Thymine DNA glycosylase (TDG) is a multifaceted protein involved in base-excision repair, DNA demethylation and transcriptional regulation, with key roles in embryonic development and tumorigenesis. However, the mechanisms underlying its role in cancer progression and the therapeutic applications targeting TDG remain largely unknown. Here we demonstrate that targeting TDG induces synthetic lethality in p53-deficient cancers. We developed C-271, a first-in-class, small-molecule inhibitor that covalently binds to TDG, disrupting its DNA-binding capability. C-271 exhibits potent therapeutic efficacy in suppressing p53-deficient tumors. Mechanistically, TDG and p53 redundantly promote the transcription of DHX9, an RNA helicase that resolves double-stranded RNA (dsRNA). TDG inhibition in p53-deficient cancer cells leads to DHX9 downregulation and, thus, aberrant dsRNA accumulation, which activates the RIG-I/MDA5–MAVS sensing pathway, resulting in tumor suppression and enhanced antitumor immunity. These findings highlight the synthetic lethality between TDG and p53, positioning TDG inhibition as a promising therapeutic strategy for p53-deficient cancers. Thymine DNA glycosylase (TDG) was identified as a synthetic-lethal target in p53-deficient cancers through alterations in double-stranded RNA accumulation. A covalent small-molecule inhibitor of TDG exhibited potent efficacy against p53-deficient tumors.

ABSTRACT Synthetic circular mRNA (hereafter referred to as circRNA) reduces susceptibility to exonuclease‐mediated degradation by its covalently closed circular structure, enabling prolonged protein expression for therapeutic applications. In this circular format, protein expression from engineered circRNAs is achieved mainly through cap‐independent translation initiation, commonly mediated by internal ribosome entry site (IRES) elements whose activity is influenced by RNA structure. Consequently, the coding sequence (CDS) and other elements should be designed with consideration of inter‐region base pairing that can shift IRES folding, a constraint not explicitly addressed by existing linear mRNA CDS optimization algorithms. Here, we present circ Design, an algorithm that explicitly incorporates IRES structural deviation into circRNA sequence design while jointly optimizing codon adaptation and thermodynamic stability. In a rabies virus glycoprotein (RABV‐G) vaccine model, circ Design‐generated circRNAs showed improved stability, translation efficiency, and vaccine immunogenicity compared with benchmark sequences optimized using conventional linear mRNA CDS design strategies, with CR3 achieving a 3.5‐fold increase in neutralizing antibody titers. Polysome profiling and targeted IRES‐disruption experiments support IRES structural integrity as a critical determinant of circRNA translation performance. Together, these results establish IRES structural preservation as a mechanistic design principle for circRNA engineering and position circ Design as a rational framework for therapeutic circRNA development.

Although dysregulation of iron metabolism is known to contribute to tumorigenesis, the upstream mechanisms driving iron overload in gastric cancer (GC) remain poorly understood. This research aims to elucidate the oncogenic role of eukaryotic translation initiation factor 3 subunit I (EIF3I) in GC and its impact on iron metabolism. EIF3I expression in GC was evaluated using clinical tissue samples and correlated with clinicopathological features. The functional impact of EIF3I on tumor progression was evaluated through both in vitro and in vivo assays. Molecular mechanisms were investigated using protein interaction studies, ubiquitination assays, and transcriptional analyses. A therapeutic peptide targeting EIF3I– was screened and tested in zebrafish and mice models. EIF3I was significantly overexpressed in GC and associated with aggressive tumor phenotypes. EIF3I depletion suppressed GC cell proliferation and tumor growth, while its overexpression enhanced malignancy. Mechanistically, EIF3I stabilized yes-associated protein 1 (YAP) by directly interacting with it and preventing its ubiquitination and degradation by STUB1. This stabilization promoted YAP nuclear liquid–liquid phase separation and transcriptional activity, leading to increased expression of transferrin receptor 1 (TFR1), and consequent intracellular iron accumulation. Targeting the EIF3I interface on YAP with a screened peptide restored STUB1-mediated YAP degradation and significantly inhibited tumor growth in vitro and in vivo. EIF3I promotes GC progression by stabilizing YAP and reprogramming iron metabolism through the EIF3I–YAP-iron axis. Targeting this interaction represents a promising therapeutic strategy for GC.

Bacteriophages possess a wide array of DNA modifications, with many acting as molecular camouflage to evade host immune defenses. Sequence databases contain numerous bacteriophage enzymes of unknown function, with some potentially involved in yet to be identified DNA modifications. Here we report the discovery of a DNA cytosine C5-carboxymethyltransferase (CmoX) in Synechococcus phage S-B43, which catalyzes the formation of a 5-carboxymethylcytosine (5cxmC), previously reported as an unnatural DNA modification formed by an engineered cytosine methyltransferase. The carboxy-S-adenosyl-L-methionine (Cx-SAM) cofactor required by CmoX is provided by a phage-encoded Cx-SAM synthase (CmoA), a homolog of the bacterial CmoA involved in tRNA modification. A crystal structure of CmoX in complex with Cx-SAM revealed the basis for its substrate selectivity, involving a key Arg residue interacting with the substrate carboxy group. In addition, we characterize a phage-encoded ATP-dependent amide ligase, CmoY that catalyzes the formation of 5cxmC-glycine amide. CmoA is present in many bacteriophage genomes, typically alongside CmoX and homologs of CmoY, suggesting that 5cxmC modification is a widespread naturally occurring DNA modification serving as a handle for further hypermodifications in bacteriophages. Our study underscores the ability of bacteriophages to repurpose RNA modification enzymes to expand their repertoire of DNA modifications.

Cancer cells exploit altered metabolic pathways to dynamically regulate epigenetic methylation and thus promote tumorigenesis and metastasis. In various human cancers, such as lung adenocarcinoma, the level of a key cellular metabolite, S-adenosylmethionine (SAM), is prominently upregulated for RNA hypermethylation as the methyl donor. However, the specific mechanisms by which cancer cells produce SAM to sustain RNA methylation remain elusive. Here, we demonstrate that PRPS2, a phosphoribosyl pyrophosphate synthetase isoform involved in the first and rate-limiting step of the purine biosynthesis pathway, exhibits distinct oncogenic functionality in regulating RNA methylation, unlike its homolog PRPS1. PRPS2 utilizes four non-conserved key residues to bypass the typical ADP/GDP allosteric feedback inhibition, enabling sustained excess production of newly synthesized ATP. Moreover, PRPS2 stabilizes methionine adenosyltransferase 2 A (MAT2A) through direct interactions to positively stimulate ATP utilization and SAM synthesis for RNA m 6 A specific methylation via the WTAP/METTL3/METTL14 methyltransferase complex, thereby promoting lung tumorigenesis. Our study links nucleotide biosynthesis with RNA epigenetics in cancer progression through the PRPS2-MAT2A-WTAP/METTL3/METTL14 axis, and elucidates both enzyme-dependent and independent functions of PRPS2. These findings have significant implications for developing targeted therapies for cancers associated with PRPS2 abnormalities.

beta-Ketoacyl-ACP synthases (KAS) catalyze carbon skeleton extension in numerous metabolic routes such as the fatty acid biosynthesis pathway (FAS), among which FabH is the only known member that links the initiation stage to the elongation cycle of type-II FAS (FAS-II) by catalyzing condensation between acetyl-CoA and malonyl-ACP for the first beta-keto-ACP intermediate acetoacetyl-ACP formation. Here, we reveal the substrate selection and condensation mechanisms of FabH from Escherichia coli. We demonstrate that EcFabH binds CoA and ACP using distinct regions in an irreversible compulsory order. The malonyl moiety is then delivered to a hydrophobic cage near the catalytic triad residues through front and middle door residues in the tunnel, and the substrate length is selected by a backdoor residue Phe87, ensuring the preferential recognition of EcFabH on acetyl moiety carried by CoA rather than longer substrates. Moreover, the malonyl moiety is locked in the cage by the acetylated Cys112 from the transacylation reaction, triggering the subsequent decarboxylation and condensation catalysis. Our study provides fundamental mechanistic insights into the initial extension of carbon skeletons catalyzed by FabH and homologues in FAS, PKS, and biotin biosynthesis pathways and may facilitate protein engineering and optimization for synthetic biological and pharmaceutical industry, as well as antibacterial drug development.

Natural product pseudolaric acid A (PAA), the main bioactive component from Traditional Chinese Medicine Pseudolarix cortex ("tujingpi"), is a promising anticancer agent. However, its potential molecular targets are not clear and this hinders its development. In this study, chemical proteomics approaches including activity-based protein profiling (ABPP) and drug affinity responsive target stability (DARTS) technology, followed by quantitative proteomics, were combined to reveal the target of PAA. Target validation was performed by NMR techniques and surface plasmon resonance. Methylenetetrahydrofolate dehydrogenase 1-like (MTHFD1L) was identified and further confirmed to be the target of PAA. The direct interaction and binding mode between MTHFD1L and PAA were elaborated. PAA induced the accumulation of the reactive oxygen species (ROS) which mediates the antitumor effect. Transcriptome and network pharmacology analysis reveals the effects of PAA on the gene expressions of the associated pathways. Taken together, our findings proposed a new target that could be used for structure-based rational design and modifications of PAA.

Unsaturated fatty acids (UFAs) are essential for the membrane function in most bacteria. In Helicobacter pylori (H. pylori), a gastric pathogen, UFA biosynthesis depends on the bifunctional dehydrogenase/isomerase FabX, a promising target against H. pylori. Herein, we report the first FabX inhibitor, P61G11 (compound 1, IC50 = 3.7 ± 0.2 μM), identified via high-throughput screening and featuring a 1,3,4-thiadiazole sulfonamide scaffold. The costructure of FabX-1 reveals occupancy of the L-shaped substrate-binding tunnel via hydrophobic interactions and hydrogen bonds. Structure-based optimization led to more potent derivatives, among which compound 47 showed potent inhibition (IC50 = 0.128 ± 0.002 μM), representing a 29-fold improvement. Compound 47 also demonstrated strong in vitro antibacterial activity (MIC = 0.5-1 μg/mL), when combined with membrane permeabilizers, efflux pump inhibitors, and clarithromycin, and exhibited narrow-spectrum efficacy against H. pylori, providing a novel strategy for anti-H. pylori therapy.

The β-ketoacyl-acyl carrier protein (ACP) synthases are pivotal elongation enzymes that catalyze the condensation of acyl-CoA or acyl-ACP with malonyl-ACP to produce β-ketoacyl-ACP. Among these, the homologous enzymes FabH (β-ketoacyl-ACP synthase III) and the recently characterized BioZ play crucial roles, initiating the biosynthetic pathways for fatty acids and biotin, respectively. FabH primarily utilizes acetyl-CoA as the primer substrate, whereas BioZ exclusively condenses the longer glutaryl-CoA, which contains a charged ω-carboxyl group. Despite their similar catalytic mechanisms, the molecular bases for the strict substrate specificities remain undetermined. Here, we report crystal structures of the BioZ: glutaryl-CoA cocrystalized complexes and demonstrate the ability to swap the physiological functions and substrate specificities between FabH and BioZ. This functional interchange was achieved by grafting the β8-α9 loop plus residue Ala317 of Agrobacterium tumefaciens BioZ to Escherichia coli FabH, resulting in a shift in substrate preference from acetyl-CoA to glutaryl-CoA. The reverse manipulations of BioZ resulted in FabH activity. These data identify the structural elements as the minimal determinants of substrate specificity and enzyme function. These findings provide valuable insights into the molecular mechanisms of substrate recognition and catalysis by FabH and BioZ and offer a foundation for the development of targeted therapeutic strategies against these enzymes.

Pseudomonas aeruginosa is an opportunistic pathogen that frequently resides in multispecies communities. During chronic infections, P. aeruginosa employs a diverse arsenal of antibacterial weapons to compete with other bacteria for resources and space. Using genetic and biochemical approaches, we identified a type VI secretion system-dependent antibacterial effector-immunity pair, PseM (P. aeruginosa secreted endolytic muramidase) and PA0990 in P. aeruginosa. Our findings demonstrate that PseM functions as an endolytic muramidase, targeting prey bacteria by hydrolytically cleaving cell-wall peptidoglycan, whereas its immunity partner PA0990 provides self-protection. The X-ray crystal structure of PseM reveals a homodimeric configuration, with its active site formed by segments from both monomers. Through structural analysis and macromolecular docking simulations, we further elucidate the substrate-binding residues critical for the activity of PseM. Importantly, we show that PseM contributes to P. aeruginosa growth among bacterial competition. Together, these results uncover a novel antibacterial mechanism mediated by PseM, highlighting the dynamic nature of interspecies and intraspecies competition within bacterial populations.

Circular mRNA (mRNA) exhibits promising potential in mRNA therapy due to its increased stability and extended duration of protein translation, which has sparked an urgent demand for efficient methods to prepare circular RNAs in vitro. Here, we present a versatile self-circularization strategy that employs simple motifs to synthesize circular RNAs, achieving robust efficiencies for sequences ranging from dozens to thousands of nucleotides. By leveraging an automated computational program, we optimized highly specific lock-key structures to maximize circularization efficiency, particularly for long RNA substrates. Furthermore, the shared sequence and functionality between linear precursor RNAs and circular products eliminate the need for additional purification steps to remove excess nucleic acid components, simplifying the production process. This approach also yields circular RNAs with superior stability and translation efficiency, enabling sustained protein expression in vitro and in vivo. Our computationally optimized, purification-free method holds immense promise for scalable circular RNA production and the development of advanced RNA therapeutics, significantly advancing mRNA therapy.

Abstract Helicobacter pylori (H. pylori) is a group‐1 definite pathogenic carcinogen that infects approximately half of the global population, yet no species‐specific chemotherapy has yet been developed. It is previously discovered that H. pylori encodes an atypical dehydrogenase/isomerase FabX in the Type‐II fatty acid biosynthesis pathway to produce unsaturated fatty acids (UFA) as well as superoxide (ROS). Here, it is demonstrated that FabX is essential for H. pylori growth and gastric colonization by retaining UFA synthesis and producing ROS, respectively, and is a species‐specific anti‐H. pylori drug target. The first small molecule inhibitor FBX‐1991 against FabX, which inhibits the enzymatic activity with an IC50 value of 0.158 × 10−6 m in vitro, is developed. FBX‐1991 binds inside the catalytic tunnel of FabX, disrupts the conformation of the key catalytic loop, and prevents the insertion of the acyl substrate for catalysis. Further in vivo studies suggest that FBX‐1991 inhibits the H. pylori growth by partially inhibiting UFA synthesis and ROS excretion through targeting FabX. This study identifies a species‐specific anti‐H. pylori drug target, FabX, and discovers the first highly potent and selective FabX inhibitor against H. pylori infection, which provides the molecular basis for developing species‐specific anti‐H. pylori chemotherapy.

Understanding how amino acids influence protein expression is crucial for advancements in biotechnology and synthetic biology. In this study, we introduce Venus-TIGER, a deep learning model designed to accurately identify amino acids critical for expression. By constructing a two-dimensional matrix that links model representations to experimental fitness, Venus-TIGER achieves improved predictive accuracy and enhanced extrapolation capability. We validated our approach on both public deep mutational scanning datasets and low-throughput experimental datasets, demonstrating notable performance compared to traditional methods. Venus-TIGER exhibits robust transferability in zero-shot predicting scenarios and enhanced predictive performance in few-shot learning, even with limited experimental data. This capability is particularly valuable for protein design aimed at enhancing expression, where generating large datasets can be costly and time-consuming. Additionally, we conducted a statistical analysis to identify expression-associated features, such as sequence and structural preferences, distinguishing between those linked to high and low expression. Our investigation also revealed a correlation among stability, activity and expression, providing insight into their interconnected roles and underlying mechanisms. ### Competing Interest Statement The authors have declared no competing interest.

N 1 -methyladenosine (m 1 A) is a prevalent post-transcriptional RNA modification, and the distribution and dynamics of the modification play key epitranscriptomic roles in cell development. At present, the human AlkB Fe(II)/α-ketoglutarate-dependent dioxygenase family member ALKBH3 is the only known mRNA m 1 A demethylase, but its catalytic mechanism remains unclear. Here, we present the structures of ALKBH3-oligo crosslinked complexes obtained with the assistance of a synthetic antibody crystallization chaperone. Structural and biochemical results showed that ALKBH3 utilized two β-hairpins (β4-loop-β5 and β′-loop-β′′) and the α2 helix to facilitate single-stranded substrate binding. Moreover, a bubble-like region around Asp194 and a key residue inside the active pocket (Thr133) enabled specific recognition and demethylation of m 1 A- and 3-methylcytidine (m 3 C)-modified substrates. Mutation of Thr133 to the corresponding residue in the AlkB Fe(II)/α-ketoglutarate-dependent dioxygenase family members FTO or ALKBH5 converted ALKBH3 substrate selectivity from m 1 A to N 6 -methyladenosine (m 6 A), as did Asp194 deletion. Our findings provide a molecular basis for understanding the mechanisms of substrate recognition and m 1 A demethylation by ALKBH3. This study is expected to aid structure-guided design of chemical probes for further functional studies and therapeutic applications.

The human fatty acid synthase (hFASN) produces fatty acids for cellar membrane construction, energy storage, biomolecule modifications and signal transduction. Abnormal expression and functions of hFASN highly associate with numerous human diseases such as obesity, diabetes, and cancers, and thereby it has been considered as a valuable potential drug target. So far, the structural and catalytic mechanisms of most of the hFASN enzymatic modules have been extensively studied, except the key dehydratase module (hDH). Here we presented the enzymatic characterization and the high‐resolution crystal structure of hDH. We demonstrated that the hDH preferentially catalyzes the acyl substrates with short lengths between 4 to 8‐carbons, and exhibits much lower enzymatic activity on longer substrates. Subsequent structural study showed that hDH displays a pseudo‐dimeric organization with a single L‐shaped composite hydrophobic catalytic tunnel as well as an atypical ACP binding site nearby, indicating that hDH achieves distinct substrate recognition and dehydration mechanisms compared to the conventional bacterial fatty acid dehydratases identified. Our findings laid the foundation for understanding the biological and pathogenic functions of hFASN, and may facilitate therapeutical drug development against diseases with abnormal functionality of hFASN.

The eukaryotic epigenetic modifications 5-methyldeoxycytosine (5mC) and N6-methyldeoxyadenine (6mA) have indispensable regulatory roles in gene expression and embryonic development. We recently identified an atypical bifunctional dioxygenase CcTet from Coprinopsis cinerea that works on both 5mC and 6mA demethylation. The nonconserved residues Gly331 and Asp337 of CcTet facilitate 6mA accommodation, while D337F unexpectedly abolishes 5mC oxidation activity without interfering 6mA demethylation, indicating a prominent distinct but unclear 5mC oxidation mechanism to the conventional Tet enzymes. Here, we assessed the molecular mechanism of CcTet in catalyzing 5mC oxidation by representing the crystal structure of CcTet-5mC-dsDNA complex. We identified the distinct mechanism by which CcTet recognizes 5mC-dsDNA compared to 6mA-dsDNA substrate. Moreover, Asp337 was found to have a central role in compensating for the loss of a critical 5mC-stablizing H-bond observed in conventional Tet enzymes, and stabilizes 5mC and subsequent intermediates through an H-bond with the N4 atom of the substrates. These findings improve our understanding of Tet enzyme functions in the dsDNA 5mC and 6mA demethylation pathways, and provide useful information for future discovery of small molecular probes targeting Tet enzymes in DNA active demethylation processes.

m6A modification is best known for its critical role in controlling multiple post-transcriptional processes of the mRNAs. Here, we discovered elevated levels of m6A modification on centromeric RNA (cenRNA) in cancerous cells compared with non-cancerous cells. We then identified CENPA, an H3 variant, as an m6A reader of cenRNA. CENPA is localized at centromeres and is essential in preserving centromere integrity and function during mitosis. The m6A-modified cenRNA stabilizes centromeric localization of CENPA in cancer cells during the S phase of the cell cycle. Mutations of CENPA at the Leu61 and the Arg63 or removal of cenRNA m6A modification lead to loss of centromere-bound CENPA during S phase. This in turn results in compromised centromere integrity and abnormal chromosome separation and hinders cancer cell proliferation and tumor growth. Our findings unveil an m6A reading mechanism by CENPA that epigenetically governs centromere integrity in cancer cells, providing potential targets for cancer therapy.