Accurate de novo design of membrane-traversing macrocycles.

We use computational design coupled with experimental characterization to systematically investigate the design principles for macrocycle membrane permeability and oral bioavailability. We designed 184 6-12 residue macrocycles with a wide range of predicted structures containing noncanonical backbone modifications and experimentally determined structures of 35; 29 are very close to the computational models. With such control, we show that membrane permeability can be systematically achieved by ensuring all amide (NH) groups are engaged in internal hydrogen bonding interactions. 84 designs over the 6-12 residue size range cross membranes with an apparent permeability greater than 1 × 10-6 cm/s. Designs with exposed NH groups can be made membrane permeable through the design of an alternative isoenergetic fully hydrogen-bonded state favored in the lipid membrane. The ability to robustly design membrane-permeable and orally bioavailable peptides with high structural accuracy should contribute to the next generation of designed macrocycle therapeutics.

Copper-catalysed dehydrogenation or lactonization of C(sp3)-H bonds.

Cytochrome P450 enzymes are known to catalyse bimodal oxidation of aliphatic acids via radical intermediates, which partition between pathways of hydroxylation and desaturation1,2. Developing analogous catalytic systems for remote C-H functionalization remains a significant challenge3-5. Here, we report the development of Cu(I)-catalysed bimodal dehydrogenation/lactonization reactions of synthetically common N-methoxyamides through radical abstractions of the γ-aliphatic C-H bonds. The feasibility of switching from dehydrogenation to lactonization is also demonstrated by altering reaction conditions. The use of a readily available amide as both radical precursor and internal oxidant allows for the development of redox-neutral C-H functionalization reactions with methanol as the sole side product. These C-H functionalization reactions using a Cu(I) catalyst with loading as low as 0.5 mol.% is applied to the diversification of a wide range of aliphatic acids including drug molecules and natural products. The exceptional compatibility of this catalytic system with a wide range of oxidatively sensitive functionality demonstrates the unique advantage of using a simple amide substrate as a mild internal oxidant.

Reverse metabolomics for the discovery of chemical structures from humans.

Determining the structure and phenotypic context of molecules detected in untargeted metabolomics experiments remains challenging. Here we present reverse metabolomics as a discovery strategy, whereby tandem mass spectrometry spectra acquired from newly synthesized compounds are searched for in public metabolomics datasets to uncover phenotypic associations. To demonstrate the concept, we broadly synthesized and explored multiple classes of metabolites in humans, including N-acyl amides, fatty acid esters of hydroxy fatty acids, bile acid esters and conjugated bile acids. Using repository-scale analysis1,2, we discovered that some conjugated bile acids are associated with inflammatory bowel disease (IBD). Validation using four distinct human IBD cohorts showed that cholic acids conjugated to Glu, Ile/Leu, Phe, Thr, Trp or Tyr are increased in Crohn's disease. Several of these compounds and related structures affected pathways associated with IBD, such as interferon-γ production in CD4+ T cells3 and agonism of the pregnane X receptor4. Culture of bacteria belonging to the Bifidobacterium, Clostridium and Enterococcus genera produced these bile amidates. Because searching repositories with tandem mass spectrometry spectra has only recently become possible, this reverse metabolomics approach can now be used as a general strategy to discover other molecules from human and animal ecosystems.

Enantioconvergent Cu-catalysed N-alkylation of aliphatic amines.

Chiral amines are commonly used in the pharmaceutical and agrochemical industries1. The strong demand for unnatural chiral amines has driven the development of catalytic asymmetric methods1,2. Although the N-alkylation of aliphatic amines with alkyl halides has been widely adopted for over 100 years, catalyst poisoning and unfettered reactivity have been preventing the development of a catalyst-controlled enantioselective version3-5. Here we report the use of chiral tridentate anionic ligands to enable the copper-catalysed chemoselective and enantioconvergent N-alkylation of aliphatic amines with α-carbonyl alkyl chlorides. This method can directly convert feedstock chemicals, including ammonia and pharmaceutically relevant amines, into unnatural chiral α-amino amides under mild and robust conditions. Excellent enantioselectivity and functional-group tolerance were observed. The power of the method is demonstrated in a number of complex settings, including late-stage functionalization and in the expedited synthesis of diverse amine drug molecules. The current method indicates that multidentate anionic ligands are a general solution for overcoming transition-metal-catalyst poisoning.

Coronavirus RNA Proofreading: Molecular Basis and Therapeutic Targeting.

The coronavirus disease 2019 (COVID-19) that is wreaking havoc on worldwide public health and economies has heightened awareness about the lack of effective antiviral treatments for human coronaviruses (CoVs). Many current antivirals, notably nucleoside analogs (NAs), exert their effect by incorporation into viral genomes and subsequent disruption of viral replication and fidelity. The development of anti-CoV drugs has long been hindered by the capacity of CoVs to proofread and remove mismatched nucleotides during genome replication and transcription. Here, we review the molecular basis of the CoV proofreading complex and evaluate its potential as a drug target. We also consider existing nucleoside analogs and novel genomic techniques as potential anti-CoV therapeutics that could be used individually or in combination to target the proofreading mechanism.

Skeletal editing through direct nitrogen deletion of secondary amines.

Synthetic chemistry aims to build up molecular complexity from simple feedstocks1. However, the ability to exert precise changes that manipulate the connectivity of the molecular skeleton itself remains limited, despite possessing substantial potential to expand the accessible chemical space2,3. Here we report a reaction that 'deletes' nitrogen from organic molecules. We show that N-pivaloyloxy-N-alkoxyamides, a subclass of anomeric amides, promote the intermolecular activation of secondary aliphatic amines to yield intramolecular carbon-carbon coupling products. Mechanistic experiments indicate that the reactions proceed via isodiazene intermediates that extrude the nitrogen atom as dinitrogen, producing short-lived diradicals that rapidly couple to form the new carbon-carbon bond. The reaction shows broad functional-group tolerance, which enables the translation of routine amine synthesis protocols into a strategy for carbon-carbon bond constructions and ring syntheses. This is highlighted by the use of this reaction in the syntheses and skeletal editing of bioactive compounds.

Photoinduced copper-catalysed asymmetric amidation via ligand cooperativity.

The substitution of an alkyl electrophile by a nucleophile is a foundational reaction in organic chemistry that enables the efficient and convergent synthesis of organic molecules. Although there has been substantial recent progress in exploiting transition-metal catalysis to expand the scope of nucleophilic substitution reactions to include carbon nucleophiles1-4, there has been limited progress in corresponding reactions with nitrogen nucleophiles5-8. For many substitution reactions, the bond construction itself is not the only challenge, as there is a need to control stereochemistry at the same time. Here we describe a method for the enantioconvergent substitution of unactivated racemic alkyl electrophiles by a ubiquitous nitrogen-containing functional group, an amide. Our method uses a photoinduced catalyst system based on copper, an Earth-abundant metal. This process for asymmetric N-alkylation relies on three distinct ligands-a bisphosphine, a phenoxide and a chiral diamine. The ligands assemble in situ to form two distinct catalysts that act cooperatively: a copper/bisphosphine/phenoxide complex that serves as a photocatalyst, and a chiral copper/diamine complex that catalyses enantioselective C-N bond formation. Our study thus expands enantioselective N-substitution by alkyl electrophiles beyond activated electrophiles (those bearing at least one sp- or sp2-hybridized substituent on the carbon undergoing substitution)8-13 to include unactivated electrophiles.

Discovery, characterization and engineering of ligases for amide synthesis.

Coronatine and related bacterial phytotoxins are mimics of the hormone jasmonyl-L-isoleucine (JA-Ile), which mediates physiologically important plant signalling pathways1-4. Coronatine-like phytotoxins disrupt these essential pathways and have potential in the development of safer, more selective herbicides. Although the biosynthesis of coronatine has been investigated previously, the nature of the enzyme that catalyses the crucial coupling of coronafacic acid to amino acids remains unknown1,2. Here we characterize a family of enzymes, coronafacic acid ligases (CfaLs), and resolve their structures. We found that CfaL can also produce JA-Ile, despite low similarity with the Jar1 enzyme that is responsible for ligation of JA and L-Ile in plants5. This suggests that Jar1 and CfaL evolved independently to catalyse similar reactions-Jar1 producing a compound essential for plant development4,5, and the bacterial ligases producing analogues toxic to plants. We further demonstrate how CfaL enzymes can be used to synthesize a diverse array of amides, obviating the need for protecting groups. Highly selective kinetic resolutions of racemic donor or acceptor substrates were achieved, affording homochiral products. We also used structure-guided mutagenesis to engineer improved CfaL variants. Together, these results show that CfaLs can deliver a wide range of amides for agrochemical, pharmaceutical and other applications.

A facile approach for incorporating tyrosine esters to probe ion-binding sites and backbone hydrogen bonds.

Amide-to-ester substitutions are used to study the role of the amide bonds of the protein backbone in protein structure, function, and folding. An amber suppressor tRNA/synthetase pair has been reported for incorporation of p-hydroxy-phenyl-L-lactic acid (HPLA), thereby introducing ester substitution at tyrosine residues. However, the application of this approach was limited due to the low yields of the modified proteins and the high cost of HPLA. Here we report the in vivo generation of HPLA from the significantly cheaper phenyl-L-lactic acid. We also construct an optimized plasmid with the HPLA suppressor tRNA/synthetase pair that provides higher yields of the modified proteins. The combination of the new plasmid and the in-situ generation of HPLA provides a facile and economical approach for introducing tyrosine ester substitutions. We demonstrate the utility of this approach by introducing tyrosine ester substitutions into the K+ channel KcsA and the integral membrane enzyme GlpG. We introduce the tyrosine ester in the selectivity filter of the M96V mutant of the KcsA to probe the role of the second ion binding site in the conformation of the selectivity filter and the process of inactivation. We use tyrosine ester substitutions in GlpG to perturb backbone H-bonds to investigate the contribution of these H-bonds to membrane protein stability. We anticipate that the approach developed in this study will facilitate further investigations using tyrosine ester substitutions.

Using 2D-IR Spectroscopy to Measure the Structure, Dynamics, and Intermolecular Interactions of Proteins in H2O.

Infrared (IR) spectroscopy probes molecular structure at the level of the chemical bond or functional group. In the case of proteins, the most informative band in the IR spectrum is the amide I band, which arises predominantly from the C═O stretching vibration of the peptide link. The folding of proteins into secondary and tertiary structures leads to vibrational coupling between peptide units, generating specific amide I spectral signatures that provide a fingerprint of the macromolecular conformation. Ultrafast two-dimensional IR (2D-IR) spectroscopy allows the amide I band of a protein to be spread over a second frequency dimension in a way that mirrors 2D-NMR methods. This means that amide I 2D-IR spectroscopy produces a spectral map that is exquisitely sensitive to protein structure and dynamics and so provides detailed insights that cannot be matched by IR absorption spectroscopy. As a result, 2D-IR spectroscopy has emerged as a powerful tool for probing protein structure and dynamics over a broad range of time and length scales in the solution phase at room temperature. However, the protein amide I band coincides with an IR absorption from the bending vibration of water (δHOH), the natural biological solvent. To circumvent this problem, protein IR studies are routinely performed in D2O solutions because H/D substitution shifts the solvent bending mode (δDOD) to a lower frequency, revealing the amide I band. While effective, this method raises fundamental questions regarding the impact of the change in solvent mass on the structural or solvation dynamics of the protein and the removal of the energetic resonance between solvent and solute.In this Account, a series of studies applying 2D-IR to study the spectroscopy and dynamics of proteins in H2O-rich solvents is reviewed. A comparison of IR absorption spectroscopy and 2D-IR spectroscopy of protein-containing fluids is used to demonstrate the basis of the approach before a series of applications is presented. These range from measurements of fundamental protein biophysics to recent applications of machine learning to gain insight into protein-drug binding in complex mixtures. An outlook is presented, considering the potential for 2D-IR measurements to contribute to our understanding of protein behavior under near-physiological conditions, along with an evaluation of the obstacles that still need to be overcome.

Ynamide Coupling Reagents: Origin and Advances.

Since the pioneering work of Curtius and Fischer, chemical peptide synthesis has witnessed a century's development and evolved into a routine technology. However, it is far from perfect. In particular, it is challenged by sustainable development because the state-of-the-art of peptide synthesis heavily relies on legacy reagents and technologies developed before the establishment of green chemistry. Over the past three decades, a broad range of efforts have been made for greening peptide synthesis, among which peptide synthesis using unprotected amino acid represents an ideal and promising strategy because it does not require protection and deprotection steps. Unfortunately, C → N peptide synthesis employing unprotected amino acids has been plagued by undesired polymerization, while N → C inverse peptide synthesis with unprotected amino acids is retarded by severe racemization/epimerization owing to the iterative activation and aminolysis of high racemization/epimerization susceptible peptidyl acids. Consequently, there is an urgent need to develop innovative coupling reagents and strategies with novel mechanisms that can address the long-standing notorious racemization/epimerization issue of peptide synthesis.This Account will describe our efforts in discovery of ynamide coupling reagents and their application in greening peptide synthesis. Over an eight-year journey, ynamide coupling reagents have evolved into a class of general coupling reagents for both amide and ester bond formation. In particular, the superiority of ynamide coupling reagents in suppressing racemization/epimerization enabled them to be effective for peptide fragment condensation, and head-to-tail cyclization, as well as precise incorporation of thioamide substitutions into peptide backbones. The first practical inverse peptide synthesis using unprotected amino acids was successfully accomplished by harnessing such features and taking advantage of a transient protection strategy. Ynamide coupling reagent-mediated ester bond formation enabled efficient intermolecular esterification and macrolactonization with preservation of α-chirality and the configuration of the conjugated α,ß-C-C double bond. To make ynamide coupling reagents readily available with reasonable cost and convenience, we have developed a scalable one-step synthetic method from cheap starting materials. Furthermore, a water-removable ynamide coupling reagent was developed, offering a column-free purification of the target coupling product. In addition, the recycle of ynamide coupling reagent was accomplished, thereby paving the way for their sustainable industrial application.As such, this Account presents the whole story of the origin, mechanistic insights, preparation, synthetic applications, and recycle of ynamide coupling reagents with a perspective that highlights their future impact on peptide synthesis.

Straightforward access to N-trifluoromethyl amides, carbamates, thiocarbamates and ureas.

Amides and related carbonyl derivatives are of central importance across the physical and life sciences1,2. As a key biological building block, the stability and conformation of amides affect the structures of peptides and proteins as well as their biological function. In addition, amide-bond formation is one of the most frequently used chemical transformations3,4. Given their ubiquity, a technology that is capable of modifying the fundamental properties of amides without compromising on stability may have considerable potential in pharmaceutical, agrochemical and materials science. In order to influence the physical properties of organic molecules-such as solubility, lipophilicity, conformation, pKa and (metabolic) stability-fluorination approaches have been widely adopted5-7. Similarly, site-specific modification with isosteres and peptidomimetics8, or in particular by N-methylation9, has been used to improve the stability, physical properties, bioactivities and cellular permeabilities of compounds. However, the N-trifluoromethyl carbonyl motif-which combines both N-methylation and fluorination approaches-has not yet been explored, owing to a lack of efficient methodology to synthesize it. Here we report a straightforward method to access N-trifluoromethyl analogues of amides and related carbonyl compounds. The strategy relies on the operationally simple preparation of bench-stable carbamoyl fluoride building blocks, which can be readily diversified to the corresponding N-CF3 amides, carbamates, thiocarbamates and ureas. This method tolerates rich functionality and stereochemistry, and we present numerous examples of highly functionalized compounds-including analogues of widely used drugs, antibiotics, hormones and polymer units.

In vitro investigation of protein assembly by combined microscopy and infrared spectroscopy at the nanometer scale.

The nanoscale structure and dynamics of proteins on surfaces has been extensively studied using various imaging techniques, such as transmission electron microscopy and atomic force microscopy (AFM) in liquid environments. These powerful imaging techniques, however, can potentially damage or perturb delicate biological material and do not provide chemical information, which prevents a fundamental understanding of the dynamic processes underlying their evolution under physiological conditions. Here, we use a platform developed in our laboratory that enables acquisition of infrared (IR) spectroscopy and AFM images of biological material in physiological liquids with nanometer resolution in a cell closed by atomically thin graphene membranes transparent to IR photons. In this work, we studied the self-assembly process of S-layer proteins at the graphene-aqueous solution interface. The graphene acts also as the membrane separating the solution containing the proteins and Ca2+ ions from the AFM tip, thus eliminating sample damage and contamination effects. The formation of S-layer protein lattices and their structural evolution was monitored by AFM and by recording the amide I and II IR absorption bands, which reveal the noncovalent interaction between proteins and their response to the environment, including ionic strength and solvation. Our measurement platform opens unique opportunities to study biological material and soft materials in general.

Acyl Capping Group Identity Effects on α-Helicity: On the Importance of Amide·Water Hydrogen Bonds to α-Helix Stability.

Acyl capping groups stabilize α-helices relative to free N-termini by providing one additional C═Oi···Hi+4-N hydrogen bond. The electronic properties of acyl capping groups might also directly modulate α-helix stability: electron-rich N-terminal acyl groups could stabilize the α-helix by strengthening both i/i + 4 hydrogen bonds and i/i + 1 n → π* interactions. This hypothesis was tested in peptides X-AKAAAAKAAAAKAAGY-NH2, where X = different acyl groups. Surprisingly, the most electron-rich acyl groups (pivaloyl and iso-butyryl) strongly destabilized the α-helix. Moreover, the formyl group induced nearly identical α-helicity to that of the acetyl group, despite being a weaker electron donor for hydrogen bonds and for n → π* interactions. Other acyl groups exhibited intermediate α-helicity. These results indicate that the electronic properties of the acyl carbonyl do not directly determine the α-helicity in peptides in water. In order to understand these effects, DFT calculations were conducted on α-helical peptides. Using implicit solvation, α-helix stability correlated with acyl group electronics, with the pivaloyl group exhibiting closer hydrogen bonds and n → π* interactions, in contrast to the experimental results. However, DFT and MD calculations with explicit water solvation revealed that hydrogen bonding to water was impacted by the sterics of the acyl capping group. Formyl capping groups exhibited the closest water-amide hydrogen bonds, while pivaloyl groups exhibited the longest. In α-helices in the PDB, the highest frequency of close amide-water hydrogen bonds is observed when the N-cap residue is Gly. The combination of experimental and computational results indicates that solvation (hydrogen bonding of water) to the N-terminal amide groups is a central determinant of α-helix stability.

Regiochemical Analysis of the ProTide Activation Mechanism.

ProTides are nucleotide analogues used for the treatment of specific viral infections. These compounds consist of a masked nucleotide that undergoes in vivo enzymatic and spontaneous chemical transformations to generate a free mononucleotide that is ultimately transformed to the pharmaceutically active triphosphorylated drug. The three FDA approved ProTides are composed of a phosphoramidate (P-N) core coupled with a nucleoside analogue, phenol, and an l-alanyl carboxylate ester. The previously proposed mechanism of activation postulates the existence of an unstable 5-membered mixed anhydride cyclic intermediate formed from the direct attack of the carboxylate group of the l-alanyl moiety with expulsion of phenol. The mixed anhydride cyclic intermediate is further postulated to undergo spontaneous hydrolysis to form a linear l-alanyl phosphoramidate product. In the proposed mechanism of activation, the 5-membered mixed anhydride intermediate has been detected previously using mass spectrometry, but the specific site of nucleophilic attack by water (P-O versus C-O) has not been determined. To further interrogate the mechanism for hydrolysis of the putative 5-membered cyclic intermediate formed during ProTide activation, the reaction was conducted in 18O-labeled water using a ProTide analogue that could be activated by carboxypeptidase Y. Mass spectrometry and 31P NMR spectroscopy were used to demonstrate that the hydrolysis of the mixed anhydride 5-membered intermediate occurs with exclusive attack at the phosphorus center.

Identification and Characterization of a Bacterial Periplasmic Solute Binding Protein That Binds l-Amino Acid Amides.

Periplasmic solute-binding proteins (SBPs) are key ligand recognition components of bacterial ATP-binding cassette (ABC) transporters that allow bacteria to import nutrients and metabolic precursors from the environment. Periplasmic SBPs comprise a large and diverse family of proteins, of which only a small number have been empirically characterized. In this work, we identify a set of 610 unique uncharacterized proteins within the SBP_bac_5 family that are found in conserved operons comprising genes encoding (i) ABC transport systems and (ii) putative amidases from the FmdA_AmdA family. From these uncharacterized SBP_bac_5 proteins, we characterize a representative periplasmic SBP from Mesorhizobium sp. A09 (MeAmi_SBP) and show that MeAmi_SBP binds l-amino acid amides but not the corresponding l-amino acids. An X-ray crystal structure of MeAmi_SBP bound to l-serinamide highlights the residues that impart distinct specificity for l-amino acid amides and reveals a structural Ca2+ binding site within one of the lobes of the protein. We show that the residues involved in ligand and Ca2+ binding are conserved among the 610 SBPs from experimentally uncharacterized FmdA_AmdA amidase-associated ABC transporter systems, suggesting these homologous systems are also likely to be involved in the sensing, uptake, and metabolism of l-amino acid amides across many Gram-negative nitrogen-fixing soil bacteria. We propose that MeAmi_SBP is involved in the uptake of such solutes to supplement pathways such as the citric acid cycle and the glutamine synthetase-glutamate synthase pathway. This work expands our currently limited understanding of microbial interactions with l-amino acid amides and bacterial nitrogen utilization.

Artificial Intelligence-based Amide-II Infrared Spectroscopy Simulation for Monitoring Protein Hydrogen Bonding Dynamics.

The structurally sensitive amide II infrared (IR) bands of proteins provide valuable information about the hydrogen bonding of protein secondary structures, which is crucial for understanding protein dynamics and associated functions. However, deciphering protein structures from experimental amide II spectra relies on time-consuming quantum chemical calculations on tens of thousands of representative configurations in solvent water. Currently, the accurate simulation of amide II spectra for whole proteins remains a challenge. Here, we present a machine learning (ML)-based protocol designed to efficiently simulate the amide II IR spectra of various proteins with an accuracy comparable to experimental results. This protocol stands out as a cost-effective and efficient alternative for studying protein dynamics, including the identification of secondary structures and monitoring the dynamics of protein hydrogen bonding under different pH conditions and during protein folding process. Our method provides a valuable tool in the field of protein research, focusing on the study of dynamic properties of proteins, especially those related to hydrogen bonding, using amide II IR spectroscopy.

General Installation of (4H)-Imidazolone cis-Amide Bioisosteres Along the Peptide Backbone.

Imidazolones represent an important class of heterocycles present in a wide range of pharmaceuticals, metabolites, and bioactive natural products and serve as the active chromophore in green fluorescent protein. Recently, imidazolones have received attention for their ability to act as a nonaromatic amide bond bioisotere which improves pharmacological properties. Herein, we present a tandem amidine installation and cyclization with an adjacent ester to yield (4H)-imidazolone products. Using amino acid building blocks, we can access the first examples of α-chiral imidazolones that have been previously inaccessible. Additionally, our method is amenable to on-resin installation which can be seamlessly integrated into existing solid-phase peptide synthesis protocols. Finally, we show that peptide imidazolones are potent cis-amide bond surrogates that preorganize linear peptides for head-to-tail macrocyclization. This work represents the first general approach to the backbone and side-chain insertion of imidazolone bioisosteres at various positions in linear and cyclic peptides.

Efficient Screening of Pharmaceutical Cocrystals by Microspacing In-Air Sublimation.

Cocrystal screening and single-crystal growth remain the primary obstacles in the development of pharmaceutical cocrystals. Here, we present a new approach for cocrystal screening, microspacing in-air sublimation (MAS), to obtain new cocrystals and grow high-quality single crystals of cocrystals within tens of minutes. The method possesses the advantages of strong designable ability of devices, user-friendly control, and compatibility with materials, especially for the thermolabile molecules. A novel drug-drug cocrystal of favipiravir (FPV) with salicylamide (SAA) was first discovered by this method, which shows improved physiochemical properties. Furthermore, this method proved effective in cultivating single crystals of FPV-isonicotinamide (FPV-INIA), FPV-urea, FPV-nicotinamide (FPV-NIA), and FPV-tromethamine (FPV-Tro) cocrystals, and the structures of these cocrystals were determined for the first time. By adjusting the growth temperature and growth distance precisely, we also achieved single crystals of 10 different paracetamol (PCA) cocrystals and piracetam (PIR) cocrystals, which underscores the versatility and efficiency of this method in pharmaceutical cocrystal screening.

Competing Reaction Mechanisms of Peptide Bond Formation in Water Revealed by Deep Potential Molecular Dynamics and Path Sampling.

The formation of an amide bond is an essential step in the synthesis of materials and drugs, and in the assembly of amino acids to form peptides. The mechanism of this reaction has been studied extensively, in particular to understand how it can be catalyzed, but a representation capable of explaining all the experimental data is still lacking. Numerical simulation should provide the necessary molecular description, but the solvent involvement poses a number of challenges. Here, we combine the efficiency and accuracy of neural network potential-based reactive molecular dynamics with the extensive and unbiased exploration of reaction pathways provided by transition path sampling. Using microsecond-scale simulations at the density functional theory level, we show that this method reveals the presence of two competing distinct mechanisms for peptide bond formation between alanine esters in aqueous solution. We describe how both reaction pathways, via a general base catalysis mechanism and via direct cleavage of the tetrahedral intermediate respectively, change with pH. This result contrasts with the conventional mechanism involving a single pathway in which only the barrier heights are affected by pH. We show that this new proposal involving two competing mechanisms is consistent with the experimental data, and we discuss the implications for peptide bond formation under prebiotic conditions and in the ribosome. Our work shows that integrating deep potential molecular dynamics with path sampling provides a powerful approach for exploring complex chemical mechanisms.