Impact of Cystinosin Glycosylation on Protein Stability by Differential Dynamic Stable Isotope Labeling by Amino Acids in Cell Culture (SILAC)

Novel mechanism for tubular injury in nephropathic cystinosis

Understanding the unique susceptibility of the human kidney to pH dysfunction and injury in cystinosis is paramount to developing new therapies to preserve renal function. Renal proximal tubular epithelial cells (RPTECs) and fibroblasts isolated from patients with cystinosis were transcriptionally profiled. Lysosomal fractionation, immunoblotting, confocal microscopy, intracellular pH, TEM, and mitochondrial stress test were performed for validation. CRISPR, CTNS -/- RPTECs were generated. Alterations in cell stress, pH, autophagic turnover, and mitochondrial energetics highlighted key changes in the V-ATPases in patient-derived and CTNS-/- RPTECs. ATP6V0A1 was significantly downregulated in cystinosis and highly co-regulated with loss of CTNS. Correction of ATP6V0A1 rescued cell stress and mitochondrial function. Treatment of CTNS -/- RPTECs with antioxidants ATX induced ATP6V0A1 expression and improved autophagosome turnover and mitochondrial integrity. Our exploratory transcriptional and in vitro cellular and functional studies confirm that loss of Cystinosin in RPTECs, results in a reduction in ATP6V0A1 expression, with changes in intracellular pH, mitochondrial integrity, mitochondrial function, and autophagosome-lysosome clearance. The novel findings are ATP6V0A1's role in cystinosis-associated renal pathology and among other antioxidants, ATX specifically upregulated ATP6V0A1, improved autophagosome turnover or reduced autophagy and mitochondrial integrity. This is a pilot study highlighting a novel mechanism of tubular injury in cystinosis.

A multi-species benchmark for training and validating mass spectrometry proteomics machine learning models

Training machine learning models for tasks such as de novo sequencing or spectral clustering requires large collections of confidently identified spectra. Here we describe a dataset of 2.8 million high-confidence peptide-spectrum matches derived from nine different species. The dataset is based on a previously described benchmark but has been re-processed to ensure consistent data quality and enforce separation of training and test peptides.

Residual Cystine Transport Activity for Specific Infantile and Juvenile CTNS Mutations in a PTEC-Based Addback Model

Cystinosis is a rare, autosomal recessive, lysosomal storage disease caused by mutations in the gene CTNS, leading to cystine accumulation in the lysosomes. While cysteamine lowers the cystine levels, it does not cure the disease, suggesting that CTNS exerts additional functions besides cystine transport. This study investigated the impact of infantile and juvenile CTNS mutations with discrepant genotype/phenotype correlations on CTNS expression, and subcellular localisation and function in clinically relevant cystinosis cell models to better understand the link between genotype and CTNS function. Using CTNS-depleted proximal tubule epithelial cells and patient-derived fibroblasts, we expressed a selection of CTNSmutants under various promoters. EF1a-driven expression led to substantial overexpression, resulting in CTNS protein levels that localised to the lysosomal compartment. All CTNSmutants tested also reversed cystine accumulation, indicating that CTNSmutants still exert transport activity, possibly due to the overexpression conditions. Surprisingly, even CTNSmutants expression driven by the less potent CTNS and EFS promoters reversed the cystine accumulation, contrary to the CTNSG339R missense mutant. Taken together, our findings shed new light on CTNS mutations, highlighting the need for robust assessment methodologies in clinically relevant cellular models and thus paving the way for better stratification of cystinosis patients, and advocating for the development of more personalized therapy.

Bidirectional de novo peptide sequencing using a transformer model

In proteomics, a crucial aspect is to identify peptide sequences. De novo sequencing methods have been widely employed to identify peptide sequences, and numerous tools have been proposed over the past two decades. Recently, deep learning approaches have been introduced for de novo sequencing. Previous methods focused on encoding tandem mass spectra and predicting peptide sequences from the first amino acid onwards. However, when predicting peptides using tandem mass spectra, the peptide sequence can be predicted not only from the first amino acid but also from the last amino acid due to the coexistence of b-ion (or a- or c-ion) and y-ion (or x- or z-ion) fragments in the tandem mass spectra. Therefore, it is essential to predict peptide sequences bidirectionally. Our approach, called NovoB, utilizes a Transformer model to predict peptide sequences bidirectionally, starting with both the first and last amino acids. In comparison to Casanovo, our method achieved an improvement of the average peptide-level accuracy rate of approximately 9.8% across all species.

Evaluation of the efficacy of cystinosin supplementation through CTNS mRNA delivery in experimental models for cystinosis

Messenger RNA (mRNA) therapies are emerging in different disease areas, but have not yet reached the kidney field. Our aim was to study the feasibility to treat the genetic defect in cystinosis using synthetic mRNA in cell models and ctns-/- zebrafish embryos. Cystinosis is a prototype lysosomal storage disorder caused by mutations in the CTNS gene, encoding the lysosomal cystine-H+ symporter cystinosin, and leading to cystine accumulation in all cells of the body. The kidneys are the first and the most severely affected organs, presenting glomerular and proximal tubular dysfunction, progressing to end-stage kidney failure. The current therapeutic standard cysteamine, reduces cystine levels, but has many side effects and does not restore kidney function. Here, we show that synthetic mRNA can restore lysosomal cystinosin expression following lipofection into CTNS-/- kidney cells and injection into ctns-/- zebrafish. A single CTNS mRNA administration decreases cellular cystine accumulation for up to 14 days in vitro. In the ctns-/- zebrafish, CTNS mRNA therapy improves proximal tubular reabsorption, reduces proteinuria, and restores brush border expression of the multi-ligand receptor megalin. Therefore, this proof-of-principle study takes the first steps in establishing an mRNA-based therapy to restore cystinosin expression, resulting in cystine reduction in vitro and in the ctns-/- larvae, and restoration of the zebrafish pronephros function.

ER-associated degradation in cystinosis pathogenesis and the prospects of precision medicine

Cystinosis is a lysosomal storage disease that is characterized by the accumulation of dipeptide cystine within the lumen. It is caused by mutations in the cystine exporter, cystinosin. Most of the clinically reported mutations are due to the loss of transporter function. In this study, we identified a rapidly degrading disease variant, referred to as cystinosin(7Δ). We demonstrated that this mutant is retained in the ER and degraded via the ER-associated degradation (ERAD) pathway. Using genetic and chemical inhibition methods, we elucidated the roles of HRD1, p97, EDEMs, and the proteasome complex in cystinosin(7Δ) degradation pathway. Having understood the degradation mechanisms, we tested some chemical chaperones previously used for treating CFTR F508Δ and demonstrated that they could facilitate the folding and trafficking of cystinosin(7Δ). Strikingly, chemical chaperone treatment can reduce the lumenal cystine level by approximately 70%. We believe that our study conclusively establishes the connection between ERAD and cystinosis pathogenesis and demonstrates the possibility of using chemical chaperones to treat cystinosin(7Δ).

Abnormal expression of lysosomal glycoproteins in patients with congenital disorders of glycosylation

OBJECTIVE

The study of the impact of some inherited defects in glycosylation on the biosynthesis of some lysosomal glycoproteins. Results description: Whole-exome sequencing revealed a homozygous variant; 428G > A; p. (R143K) in SRD5A3 in one patient and a heterozygous one c.46G > A p. (Gly16Arg) in SLC35A2 in the other patient. Both variants were predicted to be likely pathogenic. Lysosome-associated membrane glycoprotein 2 (LAMP2) immunodetection in both cases showed a truncated form of the protein. Cystinosin (CTN) protein appeared as normal and truncated forms in both patients in ratios of the mature to truncated forms of CTN were lower than the control. The levels of the truncated forms of both cellular proteins were higher in the SRD5A3-CDG case compared to the SLC35A2-CDG case. The tetrameric form of cathepsin C (CTSC) was expressed at low levels in both cases with congenital disorder of glycosylation (CDG). SLC35A2-CDG patient had one extra-unknown band while SRD5A3-CDG patient had a missing band of CTSC forms. The expression patterns of lysosomal glycoproteins could be different between different types of CDG.

Critical evaluation of the use of artificial data for machine learning based de novo peptide identification

Proteins are essential components of all living cells and so the study of their in situ expression, proteomics, has wide reaching applications. Peptide identification in proteomics typically relies on matching high resolution tandem mass spectra to a protein database but can also be performed de novo. While artificial spectra have been successfully incorporated into database search pipelines to increase peptide identification rates, little work has been done to investigate the utility of artificial spectra in the context of de novo peptide identification. Here, we perform a critical analysis of the use of artificial data for the training and evaluation of de novo peptide identification algorithms. First, we classify the different fragment ion types present in real spectra and then estimate the number of spurious matches using random peptides. We then categorise the different types of noise present in real spectra. Finally, we transfer this knowledge to artificial data and test the performance of a state-of-the-art de novo peptide identification algorithm trained using artificial spectra with and without relevant noise addition. Noise supplementation increased artificial training data performance from 30% to 77% of real training data peptide recall. While real data performance was not fully replicated, this work provides the first steps towards an artificial spectrum framework for the training and evaluation of de novo peptide identification algorithms. Further enhanced artificial spectra may allow for more in depth analysis of de novo algorithms as well as alleviating the reliance on database searches for training data.

Application of a Novel Hybrid CNN-GNN for Peptide Ion Encoding

Almost all state-of-the-art de novo peptide sequencing algorithms now use machine learning models to encode fragment peaks and hence identify amino acids in mass spectrometry (MS) spectra. Previous work has highlighted how the inherent MS challenges of noise and missing peptide peaks detrimentally affect the performance of these models. In the present research we extracted and evaluated the encoding modules from 3 state-of-the-art de novo peptide sequencing algorithms. We also propose a convolutional neural network-graph neural network machine learning model for encoding peptide ions in tandem MS spectra. We compared the proposed encoding module to those used in the state-of-the-art de novo peptide sequencing algorithms by assessing their ability to identify b-ions and y-ions in MS spectra. This included a comprehensive evaluation in both real and artificial data across various levels of noise and missing peptide peaks. The proposed model performed best across all data sets using two different metrics (area under the receiver operating characteristic curve (AUC) and average precision). The work also highlighted the effect of including additional features such as intensity rank in these encoding modules as well as issues with using the AUC as a metric. This work is of significance to those designing future de novo peptide identification algorithms as it is the first step toward a new approach.

Structure and mechanism of human cystine exporter cystinosin

Lysosomal amino acid efflux by proton-driven transporters is essential for lysosomal homeostasis, amino acid recycling, mTOR signaling, and maintaining lysosomal pH. To unravel the mechanisms of these transporters, we focus on cystinosin, a prototypical lysosomal amino acid transporter that exports cystine to the cytosol, where its reduction to cysteine supplies this limiting amino acid for diverse fundamental processes and controlling nutrient adaptation. Cystinosin mutations cause cystinosis, a devastating lysosomal storage disease. Here, we present structures of human cystinosin in lumen-open, cytosol-open, and cystine-bound states, which uncover the cystine recognition mechanism and capture the key conformational states of the transport cycle. Our structures, along with functional studies and double electron-electron resonance spectroscopic investigations, reveal the molecular basis for the transporter's conformational transitions and protonation switch, show conformation-dependent Ragulator-Rag complex engagement, and demonstrate an unexpected activation mechanism. These findings provide molecular insights into lysosomal amino acid efflux and a potential therapeutic strategy.

pNovo 3: precise de novo peptide sequencing using a learning-to-rank framework

MOTIVATION

De novo peptide sequencing based on tandem mass spectrometry data is the key technology of shotgun proteomics for identifying peptides without any database and assembling unknown proteins. However, owing to the low ion coverage in tandem mass spectra, the order of certain consecutive amino acids cannot be determined if all of their supporting fragment ions are missing, which results in the low precision of de novo sequencing.

RESULTS

In order to solve this problem, we developed pNovo 3, which used a learning-to-rank framework to distinguish similar peptide candidates for each spectrum. Three metrics for measuring the similarity between each experimental spectrum and its corresponding theoretical spectrum were used as important features, in which the theoretical spectra can be precisely predicted by the pDeep algorithm using deep learning. On seven benchmark datasets from six diverse species, pNovo 3 recalled 29-102% more correct spectra, and the precision was 11-89% higher than three other state-of-the-art de novo sequencing algorithms. Furthermore, compared with the newly developed DeepNovo, which also used the deep learning approach, pNovo 3 still identified 21-50% more spectra on the nine datasets used in the study of DeepNovo. In summary, the deep learning and learning-to-rank techniques implemented in pNovo 3 significantly improve the precision of de novo sequencing, and such machine learning framework is worth extending to other related research fields to distinguish the similar sequences.

AVAILABILITY AND IMPLEMENTATION

pNovo 3 can be freely downloaded from http://pfind.ict.ac.cn/software/pNovo/index.html.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Interaction between galectin-3 and cystinosin uncovers a pathogenic role of inflammation in kidney involvement of cystinosis

Inflammation is involved in the pathogenesis of many disorders. However, the underlying mechanisms are often unknown. Here, we test whether cystinosin, the protein involved in cystinosis, is a critical regulator of galectin-3, a member of the β-galactosidase binding protein family, during inflammation. Cystinosis is a lysosomal storage disorder and, despite ubiquitous expression of cystinosin, the kidney is the primary organ impacted by the disease. Cystinosin was found to enhance lysosomal localization and degradation of galectin-3. In Ctns-/- mice, a mouse model of cystinosis, galectin-3 is overexpressed in the kidney. The absence of galectin-3 in cystinotic mice ameliorates pathologic renal function and structure and decreases macrophage/monocyte infiltration in the kidney of the Ctns-/-Gal3-/- mice compared to Ctns-/- mice. These data strongly suggest that galectin-3 mediates inflammation involved in kidney disease progression in cystinosis. Furthermore, galectin-3 was found to interact with the pro-inflammatory cytokine Monocyte Chemoattractant Protein-1, which stimulates the recruitment of monocytes/macrophages, and proved to be significantly increased in the serum of Ctns-/- mice and also patients with cystinosis. Thus, our findings highlight a new role for cystinosin and galectin-3 interaction in inflammation and provide an additional mechanistic explanation for the kidney disease of cystinosis. This may lead to the identification of new drug targets to delay cystinosis progression.

Postnovo: Postprocessing Enables Accurate and FDR-Controlled de Novo Peptide Sequencing

De novo sequencing offers an alternative to database search methods for peptide identification from mass spectra. Since it does not rely on a predetermined database of expected or potential sequences in the sample, de novo sequencing is particularly appropriate for samples lacking a well-defined or comprehensive reference database. However, the low accuracy of many de novo sequence predictions has prevented the widespread use of the variety of sequencing tools currently available. Here, we present a new open-source tool, Postnovo, that postprocesses de novo sequence predictions to find high-accuracy results. Postnovo uses a predictive model to rescore and rerank candidate sequences in a manner akin to database search postprocessing tools such as Percolator. Postnovo leverages the output from multiple de novo sequencing tools in its own analyses, producing many times the length of amino acid sequence information (including both full- and partial-length peptide sequences) at an equivalent false discovery rate (FDR) compared to any individual tool. We present a methodology to reliably screen the sequence predictions to a desired FDR given the Postnovo sequence score. We validate Postnovo with multiple data sets and demonstrate its ability to identify proteins that are missed by database search even in samples with paired reference databases.

Predicting the Structures of Glycans, Glycoproteins, and Their Complexes

Complex carbohydrates are ubiquitous in nature, and together with proteins and nucleic acids they comprise the building blocks of life. But unlike proteins and nucleic acids, carbohydrates form nonlinear polymers, and they are not characterized by robust secondary or tertiary structures but rather by distributions of well-defined conformational states. Their molecular flexibility means that oligosaccharides are often refractory to crystallization, and nuclear magnetic resonance (NMR) spectroscopy augmented by molecular dynamics (MD) simulation is the leading method for their characterization in solution. The biological importance of carbohydrate-protein interactions, in organismal development as well as in disease, places urgency on the creation of innovative experimental and theoretical methods that can predict the specificity of such interactions and quantify their strengths. Additionally, the emerging realization that protein glycosylation impacts protein function and immunogenicity places the ability to define the mechanisms by which glycosylation impacts these features at the forefront of carbohydrate modeling. This review will discuss the relevant theoretical approaches to studying the three-dimensional structures of this fascinating class of molecules and interactions, with reference to the relevant experimental data and techniques that are key for validation of the theoretical predictions.

De novo peptide sequencing by deep learning

De novo peptide sequencing from tandem MS data is the key technology in proteomics for the characterization of proteins, especially for new sequences, such as mAbs. In this study, we propose a deep neural network model, DeepNovo, for de novo peptide sequencing. DeepNovo architecture combines recent advances in convolutional neural networks and recurrent neural networks to learn features of tandem mass spectra, fragment ions, and sequence patterns of peptides. The networks are further integrated with local dynamic programming to solve the complex optimization task of de novo sequencing. We evaluated the method on a wide variety of species and found that DeepNovo considerably outperformed state of the art methods, achieving 7.7-22.9% higher accuracy at the amino acid level and 38.1-64.0% higher accuracy at the peptide level. We further used DeepNovo to automatically reconstruct the complete sequences of antibody light and heavy chains of mouse, achieving 97.5-100% coverage and 97.2-99.5% accuracy, without assisting databases. Moreover, DeepNovo is retrainable to adapt to any sources of data and provides a complete end-to-end training and prediction solution to the de novo sequencing problem. Not only does our study extend the deep learning revolution to a new field, but it also shows an innovative approach in solving optimization problems by using deep learning and dynamic programming.