Precious2GPT: the combination of multiomics pretrained transformer and conditional diffusion for artificial multi-omics multi-species multi-tissue sample generation.

Synthetic data generation in omics mimics real-world biological data, providing alternatives for training and evaluation of genomic analysis tools, controlling differential expression, and exploring data architecture. We previously developed Precious1GPT, a multimodal transformer trained on transcriptomic and methylation data, along with metadata, for predicting biological age and identifying dual-purpose therapeutic targets potentially implicated in aging and age-associated diseases. In this study, we introduce Precious2GPT, a multimodal architecture that integrates Conditional Diffusion (CDiffusion) and decoder-only Multi-omics Pretrained Transformer (MoPT) models trained on gene expression and DNA methylation data. Precious2GPT excels in synthetic data generation, outperforming Conditional Generative Adversarial Networks (CGANs), CDiffusion, and MoPT. We demonstrate that Precious2GPT is capable of generating representative synthetic data that captures tissue- and age-specific information from real transcriptomics and methylomics data. Notably, Precious2GPT surpasses other models in age prediction accuracy using the generated data, and it can generate data beyond 120 years of age. Furthermore, we showcase the potential of using this model in identifying gene signatures and potential therapeutic targets in a colorectal cancer case study.

Non-human peptides revealed in blood reflect the composition of intestinal microbiota.

BACKGROUND: The previously underestimated effects of commensal gut microbiota on the human body are increasingly being investigated using omics. The discovery of active molecules of interaction between the microbiota and the host may be an important step towards elucidating the mechanisms of symbiosis. RESULTS: Here, we show that in the bloodstream of healthy people, there are over 900 peptides that are fragments of proteins from microorganisms which naturally inhabit human biotopes, including the intestinal microbiota. Absolute quantitation by multiple reaction monitoring has confirmed the presence of bacterial peptides in the blood plasma and serum in the range of approximately 0.1 nM to 1 µM. The abundance of microbiota peptides reaches its maximum about 5 h after a meal. Most of the peptides correlate with the bacterial composition of the small intestine and are likely obtained by hydrolysis of membrane proteins with trypsin, chymotrypsin and pepsin - the main proteases of the gastrointestinal tract. The peptides have physicochemical properties that likely allow them to selectively pass the intestinal mucosal barrier and resist fibrinolysis. CONCLUSIONS: The proposed approach to the identification of microbiota peptides in the blood, after additional validation, may be useful for determining the microbiota composition of hard-to-reach intestinal areas and monitoring the permeability of the intestinal mucosal barrier.

Biomedical generative pre-trained based transformer language model for age-related disease target discovery.

Target discovery is crucial for the development of innovative therapeutics and diagnostics. However, current approaches often face limitations in efficiency, specificity, and scalability, necessitating the exploration of novel strategies for identifying and validating disease-relevant targets. Advances in natural language processing have provided new avenues for predicting potential therapeutic targets for various diseases. Here, we present a novel approach for predicting therapeutic targets using a large language model (LLM). We trained a domain-specific BioGPT model on a large corpus of biomedical literature consisting of grant text and developed a pipeline for generating target prediction. Our study demonstrates that pre-training of the LLM model with task-specific texts improves its performance. Applying the developed pipeline, we retrieved prospective aging and age-related disease targets and showed that these proteins are in correspondence with the database data. Moreover, we propose CCR5 and PTH as potential novel dual-purpose anti-aging and disease targets which were not previously identified as age-related but were highly ranked in our approach. Overall, our work highlights the high potential of transformer models in novel target prediction and provides a roadmap for future integration of AI approaches for addressing the intricate challenges presented in the biomedical field.

Precious1GPT: multimodal transformer-based transfer learning for aging clock development and feature importance analysis for aging and age-related disease target discovery.

Aging is a complex and multifactorial process that increases the risk of various age-related diseases and there are many aging clocks that can accurately predict chronological age, mortality, and health status. These clocks are disconnected and are rarely fit for therapeutic target discovery. In this study, we propose a novel approach to multimodal aging clock we call Precious1GPT utilizing methylation and transcriptomic data for interpretable age prediction and target discovery developed using a transformer-based model and transfer learning for case-control classification. While the accuracy of the multimodal transformer is lower within each individual data type compared to the state of art specialized aging clocks based on methylation or transcriptomic data separately it may have higher practical utility for target discovery. This method provides the ability to discover novel therapeutic targets that hypothetically may be able to reverse or accelerate biological age providing a pathway for therapeutic drug discovery and validation using the aging clock. In addition, we provide a list of promising targets annotated using the PandaOmics industrial target discovery platform.

Diplocloster agilis gen. nov., sp. nov. and Diplocloster modestus sp. nov., two novel anaerobic fermentative members of Lachnospiraceae isolated from human faeces.

Three novel strains of Gram-stain-negative, obligately anaerobic, spore-forming straight or slightly curved rods with pointed ends occurring singly or in pairs were isolated from the faeces of healthy human children. The strains were characterized by mesophilic fermentative metabolism and production of acetate, ethanol and H2 as the end metabolic products. Strains ASD3451 and ASD5720T were motile, fermented lactose and raffinose, and weakly fermented maltose. Strain ASD4241T was non-motile and did not ferment the carbohydrates listed above but fermented starch. Strains ASD3451 and ASD5720T shared average nucleotide identity higher than 98.5 % with each other, while ASD4241T had only 88.5-89 % identity to them. Based on phylogenetic and chemotaxonomic analyses, we propose Diplocloster agilis gen. nov., sp. nov. (ASD5720T=JCM 34353T=VKM B-3497T) and Diplocloster modestus sp. nov. (ASD4241T=JCM 34351T=VKM B-3498T) within the family Lachnospiraceae.

All-d-Enantiomeric Peptide D3 Designed for Alzheimer's Disease Treatment Dynamically Interacts with Membrane-Bound Amyloid-ß Precursors.

Alzheimer's disease (AD) is a severe neurodegenerative pathology with no effective treatment known. Toxic amyloid-ß peptide (Aß) oligomers play a crucial role in AD pathogenesis. All-d-Enantiomeric peptide D3 and its derivatives were developed to disassemble and destroy cytotoxic Aß aggregates. One of the D3-like compounds is approaching phase II clinical trials; however, high-resolution details of its disease-preventing or pharmacological actions are not completely clear. We demonstrate that peptide D3 stabilizing Aß monomer dynamically interacts with the extracellular juxtamembrane region of a membrane-bound fragment of an amyloid precursor protein containing the Aß sequence. MD simulations based on NMR measurement results suggest that D3 targets the amyloidogenic region, not compromising its α-helicity and preventing intermolecular hydrogen bonding, thus creating prerequisites for inhibition of early steps of Aß conversion into ß-conformation and its toxic oligomerization. An enhanced understanding of the D3 action molecular mechanism facilitates development of effective AD treatment and prevention strategies.

Structural Studies Providing Insights into Production and Conformational Behavior of Amyloid-ß Peptide Associated with Alzheimer's Disease Development.

Alzheimer's disease is the most common type of neurodegenerative disease in the world. Genetic evidence strongly suggests that aberrant generation, aggregation, and/or clearance of neurotoxic amyloid-ß peptides (Aß) triggers the disease. Aß accumulates at the points of contact of neurons in ordered cords and fibrils, forming the so-called senile plaques. Aß isoforms of different lengths are found in healthy human brains regardless of age and appear to play a role in signaling pathways in the brain and to have neuroprotective properties at low concentrations. In recent years, different substances have been developed targeting Aß production, aggregation, interaction with other molecules, and clearance, including peptide-based drugs. Aß is a product of sequential cleavage of the membrane glycoprotein APP (amyloid precursor protein) by ß- and γ-secretases. A number of familial mutations causing an early onset of the disease have been identified in the APP, especially in its transmembrane domain. The mutations are reported to influence the production, oligomerization, and conformational behavior of Aß peptides. This review highlights the results of structural studies of the main proteins involved in Alzheimer's disease pathogenesis and the molecular mechanisms by which perspective therapeutic substances can affect Aß production and nucleation.

Dimeric states of transmembrane domains of insulin and IGF-1R receptors: Structures and possible role in activation.

Despite the biological significance of insulin signaling, the molecular mechanisms of activation of the insulin receptor (IR) and other proteins from its family remain elusive. Current hypothesis on signal transduction suggests ligand-triggered structural changes in the extracellular domain followed by transmembrane (TM) domains closure and dimerization leading to trans-autophosphorylation and kinase activity in intracellular segments of the receptor. Using NMR spectroscopy, we detected dimerization of isolated TM segments of IR in different membrane-mimicking environments and observed multiple signals of NH groups of protein backbone possibly corresponding to several dimer conformations. Taking available experimental data as constraints, several atomistic models of dimeric TM domains of IR and insulin-like growth factor 1 (IGF-1R) receptors were elaborated. Molecular dynamics simulations of IR ectodomain revealed noticeable collective movements potentially responsible for closure of the C-termini of FnIII-3 domains and spatial approaching of TM helices upon insulin-induced receptor activation. In addition, we demonstrated that the intracellular part of the receptor does not impose restrictions on the positioning of TM helices in the membrane. Finally, we used two independent structure prediction methods to generate a series of dimer conformations followed by their cluster analysis and dimerization free energy estimation to select the best dimer models. Biological relevance of the later was further tested via comparison of the hydrophobic organization of TM helices for both wild-type receptors and their mutants. Based on these data, the ability of several segments from other proteins to functionally replace IR and/or IGF-1R TM domains was explained.

Backbone and side-chain chemical shift assignments for the ribosome-inactivating protein trichobakin (TBK).

Trichobakin (TBK) is a type-I ribosome-inactivating protein (RIP-I), acting as an extremely potent inhibitor of protein synthesis in the cell-free translation system of rabbit reticulocyte lysate (IC50: 3.5 pM). In this respect, TBK surpasses the well-studied highly homologous RIP-I trichosanthin (IC50: 20-27 pM), therefore creation of recombinant toxins based on it is of great interest. TBK needs to penetrate into cytosol through the cell membrane and specifically bind to α-sarcin/ricin loop of 28S ribosome RNA to perform the function of specific RNA depurination. At the moment, there is no detailed structural-dynamic information in solution about diverse states RIP-I can adopt at different stages on the way to protein synthesis inhibition. In this work, we report a near-complete assignment of 1H, 13C, and 15N TBK (27.3 kDa) resonances and analysis of the secondary structure based on the experimental chemical shifts data. This work will serve as a basis for further investigations of the structure, dynamics and interactions of the TBK with its molecular partners using NMR techniques.

Transmembrane Peptides as a New Strategy to Inhibit Neuraminidase-1 Activation.

Sialidases, or neuraminidases, are involved in several human disorders such as neurodegenerative, infectious and cardiovascular diseases, and cancers. Accumulative data have shown that inhibition of neuraminidases, such as NEU1 sialidase, may be a promising pharmacological target, and selective inhibitors of NEU1 are therefore needed to better understand the biological functions of this sialidase. In the present study, we designed interfering peptides (IntPep) that target a transmembrane dimerization interface previously identified in human NEU1 that controls its membrane dimerization and sialidase activity. Two complementary strategies were used to deliver the IntPep into cells, either flanked to a TAT sequence or non-tagged for solubilization in detergent micelles. Combined with molecular dynamics simulations and heteronuclear nuclear magnetic resonance (NMR) studies in membrane-mimicking environments, our results show that these IntPep are able to interact with the dimerization interface of human NEU1, to disrupt membrane NEU1 dimerization and to strongly decrease its sialidase activity at the plasma membrane. In conclusion, we report here new selective inhibitors of human NEU1 of strong interest to elucidate the biological functions of this sialidase.

Familial L723P Mutation Can Shift the Distribution between the Alternative APP Transmembrane Domain Cleavage Cascades by Local Unfolding of the Ε-Cleavage Site Suggesting a Straightforward Mechanism of Alzheimer's Disease Pathogenesis.

Alzheimer's disease is an age-related pathology associated with accumulation of amyloid-ß peptides, products of enzymatic cleavage of amyloid-ß precursor protein (APP) by secretases. Several familial mutations causing early onset of the disease have been identified in the APP transmembrane (TM) domain. The mutations influence production of amyloid-ß, but the molecular mechanisms of this effect are unclear. The "Australian" (L723P) mutation located in the C-termini of APP TM domain is associated with autosomal-dominant, early onset Alzheimer's disease. Herein, we describe the impact of familial L723P mutation on the structural-dynamic behavior of APP TM domain studied by high-resolution NMR in membrane-mimicking micelles and augmented by molecular dynamics simulations in explicit lipid bilayer. We found L723P mutation to cause local unfolding of the C-terminal turn of the APP TM domain helix and increase its accessibility to water required for cleavage of the protein backbone by γ-secretase in the ε-site, thus switching between alternative ("pathogenic" and "non-pathogenic") cleavage cascades. These findings suggest a straightforward mechanism of the pathogenesis associated with this mutation, and are of generic import for understanding the molecular-level events associated with APP sequential proteolysis resulting in accumulation of the pathogenic forms of amyloid-ß. Moreover, age-related onset of Alzheimer's disease can be explained by a similar mechanism, where the effect of mutation is emulated by the impact of local environmental factors, such as oxidative stress and/or membrane lipid composition. Knowledge of the mechanisms regulating generation of amyloidogenic peptides of different lengths is essential for development of novel treatment strategies of the Alzheimer's disease.

Atomistic mechanism of the constitutive activation of PDGFRA via its transmembrane domain.

Single-point mutations in the transmembrane (TM) region of receptor tyrosine kinases (RTKs) can lead to abnormal ligand-independent activation. We use a combination of computational modeling, NMR spectroscopy and cell experiments to analyze in detail the mechanism of how TM domains contribute to the activation of wild-type (WT) PDGFRA and its oncogenic V536E mutant. Using a computational framework, we scan all positions in PDGFRA TM helix for identification of potential functional mutations for the WT and the mutant and reveal the relationship between the receptor activity and TM dimerization via different interfaces. This strategy also allows us design a novel activating mutation in the WT (I537D) and a compensatory mutation in the V536E background eliminating its constitutive activity (S541G). We show both computationally and experimentally that single-point mutations in the TM region reshape the TM dimer ensemble and delineate the structural and dynamic determinants of spontaneous activation of PDGFRA via its TM domain. Our atomistic picture of the coupling between TM dimerization and PDGFRA activation corroborates the data obtained for other RTKs and provides a foundation for developing novel modulators of the pathological activity of PDGFRA.

Structural basis of the signal transduction via transmembrane domain of the human growth hormone receptor.

BACKGROUND: Prior studies of the human growth hormone receptor (GHR) revealed a distinct role of spatial rearrangements of its dimeric transmembrane domain in signal transduction across membrane. Detailed structural information obtained in the present study allowed elucidating the bases of such rearrangement and provided novel insights into receptor functioning. METHODS: We investigated the dimerization of recombinant TMD fragment GHR254-294 by means of high-resolution NMR in DPC micelles and molecular dynamics in explicit POPC membrane. RESULTS: We resolved two distinct dimeric structures of GHR TMD coexisting in membrane-mimicking micellar environment and providing left- and right-handed helix-helix association via different dimerization motifs. Based on the available mutagenesis data, the conformations correspond to the dormant and active receptor states and are distinguished by cis-trans isomerization of Phe-Pro266 bond in the transmembrane helix entry. Molecular dynamic relaxations of the structures in lipid bilayer revealed the role of the proline residue in functionally significant rearrangements of the adjacent juxtamembrane region supporting alternation between protein-protein and protein-lipid interactions of this region that can be triggered by ligand binding. Also, the importance of juxtamembrane SS bonding for signal persistency, and somewhat unusual aspects of transmembrane region interaction with water molecules were demonstrated. CONCLUSIONS: Two alternative dimeric structures of GHR TMD attributed to dormant and active receptor states interchange via allosteric rearrangements of transmembrane helices and extracellular juxtamembrane regions that support coordination between protein-protein and protein-lipid interactions. GENERAL SIGNIFICANCE: This study provides a holistic vision of GHR signal transduction across the membrane emphasizing the role of protein-lipid interactions.

The Physcomitrella patens Chloroplast Proteome Changes in Response to Protoplastation.

Plant protoplasts are widely used for genetic manipulation and functional studies in transient expression systems. However, little is known about the molecular pathways involved in a cell response to the combined stress factors resulted from protoplast generation. Plants often face more than one type of stress at a time, and how plants respond to combined stress factors is therefore of great interest. Here, we used protoplasts of the moss Physcomitrella patens as a model to study the effects of short-term stress on the chloroplast proteome. Using label-free comparative quantitative proteomic analysis (SWATH-MS), we quantified 479 chloroplast proteins, 219 of which showed a more than 1.4-fold change in abundance in protoplasts. We additionally quantified 1451 chloroplast proteins using emPAI. We observed degradation of a significant portion of the chloroplast proteome following the first hour of stress imposed by the protoplast isolation process. Electron-transport chain (ETC) components underwent the heaviest degradation, resulting in the decline of photosynthetic activity. We also compared the proteome changes to those in the transcriptional level of nuclear-encoded chloroplast genes. Globally, the levels of the quantified proteins and their corresponding mRNAs showed limited correlation. Genes involved in the biosynthesis of chlorophyll and components of the outer chloroplast membrane showed decreases in both transcript and protein abundance. However, proteins like dehydroascorbate reductase 1 and 2-cys peroxiredoxin B responsible for ROS detoxification increased in abundance. Further, genes such as thylakoid ascorbate peroxidase were induced at the transcriptional level but down-regulated at the proteomic level. Together, our results demonstrate that the initial chloroplast reaction to stress is due changes at the proteomic level.

Cell-free expression of the APP transmembrane fragments with Alzheimer's disease mutations using algal amino acid mixture for structural NMR studies.

Structural investigations need ready supply of the isotope labeled proteins with inserted mutations n the quantities sufficient for the heteronuclear NMR. Though cell-free expression system has been widely used in the past years, high startup cost and complex compound composition prevent many researches from the developing this technique, especially for membrane protein production. Here we demonstrate the utility of a robust, cost-optimized cell-free expression technique for production of the physiologically important transmembrane fragment of amyloid precursor protein, APP686-726, containing Alzheimer's disease mutations in the juxtamembrane (E693G, Arctic form) and the transmembrane parts (V717G, London form, or L723P, Australian form). The protein cost was optimized by varying the FM/RM ratio as well as the amino acid concentration. We obtained the wild-type and mutant transmembrane fragments in the pellet mode of continuous exchange cell-free system consuming only commercial algal mixture of the (13)C,(15)N-labeled amino acids. Scaling up analytical tests, we achieved milligram quantity yields of isotope labeled wild-type and mutant APP686-726 for structural studies by high resolution NMR spectroscopy in membrane mimicking environment. The described approach has from 5 to 23-fold cost advantage over the bacterial expression methods described earlier and 1.5 times exceeds our previous result obtained with the longer APP671-726WT fragment.