong analyte-specific optical response to these sensors. Here, near-infrared (nIR) fluorescence response data sets for ∼100 DNA-SWNT conjugates, narrowed down by a selective evolution protocol starting from a pool of ∼1010 unique DNA-SWNT candidates, are used to train machine learning (ML) models to predict DNA sequences with strong optical response to neurotransmitter serotonin. First, classifier models based on convolutional neural networks (CNN) are trained on sequence features to classify DNA ligands as either high response or low response to serotonin. Second, support vector machine (SVM) regression models are trained to predict relative optical response values for DNA sequences. Finally, we demonstrate with validation experiments that integrating the predictions of ensembles of the highest quality neural network classifiers (convolutional or artificial) and SVM regression models leads to the best predictions of both high and low response sequences. With our ML approaches, we discovered five DNA-SWNT sensors with higher fluorescence intensity response to serotonin than obtained previously. Overall, the explored ML approaches, shown to predict useful DNA sequences, can be used for discovery of DNA-based sensors and nanobiotechnologies.

Computational Modeling of the Virucidal Inhibition Mechanism for Broad-Spectrum Antiviral Nanoparticles and HPV16 Capsid Segments.

Solid core nanoparticles (NPs) coated with sulfonated ligands that mimic heparan sulfate proteoglycans (HSPGs) can exhibit virucidal activity against many viruses that utilize HSPG interactions with host cells for the initial stages of infection. How the interactions of these NPs with large capsid segments of HSPG-interacting viruses lead to their virucidal activity has been unclear. Here, we describe the interactions between sulfonated NPs and segments of the human papilloma virus type 16 (HPV16) capsids using atomistic molecular dynamics simulations. The simulations demonstrate that the NPs primarily bind at the interfaces of two HPV16 capsid proteins. After equilibration, the distances and angles between capsid proteins in the capsid segments are larger for the systems in which the NPs bind at the interfaces of capsid proteins. Over time, NP binding can lead to breaking of contacts between two neighboring proteins. The revealed mechanism of NPs targeting the interfaces between pairs of capsid proteins can be utilized for designing new generations of virucidal materials and contribute to the development of new broad-spectrum non-toxic virucidal materials.

Genetically Encoded Fragment-Based Discovery from Phage-Displayed Macrocyclic Libraries with Genetically Encoded Unnatural Pharmacophores.

Genetically encoded macrocyclic peptide libraries with unnatural pharmacophores are valuable sources for the discovery of ligands for many targets of interest. Traditionally, generation of such libraries employs "early stage" incorporation of unnatural building blocks into the chemically or translationally produced macrocycles. Here, we describe a divergent late-stage approach to such libraries starting from readily available starting material: genetically encoded libraries of peptides. A diketone linchpin 1,5-dichloropentane-2,4-dione converts peptide libraries displayed on phage to 1,3-diketone bearing macrocyclic peptides (DKMP): shelf-stable precursors for Knorr pyrazole synthesis. Ligation of diverse hydrazine derivatives onto DKMP libraries displayed on phage that carries silent DNA-barcodes yields macrocyclic libraries in which the amino acid sequence and the pharmacophore are encoded by DNA. Selection of this library against carbonic anhydrase enriched macrocycles with benzenesulfonamide pharmacophore and nanomolar Kd. The methodology described in this manuscript can graft diverse pharmacophores into many existing genetically encoded phage libraries and significantly increase the value of such libraries in molecular discoveries.

Adsorption of poly(ethylene succinate) chain onto graphene nanosheets: A molecular simulation.

Understanding the interaction between single polymer chain and graphene nanosheets at local and global length scales is essential for it underlies the mesoscopic properties of polymer nanocomposites. A computational attempt was then performed using atomistic molecular dynamics simulation to gain physical insights into behavior of a model aliphatic polyester, poly(ethylene succinate), single chain near graphene nanosheets, where the effects of the polymer chain length, graphene functionalization, and temperature on conformational properties of the polymer were studied comparatively. Graphene functionalization was carried out through extending the parameters set of an all-atom force field. The results showed a significant conformational transition of the polymer chain from three-dimensional statistical coil, in initial state, to two-dimensional fold, in final state, during adsorption on graphene. The conformational order, overall shape, end-to-end separation statistics, and mobility of the polymer chain were found to be influenced by the graphene functionalization, temperature, and polymer chain length. Furthermore, the polymer chain dynamics mode during adsorption on graphene was observed to transit from normal diffusive to slow subdiffusive mode. The findings from this computational study could shed light on the physics of the early stages of aliphatic polyester chain organization induced by graphene.