The pro-drug C13 activates AMPK by two distinct mechanisms.

The AMP-activated protein kinase (AMPK) is a sensor of cellular energy status that is expressed in almost all eukaryotic cells. In the canonical activation mechanism, it is activated by increases in AMP:ATP and ADP:ATP ratios that signify declining cellular energy status. Once activated, AMPK phosphorylates numerous targets that promote catabolic pathways generating ATP, while inhibiting anabolic and other processes that consume ATP, thus acting to restore energy homeostasis. Pharmacological agents that activate AMPK have been useful in identifying downstream targets and have potential as drugs for treatment of metabolic disorders such as Type 2 diabetes and non-alcoholic fatty liver disease. One such agent is C13, a pro-drug with a phosphonate bis(isobutyryloxymethyl) ester moiety, with the isobutyryloxymethyl groups increasing membrane permeability. Following cellular uptake, C13 is cleaved to release C2, an AMP analogue and potent AMPK activator that is specific for complexes containing the α1 (but not the α2) catalytic subunit isoform. This has previously been assumed to be the sole mechanism by which C13 activates AMPK, with potential roles for the isobutyryloxymethyl groups being ignored. We now report that, following cleavage from C13, these protective groups are metabolized to formaldehyde, an agent that inhibits mitochondrial function and increases cellular AMP:ATP ratios, thus providing additional AMPK activation by the canonical mechanism.

Structural basis of adenine nucleotides regulation and neurodegenerative pathology in ClC-3 exchanger.

The ClC-3 chloride/proton exchanger is both physiologically and pathologically critical, as it is potentiated by ATP to detect metabolic energy level and point mutations in ClC-3 lead to severe neurodegenerative diseases in human. However, why this exchanger is differentially modulated by ATP, ADP or AMP and how mutations caused gain-of-function remains largely unknow. Here we determine the high-resolution structures of dimeric wildtype ClC-3 in the apo state and in complex with ATP, ADP and AMP, and the disease-causing I607T mutant in the apo and ATP-bounded state by cryo-electron microscopy. In combination with patch-clamp recordings and molecular dynamic simulations, we reveal how the adenine nucleotides binds to ClC-3 and changes in ion occupancy between apo and ATP-bounded state. We further observe I607T mutation induced conformational changes and augments in current. Therefore, our study not only lays the structural basis of adenine nucleotides regulation in ClC-3, but also clearly indicates the target region for drug discovery against ClC-3 mediated neurodegenerative diseases.

Acetyl-CoA synthetase activity is enzymatically regulated by lysine acetylation using acetyl-CoA or acetyl-phosphate as donor molecule.

The AMP-forming acetyl-CoA synthetase is regulated by lysine acetylation both in bacteria and eukaryotes. However, the underlying mechanism is poorly understood. The Bacillus subtilis acetyltransferase AcuA and the AMP-forming acetyl-CoA synthetase AcsA form an AcuA•AcsA complex, dissociating upon lysine acetylation of AcsA by AcuA. Crystal structures of AcsA from Chloroflexota bacterium in the apo form and in complex with acetyl-adenosine-5'-monophosphate (acetyl-AMP) support the flexible C-terminal domain adopting different conformations. AlphaFold2 predictions suggest binding of AcuA stabilizes AcsA in an undescribed conformation. We show the AcuA•AcsA complex dissociates upon acetyl-coenzyme A (acetyl-CoA) dependent acetylation of AcsA by AcuA. We discover an intrinsic phosphotransacetylase activity enabling AcuA•AcsA generating acetyl-CoA from acetyl-phosphate (AcP) and coenzyme A (CoA) used by AcuA to acetylate and inactivate AcsA. Here, we provide mechanistic insights into the regulation of AMP-forming acetyl-CoA synthetases by lysine acetylation and discover an intrinsic phosphotransacetylase allowing modulation of its activity based on AcP and CoA levels.

Inhibition of Vascular Smooth Muscle Cell Proliferation by ENPP1: The Role of CD73 and the Adenosine Signaling Axis.

The Ectonucleotide Pyrophosphatase/Phosphodiesterase 1 (ENPP1) ectoenzyme regulates vascular intimal proliferation and mineralization of bone and soft tissues. ENPP1 variants cause Generalized Arterial Calcification of Infancy (GACI), a rare genetic disorder characterized by ectopic calcification, intimal proliferation, and stenosis of large- and medium-sized arteries. ENPP1 hydrolyzes extracellular ATP to pyrophosphate (PPi) and AMP. AMP is the precursor of adenosine, which has been implicated in the control of neointimal formation. Herein, we demonstrate that an ENPP1-Fc recombinant therapeutic inhibits proliferation of vascular smooth muscle cells (VSMCs) in vitro and in vivo. Addition of ENPP1 and ATP to cultured VSMCs generated AMP, which was metabolized to adenosine. It also significantly decreased cell proliferation. AMP or adenosine alone inhibited VSMC growth. Inhibition of ecto-5'-nucleotidase CD73 decreased adenosine accumulation and suppressed the anti-proliferative effects of ENPP1/ATP. Addition of AMP increased cAMP synthesis and phosphorylation of VASP at Ser157. This AMP-mediated cAMP increase was abrogated by CD73 inhibitors or by A2aR and A2bR antagonists. Ligation of the carotid artery promoted neointimal hyperplasia in wild-type mice, which was exacerbated in ENPP1-deficient ttw/ttw mice. Prophylactic or therapeutic treatments with ENPP1 significantly reduced intimal hyperplasia not only in ttw/ttw but also in wild-type mice. These findings provide the first insight into the mechanism of the anti-proliferative effect of ENPP1 and broaden its potential therapeutic applications beyond enzyme replacement therapy.

Remdesivir alleviates skin fibrosis by suppressing TGF-ß1 signaling pathway.

Fibrotic skin diseases, such as keloids, are pathological results of aberrant tissue healing and are characterized by overgrowth of dermal fibroblasts. Remdesivir (RD), an antiviral drug, has been reported to have pharmacological activities in a wide range of fibrotic diseases. However, whether RD function on skin fibrosis remains unclear. Therefore, in our study, we explored the potential effect and mechanisms of RD on skin fibrosis both in vivo and in vitro. As expected, the results demonstrated that RD alleviated BLM-induced skin fibrosis and attenuates the gross weight of keloid tissues in vivo. Further studies suggested that RD suppressed fibroblast activation and autophagy both in vivo and in vitro. In addition, mechanistic research showed that RD attenuated fibroblasts activation by the TGF-ß1/Smad signaling pathway and inhibited fibroblasts autophagy by the PI3K/Akt/mTOR signaling pathway. In summary, our results demonstrate therapeutic potential of RD for skin fibrosis in the future.

Kinetic and structural insights into the requirement of fungal tRNA ligase for a 2'-phosphate end.

Fungal RNA ligase (LIG) is an essential tRNA splicing enzyme that joins 3'-OH,2'-PO4 and 5'-PO4 RNA ends to form a 2'-PO4,3'-5' phosphodiester splice junction. Sealing entails three divalent cation-dependent adenylate transfer steps. First, LIG reacts with ATP to form a covalent ligase-(lysyl-Nζ)-AMP intermediate and displace pyrophosphate. Second, LIG transfers AMP to the 5'-PO4 RNA terminus to form an RNA-adenylate intermediate (A5'pp5'RNA). Third, LIG directs the attack of an RNA 3'-OH on AppRNA to form the splice junction and displace AMP. A defining feature of fungal LIG vis-à-vis canonical polynucleotide ligases is the requirement for a 2'-PO4 to synthesize a 3'-5' phosphodiester bond. Fungal LIG consists of an N-terminal adenylyltransferase domain and a unique C-terminal domain. The C-domain of Chaetomium thermophilum LIG (CthLIG) engages a sulfate anion thought to be a mimetic of the terminal 2'-PO4 Here, we interrogated the contributions of the C-domain and the conserved sulfate ligands (His227, Arg334, Arg337) to ligation of a pRNA2'p substrate. We find that the C-domain is essential for end-joining but dispensable for ligase adenylylation. Mutations H227A, R334A, and R337A slowed the rate of step 2 RNA adenylation by 420-fold, 120-fold, and 60-fold, respectively, vis-à-vis wild-type CthLIG. An R334A-R337A double-mutation slowed step 2 by 580-fold. These results fortify the case for the strictly conserved His-Arg-Arg triad as the enforcer of the 2'-PO4 end-specificity of fungal tRNA ligases and as a target for small molecule interdiction of fungal tRNA splicing.

Modulation of bitter taste receptors by yeast extracts.

Yeast extracts (YEs) are used in foods because of their flavour properties and ability to reduce bitterness. The adenosine 5'-monophosphate (AMP) found in YEs is known to decrease the bitterness of some compounds. This study aimed to investigate the ability of YEs to inhibit bitter taste receptors (TAS2Rs) using in vitro cell-based assays. A screen of TAS2Rs activated by AMP and YEs revealed that AMP and the AMP-rich YE activated more TAS2Rs. The inhibitory effect of the AMP-rich YE on seven TAS2Rs activated by bitter agonists was studied. YE reduced TAS2R activation, increased the EC50 value and decreased the maximum amplitude, demonstrating competitive and non-competitive inhibitions. Amongst the nineteen TAS2Rs tested, seven showed 40 % or greater inhibition after treatment of AMP-rich YE. Our data provide a better understanding of the TAS2R inhibition mechanism of AMP-rich YEs and promote their use as a strategy to reduce bitterness in foods and medicines.

Adenosine in Interventional Cardiology: Physiopathologic and Pharmacologic Effects in Coronary Artery Disease.

This review article focuses on the role of adenosine in coronary artery disease (CAD) diagnosis and treatment. Adenosine, an endogenous purine nucleoside, plays crucial roles in cardiovascular physiology and pathology. Its release and effects, mediated by specific receptors, influence vasomotor function, blood pressure regulation, heart rate, and platelet activity. Adenosine therapeutic effects include treatment of the no-reflow phenomenon and paroxysmal supraventricular tachycardia. The production of adenosine involves complex cellular pathways, with extracellular and intracellular synthesis mechanisms. Adenosine's rapid metabolism underscores its short half-life and physiological turnover. Furthermore, adenosine's involvement in side effects of antiplatelet therapy, particularly ticagrelor and cangrelor, highlights its clinical significance. Moreover, adenosine serves as a valuable tool in CAD diagnosis, aiding stress testing modalities and guiding intracoronary physiological assessments. Its use in assessing epicardial stenosis and microvascular dysfunction is pivotal for treatment decisions. Overall, understanding adenosine's mechanisms and clinical implications is essential for optimizing CAD management strategies, encompassing both therapeutic interventions and diagnostic approaches.

Development of a Menglà virus minigenome and comparison of its polymerase complexes with those of other filoviruses.

Ebola virus (EBOV) and Marburg virus (MARV), members of the Filoviridae family, are highly pathogenic and can cause hemorrhagic fevers, significantly impacting human society. Bats are considered reservoirs of these viruses because related filoviruses have been discovered in bats. However, due to the requirement for maximum containment laboratories when studying infectious viruses, the characterization of bat filoviruses often relies on pseudoviruses and minigenome systems. In this study, we used RACE technology to sequence the 3'-leader and 5'-trailer of Menglà virus (MLAV) and constructed a minigenome. Similar to MARV, the transcription activities of the MLAV minigenome are independent of VP30. We further assessed the effects of polymorphisms at the 5' end on MLAV minigenome activity and identified certain mutations that decrease minigenome reporter efficiency, probably due to alterations in the RNA secondary structure. The reporter activity upon recombination of the 3'-leaders and 5'-trailers of MLAV, MARV, and EBOV with those of the homologous or heterologous minigenomes was compared and it was found that the polymerase complex and leader and trailer sequences exhibit intrinsic specificities. Additionally, we investigated whether the polymerase complex proteins from EBOV and MARV support MLAV minigenome RNA synthesis and found that the homologous system is more efficient than the heterologous system. Remdesivir efficiently inhibited MLAV as well as EBOV replication. In summary, this study provides new information on bat filoviruses and the minigenome will be a useful tool for high-throughput antiviral drug screening.

Differential Bioactivation Profiles of Different GS-441524 Prodrugs in Cell and Mouse Models: ProTide Prodrugs with High Cell Permeability and Susceptibility to Cathepsin A Are More Efficient in Delivering Antiviral Active Metabolites to the Lung.

We assessed factors that determine the tissue-specific bioactivation of ProTide prodrugs by comparing the disposition and activation of remdesivir (RDV), its methylpropyl and isopropyl ester analogues (MeRDV and IsoRDV, respectively), the oral prodrug GS-621763, and the parent nucleotide GS-441524 (Nuc). RDV and MeRDV yielded more active metabolite remdesivir-triphosphate (RDV-TP) than IsoRDV, GS-621763, and Nuc in human lung cell models due to superior cell permeability and higher susceptivity to cathepsin A. Intravenous administration to mice showed that RDV and MeRDV delivered significantly more RDV-TP to the lung than other compounds. Nevertheless, all four ester prodrugs exhibited very low oral bioavailability (<2%), with Nuc being the predominant metabolite in blood. In conclusion, ProTides prodrugs, such as RDV and MeRDV, are more efficient in delivering active metabolites to the lung than Nuc, driven by high cell permeability and susceptivity to cathepsin A. Optimizing ProTides' ester structures is an effective strategy for enhancing prodrug activation in the lung.

Transcriptome dynamics of the BHK21 cell line in response to human coronavirus OC43 infection.

The progress of viral infection involves numerous transcriptional regulatory events. The identification of the newly synthesized transcripts helps us to understand the replication mechanisms and pathogenesis of the virus. Here, we utilized a time-resolved technique called metabolic RNA labeling approach called thiol(SH)-linked alkylation for the metabolic sequencing of RNA (SLAM-seq) to differentially elucidate the levels of steady-state and newly synthesized RNAs of BHK21 cell line in response to human coronavirus OC43 (HCoV-OC43) infection. Our results showed that the Wnt/ß-catenin signaling pathway was significantly enriched with the newly synthesized transcripts of BHK21 cell line in response to HCoV-OC43 infection. Moreover, inhibition of the Wnt pathway promoted viral replication in the early stage of infection, but inhibited it in the later stage of infection. Furthermore, remdesivir inhibits the upregulation of the Wnt/ß-catenin signaling pathway induced by early infection with HCoV-OC43. Collectively, our study showed the diverse roles of Wnt/ß-catenin pathway at different stages of HCoV-OC43 infection, suggesting a potential target for the antiviral treatment. In addition, although infection with HCoV-OC43 induces cytopathic effects in BHK21 cells, inhibiting apoptosis does not affect the intracellular replication of the virus. Monitoring newly synthesized RNA based on such time-resolved approach is a highly promising method for studying the mechanism of viral infections.

Legionella effector LnaB is a phosphoryl-AMPylase that impairs phosphosignalling

AMPylation is a post-translational modification in which AMP is added to the amino acid side chains of proteins1,2. Here we show that, with ATP as the ligand and actin as the host activator, the effector protein LnaB of Legionella pneumophila exhibits AMPylase activity towards the phosphoryl group of phosphoribose on PRR42-Ub that is generated by the SidE family of effectors, and deubiquitinases DupA and DupB in an E1- and E2-independent ubiquitination process3-7. The product of LnaB is further hydrolysed by an ADP-ribosylhydrolase, MavL, to Ub, thereby preventing the accumulation of PRR42-Ub and ADPRR42-Ub and protecting canonical ubiquitination in host cells. LnaB represents a large family of AMPylases that adopt a common structural fold, distinct from those of the previously known AMPylases, and LnaB homologues are found in more than 20 species of bacterial pathogens. Moreover, LnaB also exhibits robust phosphoryl AMPylase activity towards phosphorylated residues and produces unique ADPylation modifications in proteins. During infection, LnaB AMPylates the conserved phosphorylated tyrosine residues in the activation loop of the Src family of kinases8,9, which dampens downstream phosphorylation signalling in the host. Structural studies reveal the actin-dependent activation and catalytic mechanisms of the LnaB family of AMPylases. This study identifies, to our knowledge, an unprecedented molecular regulation mechanism in bacterial pathogenesis and protein phosphorylation.

Kinetic analysis of T4 polynucleotide kinase via isothermal titration calorimetry.

T4 polynucleotide kinase (T4 PNK) phosphorylates the 5'-terminus of DNA and RNA substrates. It is widely used in molecular biology. Single nucleotides can serve as substrates if a 3'-phosphate group is present. In this study, the T4 PNK-catalyzed conversion of adenosine 3'-monophosphate (3'-AMP) to adenosine-3',5'-bisphosphate was characterized using isothermal titration calorimetry (ITC). Although ITC is typically used to study ligand binding, in this case the instrument was used to evaluate enzyme kinetics by monitoring the heat production due to reaction enthalpy. The reaction was initiated with a single injection of 3'-AMP substrate into the sample cell containing T4 PNK and ATP at pH 7.6 and 30 °C, and Michaelis-Menten analysis was performed on the reaction rates derived from the plot of differential power versus time. The Michaelis-Menten constant, KM, was 13 µM, and the turnover number, kcat, was 8 s-1. The effect of inhibitors was investigated using pyrophosphate (PPi). PPi caused a dose-dependent decrease in the apparent kcat and increase in the apparent KM under the conditions tested. Additionally, the intrinsic reaction enthalpy and the activation energy of the T4 PNK-catalyzed phosphorylation of 3'-AMP were determined to be -25 kJ/mol and 43 kJ/mol, respectively. ITC is seldom used as a tool to study enzyme kinetics, particularly for technically-challenging enzymes such as kinases. This study demonstrates that quantitative analysis of kinase activity can be amenable to the ITC single injection approach.

Visualizing ribonuclease digestion of RNA-like polymers produced by hot wet-dry cycles.

Polymerization of nucleotides under prebiotic conditions simulating the early Earth has been extensively studied. Several independent methods have been used to verify that RNA-like polymers can be produced by hot wet-dry cycling of nucleotides. However, it has not been shown that these RNA-like polymers are similar to biological RNA with 3'-5' phosphodiester bonds. In the results described here, RNA-like polymers were generated from 5'-monophosphate nucleosides AMP and UMP. To confirm that the polymers resemble biological RNA, ribonuclease A should catalyze hydrolysis of the 3'-5' phosphodiester bonds between pyrimidine nucleotides to each other or to purine nucleotides, but not purine-purine nucleotide bonds. Here we show AFM images of specific polymers produced by hot wet-dry cycling of AMP, UMP and AMP/UMP (1:1) solutions on mica surfaces, before and after exposure to ribonuclease A. AMP polymers were unaffected by ribonuclease A but UMP polymers disappeared. This indicates that a major fraction of the bonds in the UMP polymers is indeed 3'-5' phosphodiester bonds. Some of the polymers generated from the AMP/UMP mixture also showed clear signs of cleavage. Because ribonuclease A recognizes the ester bonds in the polymers, we show for the first time that these prebiotically produced polymers are in fact similar to biological RNA but are likely to be linked by a mixture of 3'-5' and 2'-5' phosphodiester bonds.

In situ enzymatic control of colloidal phoresis and catalysis through hydrolysis of ATP.

The ability to sense chemical gradients and respond with directional motility and chemical activity is a defining feature of complex living systems. There is a strong interest among scientists to design synthetic systems that emulate these properties. Here, we realize and control such behaviors in a synthetic system by tailoring multivalent interactions of adenosine nucleotides with catalytic microbeads. We first show that multivalent interactions of the bead with gradients of adenosine mono-, di- and trinucleotides (AM/D/TP) control both the phoretic motion and a proton-transfer catalytic reaction, and find that both effects are diminished greatly with increasing valence of phosphates. We exploit this behavior by using enzymatic hydrolysis of ATP to AMP, which downregulates multivalent interactivity in situ. This produces a sudden increase in transport of the catalytic microbeads (a phoretic jump), which is accompanied by increased catalytic activity. Finally, we show how this enzymatic activity can be systematically tuned, leading to simultaneous in situ spatial and temporal control of the location of the microbeads, as well as the products of the reaction that they catalyze. These findings open up new avenues for utilizing multivalent interaction-mediated programming of complex chemo-mechanical behaviors into active systems.

Systematic Analysis of the Effect of Genomic Knock-Out of Non-Essential Promiscuous HAD-Like Phosphatases YcsE, YitU and YwtE on Flavin and Adenylate Content in Bacillus Subtilis.

Studying the metabolic role of non-essential promiscuous enzymes is a challenging task, as genetic manipulations usually do not reveal at which point(s) of the metabolic network the enzymatic activity of such protein is beneficial for the organism. Each of the HAD-like phosphatases YcsE, YitU and YwtE of Bacillus subtilis catalyzes the dephosphorylation of 5-amino-6-ribitylamino-uracil 5'-phosphate, which is essential in the biosynthesis of riboflavin. Using CRISPR technology, we have found that the deletion of these genes, individually or in all possible combinations failed to cause riboflavin auxotrophy and did not result in significant growth changes. Analysis of flavin and adenylate content in B. subtilis knockout mutants showed that (i) there must be one or several still unidentified phosphatases that can replace the deleted proteins; (ii) such replacements, however, cannot fully restore the intracellular content of any of three flavins studied (riboflavin, FMN, FAD); (iii) whereas bacterial fitness was not significantly compromised by mutations, the intracellular balance of flavins and adenylates did show some significant changes.

Discovery and validation of a novel inhibitor of HYPE-mediated AMPylation.

Adenosyl monophosphate (AMP)ylation (the covalent transfer of an AMP from Adenosine Triphosphate (ATP) onto a target protein) is catalyzed by the human enzyme Huntingtin Yeast Interacting Partner E (HYPE)/FicD to regulate its substrate, the heat shock chaperone binding immunoglobulin protein (BiP). HYPE-mediated AMPylation of BiP is critical for maintaining proteostasis in the endoplasmic reticulum and mounting a unfolded protein response in times of proteostatic imbalance. Thus, manipulating HYPE's enzymatic activity is a key therapeutic strategy toward the treatment of various protein misfolding diseases, including neuropathy and early-onset diabetes associated with two recently identified clinical mutations of HYPE. Herein, we present an optimized, fluorescence polarization-based, high-throughput screening (HTS) assay to discover activators and inhibitors of HYPE-mediated AMPylation. After challenging our HTS assay with over 30,000 compounds, we discovered a novel AMPylase inhibitor, I2.10. We also determined a low micromolar IC50 for I2.10 and employed biorthogonal counter-screens to validate its efficacy against HYPE's AMPylation of BiP. Further, we report low cytotoxicity of I2.10 on human cell lines. We thus established an optimized, high-quality HTS assay amenable to tracking HYPE's enzymatic activity at scale, and provided the first novel small-molecule inhibitor capable of perturbing HYPE-directed AMPylation of BiP in vitro. Our HTS assay and I2.10 compound serve as a platform for further development of HYPE-specific small-molecule therapeutics.

GS-441524-Diphosphate-Ribose Derivatives as Nanomolar Binders and Fluorescence Polarization Tracers for SARS-CoV-2 and Other Viral Macrodomains.

Viral macrodomains that can bind to or hydrolyze protein adenosine diphosphate ribosylation (ADP-ribosylation) have emerged as promising targets for antiviral drug development. Many inhibitor development efforts have been directed against the severe acute respiratory syndrome coronavirus 2 macrodomain 1 (SARS-CoV-2 Mac1). However, potent inhibitors for viral macrodomains are still lacking, with the best inhibitors still in the micromolar range. Based on GS-441524, a remdesivir precursor, and our previous studies, we have designed and synthesized potent binders of SARS-CoV-2 Mac1 and other viral macrodomains including those of Middle East respiratory syndrome coronavirus (MERS-CoV), Venezuelan equine encephalitis virus (VEEV), and Chikungunya virus (CHIKV). We show that the 1'-CN group of GS-441524 promotes binding to all four viral macrodomains tested while capping the 1″-OH of GS-441524-diphosphate-ribose with a simple phenyl ring further contributes to binding. Incorporating these two structural features, the best binders show 20- to 6000-fold increases in binding affinity over ADP-ribose for SARS-CoV-2, MERS-CoV, VEEV, and CHIKV macrodomains. Moreover, building on these potent binders, we have developed two highly sensitive fluorescence polarization tracers that only require nanomolar proteins and can effectively resolve the binding affinities of nanomolar inhibitors. Our findings and probes described here will facilitate future development of more potent viral macrodomain inhibitors.

Protein feature engineering framework for AMPylation site prediction.

AMPylation is a biologically significant yet understudied post-translational modification where an adenosine monophosphate (AMP) group is added to Tyrosine and Threonine residues primarily. While recent work has illuminated the prevalence and functional impacts of AMPylation, experimental identification of AMPylation sites remains challenging. Computational prediction techniques provide a faster alternative approach. The predictive performance of machine learning models is highly dependent on the features used to represent the raw amino acid sequences. In this work, we introduce a novel feature extraction pipeline to encode the key properties relevant to AMPylation site prediction. We utilize a recently published dataset of curated AMPylation sites to develop our feature generation framework. We demonstrate the utility of our extracted features by training various machine learning classifiers, on various numerical representations of the raw sequences extracted with the help of our framework. Tenfold cross-validation is used to evaluate the model's capability to distinguish between AMPylated and non-AMPylated sites. The top-performing set of features extracted achieved MCC score of 0.58, Accuracy of 0.8, AUC-ROC of 0.85 and F1 score of 0.73. Further, we elucidate the behaviour of the model on the set of features consisting of monogram and bigram counts for various representations using SHapley Additive exPlanations.

AMP-activated protein kinase can be allosterically activated by ADP but AMP remains the key activating ligand.

The AMP-activated protein kinase (AMPK) is a sensor of cellular energy status. When activated by increases in ADP:ATP and/or AMP:ATP ratios (signalling energy deficit), AMPK acts to restore energy balance. Binding of AMP to one or more of three CBS repeats (CBS1, CBS3, CBS4) on the AMPK-γ subunit activates the kinase complex by three complementary mechanisms: (i) promoting α-subunit Thr172 phosphorylation by the upstream kinase LKB1; (ii) protecting against Thr172 dephosphorylation; (iii) allosteric activation. Surprisingly, binding of ADP has been reported to mimic the first two effects, but not the third. We now show that at physiologically relevant concentrations of Mg.ATP2- (above those used in the standard assay) ADP binding does cause allosteric activation. However, ADP causes only a modest activation because (unlike AMP), at concentrations just above those where activation becomes evident, ADP starts to cause competitive inhibition at the catalytic site. Our results cast doubt on the physiological relevance of the effects of ADP and suggest that AMP is the primary activator in vivo. We have also made mutations to hydrophobic residues involved in binding adenine nucleotides at each of the three γ subunit CBS repeats of the human α2ß2γ1 complex and examined their effects on regulation by AMP and ADP. Mutation of the CBS3 site has the largest effects on all three mechanisms of AMP activation, especially at lower ATP concentrations, while mutation of CBS4 reduces the sensitivity to AMP. All three sites appear to be required for allosteric activation by ADP.